The CFTC’s crypto rulemaking carries no Unified Agenda entry and no press release, so the OIRA queue is its only public trace. The classification there is Prerule, the stage before a notice of proposed rulemaking. No text yet exists for anyone to read. The question I am interested in comes earlier, being what the Commission could write at all on the authority the Commodity Exchange Act already gives it.

The filing

On 17 September 2026 the Commodity Futures Trading Commission sent a rulemaking to the Office of Information and Regulatory Affairs under the title Regulation Crypto Asset Transactions and Regulation Crypto Asset Markets. The entry carries RIN 3038-AF80 and records the stage as Prerule.¹ It is the only Commission item in the OIRA queue. No entry for it appears in the Unified Agenda. The Commission has issued no press release about it, so the queue itself is the submission’s only public trace.²

The difference between a prerule and a proposed rule is not a formality. Executive Order 12866 gives OIRA ten working days to review a notice of inquiry, an advance notice of proposed rulemaking or any other preliminary regulatory action taken before a notice of proposed rulemaking. Proposed and final rules get ninety calendar days.³ The submission is preliminary and not a proposed rule. While it remains under OIRA review the public entry carries no text anyone can read or comment on.

Two days earlier, on Tuesday 15 September, the Senate declined to invoke cloture on the motion to proceed to the Digital Asset Market Clarity Act. The motion was rejected 49 to 50, against the sixty votes required.⁴ Chairman Michael Selig said in the same week that the Commission would proceed under its existing statutory authorities. The chairman of the Securities and Exchange Commission said much the same for his own agency.⁵

OIRA discloses the title, the receipt date and the classification. It does not disclose the text or the statutory basis, so what follows addresses the question the title raises rather than a rule anyone has read. The first thing to look for when a text appears is whether the Commission has confined itself to leveraged retail products or claimed a wider cash market perimeter. Either way it has to identify, somewhere in the Commodity Exchange Act, a cash market it is entitled to regulate.

What the Commodity Exchange Act gives

Three provisions frame the jurisdictional question. Section 2(c)(2)(D) supplies the direct route by which certain spot transactions become subject to exchange trading requirements, with the designated contract market provisions and Part 40 of the Commission’s regulations supplying the machinery that follows.

Section 2(a)(1)(A) gives the Commission exclusive jurisdiction over accounts, agreements and transactions involving swaps or contracts of sale of a commodity for future delivery.⁶ Futures and swaps, in other words. Nothing in it touches a cash sale.

Section 6(c)(1), codified at 7 U.S.C. section 9(1), makes it unlawful to use or employ any “manipulative or deceptive device or contrivance” in connection with a “contract of sale of any commodity in interstate commerce”, in contravention of such rules as the Commission promulgates.⁷ Those last quoted words are what carry the Commission into the spot market. The Ninth Circuit held in Commodity Futures Trading Commission v Monex Credit Company that the disjunction between manipulative and deceptive is a real one, so that the Commission may sue for fraudulent conduct without also pleading manipulation.⁸ District courts in New York and Massachusetts had already held that virtual currencies are commodities within the meaning of the Act.⁹ What those decisions confirm is an enforcement authority. It permits the Commission to sue a person who defrauded a buyer of bitcoin. It does not permit the Commission to tell a venue how to hold client assets, how to sequence a matching engine or what to file when it changes a rule.

Section 2(c)(2)(D), the retail commodity transaction provision, is the operative one. It applies to any agreement in any commodity that is entered into with or offered to a person who is not an eligible contract participant or an eligible commercial entity and that is entered into or offered on a leveraged or margined basis.¹⁰ Clause (iii) then applies 7 U.S.C. sections 6(a), 6(b) and 6b, which are sections 4(a), 4(b) and 4b of the Act, to such a transaction “as if” it “was a contract of sale of a commodity for future delivery”.¹¹ Section 4(a) is the exchange-trading requirement, which is why a leveraged retail transaction in a commodity is deemed into the futures regime and has to be traded on a designated contract market.

That deeming carries everything the Commission has so far done in spot crypto. The listed spot trading initiative announced in August 2025 and the first leveraged retail spot crypto exchange, announced on 4 December 2025 for launch in the week of 8 December, both proceed on the footing that section 2(c)(2)(D) permits a registered contract market to list a spot digital asset product.¹²

The twenty-eight day problem

The deeming provision has a hole in it and the hole is where the cash market is.

Clause (ii)(III)(aa) removes from the subparagraph any contract of sale that “results in actual delivery within 28 days or such other longer period as the Commission may determine by rule or regulation based upon the typical commercial practice in cash or spot markets for the commodity involved”.¹³ A retail transaction that delivers promptly is not deemed a future. It falls outside section 4(a) and therefore outside the requirement to trade on a designated contract market, together with the registration and core principle obligations that attach to trading there.

The reach of section 2(c)(2)(D) is therefore confined to retail leveraged or margined transactions that do not deliver within twenty-eight days. An eligible contract participant buying spot bitcoin falls outside it, because clause (i) applies only where the counterparty is neither an eligible contract participant nor an eligible commercial entity. A retail buyer who pays in full and takes delivery falls outside it too. So does a venue matching unleveraged spot trades between pre-funded accounts. Under the Act those transactions stay inside the Commission’s cash market antifraud and anti-manipulation perimeter and outside the exchange trading regime that section 2(c)(2)(D) creates. A rule called Regulation Crypto Asset Markets that rests on that subparagraph regulates a leveraged corner of the market and leaves the cash market where it found it.

The Commission cannot reach the rest of that market by adjusting the delivery period. The only adjustment clause (ii)(III)(aa) permits is a period longer than twenty-eight days, never a shorter one. Lengthening it would contract the Commission’s jurisdiction. A longer period would bring more transactions within the actual delivery exclusion, so fewer leveraged retail transactions would be deemed into the futures regime. Settlement practice in digital assets runs to days or less, which favours a shorter period. The Commission said as much in its 2020 interpretation on retail commodity transactions in virtual currency. It agreed there that the delivery period should correspond to the reality of a virtual currency spot transaction, while recording that it is limited in its ability to shorten the twenty-eight days the statute specifies.¹⁴ Congress set that floor and only Congress can lower it.

What has widened the perimeter is litigation rather than rulemaking. In Monex the Ninth Circuit read actual delivery to require “at least some meaningful degree of possession or control by the customer”. It rejected an arrangement under which metals stayed in the broker’s chosen depository, never changed hands and remained subject to the broker’s exclusive control.¹⁵ That construction does not condemn custody as such. The Commission’s 2020 interpretation accepts actual delivery through a depository, on conditions. A customer must secure possession and control of the entire quantity with the ability to use it freely in commerce. Once the twenty-eight days expire neither the offeror nor the counterparty seller may retain any interest, legal right or control over the asset.¹⁶ Where a platform retains control or the customer cannot transfer or use the asset freely, delivery is not made out and the transaction stays inside the deeming provision without the Commission writing anything at all. What follows may be a wider perimeter over leveraged retail transactions and no corresponding authority over the unleveraged cash market.

What CLARITY would have supplied

That gap does not have to be inferred. The House Financial Services Committee’s section-by-section analysis of its own bill describes section 401 as giving the Commission “exclusive regulatory jurisdiction over digital commodity cash or spot markets that occur on or with new entities required to be registered”. It states that this authority “complements the CFTC’s existing anti-fraud and anti-manipulation authority over all cash or spot market commodity transactions”.¹⁷ Sections 404 and 406 would have created three registration categories to carry that jurisdiction, covering the digital commodity exchange, the digital commodity broker and the digital commodity dealer.¹⁸

That is the bill’s own sponsors, writing in support of it, describing what the Commission presently holds over cash and spot commodity transactions as anti-fraud and anti-manipulation authority. They were explaining what the bill would add. Whoever drafts the preamble to the Commission’s rule will now have to explain why the position before the bill was something more than the bill’s promoters said it was.

Section 22 and the limits of provision-specific deeming

For litigants the more immediate consequence is narrower and easier to miss. The deeming in clause (iii) is provision-specific. It applies sections 4(a), 4(b) and 4b of the Act to a leveraged retail transaction as if that transaction were a contract of sale for future delivery. It does not deem the transaction to be a future for the purposes of the Act at large.

Section 22(a)(1)(B), at 7 U.S.C. section 25(a)(1)(B), gives a private right of action to a person who made through the defendant “any contract of sale of any commodity for future delivery”.¹⁹ A customer who loses money on a listed leveraged spot product therefore faces a threshold question. Did he make a contract for future delivery within the meaning of section 22 or a spot contract that section 2(c)(2)(D) deems to be a future for three enumerated provisions and no others? Clause (iii) names 7 U.S.C. sections 6(a), 6(b) and 6b and stops there, so the text does not answer the question.

The private remedy does not end there. Section 22(b)(1) gives a separate action against a registered entity. It is available to a person who engaged in any transaction on or subject to that entity’s rules, for actual losses caused by the entity’s failure to enforce a rule it is required to enforce or by its enforcement of one.²⁰ Section 22(b)(4) requires the claimant to establish bad faith.²¹ A customer of a designated contract market listing spot digital asset products may therefore have a route against the exchange whether or not the underlying contract counts as a future under section 22(a).

The gap, such as it is, is claim-specific. A rulemaking cannot enlarge the causes of action Congress enacted. Within a validly regulated perimeter the Commission can impose registration, custody and rule-filing requirements without touching them. Nor does the narrow drafting of clause (iii) disturb the registered entity liability that section 22(b) supplies independently. A market structure rule could supply obligations relevant to such a claim, without displacing the separate requirements of bad faith, causation and actual loss that section 22(b) imposes.

Loper Bright and the shape of a challenge

It is tempting to reach for Loper Bright Enterprises v Raimondo and stop there. The decision did less than that. It removed the obligation on a court to defer to an agency’s reading of an ambiguous statute.²² It left arbitrary and capricious review untouched, that being a separate inquiry under the Administrative Procedure Act into whether the agency explained itself against the standard the statute sets. It has nothing at all to say about how many commissioners an agency has. Keeping those three apart is the difference between an argument and a slogan.

On the statutory question, the issue is not whether the Commission may regulate retail leveraged commodity transactions, which section 2(c)(2)(D) expressly permits, nor whether it may set a delivery period longer than twenty-eight days, which the same clause expressly permits. The issue is whether a registration and market-conduct regime for crypto asset markets generally fits the transactions the statute actually deems into the futures regime and whether the Commission has explained why it does. A rule confined to leveraged retail products listed on designated contract markets is comfortably inside the grant. The general rulemaking power at section 8a(5) is available to fill in the detail.²³ A rule that reaches custody, settlement or trading in the cash market, on the footing that the deemed futures character of one class of transaction licenses the Commission to govern the venue at large, is a different proposition. A reviewing court will now resolve that question for itself without deferring to the Commission’s reading of the Act.

On the reasoning question the prerule posture is the Commission’s friend. An advance notice that asks the right questions and builds a comment file is the orthodox way to produce the explanation that arbitrary and capricious review demands. The exposure grows if the Commission proceeds to a proposed rule without using the preliminary stage to expose and test its jurisdictional premise.

Commission composition is a separate question again. The Commission’s own website currently lists one commissioner, Chairman Selig.²⁴ It has continued to act with him as its sole serving member, approving a final rule on whistleblower awards on 11 September.²⁵ The consequence is one of durability, since a framework resting on a short statutory foundation can be revisited without much difficulty by the three or four commissioners who follow him.

The comparator

Four weeks earlier the Securities and Exchange Commission proposed Regulation Crypto Assets, on a firmer statutory footing than the Commodity Futures Trading Commission has available to it.²⁶ Its offering exemptions rest on the Commission’s exemptive authority. Section 18(b)(3) expressly allows it to define qualified purchaser differently for different categories of securities.

The Commission’s position under the Commodity Exchange Act is the reverse of that. Its expansive provision, section 6(c)(1), supports enforcement and not registration, while the provision that does support registration, section 2(c)(2)(D), reaches only a slice of the market. A rulemaking that produces a market structure regime for crypto assets generally has to do something that neither provision does on its own, which is the difficulty the preamble will have to meet.

What to watch

Watch the stage at which the text emerges. An advance notice carrying questions means the Commission is building a record and the framework is a 2027 proposition at the earliest. A proposed rule appearing within weeks means the prerule filing was a formality and the text already exists.

Then read the authority section. A release citing section 2(c)(2)(D) and section 8a(5), confined to leveraged retail products on registered exchanges, is defensible on its face. A release reaching into cash market conduct on the strength of section 6(c)(1) invites the challenge set out above and invites it from a direction the industry may not expect. The firms building venues want the rules. The challenge is more likely to come from a firm left outside the perimeter, from a state regulator displaced by it or from a defendant in an enforcement action taking the point collaterally.

Watch also for any proposal touching the delivery period. Lengthening it would narrow the Commission’s own perimeter, so a proposal to do so would need an explanation. The standard against which it would be measured is typical commercial practice in the cash or spot market for the commodity involved.

The Senate’s refusal to act left the Commission with an antifraud provision that reaches the whole cash market, a deeming provision that reaches a corner of it and no power to enlarge that corner by shortening the statutory delivery period. A framework can be written on that foundation. How much of the market it reaches is a question for a court reading section 2(c)(2)(D).

Notes

1. Office of Information and Regulatory Affairs, Executive Order Submissions Under Review, entry for Commodity Futures Trading Commission, RIN 3038-AF80, Regulation Crypto Asset Transactions and Regulation Crypto Asset Markets, stage Prerule, received 17 September 2026, status Pending Review. Consulted 23 September 2026 at reginfo.gov.

2. A search of the Unified Agenda for RIN 3038-AF80 at reginfo.gov returns the single line “This rule has not been published in a Unified Agenda”. The most recent Commission press release is 9302-26 of 22 September 2026, a staff advisory on mention markets. No release announces the submission. Both consulted 23 September 2026.

3. Executive Order 12866, section 6(b)(2)(A) and (B). Ten working days for a notice of inquiry, advance notice of proposed rulemaking or other preliminary regulatory action taken before a notice of proposed rulemaking. Ninety calendar days for all other regulatory actions, reduced to forty-five where OIRA has previously reviewed the information and there has been no material change, extendable once by up to thirty days with the written approval of the Director of the Office of Management and Budget or at the request of the agency head.

4. United States Senate, Cloture Motions, 119th Congress, recording a cloture motion on the motion to proceed to H.R. 3633, the Digital Asset Market Clarity Act, filed 8 August 2026 and rejected on 15 September 2026 by 49 votes to 50 against the sixty required.

5. Reported by The Block, 16 September 2026, quoting Chairman Selig as saying the Commission is “locked in and ready to ship its rules for the new frontier of finance” and Chairman Atkins as saying that “with or without legislation, we will act decisively within the SEC’s statutory authority to deliver certainty”. Both quotations are taken from posts on X and are reproduced as reported.

6. Commodity Exchange Act section 2(a)(1)(A), 7 U.S.C. section 2(a)(1)(A).

7. 7 U.S.C. section 9(1).

8. United States Commodity Futures Trading Commission v Monex Credit Company, 931 F.3d 966 (9th Cir. 2019), No. 18-55815, decided 25 July 2019.

9. Commodity Futures Trading Commission v McDonnell, 287 F. Supp. 3d 213 (E.D.N.Y. 2018). Commodity Futures Trading Commission v My Big Coin Pay, Inc., 334 F. Supp. 3d 492 (D. Mass. 2018).

10. 7 U.S.C. section 2(c)(2)(D)(i).

11. 7 U.S.C. section 2(c)(2)(D)(iii), which applies 7 U.S.C. sections 6(a), 6(b) and 6b, being sections 4(a), 4(b) and 4b of the Commodity Exchange Act. Section 6 of title 7 is headed Regulation of futures trading and foreign transactions and carries at subsection (a) the requirement that a futures transaction be made on or subject to the rules of a designated contract market.

12. CFTC Press Release 9105-25, Acting Chairman Pham Launches Listed Spot Crypto Trading Initiative, 4 August 2025, which invites comment on listing spot crypto asset contracts on a designated contract market “including section 2(c)(2)(D) of the Commodity Exchange Act, Part 40 of CFTC regulations”. CFTC Press Release 9145-25, 4 December 2025, states that listed spot cryptocurrency products “will begin trading for the first time in U.S. federally regulated markets on CFTC registered futures exchanges” and names no exchange or date. Bitnomial’s own announcement of the same date puts the launch in the week of 8 December 2025.

13. 7 U.S.C. section 2(c)(2)(D)(ii)(III)(aa).

14. Retail Commodity Transactions Involving Certain Digital Assets, 85 Fed. Reg. 37734 (24 June 2020). The Commission agrees that the actual delivery period should correspond to the reality of a virtual currency spot transaction and records that it is limited in its ability to shorten the twenty-eight days specified in section 2(c)(2)(D).

15. Ibid., on the conditions for actual delivery. The customer must secure “possession and control of the entire quantity of the commodity” and “the ability to use the entire quantity of the commodity freely in commerce” within twenty-eight days. The offeror and counterparty seller must retain no interest, legal right or control over the commodity at the expiration of that period.

16. Monex, above, on the construction of actual delivery.

17. Committee on Financial Services, United States House of Representatives, Section-by-Section, Digital Asset Market Clarity (CLARITY) Act of 2025, 10 July 2025, section 401.

18. Ibid., sections 404 and 406.

19. 7 U.S.C. section 25(a)(1)(B).

20. 7 U.S.C. section 25(b)(1), which makes a registered entity that fails to enforce a rule it is required to enforce liable for actual damages sustained by a person who engaged in any transaction on or subject to the rules of that entity.

21. 7 U.S.C. section 25(b)(4), requiring a person seeking to enforce liability under the section to establish that the registered entity acted in bad faith and that the failure or action caused the loss.

22. Loper Bright Enterprises v Raimondo (2024).

23. Commodity Exchange Act section 8a(5), 7 U.S.C. section 12a(5), empowering the Commission to make such rules and regulations as in its judgment are “reasonably necessary to effectuate any of the provisions or to accomplish any of the purposes of this chapter”.

24. Commodity Futures Trading Commission, Chairman and Commissioners, consulted 23 September 2026, listing Michael S. Selig alone, as Chairman.

25. CFTC Press Release 9297-26, CFTC Approves Final Rule Concerning Whistleblower Awards, 11 September 2026.

26. Securities and Exchange Commission, Regulation Crypto Assets, Release Nos. 33-11434 and 34-106150, File No. S7-2026-27, issued 18 August 2026, published in the Federal Register on 21 August 2026 with comments due by 20 October 2026.

Jump Trading must face the bulk of a proposed securities class action over the Terra collapse, on a ruling reported on 3 September in Patterson v TerraForm Labs Pte Ltd, No. 22-cv-03600 (N.D. Cal.).¹ Plaintiffs allege that when UST lost its dollar peg in May 2021 and Terraform’s algorithm failed to restore it, Jump agreed with Do Kwon to buy large quantities of UST at above market prices to hold the coin at $1, without disclosing the arrangement.

On the available reporting, the omission and manipulation theories do not appear to treat price support as wrongful in itself. Alleged concealment is central to the case.

The US stabilisation framework

Section 9(a)(6) of the Securities Exchange Act, 15 U.S.C. § 78i(a)(6), makes it unlawful to effect a series of transactions in any security other than a government security for the purpose of pegging, fixing or stabilising its price, but only in contravention of rules the Commission prescribes. The provision once reached only securities registered on a national securities exchange. Section 929L of the Dodd-Frank Act removed that limit in 2010.² Whether the section and an implementing rule could reach a given digital asset therefore requires analysis of the instrument and the conduct, and is not disposed of by noting that the asset trades off exchange.

Regulation M Rule 104, 17 CFR 242.104, supplies the rules.³ Stabilisation is permitted subject to conditions covering the purpose of the bid, price limits, notification and recordkeeping, together with the requirement in Rule 104(c) that a person stabilising grant priority to any independent bid at the same price, irrespective of the size of that bid. Disclosure alone does not satisfy those conditions, and silence alone does not make out a fraud or manipulation claim.

Rule 104 is built around stabilisation in connection with a securities distribution. Its price tests and disclosure machinery assume an offering context and a distribution with a beginning and an end. Ordinary peg maintenance is not necessarily tied to that structure and is designed to operate continuously.

Section 17 of the GENIUS Act, Public Law 119-27, carves a payment stablecoin issued by a permitted payment stablecoin issuer out of the definition of security in the 1933 and 1934 Acts and out of the definition of commodity in the Commodity Exchange Act.⁴ UST was not issued by a permitted payment stablecoin issuer, and the statutory exclusion does not govern the pleaded events. Where the exclusion applies, a claim cannot proceed merely by assuming that the qualifying payment stablecoin is itself a security. It would require a connection to another security or a different legal route. Terra remains a useful comparison about concealed support without carrying the cause of action across.

Disclosure and manipulation under MiCA

Within the scope fixed by Article 86, which covers crypto-assets admitted to trading or for which admission has been requested, MiCA addresses this territory from two directions. Article 88 requires issuers, offerors and persons seeking admission to trading to disclose inside information directly concerning them as soon as possible. Disclosure may be delayed where immediate disclosure would prejudice legitimate interests, delay is not likely to mislead the public and confidentiality can be ensured. Article 87 supplies the inside information test. A support arrangement could engage that duty depending on its precision, its non-public character, its likely price significance and whether it directly concerns the disclosing person.

Article 91(2)(a) prohibits, unless carried out for legitimate reasons, entering into a transaction or engaging in behaviour which gives or is likely to give false or misleading signals as to supply, demand or price, or which secures or is likely to secure the price of a crypto-asset at an abnormal or artificial level. Maintaining a token’s promised reference value is not automatically maintaining an artificial price, and the mechanism, the representations made about it and the surrounding circumstances all bear on the answer.

What Title VI does not contain is a counterpart to Article 5 of the Market Abuse Regulation, which exempts stabilisation carried out in accordance with Commission Delegated Regulation (EU) 2016/1052. That instrument sets instrument-specific periods, thirty calendar days from commencement of trading for an initial offer of shares, with separate treatment for debt, and staged disclosure under Article 6 before the offer, during the period and within a week of its close.⁵

Classification is often put the wrong way round here. Recital 41 provides that where a crypto-asset falls within the asset-referenced token or e-money token definition, the applicable title applies irrespective of the mechanism by which the issuer intends to maintain a stable value, algorithmic mechanisms included. Inability to satisfy a regime’s requirements is not the same as falling outside its scope. A token purporting to maintain a stable value by reference to one official currency is an e-money token under Article 3(1)(7), which puts a dollar-referencing design in Title IV rather than Title III.

The remaining regulatory question

The GENIUS Act framework binds permitted issuers from 18 January 2027 on the section 20 backstop, subject to the earlier commencement mechanism if final implementing rules are issued. A permitted issuer holding a reserve may face redemption pressure capable of moving the secondary price below par. Redemption at par and secondary market purchases are different responses with different consequences, and the choice between them is a live commercial question.

MiCA contains a disclosure duty and a manipulation rule capable of applying to a concealed support arrangement within Article 86. The GENIUS Act prescribes reserve, redemption and publication requirements, but does not supply an equivalent rule for undisclosed peg-support purchases. Neither regime contains a dedicated framework equivalent to Rule 104 or Article 5 for continuing peg maintenance.

Notes

¹ Reported by Law360, “Jump Trading Must Face Terraform Stablecoin Fraud Claims”, 3 September 2026, and The D&O Diary, September 2026. The order itself was not accessible. Its date, reasoning and the precise claims permitted to proceed have not been verified against the record, and the May 2021 characterisation is pleaded rather than established.

² Dodd-Frank Wall Street Reform and Consumer Protection Act, Public Law 111-203, section 929L(1).

³ 17 CFR 242.104, including Rule 104(c).

⁴ GENIUS Act, Public Law 119-27, 139 Stat. 419, section 17.

⁵ Commission Delegated Regulation (EU) 2016/1052, Articles 5 and 6.

Cecil Wright & Partners commenced proceedings in the High Court on 14 July 2026 against Nik Storonsky, the co-founder and chief executive of Revolut, claiming €17.5 million in commission on his purchase of a 102-metre Lürssen for a reported €350 million.^[1] The brokerage says that it was the effective cause of the acquisition and that a commission of 5 per cent is due on it. Storonsky denies liability. The defence reported this month introduces an unfamiliar item of evidence into a dispute otherwise governed by a century of agency law, because he says that he already knew who had owned the yacht, having found the connection through a ChatGPT search of publicly available information.^[2]

No one suggests that a generative model negotiated a €350 million purchase, nor does the pleading go near such a claim. It does something narrower and more useful, since an exchange with a chatbot is advanced as evidence of how a buyer came to know whom to approach. The question of how a buyer came by that knowledge belongs to the law of effective cause, which English judges have been applying to yacht brokers for the better part of two decades.

One yacht, two transactions

Nixie was commissioned from Lürssen by Patrick Dovigi, the Canadian founder of GFL Environmental. A contractual interest in the project appears to have been transferred during construction in 2024 to the Brazilian financier Daniel Vorcaro, then the president of Banco Master.^[3] Storonsky’s family office approached Cecil Wright in October 2024 about a new build, then wrote again on 19 July 2025 to ask about an interim vessel to use while that build proceeded. Cecil Wright identified the Lürssen project in Germany and says that its client remained interested in acquiring her.^[4]

Vorcaro was detained at Guarulhos airport on 17 November 2025 in connection with an alleged fraud of R$12.2 billion involving Banco Master, allegations he has contested.^[5] The negotiations in which Cecil Wright had been engaged came to nothing, Dovigi recovered the vessel and on 12 January 2026 he sold her to Storonsky, four months before Lürssen delivered her on 26 June 2026.^[6]

Everything turns on how that sequence is described. For Cecil Wright there was one acquisition process, interrupted by an unusual change of ownership and completed by the client without the broker. For Storonsky there were two transactions, the first of which failed after the arrest and the second of which began independently with a seller the brokerage had never introduced to him.

Whether Vorcaro ever held title to the vessel is not established on the public record. The material that can be opened suggests he did not. She was launched in February 2026 and delivered on 26 June 2026, both of them after his arrest, so there was no delivery to him at which title to a finished ship could pass. Her flag is the Marshall Islands, whose registry publishes nothing searchable, so the transcript of register that would settle the question is obtainable only on application.^[7] Brazilian reporting describes the 2024 transaction as the acquisition of a position under the building contract, records that he never went aboard and says that Dovigi recovered the vessel by invoking contractual clauses once payments were not made.^[8]

The liquidator of Banco Master has subpoenaed Lürssen in Chapter 15 proceedings in the United States for the contracts and the records of transfers, which is the conduct of an office-holder trying to establish what happened rather than one holding a conveyance.^[9] The point bears on the shape of the defence. If what passed in 2024 was a contractual position that reverted before delivery, the available material is consistent with Dovigi either having remained the relevant contractual counterparty before the January sale or having become that counterparty again. A different beneficial or contractual interest during construction is not necessarily a different seller of the completed yacht. That does not answer the causal question, since Moran shows that a long interval and fresh negotiations can break the chain whoever the seller turns out to be, although it weakens one of the two planks on which the reported defence rests.

First, was there a commission agreement?

There is a question before effective cause. What, if anything, had Storonsky agreed to pay?

No conventional written brokerage agreement has been identified in the public reporting of the filings. The material reported so far includes an email from Storonsky’s team on 19 July 2025 referring to the “commercial arrangements you would expect from this” and using the word “commission”. Cecil Wright claims 5 per cent of the €350 million purchase price, although no report of the filings traces that percentage to an agreed term.^[10]

The sequence is important, because the appeal in Moran Yacht & Ship Inc v Pisarev did not turn on effective cause. The Court of Appeal dismissed the broker’s appeal because neither the express nor the implied terms of the agreements entitled it to the commission claimed. Christopher Clarke LJ therefore found it unnecessary to decide whether Males J had been right on causation.^[11] Before asking whether Cecil Wright caused the purchase, the court will have to determine what bargain the parties made and what it entitled the brokerage to be paid.

The effective cause rule

An agent whose remuneration depends on bringing about a transaction must ordinarily show that it was the effective cause of that transaction. Males J restated the rule in Moran by reference to Chitty and to the formulation in Nahum v Royal Holloway and Bedford New College, under which the claimant has to establish that “it was his actions that really brought about the relation of buyer and seller”.^[12] The agent need not be the immediate cause, although the connection between what the agent did and the transaction that followed has to be sufficient to support the claim.

Berezovsky and Moran

Boris Berezovsky instructed Edmiston & Co on 14 May 2008 to market the 110-metre Darius on a discreet and non-exclusive basis. Merle Wood, a Florida broker, passed the details to Captain Wrigley, who showed them to Abdullah Al Futtaim on 13 July 2008. The Al Futtaim family bought the yacht through Paragon International on 22 October 2008 for €240 million.^[13]

The purchasers dealt directly and had no wish to negotiate through the brokers. Field J held that Edmiston was nonetheless the effective cause, on the footing that the Al Futtaims had no knowledge of Darius’s existence, let alone of her availability, until the information reached them by that route.^[14] The Court of Appeal upheld the finding and reduced the reasonable rate from 3 per cent to 2.5 per cent, which produced an award of €6 million.^[15]

Moran went the other way on its facts. Lidia Tsareva showed 4YOU to Alexander Miliavsky on 18 May 2010 and nothing came of the visit, which Males J found to have been short and unserious. Some 21 months later a company controlled by Miliavsky contracted to buy the vessel for €19.8 million after discussions that had arisen independently. The judge held that Moran had not been the effective cause, the earlier introduction forming part of the history without having brought the transaction about.^[16]

That holding has to be read with the appeal already described. Males J’s decision in Moran illustrates how a long interval and a fresh set of negotiations can break the chain, although the Court of Appeal expressly left that causation analysis undecided and disposed of the case on the agreements instead. Both authorities differ from this dispute in the direction of the retainer. Moran and Edmiston were each engaged on the sell side and claimed against the owner who had instructed them, whereas Cecil Wright is claiming against a buyer whose family office had approached it.

Where the search enters the chain

Cecil Wright has an argument from Berezovsky available on its own pleaded case. The client was looking for an interim yacht, the broker identified this one, brought the opportunity to his attention and remained engaged while he pursued it. That he concluded matters without the brokerage does not by itself answer whether the commission was earned.

Storonsky’s reported defence is closer to Moran. The proposed transaction with Vorcaro failed after the arrest, Dovigi resumed control of the project and the negotiations that produced the January 2026 sale are said to have begun without any introduction from Cecil Wright. The same vessel therefore stands at the centre of two competing accounts of what occurred between November 2025 and January 2026.

The ChatGPT search has a part to play only within the second account. If the route by which Storonsky came to know of Dovigi’s connection with the yacht is disputed, the interaction may support an alternative explanation of how that knowledge reached him. That is a modest function for a generative model, although it is capable of bearing on €17.5 million.

A prompt as evidence

A party who says that information came from ChatGPT raises a different evidential question from one who says that it came from a named person, a document or a database, because the model is not the underlying source. Where the point is contested, potentially relevant disclosable material could include the prompt, the response, the date and any contemporaneous step taken as a result, together with the public information from which the answer could have been drawn.

No AI response proves ownership of a yacht and none is offered here for that purpose. The proposition advanced is narrower, since the interaction would be evidence that Storonsky held particular information at a particular time and, if the record supports it, evidence of the route by which he obtained it. Much of the litigation about generative models has turned on the unreliability of what they produce, whereas the question in this case may be whether somebody asked before the broker answered.

The brokers in the background

Chris Cecil-Wright, who founded Cecil Wright & Partners in February 2013, has appeared in the reports of English superyacht commission litigation before.^[17] Field J recorded that Nicholas Edmiston reported the meeting of 14 May 2008 to his son Jamie, who managed the London office, as well as to another of his senior brokers, Chris Cecil-Wright.^[18] Eighteen years later his own firm is asking the High Court to apply the same test to another nine-figure transaction.

There is a further coincidence in the professional cast. Moran Yacht & Ship, the claimant in the authority that helps Storonsky, was the build manager for Nixie and says on its own website that it managed the project from contract through her subsequent sale.^[19] Whether that role touched the January 2026 transaction is not apparent from the public record, which is a matter the pleadings should settle.

Darius supplies the closer parallel. She too was commissioned by one owner, was sold before completion to another and produced a commission claim that reached the Court of Appeal. She is afloat today as Radiant, owned by the family that bought her in 2008. The law that decided her case in 2010 is the law that will decide this one. What has changed is the evidence by which a buyer may seek to prove that he found his own way to the seller. This time, part of that evidence is a ChatGPT search.

^[1] The claim was issued in the High Court on 14 July 2026. Dentons UK and Middle East act for Cecil Wright & Partners and Hannaford Turner for Storonsky. €17.5 million is 5 per cent of €350 million. YachtBuyer, “Cecil Wright sues Revolut founder Nik Storonsky for €17.5m over alleged superyacht commission”, 4 August 2026, records that the filings describe the vessel only as a 102-metre Lürssen that was nearing delivery, so the identification with Nixie comes from the reporting rather than from the pleadings. Sifted gives the rate as 8 per cent, which is inconsistent with the arithmetic and with every other report.

^[2] Bloomberg, “Revolut CEO Found €350 Million Yacht’s Ex-Owner Via ChatGPT, His Lawyers Say”, 16 September 2026.

^[3] SuperYachtFan, NIXIE, recording that the yacht was commissioned by Patrick Dovigi, founder of GFL Environmental, was sold during construction to a Brazilian owner and was then reacquired by Dovigi. The Brazilian buyer is identified as Daniel Vorcaro in the reporting of the claim.

^[4] Marine Industry News, as above, for the October 2024 and July 2025 approaches. Superyacht Investor, “Cecil Wright sues Revolut founder Storonsky”, 7 August 2026, gives the completion date of 12 January 2026.

^[5] Reuters, “Banco Master owner Vorcaro detained by Brazil police”; Wikipedia, Daniel Vorcaro, recording the detention at São Paulo Guarulhos on 17 November 2025 and the alleged R$12.2 billion in fraudulent transactions with Banco de Brasília.

^[6] SuperYacht Times, “102m Lürssen superyacht Nixie delivered”, giving delivery on 26 June 2026, a length of 102.4 metres and the build name Project JASSJ. YachtBuyer, “First Look: Secretive 102m Lürssen Superyacht Jassi Revealed at Launch”, 6 February 2026, covers the launch and still describes the operative sale as the 2021 contract. Boat International, 27 July 2026, records that “the project was sold by Moran Yacht & Ship in 2021”.

^[7] Flag and identity from vessel databases giving IMO 9954541, MMSI 538072853 and a Marshall Islands flag. International Registries publishes no searchable public register, and Equasis, GISIS, MarineTraffic, VesselFinder and the DNV register were each login-walled, robots-blocked or returned nothing. No record that could be opened names Vorcaro or any vehicle of his as registered owner. A transcript of register for IMO 9954541, obtainable from International Registries on application, would settle the question.

^[8] FatoNews, 8 August 2026, reporting that “sem que os pagamentos pelo barco fossem feitos, Dovigi ativou cláusulas contratuais para reaver a embarcação” and describing a protection and reversal clause allowing the original seller to resume control of the building contract with Lürssen. Seu Dinheiro and BP Money both use “posse”, meaning possession, rather than “propriedade”, and both record that Vorcaro never went aboard. None of these outlets publishes the contract, so the account is single-stranded and untested.

^[9] Revista Oeste and Jornal de Brasília, September 2026, reporting that the liquidator of Banco Master, Eduardo Felix Bianchini, subpoenaed Lürssen on 15 June 2026 for contracts, agreements, letters of intent and records of money transfers, with a deadline of 2 July, in Chapter 15 proceedings in the United States. The Nixie appears on neither of the published asset-freeze lists in the Banco Master liquidation, the only vessel on those lists being a yacht moored in Florida valued at about R$80 million.

^[10] Superyacht Investor, “Cecil Wright sues Revolut founder Storonsky”, 7 August 2026, reporting that the email of 19 July 2025 from Storonsky’s team referred to “commercial arrangements you would expect from this” and used the word “commission”. No written brokerage agreement is described in any report of the filings, and the 5 per cent is the rate claimed rather than a rate any report traces to an agreed term.

^[11] Moran Yacht & Ship Inc v Pisarev [2016] EWCA Civ 317, 11 February 2016, dismissing the appeal from [2014] EWHC 1098 (Comm). caselaw.nationalarchives.gov.uk/ewca/civ/2016/317

^[12] Moran Yacht & Ship Inc v Pisarev [2014] EWHC 1098 (Comm) at [94] and [95], citing Nahum v Royal Holloway and Bedford New College [1999] EMLR 252. caselaw.nationalarchives.gov.uk/ewhc/comm/2014/1098

^[13] Berezovsky v Edmiston & Company Ltd [2010] EWHC 1883 (Comm) at [9], [17] and [18]. caselaw.nationalarchives.gov.uk/ewhc/comm/2010/1883

^[14] Ibid at [47].

^[15] Berezovsky v Edmiston & Company Ltd [2011] EWCA Civ 431, 19 April 2011, per Jackson LJ. caselaw.nationalarchives.gov.uk/ewca/civ/2011/431

^[16] Moran at [102], where the judge found that Moran was not the effective cause, or even an effective cause, of the sale of 4YOU.

^[17] Cecil Wright & Partners Limited, company number 08412005, incorporated 20 February 2013 (Companies House).

^[18] Berezovsky [2010] EWHC 1883 (Comm) at [9].

^[19] moranyachts.com, project page for NIXIE, which lists Moran Yacht & Ship as build manager and states that it “managed the project from contract through her subsequent sale”. The page does not identify which sale is meant.

On Friday 11 September 2026 District Judge Sam Goozée, sitting at Westminster Magistrates’ Court, extended by four months an account freezing order over £10,024,041.33 standing to the credit of the Premier League, together with accrued interest. He found nothing to suggest that the League had itself engaged in unlawful conduct. He continued the freeze regardless. The judgment repays reading by anyone whose business involves taking substantial sums from a counterparty operating at the edge of a licensing regime, because it shows how little the recipient’s innocence contributes once the statutory machinery engages.[1]

The money is licence income. On 1 November 2022 the League entered into a licence agreement with Sorare SAS, the Paris registered fantasy football platform, under which Sorare became its official fantasy football NFT licensee, paying royalties on net revenue from sales of licensed products together with annual minimum guarantees for each season. The League received £15m for the 2023/24 season and five instalments of £2.5m across 2024/25 and 2025/26.[2] By the date of the agreement the Gambling Commission had been examining Sorare for more than a year. It charged the company in July 2024 with three offences of providing facilities for gambling without an operating licence under the Gambling Act 2005, and the trial is listed at Birmingham Magistrates’ Court for June 2027.[3]

The order, and its ceiling

Account freezing orders occupy Chapter 3B of Part 5 of the Proceeds of Crime Act 2002, inserted by the Criminal Finances Act 2017. An enforcement officer may apply where he has reasonable grounds for suspecting that money held in an account is recoverable property or is intended by any person for use in unlawful conduct, and the magistrates’ court may make the order on the same test. Nothing in the scheme requires an allegation against the account holder. Suspicion attaches to the money. That is why Judge Goozée could find no suggestion of wrongdoing by the League while holding that “the freezing of accounts is necessary and proportionate to prevent dissipation of funds which are suspected to be the fruits of criminal conduct and enable proper investigation by the authorities”.[4]

The order has now been extended to the full statutory limit. It remains in force until 15 January 2027. It was made on 16 January 2025 for nine months, extended in July 2025 for a further ten, varied by consent in May 2026, and extended again by consent to 14 September 2026. Section 303Z3(4) provides that the period specified may not exceed two years starting with the day the order was made.[5] The extension granted on 11 September takes it to that ceiling and no further.

The League should take limited comfort from the arithmetic. Section 303Z3(3) preserves the operation of section 303Z14, and section 303Z14(6) provides that where an application for forfeiture is made the freezing order continues to have effect until the time referred to in subsection (7)(b) or (8), which is to say until the application is determined and, on a forfeiture order, until the money has been transferred.[6] An application made before the order lapses therefore carries the freeze beyond the two-year period. Forfeiture is decided on the balance of probabilities by the same magistrates’ court, and the judge has reserved any such application to himself for judicial continuity.[7] It does not wait upon the criminal trial of anyone. The order expires in January 2027. Sorare is not tried until June.

What section 308 has already done

Section 308(1) provides that recoverable property ceases to be recoverable where the person who obtains it on a disposal does so in good faith, for value and without notice that it was recoverable property. The provision is not merely a point reserved for any forfeiture hearing. It has already altered the order once, while the judge refused to allow the League to rely on it at the investigatory stage for the later receipts.

On 8 May 2026 the order was varied by consent to reduce the frozen sum from £16,273,678.43 to its present figure. The judge records that the National Crime Agency accepted the funds released were not recoverable property by operation of section 308, because up to 4 July 2024 the League had received them in good faith, for value and without notice.[8] 4 July 2024 is the day Sorare telephoned to say it had been charged. Everything now frozen is money received after that call.

For the receipts after it, the judge found the exception could not be sustained at this investigatory stage. He applied the Court of Appeal in R (World Uyghur Congress) v National Crime Agency, where Dove J, giving the judgment of a court including the Lady Chief Justice, held that payment of market value is not by itself enough and that “[a] person who suspected that it was would not be able to rely on s.308”.[9] What defeated the League was its own paperwork. On 6 December 2024 it filed a defence against money laundering disclosure recording that it “is concerned that it may hold criminal property” and that it “retains concerns that it could be in possession of criminal property”. Asked in cross-examination whether the disclosure set out the League’s suspicions, its Director of Legal answered “Yes it is”.[10] The judge found any attempt to distance the League from that suspicion implausible, and treated the arguments about actual and constructive notice as a layer of unnecessary complexity he did not need to enter.[11]

The steps taken to protect the League, ring-fencing the money and notifying the authorities, also evidenced the suspicion that prevented it from relying on section 308 at this stage. The judge noted that the League has not sued Sorare for a further £7.5m owed for over a year, which he said gave greater credence to its suspicions.[12]

The diligence the judge criticised

The notice analysis makes the earlier history worth setting out, though it goes to the judge’s criticism rather than to any finding that the League was on notice in 2022.

On 8 October 2021 the Gambling Commission published a consumer information notice stating that Sorare.com was not licensed by the Commission, that activity by consumers in Great Britain therefore fell outside the gambling regulations a licensed operator must comply with, and that it was carrying out enquiries into whether an operating licence was required.[13] The League met Sorare four days later and asked for a revised offer and a legal opinion. In November 2021 it received three legal notes obtained by Sorare, including one from Memery Crystal Limited, to the effect that no licence was required and that the Commission had erred in its reading of the Act. The summary on which the League relied stated in terms that it should not be relied upon by any third party.[14]

The League took no independent advice. Judge Goozée found that it had undertaken some due diligence, and then observed that the opinion “clearly would have been self-serving to Sorare”, that it appeared to have been accepted at face value, and that “in view of the millions of pounds which was at stake, and a public announcement by a gambling regulatory body that Sorare was not licenced ... one might question why they did not seek their own legal opinion to protect their interests”.[15]

Advice obtained by a counterparty is prepared for that counterparty’s purposes. Where it disclaims third-party reliance, it cannot substitute for advice addressed to the party carrying the commercial and regulatory risk.

The anterior question

Underlying all of this sits an issue no court has decided. Whether Sorare’s product constitutes gambling within the 2005 Act is the anterior question. Without unlawful conduct there can be no recoverable property, although tracing and section 308 would remain separate questions in any forfeiture application. The Commission’s position is that the game structure amounts to a betting prize competition under section 11(1) of the Act irrespective of the entry method, and it has prosecuted on that footing. Sorare denies it and says its game rewards skill. Judge Goozée expressly declined to decide the point, holding only that Sorare has been prosecuted by a regulatory body in circumstances where the evidential test has been met.[16] The trial is in June 2027. The order expires five months before it.

One point of precision deserves recording, because much of the coverage has obscured it. The frozen asset is not cryptocurrency. It is a sterling balance, quantified by the court to the penny, held in an ordinary bank account. Whatever Sorare’s cards may be, what the Agency has restrained is a licensee’s payment in conventional money.

An adjacent dispute in the United States

A related classification dispute is being fought in the United States. The Third and Ninth Circuits have divided over whether the Commodity Exchange Act prevents state gambling regulators from applying their laws to sports event contracts offered by a federally regulated exchange. New Jersey petitioned for certiorari on 2 September 2026 in Flaherty v KalshiEX LLC. The petition was docketed on 8 September as No. 26-299.[17]

The statutes and procedural questions are different. The common feature is the attempt to fit a digital product within competing legal categories. Kalshi characterises its contracts as federally regulated derivatives. State regulators characterise them as sports wagers. Sorare says its game rewards skill. The Gambling Commission says its structure amounts to a betting prize competition.

The label chosen by the operator does not decide the legal classification. Until the courts do, counterparties accepting money generated by these products carry a risk that extends beyond the contract itself.

Notes

1. National Crime Agency v The Football Association Premier League Limited, District Judge (Magistrates’ Courts) Sam Goozée, Westminster Magistrates’ Court, 11 September 2026, at [1], [71], [83] and [85]. Available at https://www.judiciary.uk/wp-content/uploads/2026/09/NCA-v.-Football-Accn-Premier-League-Ltd-11.09.2026.pdf.

2. Judgment at [9], [13], [14] and [15].

3. Judgment at [18], [22] and [27]. The three charges relate to periods between February 2022 and April 2024. Sorare pleaded not guilty at Birmingham Magistrates’ Court on 4 October 2024.

4. Judgment at [83]. Sections 303Z1 and 303Z3, Proceeds of Crime Act 2002.

5. Section 303Z3(4), Proceeds of Crime Act 2002. Procedural history at judgment [1].

6. Sections 303Z3(3) and 303Z14(6) to (8), Proceeds of Crime Act 2002.

7. Judgment at [86].

8. Judgment at [1], [29] and [70].

9. R (World Uyghur Congress) v National Crime Agency [2024] EWCA Civ 715 at [46], cited at judgment [64].

10. Judgment at [35], [72], [73] and [74].

11. Judgment at [76] and [77].

12. Judgment at [84].

13. Gambling Commission, consumer information notice, Sorare.com, 8 October 2021. Judgment at [9].

14. Judgment at [10] and [42].

15. Judgment at [71].

16. Judgment at [12] and [69].

17. Flaherty v KalshiEX LLC, No. 26-299 (US), petition for a writ of certiorari filed 2 September 2026, docketed 8 September 2026. KalshiEX LLC v Flaherty (3d Cir, April 2026). KalshiEX LLC v Assad, No. 25-7516 (9th Cir, 28 August 2026).

Complaints unsealed in Manhattan federal court on 15 September charge two Robinhood engineers with commodities fraud and wire fraud over perpetual futures traded on Hyperliquid ahead of Robinhood Crypto’s token listing announcements.¹

Prosecutors did not have to establish that the underlying tokens were securities. Count One pleads that each defendant, “knowing that he had obtained material nonpublic information in breach of a duty, used that information to trade perpetual futures on a decentralized derivatives exchange for personal gain”.² That resembles the misappropriation theory approved in O’Hagan, placing the alleged deception in the employee’s misuse of entrusted information rather than in the classification of the underlying token. The same facts produce a different result in Europe and England.

The American route avoids token classification

Two features of the alleged conduct matter. The defendants are said to have traded derivatives referencing the tokens without acquiring the tokens themselves, and to have done so on a protocol that the complaints plead holds no CFTC approval to operate a futures contracts market and geofences United States IP addresses.³

The complaints plead both features without first classifying the underlying tokens as securities. Count One is brought under section 6(c)(1) of the Commodity Exchange Act and Rule 180.1, and tracks the statutory language reaching conduct “in connection with a swap, a contract of sale of a commodity in interstate commerce, and for future delivery on and subject to the rules of a registered entity”.⁴ Count Two avoids the classification question differently, by requiring a scheme to obtain money or property by fraud and the use of interstate or foreign wires. The complaints allege both, identifying the listing information as valuable confidential business information and the trading profits as money obtained from counterparties.⁵

The Office made the same move in 2022. The indictment against Ishan Wahi over Coinbase listings charged two counts of wire fraud conspiracy and two of wire fraud, and pleaded no securities count, while the Commission alleged in parallel civil proceedings that several of the tokens were securities.⁶ The classification question was not adjudicated in the criminal case, which did not require it to be.

Where Europe actually draws the line

The intuitive European answer is that a decentralised venue defeats the regime. It does not.

Title VI of the Markets in Crypto-Assets Regulation, Articles 86 to 92, has prohibited insider dealing, unlawful disclosure and market manipulation in crypto-assets since 30 December 2024, with Article 89 carrying the insider dealing prohibition.⁷ Its scope is fixed by the asset, not the venue. Where a crypto-asset is admitted to trading or admission has been requested, the prohibitions bite on conduct wherever it occurs, including dealing on a fully decentralised exchange running as a smart contract.⁸ Hyperliquid would not save anyone.

The obstacle is the classification of the instrument traded. Article 2(4)(a) excludes financial instruments from MiCA, and ESMA’s guidelines indicate that a derivative referencing a crypto-asset will fall within MiFID II where it satisfies the relevant derivative criteria.⁹ A cash-settled perpetual of the kind alleged here is therefore likely to be treated as a financial instrument rather than a crypto-asset governed by MiCA. The underlying token may remain a crypto-asset. The perpetual itself routes instead to MAR, which ordinarily requires the financial instrument to be admitted to trading, traded on an MTF or OTF, or otherwise connected through Article 2(1)(d) to an instrument within that perimeter.¹⁰ On the facts presently alleged, there is no identified Union venue or covered instrument supplying that connection.

The resulting gap begins with the classification of the perpetual and continues through MAR’s trading perimeter. It is also a live European argument, since derivatives sit outside MiCA while carrying most of the trading volume, which licensed exchanges have begun calling reverse discrimination.¹¹

England is narrower again. The criminal insider dealing offence in Part V of the Criminal Justice Act 1993 applies only to securities within Schedule 2 and is subject to its own market and territorial conditions.¹² A perpetual on a token traded through an offshore decentralised protocol is not an obvious fit. Section 4 of the Fraud Act 2006 provides a possible alternative, where an employee dishonestly abuses a position in which he is expected to safeguard, or not act against, his employer’s financial interests, intending thereby to make a gain or cause a loss.¹³ That resembles O’Hagan in locating the alleged wrong in an abuse of position, but its application to these facts would still have to satisfy the English offence’s full elements and territorial rules.

The employer documents the duty

A duty-based theory needs proof of the duty and the defendant’s knowledge of it. In both American cases, the employers’ internal arrangements supplied unusually clear evidence. The 2022 indictment records that Ishan Wahi belonged to a private Coinbase messaging channel reserved for employees involved in asset listings.⁶ The Robinhood complaints plead that listing information was restricted to “Coin Aware Individuals” with access to a confidential Slack channel. Robinhood’s policy also prohibited those employees from trading on any platform before a listing announcement and during the following 24 hours.¹⁴

The compliance point is practical. A dealing policy confined to particular instrument categories or to the firm’s own venue may leave avoidable uncertainty over the scope of the employee’s obligation. The policy does not create criminal liability or replace the statute. It can, however, provide powerful evidence that the information was confidential, that the employee knew it, and that off-platform trading breached an identified duty.

None of this turned on the Clarity Act cloture motion failing the same afternoon. Legislation dividing digital assets between the Commission and the Commodity Futures Trading Commission would establish a clearer framework for registration, disclosure and civil liability. It would not displace the theory pleaded in these complaints. The complaints are allegations, both defendants are presumed innocent, and the theory has not yet been tested on a motion to dismiss, at trial or on appeal.

Notes

¹ United States v Chai, 26 Mag. 3716, and United States v Xiang, 26 Mag. 3718 (S.D.N.Y.), sealed complaints sworn before Willis USMJ and unsealed 15 September 2026. Both plead violations of 7 U.S.C. §§ 9(1) and 13(a)(5), 17 C.F.R. § 180.1 and 18 U.S.C. §§ 1343 and 2. Hefu Chai, 36, of Menlo Park, California; Huaisong Xiang, also known as “Jerry Xiang”, 30, of Jersey City, New Jersey. See also USAO SDNY press release 26-260.

² Chai complaint, Count One; Xiang complaint, Count One. See United States v O’Hagan, 521 U.S. 642 (1997).

³ Chai complaint ¶ 5(a); Xiang complaint ¶ 5(a). The complaints add that the geofencing may be bypassed using a virtual private network.

⁴ Chai complaint, Count One; 7 U.S.C. § 9(1); 17 C.F.R. § 180.1.

⁵ Chai complaint, Count Two; Xiang complaint, Count Two; 18 U.S.C. §§ 1343 and 2.

⁶ USAO SDNY press release 22-232, 21 July 2022, announcing the unsealing of an indictment. Ishan Wahi was sentenced by Preska J to 24 months in May 2023.

⁷ Regulation (EU) 2023/1114, Title VI, Articles 86 to 92.

⁸ This reading of Title VI scope follows the analysis in Crypto-Asset Market Abuse Under EU MiCA, European Journal of Risk Regulation. It is interpretative rather than settled.

⁹ Regulation (EU) 2023/1114, Article 2(4)(a); ESMA Final Report on the conditions and criteria for the qualification of crypto-assets as financial instruments, ESMA75-453128700-1323. The guidelines apply the MiFID II criteria and do not establish that every product labelled a perpetual future qualifies.

¹⁰ Regulation (EU) No 596/2014, Articles 2(1) and 14.

¹¹ The volume estimate and the reverse discrimination characterisation come from industry commentary on the derivatives exclusion, not from an official source.

¹² Criminal Justice Act 1993, Part V and Schedule 2, as amended by the Insider Dealing (Securities and Regulated Markets) Order 2023, SI 2023/582.

¹³ Fraud Act 2006, s 4.

¹⁴ Chai complaint ¶¶ 4(e) to (h); Xiang complaint ¶¶ 4(e) to (h).

An Illinois court has been asked to stop the state’s first-in-the-nation digital asset tax before it takes effect on 1 January 2027. The twenty-six-page memorandum of law is dated 8 September. It rests principally on a statute written in 1998 for a commercial internet that did not yet have assets of its own.[1]

Two suits, one statute

The Digital Asset Tax Act reached the statute book by an unusual route. Senate Bill 3019 was introduced in January 2026 as a two-page measure on agricultural property loans. On the final day of the 2026 session two amendments replaced everything after the enacting clause and produced a 1,624-page bill covering gambling licences, vehicle weight restrictions and much else. Twenty of those pages were the Act. The bill cleared committee and both chambers in under twenty-four hours on one hour’s public notice. The Governor signed it on 16 June 2026.[2] House Bill 5798, introduced six days later, would repeal the Act outright.

Two challenges are now before the same court. The Digital Chamber filed a verified complaint on 21 July advancing six counts, including the Illinois Uniformity Clause, the proportionate penalties clause, the dormant Commerce Clause and preemption under the Internet Tax Freedom Act.[3] The Blockchain Association and the Crypto Council for Innovation filed their action on 22 August and have now moved for interim relief, represented by Gibson Dunn with Springfield local counsel.

What the preemption test actually requires

Section 1105(2)(A)(i) of the Internet Tax Freedom Act defines a discriminatory tax as one imposed on electronic commerce that is not generally imposed on transactions involving similar property, goods, services or information accomplished through other means.

The operative word is similar. In Performance Marketing Association v Hamer the comparison was straightforward. Illinois had imposed a collection obligation on retailers using online performance marketing while leaving offline marketing untouched. The Illinois Supreme Court held the statute void and unenforceable.[4] The Maryland Tax Court reached the same conclusion about digital advertising on 14 August, reasoning that Congress could not have been clearer that internet services should not be taxed unless similar services in the broader sense were taxed as well.[5]

Neither statute presented the difficulty this one does. Advertising and marketing exist off the internet, so the comparison holds. A native digital asset has no off-chain version of itself. The plaintiffs therefore pitch similarity at the level of function: Illinois exempts bullion and currency from sales and use tax and does not tax the exchange or storage of stocks and bonds, so an asset used as a medium of exchange, unit of account or store of value ought to be treated alike whatever the ledger.[6] Should the court instead require a physical analogue, the protection of a technology-neutral tax policy would be unavailable precisely to the technology that has no analogue.

Apportionment and the full-value measure

The apportionment challenge turns on internal consistency. Under the second limb of Complete Auto a state tax must be internally consistent, meaning that identical adoption of the same rule by every state would not place interstate activity at a disadvantage.[7]

Section 3-25 presumes a customer to be in Illinois on the strength of a home address, a mailing address, an Illinois IP address or other place-of-primary-use data. The burden of proving the negative falls on the broker. A resident who trades from another state and a visitor who trades while passing through are both presumptively Illinois customers. The plaintiffs argue that if every state adopted the same rule, competing address and IP presumptions could subject one transaction to more than one full-value charge. That, they say, is the internal-consistency defect identified in Wynne.

The stablecoin question

The Act borrows its definition of digital asset from the Digital Assets and Consumer Protection Act, which excludes digital representations of value having substantial value or utility beyond mere existence as a digital asset. A payment stablecoin, whose reserves must meet the federal one-for-one requirement, is on the plaintiffs’ case a strong candidate for that exclusion. The Department of Financial and Professional Regulation nonetheless proposed a rule in May treating payment stablecoins as digital assets.[8]

If that construction is maintained, transferring a dollar-backed token would be taxed at 0.2% of its value while transferring a dollar would not. That is the cleanest comparator the challenge is likely to get. From 1 January a digital asset broker that violates any provision of the Act commits a Class 3 felony under section 3-55. The collection obligation in section 3-35(a) reaches brokers maintaining a place of business in Illinois, a definition that captures $100,000 of Illinois gross receipts over the preceding twelve months. The registration, remittance and recordkeeping provisions are not limited consistently by that threshold, which is one of the vagueness complaints. Those provisions take effect on 1 January unless the court grants relief or the legislature repeals the Act.

Notes to the text

1. Blockchain Association and Crypto Council for Innovation v Harris, Raoul and Milhiser, No. 2026 MR 000312 (Cir. Ct. 7th Jud. Cir., Sangamon County), memorandum of law in support of plaintiffs’ motion for preliminary injunction, dated 8 September 2026. The Crypto Council for Innovation announced the motion on 9 September 2026.

2. SB 3019, 104th Gen. Assemb., enacted as Public Act 104-0468. The Digital Asset Tax Act is codified at 35 ILCS 195/3-5 et seq.

3. Chamber of Digital Commerce v Harris, Circuit Court of Sangamon County, verified complaint filed 21 July 2026.

4. Performance Marketing Association, Inc. v Hamer, 2013 IL 114496 (18 October 2013).

5. Google LLC v Comptroller of Maryland, No. 23-DA-OO-0649 (Md. Tax Ct., 14 August 2026), decided with Apple Inc. v Comptroller and Peacock TV, LLC v Comptroller.

6. 35 ILCS 105/3-5(8), 35 ILCS 120/2-5(18) and 86 Ill. Admin. Code 130.120(a).

7. Complete Auto Transit, Inc. v Brady, 430 U.S. 274 (1977), Comptroller of the Treasury of Maryland v Wynne, 575 U.S. 542 (2015) and Oklahoma Tax Commission v Jefferson Lines, Inc., 514 U.S. 175 (1995).

8. 205 ILCS 731/1-5(a). See also Illinois Department of Financial and Professional Regulation, proposed rule, 50 Ill. Reg. 6746, 6752-55 (15 May 2026).

Judge Rakoff ruled in February that a criminal defendant’s exchanges with a consumer AI chatbot were protected by neither the attorney-client privilege nor the work product doctrine.¹ Four months later two state courts reached different work-product results under different procedural rules. A stayed order in Connecticut raises a further question that none of them addresses.

What Heppner decided

Bradley Heppner, indicted in the Southern District of New York on securities and wire fraud counts arising from his conduct at GWG Holdings, used the consumer version of Anthropic’s Claude after receiving a grand jury subpoena and retaining counsel. He put information learned from his attorneys into the tool, generated reports on possible defences, then gave those reports to his lawyers. The court held that the privilege did not attach, the provider’s terms having permitted disclosure of user data and the use of prompts and outputs for training. It treated the reports as Heppner’s own research rather than as material prepared at counsel’s direction, which disposed of the work product claim.²

The more tentative observation concerned what might have followed had counsel instructed the use of the tool. On that footing, the court suggested, the tool might have functioned as a lawyer’s agent in the manner contemplated by United States v. Kovel.³ Kovel extended the privilege to an accountant engaged by counsel to make the client’s information intelligible to the lawyer. Whether a model can occupy that role is untested. For in-house teams the practical consequence is that the instruction trail may do more work than the procurement decision.

The June divergence

The two state decisions rest on different rules. Tate Group turned on the breadth of the Texas provision, which reaches material prepared and mental impressions developed in anticipation of litigation by or for a party or a party’s representatives. That language covers a party principal who is neither a lawyer nor self-represented. Assini applied a narrower New York standard, since CPLR 3101(d) attaches only where the material was prepared solely in anticipation of litigation.⁴ Justice Fischer quashed both subpoenas served on OpenAI OpCo, LLC in their entirety. She then referred the litigant to 22 NYCRR Part 161 with a sanctions warning, so the protection came with a supervisory direction attached.

Where the English position differs

An English practitioner should not assume that the American answers transfer. Three Rivers (No 5) defines the corporate client narrowly, but Aabar Holdings v Glencore held that legal advice privilege may also cover documents created or circulated within the client group for the dominant purpose of seeking legal advice.⁵ Picken J gave preparatory notes for a meeting with counsel as an example. A standalone AI chat therefore does not fail merely because no lawyer participated in it. Whether it is protected will depend on who created it, its purpose, its intended use and whether it remained confidential. Material created by an employee outside the client group remains more difficult. Litigation privilege is the alternative route, requiring after SFO v Eurasian Natural Resources Corporation that litigation be in reasonable contemplation and that the dominant purpose of the document be conducting it.⁶

Expert use raises a separate question. In Conservation Law Foundation, Inc. v. Shell Oil Co., Magistrate Judge Thomas O. Farrish ordered production of the GPT-4o prompts used to cull the defendants’ document production for an expert report, holding that the culling process formed part of methodology rather than protected notes or drafts.⁷ The order was stayed pending the district judge’s ruling on a Rule 72(a) objection. It therefore remains subject to review. The order should therefore be treated cautiously. English procedure approaches the point differently. CPR 35.10 requires the report to state the substance of all material instructions. Those instructions are not privileged under CPR 35.10(4), though the court will not order disclosure or permit questioning unless there are reasonable grounds to consider the statement inaccurate or incomplete. Undisclosed use of a model to select the documents considered by the expert may provide grounds for arguing that the statement of material instructions is incomplete. Whether the prompts themselves would then be ordered disclosed has not been decided in England.

What to do about it

Segregate litigation work from general enquiries and record any instruction to use the tool. Where an expert uses AI, preserve the prompts and outputs, establish what role they played and ensure that the report accurately describes the material methodology. The present authorities are too fact-specific to justify assuming either protection or disclosure.

¹ United States v. Heppner, No. 1:25-cr-00503 (S.D.N.Y.), bench ruling 10 February 2026, written opinion 17 February 2026, 2026 WL 436479, 2026 U.S. Dist. LEXIS 32697.

² The two grounds are summarised in the same terms in the contemporaneous firm reporting of the bench ruling and of the written opinion.

³ United States v. Kovel, 296 F.2d 918 (2d Cir. 1961).

⁴ Assini v Hayward, 2026 NY Slip Op 26086, Index No. 607683/2024 (Sup Ct, Nassau County, 4 June 2026) (Fischer, J.). The “solely in anticipation of litigation” requirement is taken from Agovino v. Taco Bell 5083, 225 AD2d 569 (2d Dept 1996). The substantive reasoning adopted is that of Morgan v. V2X, Inc., 2026 WL 864223, 2026 U.S. Dist. LEXIS 67939 (D. Colo. 30 March 2026), which framed the question as whether disclosure substantially increased the likelihood that an adversary would obtain the material. Warner v. Gilbarco, Inc., No. 2:24-cv-12333 (E.D. Mich. 10 February 2026), ECF No. 94, 2026 WL 373043 (Patti, M.J.), is to the same effect on waiver.

⁵ Aabar Holdings S.A.R.L. and others v Glencore plc, Ivan Glasenberg and Steven Frank Kalmin [2026] EWHC 877 (Comm), Case Nos FL-2022-000024 and others, Commercial Court (KBD), Financial List, 16 April 2026, per Picken J. The holding is at paragraph 16. Three Rivers District Council v Governor and Company of the Bank of England (No 5) [2003] EWCA Civ 474, [2003] QB 1556, was distinguished as concerned with material outside the client group. Support was drawn from R (Jet2.com Ltd) v Civil Aviation Authority [2020] EWCA Civ 35. The decision is one of first instance in the Commercial Court.

⁶ Director of the Serious Fraud Office v Eurasian Natural Resources Corporation Ltd [2018] EWCA Civ 2006.

⁷ Conservation Law Foundation, Inc. v. Shell Oil Co., No. 3:21-cv-00933 (D. Conn.), ECF No. 970 (18 May 2026). The order was stayed by Judge Vernon D. Oliver at ECF No. 977 on 3 June 2026 pending consideration of the objection at ECF No. 975 and the stay motion at ECF No. 974. Briefing closed on 29 June 2026 at ECF No. 986. The docket was checked on 10 September 2026, at which date the objection remained undecided.

The Seattle Times and Newsday sued OpenAI and Microsoft in the Southern District of New York on 4 September.¹ Seven counts, of which three are trademark dilution. Those three make the pleading worth reading closely because a ruling on Copyright Act fair use would not decide them.

What the rulings so far have decided

The action arrives against a body of rulings from Judge Sidney Stein in the consolidated Times, Daily News and Center for Investigative Reporting actions. On 4 April 2025 he gave judgment on the motions to dismiss in all three.² He dismissed the section 1202(b)(3) claims against every defendant in all three actions. He dismissed the section 1202(b)(1) claims against Microsoft in all three. Each was dismissed without prejudice. Section 1202(b)(1) claims against OpenAI survived in the Daily News and CIR actions but not in the Times action. Section 1202(b) carries what the court described as a double scienter requirement. Those pleadings failed on the second limb, knowledge that the conduct would induce, enable, facilitate or conceal infringement.

On the trademark side he went the other way. He denied the motions to dismiss the federal and New York state dilution claims in the Daily News action. The federal claim survived because the diluted marks were plausibly alleged to be famous. The New York claim survived a different attack. Microsoft’s challenge to section 360-l was brought under the dormant Commerce Clause rather than on the merits of dilution. At section VIII.B of the opinion the court held that the New York statute does not offend the Clause.

Counts III and IV of the new complaint are pleaded against the defendants without limitation. Counts V, VI and VII are dilution counts. The DMCA counts repeat theories that the April opinion rejected. The dilution counts follow theories it allowed to proceed. As at the 4 September docket update, no judge was shown and the action had not been tagged into the multidistrict litigation. Whether these issues will come before Judge Stein remains to be seen.

The untested exclusions

Section 1125(c)(3) excludes three categories of conduct from federal dilution liability. The first is fair use, other than as a designation of source for the defendant’s own goods or services, including nominative or descriptive use and certain parody, criticism or commentary. The other two are all forms of news reporting and news commentary, together with any noncommercial use of a mark. None of the three is addressed in the complaint, which is ordinary at the pleading stage but leaves the whole of Count V contingent on how a court reads statutory language drafted with a human speaker in mind.

The federal count presents a different threshold question. Section 1125(c)(1) concerns a defendant’s use of a mark or trade name in commerce that is likely to cause dilution. Here the alleged use is a masthead appearing in generated output rather than a mark adopted to identify the defendants’ own services. Whether that output satisfies the statutory requirements remains unresolved in these cases, as does whether the reporting or noncommercial exclusions apply.

The defendants could invoke the news-reporting exclusion where an output uses a masthead while summarising a news event. Harder is whether the statutory description attaches to the particular output or whether the court should examine the commercial service through which the output was produced. On that sequencing question the text is silent. In Jack Daniel’s Properties, Inc. v. VIP Products LLC the Supreme Court held that the noncommercial use exclusion did not shield parody or commentary where the challenged mark was being used to designate the source of the defendant’s own goods.³ The Court described its decision as narrow and did not determine the wider reach of the noncommercial use exclusion.

Paragraph 120 pleads dilution by tarnishment under section 1125(c)(2)(C) on the footing that the products hallucinate content and misattribute it to the plaintiffs. Tarnishment doctrine developed around unsavoury association, typically a famous mark placed alongside sexual or illicit material. Applied here it becomes a claim about model error, which is a claim about a defect in the product rather than about what the defendant chose to associate with the mark.

The state counts and why they were pleaded separately

The state counts offer alternative routes around the federal fame threshold.

The Washington count avoids the federal nationwide fame standard, but it does not avoid every federal difficulty. RCW 19.77.160 contains its own exclusions for specified fair use, noncommercial use and all forms of reporting and news commentary.

The complaint pleads the RCW 19.77.160 factors in terms at paragraphs 126 to 129. Continuous and prominent use since at least 1896, an audience of over one million readers monthly, eleven Pulitzer Prizes and registration on the Principal Register. That is a regional publisher pleading fame within Washington, an alternative unavailable under the federal nationwide fame standard. Section 360-l expressly provides injunctive relief. Its value here is the distinctiveness threshold it applies rather than a federal requirement of fame among the general consuming public.

The sequencing problem

If the action is transferred into MDL 3143 and stayed with the other newly tagged cases, none of this is likely to be decided soon. Judge Stein has issued a run of orders in that multidistrict litigation requiring newly tagged plaintiffs to show cause why their actions should not be stayed pending resolution of summary judgment in the other active cases. The pattern is visible at ECF 1556 on 18 May 2026, ECF 1599 on 29 June 2026 and ECF 1679 entered on 31 August 2026.⁴ The Department of Justice filed a statement of interest in the same multidistrict litigation on 1 September supporting the position that training is fair use under the Copyright Act.

The stays are being ordered by reference to summary judgment on infringement. Section 107 is a defence to copyright infringement. It does not decide claims under section 1125(c), RCW 19.77.160 or section 360-l. The federal and Washington provisions contain their own express exclusions. The New York provision does not. A ruling that training on news content is fair use under section 107 would leave Counts V, VI and VII intact. A ruling the other way would not decide them either, since neither outcome speaks to whether a masthead reproduced in generated output is a use in commerce.

For practitioners advising publishers, the drafting point is clear enough. Copyright and dilution claims do not rise or fall on the same legal question. If this action is transferred into MDL 3143 and stayed, the trademark counts may wait behind the copyright bellwethers even though a ruling on section 107 would not decide them. The state counts also avoid the federal nationwide fame threshold, although the Washington statute contains parallel exclusions and the New York provision is directed principally to injunctive relief.

¹ Complaint, The Seattle Times Company and Newsday LLC v. OpenAI, Inc., OpenAI GP, LLC, OpenAI, LLC, OpenAI OpCo, LLC, OpenAI Global LLC, OAI Corporation, OpenAI Holdings, LLC, OpenAI Foundation, OpenAI Group PBC and Microsoft Corporation, No. 1:26-cv-07644 (S.D.N.Y., filed 4 September 2026), ECF No. 1. Thirty-eight pages. Counsel Klaris Law PLLC. Jury demanded.

² Opinion, The New York Times Company v. Microsoft Corporation, No. 1:23-cv-11195-SHS-OTW (S.D.N.Y.), ECF No. 514 (signed 4 April 2025), filed in associated cases 1:24-cv-03285-SHS-OTW and 1:24-cv-04872-SHS-OTW. The famousness holding is at section VII.B. The New York ruling is at section VIII.B and rejects a dormant Commerce Clause challenge.

³ Jack Daniel’s Properties, Inc. v. VIP Products LLC, 599 U.S. 140 (2023).

⁴ In re: OpenAI, Inc., Copyright Infringement Litigation, No. 1:25-md-03143-SHS-OTW (S.D.N.Y.). ECF No. 1679 was signed on 28 August 2026 and entered on 31 August 2026, with a show cause response due 11 September 2026, in associated case 1:26-cv-07171-SHS.

Sony Music Publishing and Warner Chappell have sued Anthropic PBC and named Dario Amodei and Benjamin Mann as individual defendants.¹ The corporate defendant is the obvious target. The naming of two individuals is the development that will occupy general counsel this week.

A pleading that travelled

Paragraph 79 of the complaint cites Benjamin Mann’s answer in the second Concord action.² In that answer Mann pleads that he discussed acquiring LibGen data with Amodei and other Anthropic colleagues, and that “Dr. Amodei approved”. The complaint also alleges that Anthropic’s internal policy required the chief executive’s express approval before a new dataset could be obtained or used in model development.

An answer of that kind is ordinarily protective. It places the employee’s conduct inside a corporate authorisation structure and invites the court to treat the company as the actor. Read across into a fresh complaint, the same paragraph identifies the person who authorised the act. Sony and Warner Chappell have taken an internal approval requirement, which existed to control the company’s data practice, and used it as the evidential route to the officer who operated it.

The complaint pairs that with material unsealed in Bartz: Mann’s description of LibGen as “sketchy AF”, and Anthropic’s own Archive Team recording that the site was a “blatant violation of copyright”.³ The knowledge element for contributory liability is being assembled from the defendants’ own documents.

The sequence

The English comparison

An English claimant would face a different structure and, on these facts, might find it easier rather than harder.

Accessory liability against a director tightened considerably in Lifestyle Equities CV v Ahmed.⁴ The Supreme Court held that a director who procures a company’s tort is liable as an accessory only if he knows the essential facts that make the act wrongful, and that this holds even where the underlying tort is one of strict liability. Trade mark infringement did not become a route to personal liability simply because it required no knowledge as against the company.

Copyright is differently drafted. Section 16(2) of the Copyright, Designs and Patents Act 1988 provides that copyright is infringed by a person who, without the licence of the copyright owner, does or authorises another to do any of the acts restricted by the copyright. Authorisation is infringement in its own terms rather than accessory liability grafted onto the company’s wrong, so the Lifestyle Equities knowledge requirement does not govern it. The constraint comes instead from CBS Songs Ltd v Amstrad Consumer Electronics plc, where the House of Lords held that to authorise means to grant or purport to grant, expressly or by implication, the right to do the act complained of.⁵

On the facts pleaded, an approval given under a policy that required it before the dataset could be acquired sits close to the centre of that definition. Amstrad escaped because it sold equipment and had no control over what purchasers chose to record. A chief executive whose required approval preceded the download has no comparable distance from the act.

What this changes for the developer

Personal exposure therefore begins to affect the case before liability is determined.

Naming the founders immediately raises questions about indemnification, defence costs and the scope of any directors’ insurance. Much will depend on the policy wording, particularly whether a conduct exclusion operates only after a final adjudication. Their position may therefore complicate a settlement negotiated principally for the company.

The second consequence is evidential discipline. Anthropic is defending several publisher actions arising from the same acquisition conduct. A pleading served in one is available in the next, and an answer drafted to establish corporate authority is a document that names the officer who gave it.

The Bartz litigation showed that the provenance of a training corpus, quite apart from the fact of training, can create ten-figure exposure. This complaint seeks to attach that exposure to named individuals.

Anthropic has said it disagrees with the claims and intends to defend itself robustly. No defendant has yet responded to the complaint.

Notes

1. Sony Music Publishing (US) LLC v. Anthropic PBC, No. 5:26-cv-09217 (N.D. Cal., commenced 28 August 2026). Thirty-five plaintiff entities. Count I is pleaded against all three defendants, Count II against Amodei and Mann, and Counts III and IV against Anthropic alone.

2. Concord Music Group, Inc. v. Anthropic PBC, No. 5:26-cv-00880-EKL (N.D. Cal.), ECF No. 106 (Answer to Amended Complaint by Benjamin Mann), at ¶ 71, cited in the complaint at ¶ 79.

3. Bartz v. Anthropic PBC, 791 F. Supp. 3d 1038, 1064 (N.D. Cal. 2025). The internal material is cited in the complaint at ¶¶ 78 and 79 to the Bartz docket, ECF No. 325 at 8 and ECF No. 337-1 at 144:4-13 and 396:3-13.

4. Lifestyle Equities CV v Ahmed [2024] UKSC 17.

5. CBS Songs Ltd v Amstrad Consumer Electronics plc [1988] AC 1013 (HL).

On 19 August 2026, the Commodity Futures Trading Commission issued a Request for Comment on the Listing of Compute Derivatives Contracts. It appeared two days later at 91 FR 54259 under RIN 3038-AF77. Comments close on 20 October. The voting record looks odd only until one checks the Commission’s composition. Chairman Michael S. Selig voted for the notice and nobody voted against it. Selig is the CFTC’s sole sitting Commissioner.[1]

The following afternoon, at the first meeting of the Innovation Advisory Committee, Selig put the request into a broader policy programme headed “Winning the AI Race: Roadmap for Compute Market Dominance”. The CFTC is working with the Department of Commerce. Its stated reference point is the White House AI Action Plan of July 2025. Its intended destination is what Selig called a “gold standard regulatory framework” for the commodity that will power the intelligence economy.[2]

The notice is not a proposed rule and commits the Commission to no particular result. Its significance is more immediate. Exchanges have already designed the products. The CFTC is asking whether the markets and benchmarks beneath them are sufficiently visible to support lawful listing.

What is being traded

Not chips. AI labs and enterprises rent processing capacity from cloud providers and data-centre operators, commonly priced by the GPU-hour. The contracts discussed here reference the price of access, not ownership of the hardware.

The distinction is important. Access is time-limited. An hour on one machine is not interchangeable with an hour on another once provider, region, interconnect, network and term are taken into account. Capacity cannot be stored. Electricity is a closer analogy than oil, although neither is exact. Supply is concentrated and much of the price formation takes place in private negotiation.

The statutory category

Compute fits within the Commodity Exchange Act’s definition of commodity. Section 1a(9) extends beyond goods and articles to “all services, rights, and interests ... in which contracts for future delivery are presently or in the future dealt in”.[3] Rental access to processing capacity falls naturally within the language of services, rights and interests.

Selig traced that breadth to Congress’s decision to create a federal regime capable of accommodating new underlyings. He quoted the Senate Agriculture Committee’s observation that “[t]he nature of the underlying commodity is not an adequate basis to divide regulatory authority” and cited the description of an “essentially unbounded field of potential commodities”.[4]

The harder question is not whether compute can be a commodity. It is whether a particular contract and its settlement benchmark satisfy the conditions for listing.

Five venues, two instruments

Kalshi launched live compute forward curves on 14 July 2026 for the Nvidia B200, H200 and A100, derived from prices in its weekly and monthly event contracts. CME Group and Silicon Data plan to launch H100 and B200 Rental Index Futures on NYMEX on 5 October 2026, subject to completion of the regulatory review periods. ICE and Ornn announced US-dollar, cash-settled GPU compute futures on the Ornn Compute Price Index on 19 May, subject to regulatory approval. PMEX Markets and PMEX Clearing, associated with Liquid Compute, remain pending applicants for designation as a contract market and clearing organisation.[5]

Polymarket US chose another route. On 20 August it self-certified a specific Compute Index Value Contract under 17 CFR § 40.2(a), rather than using a broad template. The contract asks whether the Ornn Data GPU Price Index for the H100 SXM will be at least $2.50 per GPU-hour on 15 March 2027. A class certification under § 40.2(d) followed on the same date. The exchange certified the class as swaps based on an excluded commodity and stated that listing could begin on 24 August.[6]

The economic exposure is related, but the legal forms are not. CME and ICE propose futures. Polymarket US certified swaps. Instrument characterisation affects the statutory route and the applicable listing analysis. Filing strategy carries its own consequences. Self-certification is not Commission approval.

Eight days later, the Ninth Circuit decided KalshiEX, LLC v Assad. It affirmed the dissolution of a preliminary injunction against Nevada’s enforcement of state gaming law as it applied to Kalshi’s sports contracts. On the court’s reading, at that preliminary stage, the sports contracts were likely not swaps under the CEA. A broad reading of the swap definition was textually inferior and lacked a limiting principle. The court declined to treat the CEA as having handed the Commission authority over gaming nationwide absent clearer language.[7]

Assad does not decide the status of compute contracts. A GPU price index has an obvious financial and commercial connection that a sporting result may lack. The decision is nevertheless important because it confirms that a court may examine the substance of the product for itself. Certification does not settle characterisation.

The Polymarket class also extends beyond capacity. It includes an Ornn Token Price Index, expressed in dollars per million tokens, for Anthropic, OpenAI, Google and DeepSeek. That is a price for inference rather than GPU rental. It deserves separate treatment.

The surveillance problem

The CFTC’s concern is stated plainly. Compute pricing occurs through “opaque bilateral transactions”, limiting current and historical price data. The notice asks whether the market has sufficient fungibility, standardisation and liquidity. It also raises the pricing power of dominant participants and the resulting risk of manipulation.[8]

The most difficult question is whether trading should be allowed when settlement depends on data the Commission “may not be able to observe, verify, or surveil, in whole or in part”.[8]

A cash-settled contract pays against a benchmark at expiry. Surveillance normally permits the regulator to compare the derivative with the underlying market and to test activity in each against the other. If the cash-market inputs cannot be seen or checked, part of that comparison disappears. Exchange diligence cannot by itself restore data that neither exchange nor regulator can obtain.

Core Principle 3

Core Principle 3 supplies the legal constraint. A designated contract market may list only contracts that are not readily susceptible to manipulation.[9]

In a concentrated market, a provider may be able to alter posted rates, redirect capacity or decide whether to transact during a settlement window. Any of those choices may affect an index without amounting to misconduct. The vulnerability lies in the benchmark design.

Methodology is therefore central, but it is not the whole answer. The Commission will also need to consider access to raw transaction data, contributor concentration, the treatment of related-party trades, minimum volume requirements, fallbacks, settlement windows and surveillance arrangements between the exchange and the index provider. The request leaves those questions open.

Self-certification and the stay

Under 17 CFR § 40.2, a designated contract market may self-certify a product rather than seek approval under § 40.3. That does not turn the filing into an adjudication of compliance. The Commission may stay a self-certification under § 40.2(c).

It announced that step on 9 July 2026 in relation to CME’s self-certified proposal for 24/7 crude-oil futures trading. The filing arrived while the CFTC’s own comment process on round-the-clock trading was still open. Selig called CME’s decision to proceed “wholly inappropriate”. The CFTC stayed the § 40.2 filing while continuing review under § 40.3.[10]

That episode does not forecast the treatment of compute. It shows that the stay power is real and that an open policy review may affect the Commission’s response to a contemporaneous filing. Any venue planning a launch before 20 October must price that procedural risk.

The industry has challenged the uneven pace. At the Innovation Advisory Committee meeting, CME chairman and chief executive Terrence Duffy said that about 2,500 event contracts had been self-certified since January 2025 without opposition from the Commission, including products that he believed offended Core Principle 3. Selig answered that the two examples Duffy identified had been listed offshore. Duffy also contrasted the review of CME’s compute futures with the quicker appearance of prediction-market exposure.[11]

That exchange between Duffy and Selig does not prove that event contracts receive a legal preference. It does expose a difference in process. A product filed by self-certification may reach the market before a comparable product submitted for review. The speed follows the route chosen, not simply the label attached to the instrument.

The calendar

Comments close on 20 October. CME has announced 5 October as its intended launch date, subject to regulatory review. Polymarket US stated that its first contract could list from 24 August. The market may therefore develop before the comment file closes. That sequence is permitted by the listing framework. It also explains why the central question is practical rather than theoretical. Can the CFTC surveil a benchmark once contracts tied to it are already trading?

Notes

1. CFTC, Request for Comment on the Listing of Compute Derivatives Contracts, 91 FR 54259 (21 August 2026), RIN 3038-AF77; CFTC Release 9286-26 (19 August 2026); CFTC, Chairman & Commissioners, listing Michael S. Selig as the sole current Commissioner and recording that he was sworn in on 22 December 2025.

2. Michael S. Selig, Remarks at Innovation Advisory Committee Conference, Washington, DC (20 August 2026); The White House, Winning the Race: America’s AI Action Plan (23 July 2025).

3. 7 U.S.C. § 1a(9).

4. Selig, Remarks (n 2), citing Report on S. 2391 of the Senate Agriculture, Nutrition, and Forestry Committee, S. Rep. No. 95-850, 95th Cong., 2d Sess. 22-23 (15 May 1978), and Johnson and Hazen, Derivatives Regulation § 2.03.

5. Kalshi, Kalshi Announces Compute Forward Curves (14 July 2026); CME Group, CME Group and Silicon Data to Launch Compute Futures on October 5 (11 August 2026); Intercontinental Exchange, ICE and Ornn to Launch GPU Compute Futures Contracts (19 May 2026); CFTC registers of pending DCM and DCO applications for PMEX Markets and PMEX Clearing.

6. QCX LLC d/b/a Polymarket US, Compute Index Value Contracts, specific-product certification under 17 CFR § 40.2(a) and class certification under § 40.2(d), 20 August 2026; CFTC Staff Letter No. 26-22 (24 July 2026).

7. KalshiEX, LLC v Assad, No. 25-7516 (9th Cir., 28 August 2026), affirming in part and remanding in part an appeal from No. 2:25-cv-00575 (D. Nev.). The CFTC appeared as amicus curiae.

8. 91 FR 54259 (n 1).

9. Commodity Exchange Act § 5(d)(3), 7 U.S.C. § 7(d)(3); 17 CFR § 38.200.

10. CFTC Release 9265-26 (9 July 2026); 17 CFR § 40.2(c).

11. Terrence A. Duffy, remarks at the CFTC Innovation Advisory Committee meeting (20 August 2026), livestreamed by the Commission. Duffy’s criticism of the relative pace of review is reported in Arthur Bautzer, “CFTC accused of ‘double standards’ on compute futures”, Risk.net (21 August 2026). The figure of approximately 2,500 unopposed self-certifications since January 2025 is reported, with direct quotation of the exchange between Duffy and Selig, in The Crypto Times (21 August 2026). CNBC and The Block carried the same figure on 20 August 2026. The figures and criticisms are attributed to Duffy and are not independently established facts.

The Ninth Circuit has vacated the injunction preventing Perplexity’s shopping agent from operating on Amazon.com.¹ The panel did not announce a general rule for AI agents. Its significance lies in the interaction between statutory access, system architecture and criminal law ambiguity. What follows is the interpretive tradition the panel drew on, together with what happens when the same architecture is put to an English court.

Fourteen years of resistance to expansive readings

Over the past fourteen years the Ninth Circuit and the Supreme Court have resisted expansive readings of the CFAA. Nosal I warned in 2012 against constructions that “would transform the CFAA from an anti-hacking statute into an expansive misappropriation statute”.² Van Buren fixed access, in 2021, as “the act of entering a computer system itself or a particular part of a computer system”.³ hiQ described the statute’s purpose in 2022 as preventing “intentional intrusion onto someone else’s computer”.⁴ Brekka supplies the applicable interpretive rule. Because the CFAA is primarily a criminal statute and its provisions are construed identically in civil and criminal cases, ambiguity is resolved against liability.⁵

A consequence of Amazon’s construction went unanswered. A Comet user who asked the Assistant to buy something could face criminal exposure on a conspiracy or aiding and abetting theory for facilitating Perplexity’s supposed unauthorised access. That potential expansion of criminal liability reinforced the panel’s refusal to adopt Amazon’s reading on this record.

How English law frames the same conduct

The English statute frames the conduct differently. Its starting point is not whether the defendant entered the target system, but whether the defendant caused a computer to perform a function with the required intention and knowledge. Section 1(1) of the Computer Misuse Act 1990 applies where a person causes a computer to perform any function with intent to secure access to any program or data held in any computer, where the access intended is unauthorised and the person knows at the time that it is.⁶ Section 17(2) defines securing access to include causing a computer to perform a function which alters, erases, copies, moves, uses or outputs the program or data.⁷ Instructions transmitted from Perplexity’s servers could therefore present a causation question even though those servers never communicated directly with Amazon’s servers.

An enabling limb once looked apt for a case of this kind and is no longer available in England. The Police and Justice Act 2006 would have extended section 1(1) to a person who causes a computer to perform a function intending “to enable any such access to be secured”. Those words were commenced for Scotland on 1 October 2007. The inserting provision was repealed for England, Wales and Northern Ireland on 1 October 2008 by section 61 of the Serious Crime Act 2007, headed “Repeal of offence of enabling unauthorised access to computer material”, so that enabling conduct would be addressed by the encouraging or assisting offences in Part 2 of that Act instead.⁸ A developer in Perplexity’s position might therefore fall to be considered under Part 2 of the 2007 Act if its conduct encouraged or assisted a section 1 offence and the applicable mental elements were established. That route would not necessarily displace the argument that the developer itself caused the relevant computer function and was a principal offender under section 1.

None of this establishes liability. Section 17(5) asks who was entitled to control access of the relevant kind to the particular program or data. The customer authorised use of the account. Amazon controlled the underlying systems and had expressly rejected agentic access. Under Allison, control means entitlement to authorise or forbid the particular kind of access to the actual program or data. Technical ability or general access is insufficient.⁹ English authority does not resolve that conflict on facts of this kind. Intention, knowledge and the Act’s territorial application would each fall to be established as well. The 1990 Act also creates criminal offences and provides no direct English analogue to the civil action Amazon has brought.

What remains after the appeal

Amazon’s § 1030(a)(4) claim remains pleaded because the district court’s written order did not address it and Amazon did not argue it on appeal.¹⁰ The panel therefore did not decide that claim, although its reasoning on access will be difficult to avoid, since subsection (a)(4) also requires access without authorisation. The opinion leaves broader questions of responsibility for agentic conduct outside the CFAA and CDAFA unanswered. Amazon remains free to regulate use of its site through private terms, but this action does not presently include contract or tort causes of action. The complaint pleads only two statutory counts.¹¹

One fact also remains live. Whether Perplexity knowingly altered the Assistant’s user-agent string after Amazon began blocking it is disputed on the record.¹² That question bears on every deception-based theory that might follow.

For developers, the practical consequence is that system architecture now has legal significance it was never designed to possess. Relaying through the user’s device kept Perplexity outside section 1030(a)(2) on this record. An agent whose own servers call the target site may fare differently. Whether that distinction will survive the pressure the market is about to place on it is a question for another case.

Notes

1. Amazon.com Services, LLC v. Perplexity AI, Inc., No. 26-1444 (9th Cir. 4 August 2026), vacating the preliminary injunction entered in No. 3:25-cv-09514-MMC (N.D. Cal.) (Chesney, J.) and remanding. Argued and submitted 11 June 2026, Seattle. Panel: Milan D. Smith Jr. and Eric C. Tung, Circuit Judges, and John C. Hinderaker, District Judge (D. Ariz.), sitting by designation. Opinion by Judge Smith.

2. United States v. Nosal (Nosal I), 676 F.3d 854, 857 (9th Cir. 2012) (en banc), quoted at slip op. 10.

3. Van Buren v. United States, 593 U.S. 374, 388 & n.6 (2021), quoted at slip op. 15.

4. hiQ Labs, Inc. v. LinkedIn Corp., 31 F.4th 1180, 1196 (9th Cir. 2022), quoted at slip op. 5 and 16.

5. LVRC Holdings LLC v. Brekka, 581 F.3d 1127, 1132 and 1134-35 (9th Cir. 2009).

6. Computer Misuse Act 1990, s.1(1)(a) to (c), as it has effect in England and Wales.

7. Computer Misuse Act 1990, s.17(2)(a) to (d).

8. Police and Justice Act 2006, s.35(2)(a) and (b), commenced for Scotland by S.S.I. 2007/434, art. 2; repealed for England, Wales and Northern Ireland by Serious Crime Act 2007, ss.61(2), 92, 94 and Sch. 14, commenced by S.I. 2008/2504, art. 2(a)(i)(viii). See the Explanatory Notes to the 2007 Act at paragraph 225.

9. R v Bow Street Metropolitan Stipendiary Magistrate, ex p Government of the United States of America (Re Allison) [1999] UKHL 31, [2000] 2 AC 216 (HL), disapproving the glosses placed on s.17(5) in DPP v Bignell.

10. Slip op. 8 n.2.

11. Complaint, Amazon.com Services LLC v. Perplexity AI, Inc., No. 3:25-cv-09514 (N.D. Cal., 4 November 2025), Count One (18 U.S.C. § 1030(a)(2) and (a)(4)) and Count Two (Cal. Penal Code § 502(c)(2), (c)(3), (c)(4) and (c)(7)). The panel’s CDAFA analysis addressed s.502(c)(7). See also slip op. 21 n.5 on private terms.

12. Slip op. 8 n.1.

On 24 August 2026, a federal jury in San Francisco convicted Japheth Dillman, the founder of cryptocurrency investment fund Block Bits, of wire fraud and conspiracy to commit wire fraud. The prosecution was framed as a crypto case. Its more important feature was simpler. Investors were told that a proprietary automated trading system existed, worked and was producing returns. The government proved that it did not.¹

The verdict is an early and unusually clean example of what now appears across artificial intelligence markets as “AI-washing”: the sale of an investment proposition by attributing capability, activity or performance to technology that is absent, unfinished or materially different from the system described. Technology changes. The legal problem does not. A statement about an algorithm is a statement of fact when it asserts that the algorithm has been built, deployed or used to generate results.

Block Bits therefore reaches further than the modest sum raised. The case sets out the evidential anatomy of technology fraud. The enquiry separates existence, capability, deployment and performance attribution. A promoter who compresses those four questions into the word “proprietary” does not make them less susceptible to proof.

The representation was the product

Dillman and his co-founder, David Mata, began soliciting investment in 2017. According to the Securities and Exchange Commission’s civil complaint, the fund was marketed as “An Automated Cryptocurrency Fund”. Its proposed advantage was an in-house trading bot capable of arbitrage across approximately 100 cryptocurrencies and 30 exchanges. The SEC alleges that Dillman told prospective investors that the system had been built and, on 25 September 2017, that “We have finished the Arbitrage AutoTrader”.²

Those details come from a civil pleading and remain allegations in that proceeding. The criminal verdict establishes the central proposition on a different and higher standard of proof. The Department of Justice said the trial evidence showed that Dillman raised nearly $1 million from more than 20 investors while representing that the Autotrader was working and complete. He knew it was not working and that investor money could not be used in the manner promised.³

A conviction is not proof of every allegation made by another agency in a parallel case. It is proof beyond reasonable doubt of the offences submitted to the jury. The public verdict announcement records convictions for wire fraud and conspiracy to commit wire fraud after a ten-day trial before U.S. District Judge Richard Seeborg. Law360 reported that the jury returned guilty verdicts on four wire-fraud counts and one conspiracy count after approximately six hours of deliberation over two days.⁴

Mata testified for the prosecution under a plea agreement. He had pleaded guilty in June 2022 to one count of wire fraud arising from the Block Bits scheme.⁵ Co-operator evidence often becomes the bridge between promotional language and internal knowledge. In a technology case, that bridge is particularly important. The external claim may be expressed through a pitch deck, an email or a performance report. The system’s actual condition appears in the source code, the testing records, the exchange accounts, the development messages and the evidence of the person expected to make it work.

Four falsifiable claims

“We use an algorithm” sounds general. In an investment solicitation it can contain at least four distinct representations.

The first concerns existence. Was there identifiable software rather than an aspiration, specification or unfinished development project? The second concerns capability and asks whether the software had been tested sufficiently to support the functions attributed to it. Deployment comes third. Was the system connected to live accounts and used to execute the relevant trades? The fourth question is attribution and asks whether reported returns were generated by that system rather than by manual trading, market appreciation, selective accounting or an unrelated investment.

The SEC’s allegations illustrate the separation. On its case no functional bot was ever tested or deployed and Mata conducted the fund’s trading manually. Investor reports nevertheless continued to describe automated arbitrage and remarkable returns. If proved, the problem would not be that the code performed poorly. It would be that the claimed technological process and the actual investment process were different things.⁶

This is why the familiar defence that startups sell a future is incomplete. There is a legal and factual difference between a development plan and a statement that development is complete. “We intend to build” describes a proposal. “We have finished” describes an existing fact. The first may fail commercially. The second may be fraudulent when it is knowingly false, material and used to obtain money. Nor is a statement of intention immune. It may be fraudulent where the speaker did not hold the stated intention, or knowingly presented the proposed outcome as attainable when there was no basis for doing so. Someone who says he intends to build, knowing there is neither funding nor a developer, stands where the promoter who says he has finished stands.

The distinction also answers a recurring objection in cases involving opaque systems. A jury need not evaluate the elegance of the code or determine whether one technical architecture was preferable to another. Jurors can compare what investors were told with what existed, what was deployed and what happened to their money. Complexity may enlarge the evidential record. It does not displace the comparison.

Legacy fraud law, modern evidence

Federal wire fraud does not require a special crypto or algorithmic offence. The prosecution must prove a scheme to defraud resting on a material misrepresentation, use of interstate wires in furtherance of that scheme and specific intent to defraud. Intent may be established through circumstantial evidence.⁷ Conspiracy under 18 U.S.C. § 1349 addresses the agreement to commit the underlying offence and, unlike a general conspiracy, requires no overt act and carries the same maximum penalty as the substantive offence.⁸

On the government’s account, the relevant falsehoods were material because automation was not a peripheral feature of the offering. It was the stated investment method. Block Bits was not merely buying digital assets while using ordinary software in the background. Its claimed advantage was the capacity to identify and execute arbitrage opportunities across fragmented markets without relying on manual intervention.

The use of old law is a strength rather than an embarrassment. Technology-specific legislation is necessary when the conduct creates a new interest or harm. It is unnecessary when novel terminology is used to obtain familiar property by familiar deceit. No statutory definition of an autotrader is needed before a jury may decide whether a representation that it was complete was false.

The evidence, however, is technologically specific. A serious diligence or enforcement exercise should preserve the development repository, release history, test outputs, application logs, API connections, exchange credentials, trade timestamps and model or strategy documentation. Promotional claims should then be mapped against those records. The decisive document may be neither the source code nor the pitch deck, but the version history showing that the represented function did not exist on the date money was solicited.

Performance attribution requires the same discipline. If a manager says a system generated a return, the assertion should be capable of reconciliation to live-account logs, executed orders, fees, cash movements and valuation records. A back-test is not live trading. A gross return is not a net investor return. Manual selection of a favourable transaction produces neither an automated output nor a representative one. A model that recommends a trade may sit some distance from the system that executes it.

From crypto-washing to AI-washing

The Block Bits conduct predates the present investment fashion for artificial intelligence, but the representational pattern is current. In March 2024, the SEC settled enforcement proceedings against two investment advisers, Delphia (USA) Inc. and Global Predictions Inc., over false or misleading statements about their use of AI. The Commission described the conduct as AI-washing. The firms agreed to pay a combined $400,000 in civil penalties without admitting or denying the findings.⁹

The analogy should not be overstated. Those were civil and administrative settlements. Dillman has been convicted by a jury of federal crimes. The underlying compliance lesson is nevertheless common. A technology claim must be supported at the level at which it is made. If marketing says that AI drives portfolio construction, it is insufficient that the firm owns analytical software, has experimented with a model or intends to incorporate machine learning later.

Prosecutors have since brought a criminal case of the same kind on the artificial intelligence side. In April 2025 the Securities and Exchange Commission sued Albert Saniger, founder and former chief executive of Nate, Inc., over a shopping application marketed as completing purchases through proprietary artificial intelligence. The Commission alleges that he raised more than $42 million from investors while the application relied in large part on contract employees who entered orders by hand. The Department of Justice charged him in parallel in the Southern District of New York. Block Bits and Nate share a structure. Each rests on a claimed automated system, on human execution behind it and on investor money raised on the representation.¹⁰

The same four questions apply. Does the model exist? Does it possess the described capability? Was it used in the relevant decision process? Did it generate the result attributed to it? For generative AI, the enquiry extends to which model and version operated at the relevant time, what data and tools were available to it, whether a human decision was presented as an autonomous output and whether favourable examples were selected from a wider set of failed or discarded ones.

These are not semantic objections. They determine whether the buyer received the process for which capital was sought. They also determine scienter. Internal model cards, validation reports, deployment approvals and incident logs may show uncertainty or limitation. That does not itself establish fraud. The sharper evidence is a contradiction between those internal records and a confident external assertion that the limitation does not exist.

The money trail still decides cases

Technology representations do not replace conventional financial analysis. They direct it. The Department of Justice said evidence showed that investor money was used to pay the founders and to pursue risky and speculative crypto ventures, while investors received misleading reports that concealed losses.¹¹

The SEC’s complaint supplies further allegations. On the Commission’s case approximately $960,000 was raised from approximately 22 retail investors and funds Dillman described as sitting in risk-free cold storage were in fact placed in high-risk loans and token investments, among them the AML Bitcoin offering the Commission has separately alleged to be fraudulent. It alleges that approximately $300,000 was transferred to the general partner, including $67,500 that reached Dillman personally. The Commission characterises much of the compensation as unauthorised under the offering terms.¹² Those propositions remain for determination in the civil case unless resolved by agreement or otherwise disposed of.

The civil proceedings have been stayed while the criminal case proceeds. An order of 8 July 2022 required the parties to request a status hearing within thirty days of the resolution of the criminal case.¹³ The verdict materially changes the posture of the SEC case and may narrow the issues remaining for determination, but it does not automatically decide the civil proceedings. Sentencing is listed for 8 December 2026. Post-trial motions and any appeal remain possible.¹⁴

The compliance consequence

Boards, investment committees and counsel should treat material technology claims as controlled statements. Each claim should have an owner, a dated evidential basis and a defined scope. “Uses AI”, “automated”, “real time” and “proprietary” are not harmless adjectives. They are compressed factual propositions.

The control should extend beyond marketing approval. Development and investment records must use the same vocabulary. If engineers describe a system as a prototype while fundraising materials call it operational, the discrepancy requires resolution before publication. If performance reports attribute returns to an algorithm, the attribution requires account-level support. If human intervention is substantial, the disclosure should say so.

Block Bits does not extend fraud law into an unregulated technological frontier. It shows that no frontier needed to be crossed. An unfinished trading bot was presented as a completed investment engine. Investors supplied money. The wires carried the representations. The internal and financial records supplied the contradiction.

A promoter can obscure a product behind the vocabulary of innovation for a time. He cannot change the legal character of a false statement about what has been built, what has been used and what has produced the return.

Notes

1. U.S. Attorney’s Office, Northern District of California, “Founder of Cryptocurrency Trading Fund Convicted of Defrauding Investors” (24 August 2026), https://www.justice.gov/usao-ndca/pr/founder-cryptocurrency-trading-fund-convicted-defrauding-investors.

2. Complaint, SEC v Block Bits Capital, LLC et al., No. 3:22-cv-02563 (N.D. Cal., filed 28 April 2022), paras 1–5, 14–15, 18, https://www.sec.gov/files/litigation/complaints/2022/comp25376-blockbits.pdf.

3. U.S. Attorney’s Office, note 1.

4. Bonnie Eslinger, “Founder Of Crypto Startup Block Bits Convicted Of Fraud”, Law360 (24 August 2026), https://www.law360.com/fintech/articles/2516827; United States v Japheth Dillman, No. 3:23-cr-00140 (N.D. Cal.).

5. Eslinger, note 4 (Mata pleaded guilty in June 2022 and testified for the prosecution). For the charge, see U.S. Attorney’s Office, Northern District of California, “San Francisco Man Charged In Alleged Cryptocurrency Investor Fraud Scheme” (27 April 2022), https://www.justice.gov/usao-ndca/pr/san-francisco-man-charged-alleged-cryptocurrency-investor-fraud-scheme; SEC, “SEC Charges Promoters of ‘Automated Cryptocurrency Fund’ with Fraud and Registration Violations”, Litigation Release No. 25376 (28 April 2022), https://www.sec.gov/enforcement-litigation/litigation-releases/lr-25376. Mata was charged by separate information on the same day.

6. SEC Complaint, note 2, paras 3–5, 17, 20–21.

7. 18 U.S.C. § 1343; United States v French, 748 F.3d 922, 935 (9th Cir. 2014) (scheme, wires and specific intent; intent may be circumstantial), https://cdn.ca9.uscourts.gov/datastore/opinions/2014/04/07/12-10185.pdf; Neder v United States, 527 U.S. 1, 25 (1999) (materiality is an element of the federal fraud statutes).

8. 18 U.S.C. § 1349, which provides the same penalties as the substantive offence. A general conspiracy under 18 U.S.C. § 371 requires an overt act. Where a conspiracy statute is silent as to an overt act, none is required. See Whitfield v United States, 543 U.S. 209, 213–214 (2005), applying United States v Shabani, 513 U.S. 10 (1994).

9. SEC, “SEC Charges Two Investment Advisers with Making False and Misleading Statements About Their Use of Artificial Intelligence” (18 March 2024), Release No. 2024-36, https://www.sec.gov/newsroom/press-releases/2024-36.

10. SEC, “SEC Charges Founder and Former CEO of Artificial Intelligence Startup with Misleading Investors”, Litigation Release No. 26282 (11 April 2025), https://www.sec.gov/enforcement-litigation/litigation-releases/lr-26282; SEC v Saniger Mantinan, No. 1:25-cv-02937 (S.D.N.Y., commenced 9 April 2025). For the criminal charges see U.S. Attorney’s Office, Southern District of New York, “Tech CEO Charged In Artificial Intelligence Investment Fraud Scheme” (9 April 2025), https://www.justice.gov/usao-sdny/pr/tech-ceo-charged-artificial-intelligence-investment-fraud-scheme (one count of securities fraud and one count of wire fraud).

11. U.S. Attorney’s Office, note 1.

12. SEC Complaint, note 2, paras 4–5, 27.

13. Stipulation and Order Re: Stay of Proceedings, SEC v Block Bits Capital, LLC et al., No. 3:22-cv-02563-RS (N.D. Cal., 18 August 2022), https://www.govinfo.gov/content/pkg/USCOURTS-cand-3_22-cv-02563/pdf/USCOURTS-cand-3_22-cv-02563-1.pdf.

14. U.S. Attorney’s Office, note 1.

Image: photograph by Anne Nygård, Unsplash.

The US Attorney’s Office for the Southern District of New York opposed Van Dyke’s motion to dismiss on 19 August in what appears to be the first criminal insider trading prosecution brought over trades on a prediction market.⁴ The indictment against Gannon Ken Van Dyke, a US Army Special Forces master sergeant, alleges that he spent approximately $33,934 across thirteen Polymarket trades between 27 December 2025 and 2 January 2026 while holding classified access to the planning of Operation Absolute Resolve, the operation that apprehended Nicolás Maduro in Caracas on 3 January 2026. His alleged total profit was approximately $409,881.¹

The government has built five counts on two contested definitions. Both are being argued elsewhere while the motion sits before Judge Margaret M. Garnett.

Chronology

The positions did not all end the same way. The “Maduro out” and “US forces in Venezuela” contracts resolved YES. Van Dyke is alleged to have sold his war powers position at a profit before resolution, and to have sold his invasion position at a profit as well, although that contract ultimately resolved NO. This is important for loss and forfeiture rather than for liability, but it also disposes of the tidy account that the trades were a single winning bet on a single outcome.

The premise beneath counts one to three

Counts one and two arise under the government-information provisions inserted into the Commodity Exchange Act by section 746 of the Dodd-Frank Act, at 7 U.S.C. §§ 6c(a)(3) and 6c(a)(4)(C). The CFTC has described its parallel civil action as its first insider trading case involving event contracts and its first use of what practitioners call the Eddie Murphy Rule.² Count three alleges commodities fraud under 7 U.S.C. §§ 9(1) and 13(a)(5), on a misappropriation theory of the kind familiar from securities insider dealing.

Each of those counts characterises the transactions as swaps. The definition at 7 U.S.C. § 1a(47)(A) reaches agreements providing for payment on the occurrence of an event “associated with a potential financial, economic, or commercial consequence”. The government’s position is that Maduro’s removal carried consequences for crude prices, Venezuelan sovereign bonds and the bolívar, which places the contracts inside the definition. The defence position is that a geopolitical bet remains a bet.

A divided Third Circuit accepted the government’s side of that argument at the preliminary-injunction stage in KalshiEX LLC v Flaherty, 172 F.4th 220 (3d Cir. 2026), holding that sports event contracts were swaps because sporting outcomes carry economic consequences for sponsors, broadcasters, franchises and local economies. The authorities below are divided. The Southern District of Ohio denied Kalshi an injunction on the footing that its contracts were likely not swaps, while the Middle District of Tennessee granted one. A Sixth Circuit panel heard the consolidated merits appeals on 30 July 2026 and has not ruled.³

Neither appeal binds Judge Garnett, and neither concerns these contracts. The Kalshi litigation is about sports event contracts listed on a designated contract market, in a pre-emption dispute with state gaming regulators. Van Dyke concerns Venezuela contracts on the platform the indictment describes as operated by Blockratize, Inc., in a criminal prosecution. A Sixth Circuit judgment against Kalshi would deepen an appellate disagreement about the reach of the swap definition rather than resolve the question in this case. The parallel civil action, CFTC v. Van Dyke, No. 1:26-cv-03369 (S.D.N.Y.), was stayed on 10 August 2026 pending completion of the criminal case, so the definitional question will now be answered first in a criminal court.

Property after Blaszczak and Chastain

Count four charges wire fraud under 18 U.S.C. § 1343. Money or property must be an object of the alleged scheme. The indictment pleads that the classified information carried pecuniary value.

The Second Circuit has been contracting that term. In United States v. Blaszczak, 56 F.4th 230 (2d Cir. 2022), the court held that confidential Centers for Medicare & Medicaid Services information about forthcoming reimbursement rates was regulatory in character rather than proprietary and so fell outside the statute, notwithstanding that the information moved markets and was held in confidence. In United States v. Chastain, 145 F.4th 282 (2d Cir. 2025), decided on 31 July 2025, the court vacated the wire fraud conviction in the first non-fungible token “insider trading” prosecution and held that confidential business information constitutes property only where it carries commercial value to its owner.

Blaszczak preserved United States v. Girard, 601 F.2d 69 (2d Cir. 1979), in which records identifying Drug Enforcement Administration informants were held to be a record or thing of value under 18 U.S.C. § 641 because the information was inherently valuable to the agency’s operations. Judge Richard Sullivan, dissenting in Blaszczak, questioned how Girard could survive the majority’s reasoning. The analogy is therefore imperfect on its face, since Girard turned on the conversion statute rather than on section 1343, and prosecutors must persuade the court that operational military intelligence sits closer to the informant records in Girard than to the regulatory information in Blaszczak or the listing information in Chastain.

The pleading of pecuniary value suggests an alternative argument under Chastain, treating the price the information commanded on Polymarket as evidence of the value it held. That reasoning has an awkward shape. It makes the property status of a national security secret depend on the existence of a market prepared to price it, which is a market the government neither authorised nor controls.

What practitioners should take from this

Three points follow for anyone advising holders of confidential information.

First, most insider trading and MNPI policies are drafted by reference to securities and say nothing about event contracts. Earnings, regulatory approvals, clinical results and deal timing now have a tradable venue that sits outside the securities perimeter as most codes of conduct define it. The gap is a drafting problem rather than a doctrinal one, and it is straightforward to close.

Second, the outcome of the dismissal motion will not settle the swap question. A Sixth Circuit judgment against Kalshi would produce a split with the Third Circuit and a plausible route to the Supreme Court, with the CFTC’s own event contract rulemaking running alongside it.

Third, the property question has consequences well beyond prediction markets. If classified operational information is property under section 1343 because a betting market will pay for it, the same logic reaches any confidential government information for which a market exists. If it is not property, the wire fraud count falls away and the prosecution rests on the Commodity Exchange Act counts, which are the counts exposed to the definitional dispute.

The court has not ruled on the motion.

Notes

1. Indictment, United States v. Van Dyke, No. 1:26-cr-00156 (S.D.N.Y., returned 21 April 2026), at paragraph 10 for the aggregate stake. See also U.S. Attorney’s Office, Southern District of New York, Press Release 26-109, 23 April 2026, which gives a slightly different aggregate figure.

2. Commodity Futures Trading Commission, Release 9217-26, 23 April 2026. The government-information prohibitions were inserted by section 746 of the Dodd-Frank Wall Street Reform and Consumer Protection Act 2010.

3. Oral argument in the consolidated Ohio and Tennessee appeals was heard on 30 July 2026 before Judges Gibbons, Clay and Bloomekatz. An earlier unpublished Sixth Circuit order of 24 April 2026 upheld the Ohio injunction denial while assuming, without deciding, that the contracts were swaps.

4. The DOJ announcement does not itself claim a first. The characterisation rests on specialist commentary, which hedges it, and on the CFTC’s statement that its parallel civil action was its first insider trading case involving event contracts.

By the end of 2025, the FDA’s public list identified 1,451 AI-enabled medical devices authorised for marketing in the United States, including 295 authorised during that year.¹ Three quarters are in radiology. Most of the remainder are concentrated in cardiology, haematology, pathology, ophthalmology and a growing number of surgical specialties. Few regulated sectors reveal so starkly the distance between regulatory authorisation and the allocation of civil liability.

This is the sixth article in a series examining AI liability across industrial sectors. The previous five covered vessels, energy grids, aviation, mining and agriculture. In healthcare, the end user is a patient. The patient does not select the device, cannot interrogate the algorithm and has no contractual relationship with the developer. The liability questions are accordingly different from those arising in autonomous shipping, grid management or mine automation: who owes a duty of care to a person who never chose to interact with the system at all?

Clinical AI and the Product Liability Question

In May 2025, the US District Court for the Middle District of Florida declined to dismiss product-liability claims against Character Technologies.² At the pleading stage, the court was not prepared to hold that the Character.AI service and its outputs fell outside Florida product-liability law as a matter of law. It also rejected the defendants’ attempt to obtain dismissal on First Amendment grounds. The decision did not finally classify the chatbot as a product, but it prevented the developer from defeating that classification at the outset. The case settled in January 2026, leaving the pleading-stage order on the public record but the classification question unresolved.

Garcia concerned a consumer chatbot, not a clinical device. It does, however, show that courts may be unwilling to exclude software-based AI systems categorically from product-liability doctrine. Clinical AI presents a stronger but legally distinct case, where the software is incorporated into a device already regulated by the FDA as a medical device.

Surgical robotics illustrates the interaction between product liability and clinical negligence. Claims involving the da Vinci system have alleged both defects in robotic instruments and failures in surgical technique, training or postoperative care.³ The systems in widespread clinical use remain surgeon-controlled rather than independently autonomous, making fault allocation highly fact-sensitive.

That interaction tests the factual assumptions underlying the learned-intermediary doctrine. Under the traditional formulation, a medical-device manufacturer discharges its duty to warn by providing adequate information to the prescribing or operating clinician. The clinician, as a learned intermediary, interprets that information and exercises independent professional judgement before the product reaches the patient. The doctrine assumes that the clinician stands between the manufacturer and the patient as a filter of risk.

Clinical AI does not displace the learned-intermediary doctrine merely because it performs a defined task as accurately as a clinician. The doctrine concerns whether the manufacturer adequately warned the clinician and whether the clinician could exercise informed, independent judgement. That factual predicate weakens where a model’s limitations are undisclosed, where its output is not reasonably interpretable, where its recommendations reach the patient without meaningful clinical review, or where the decision to adopt and configure the system was made institutionally rather than by the treating clinician. The doctrine has not been formally displaced in any US jurisdiction. Its application will depend on whether the clinician performed a meaningful intermediary function, and the more autonomous the clinical output, the harder that factual case becomes.

Regulation: Authorisation Without Liability Rules

The FDA’s approach to AI-enabled medical devices rests on three instruments. The first is the Predetermined Change Control Plan, finalised in August 2025, which permits a manufacturer to obtain authorisation for specified future modifications to an AI device, together with the methods by which those modifications will be developed, validated and implemented.⁴ The PCCP responds to the problem that some machine learning models are designed to be modified after deployment. A device cleared against a defined dataset and performance profile may subsequently be modified, retrained or updated in ways that alter its diagnostic behaviour. Without a mechanism for pre-authorised modification, such a modification might otherwise require a new marketing submission, depending on its nature and effect.

The second instrument is the Clinical Decision Support guidance, finalised on 29 January 2026.⁵ Section 520(o)(1)(E) of the Federal Food, Drug, and Cosmetic Act excludes certain professional clinical-decision-support functions from the statutory definition of a device where all four statutory criteria are met. Among them, the software must enable the healthcare professional independently to review the basis for its recommendations rather than rely primarily on the software in making a diagnosis or treatment decision. The January 2026 guidance clarifies the boundary between software that meets those criteria and software that does not.

The third is enforcement. In February 2025, the FDA issued a warning letter to Exer Labs, treating its AI-based exercise analysis tool as an adulterated and misbranded device marketed without the required premarket authorisation.⁶ The action confirms that the FDA applies its existing medical-device requirements to AI-enabled products without creating a separate regulatory category.

Pre-emption adds a further layer. For devices that have received premarket approval under section 515 of the FD&C Act, the Supreme Court’s decision in Riegel v Medtronic, 552 U.S. 312 (2008), established express pre-emption of state-law claims that impose requirements different from or additional to the federal ones.⁷ Most AI-enabled medical devices on the FDA’s list have been cleared under the 510(k) pathway, where pre-emption is narrower. Whether and how pre-emption applies to AI-specific tort claims, particularly where the alleged defect lies in training-data selection or algorithmic design rather than device manufacture, is untested.

The European Union has taken a different path. The interaction between the Medical Devices Regulation and the AI Act remained uncertain until June 2025, when the Medical Device Coordination Group and the AI Board published MDCG 2025-6, a frequently asked questions document on the application of both instruments to AI-enabled medical devices.⁸ It confirms that medical-device and AI Act obligations may apply cumulatively. An AI system that is itself a medical device, or a safety component of one, will fall within Article 6(1) of the AI Act where the product is covered by Annex I harmonisation legislation and requires third-party conformity assessment.

Two subsequent developments altered the trajectory. Regulation (EU) 2026/1744, the Digital Omnibus on AI, published in the Official Journal on 24 July 2026, postponed the application of the relevant high-risk AI obligations for Annex I product systems until 2 August 2028. In October 2025, the European Commission withdrew the proposed AI Liability Directive, which would, in defined circumstances, have introduced disclosure obligations and a rebuttable presumption concerning the causal link between specified fault and an AI system’s output or failure to produce an output.⁹ The withdrawal removed the only pending EU instrument directed specifically at evidential and fault-based civil-liability problems created by AI. Member States retain their domestic product-liability and fault-based regimes.

The UK Medicines and Healthcare products Regulatory Agency published guidance on digital mental health technologies as Software as a Medical Device in February 2025 and proposed international reliance pathways in July 2025, under which the MHRA could accept regulatory decisions from trusted overseas authorities as part of its own assessment process.¹⁰ In August 2025, the MHRA joined the FDA and Health Canada in publishing joint principles on predetermined change control plans. The UK has pursued a guidance-led approach. Its civil-liability position remains grounded in existing product-liability, negligence and clinical-malpractice rules rather than an AI-specific statutory regime.

Algorithmic Bias and the Standard of Care

In January 2025, the FDA published draft guidance on pulse oximetry devices, responding to a body of evidence that optical sensors perform less accurately on patients with darker skin pigmentation.¹¹ Inaccurate oxygen saturation readings in darker-skinned patients have been linked to delayed recognition of hypoxia. The draft guidance recommends performance studies that include a broader range of skin pigmentation and more rigorous methods for evaluating differential performance.

Pulse oximetry is a hardware problem with a software dimension. The bias originates in the optical sensor’s calibration. Downstream clinical systems may then reproduce or amplify the measurement error if they treat the sensor output as reliable or were not validated adequately across affected populations. A comparable problem appears in dermatology AI systems trained predominantly on images of lighter skin, which are less accurate at identifying conditions on darker skin.¹² The training-data disparity is well documented. A model trained on an unrepresentative dataset is liable to reproduce those gaps in clinical performance. The liability question is whether a manufacturer that deploys a diagnostic AI system with a known training-data limitation has met the standard of care expected of a reasonable manufacturer in its position.

Radiology also presents scale risk. Three quarters of the AI-enabled devices on the FDA’s list are in radiology. A defect in a widely deployed chest X-ray algorithm may affect more patients and clinical pathways than a defect in a specialised planning tool. The relevant exposure therefore depends not simply on the number of authorised products, but on deployment, interoperability and reliance across institutions.

California has moved ahead of federal law on two fronts. SB 1120, effective from January 2025, requires that utilisation review and medical-necessity determinations involving AI be made by appropriately qualified healthcare professionals; the statute prevents AI from supplanting the clinician’s decision-making in coverage determinations.¹³ AB 3030, also effective from 2025, requires that when generative AI generates specified written or verbal communications concerning patient clinical information, the provider must include a disclaimer and instructions for contacting a human provider, subject to exceptions for communications reviewed and approved by a licensed healthcare provider.¹⁴ Neither statute expressly creates a standalone private cause of action in the provisions discussed here. Both address the patient-facing consequences of clinical AI deployment.

Drug Discovery and the AI-Generated Molecule

In June 2025, Insilico Medicine published Phase IIa results for rentosertib, a drug candidate identified and optimised using its AI platform, in Nature Medicine.¹⁵ The molecule moved from target identification to Phase IIa data in under four years. In March 2026, Eli Lilly entered a partnership with Insilico valued at up to $2.75 billion on a contingent-milestone basis.¹⁶

The FDA’s January 2025 draft guidance on AI in pharmaceutical development addresses the use of AI-generated information to support regulatory submissions and decisions.¹⁷ Early-stage drug discovery falls outside its scope. A drug candidate generated by an AI system is subject to the same premarket approval requirements as one identified through conventional screening. The core approval requirements do not change merely because AI contributed to target or molecule identification.

Patent law in most jurisdictions requires a human inventor. The US Patent and Trademark Office and the UK Intellectual Property Office have both confirmed that an AI system cannot be named as an inventor on a patent application.¹⁸ Where an AI system identifies a novel molecular structure, the question is whether a natural person conceived the claimed invention. Under the USPTO’s revised November 2025 guidance, conception remains the touchstone: the human inventor must possess a definite and permanent idea of the complete and operative invention. How that standard applies where an AI system generated the structure remains fact-sensitive. The inventorship issue may affect patent validity, exclusivity and therefore asset value. Liability for clinical performance follows a different path through product-liability, regulatory and professional-negligence rules; it does not turn on who identified the molecule.

The FDA applies its existing Current Good Manufacturing Practice requirements to AI use in pharmaceutical operations. A manufacturer that uses machine learning to optimise a production parameter is subject to the same validation, documentation and deviation-reporting obligations as a manufacturer that uses a conventional control system.

Strategic Outlook

Product-liability claims may reach AI-enabled systems, although Garcia leaves the classification question unresolved and the applicable rules remain state-specific. Medical negligence claims may examine clinicians’ reliance on AI and institutions’ selection, configuration and supervision of the systems they deploy. Regulatory enforcement imposes compliance obligations that define the floor of acceptable conduct, while differential performance and state-level disclosure requirements create additional sources of regulatory and litigation exposure.

Additional liability questions remain. Automation bias, where clinicians defer to algorithmic recommendations against their own clinical judgement, creates a distinct professional-liability exposure. Hospital liability for institutional procurement and configuration decisions is largely untested. Data-protection and cybersecurity obligations introduce further layers of potential non-compliance. Contractual allocation of AI-related risk among developers, distributors and deploying institutions remains at an early stage.

The regulatory divergence among the United States, European Union and United Kingdom is significant. The FDA has the most detailed device-specific authorisation process but no federal AI liability statute. The EU has built the most extensive classification system but has withdrawn its liability directive and postponed the application of its high risk obligations. The UK has proposed international reliance pathways but its civil liability position remains grounded in existing law. A manufacturer deploying a clinical AI system across all three markets faces distinct regulatory regimes and no single cross-jurisdictional standard for algorithmic transparency, bias validation or post-market surveillance.

The revised Product Liability Directive expressly includes software within the concept of a product and introduces disclosure rules and rebuttable presumptions intended to relieve some of the evidential burden created by technically complex products.¹⁹ Member States must transpose it by 9 December 2026, and the new regime applies to products placed on the market or put into service thereafter. It mitigates, but does not eliminate, the difficulty of proving defect and causation where model operation, training data and technical documentation remain largely within the producer’s control.

By the end of 2025, the FDA had identified 1,451 AI-enabled devices authorised for marketing. The civil liability rules governing their use in clinical practice have not kept pace. The hardest claims will emerge in the distance between regulatory authorisation and responsibility for clinical use.

Notes

1. US Food and Drug Administration, “Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices” (updated December 2025). The FDA maintains a public list of devices authorised for marketing that include AI/ML components. The agency notes that the list is not intended to be comprehensive. The figures cited (1,451 total, 295 in 2025, approximately 76% in radiology) are derived from the December 2025 version of the list.

2. Garcia v Character Technologies, No. 6:24-cv-01903 (M.D. Fla., May 2025). The court declined to dismiss the product-liability claim at the pleading stage, holding that it was not prepared to find as a matter of law that the AI chatbot fell outside Florida product-liability doctrine. It also declined to dismiss on First Amendment grounds. The case settled in January 2026, leaving the pleading-stage order on the public record but the classification question unresolved.

3. Claims involving the da Vinci system have alleged instrument defects, inadequate warnings, insufficient training and failures in postoperative care. See, for example, Pierre v Intuitive Surgical, Inc., No. 18-62553-CIV (S.D. Fla. 2020).

4. US Food and Drug Administration, “Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence/Machine Learning-Enabled Device Software Functions” (finalised August 2025).

5. US Food and Drug Administration, “Clinical Decision Support Software: Guidance for Industry and Food and Drug Administration Staff” (29 January 2026).

6. FDA warning letter to Exer Labs Inc (10 February 2025), treating the Exer Scan product as an adulterated and misbranded device marketed without premarket approval, clearance or authorisation.

7. Riegel v Medtronic, Inc., 552 U.S. 312 (2008). The Supreme Court held that the Medical Device Amendments expressly pre-empt state-law claims imposing requirements different from or additional to federal requirements for PMA-approved devices. Pre-emption is narrower for 510(k)-cleared devices: see Medtronic, Inc. v Lohr, 518 U.S. 470 (1996). Parallel claims alleging violation of federal requirements may survive pre-emption.

8. Medical Device Coordination Group and AI Board, MDCG 2025-6, “FAQ on the Interplay Between the Medical Devices Regulation (EU) 2017/745, the In Vitro Diagnostic Medical Devices Regulation (EU) 2017/746 and the Artificial Intelligence Act” (19 June 2025).

9. Regulation (EU) 2026/1744, the Digital Omnibus on AI, published in the Official Journal on 24 July 2026. The European Commission withdrew the proposed AI Liability Directive in October 2025.

10. UK Medicines and Healthcare products Regulatory Agency, guidance on digital mental health technologies as Software as a Medical Device (February 2025). International reliance pathways proposed July 2025. FDA, Health Canada and MHRA, joint principles on predetermined change control plans for AI/ML-enabled medical devices (August 2025).

11. US Food and Drug Administration, draft guidance on pulse oximetry device performance across skin pigmentation levels (January 2025).

12. R. Daneshjou, K. Vodrahalli, R.A. Novoa et al., “Disparities in Dermatology AI Performance on a Diverse, Curated Clinical Image Set”, Science Advances, vol. 8, no. 32 (2022), eabq6147, DOI: 10.1126/sciadv.abq6147; A.S. Adamson and A. Smith, “Machine Learning and Health Care Disparities in Dermatology”, JAMA Dermatology, vol. 154, no. 11 (2018), pp. 1247–1248, DOI: 10.1001/jamadermatol.2018.2348.

13. California SB 1120 (effective 1 January 2025). The statute requires that utilisation review and medical-necessity determinations involving AI be made by appropriately qualified healthcare professionals.

14. California AB 3030 (effective 2025). The statute applies to specified AI-generated written or verbal communications concerning patient clinical information. Communications reviewed and approved by a licensed or certified healthcare provider are exempted.

15. A. Zhavoronkov et al., “A Generative AI-Discovered TNIK Inhibitor for Idiopathic Pulmonary Fibrosis: A Randomized Phase 2a Trial”, Nature Medicine, vol. 31, no. 8 (2025), pp. 2602–2610, DOI: 10.1038/s41591-025-03743-2. Published online 3 June 2025. Rentosertib was identified and optimised using Insilico’s AI-driven drug discovery platform.

16. Eli Lilly and Insilico Medicine partnership announced March 2026. The stated value of up to $2.75 billion represents a contingent maximum including development and commercial milestones, not the upfront transaction value.

17. US Food and Drug Administration, “Considerations for the Use of Artificial Intelligence To Support Regulatory Decision-Making for Drug and Biological Products” (draft guidance, January 2025).

18. US Patent and Trademark Office, “Revised Inventorship Guidance for AI-Assisted Inventions”, 90 Fed. Reg. 54636 (28 November 2025), rescinding and replacing the guidance published at 89 Fed. Reg. 10043 (13 February 2024). The revised guidance confirms that ordinary principles of human conception apply and that AI systems cannot be inventors. The UK position was confirmed in Thaler v Comptroller-General of Patents [2023] UKSC 49.

19. Directive (EU) 2024/2853 of the European Parliament and of the Council of 23 October 2024 on liability for defective products. Member States must transpose by 9 December 2026.

What the release fixes

The figures Chair Atkins had previewed now sit on the face of a release. A one-time exemption would permit offerings of up to $5 million during a four-year period. A second would permit offerings of up to $75 million during each 12-month period, with financial statements and ongoing reporting attached to the second and principles-based narrative disclosure required under both.² A conditional safe harbour would take a crypto asset outside the term “investment contract” in the Securities Act and Exchange Act definitions of “security” where its conditions are satisfied.² Issuers relying on the exemptions would remain subject to the antifraud and antimanipulation provisions of the federal securities laws.²

The release carries Release Nos. 33-11434 and 34-106150 and File No. S7-2026-27, under RIN 3235-AN38. The comment period runs for 60 days from publication in the Federal Register, which had not occurred when this was written.²

Subpart E

The caps have taken most of the coverage. Subpart E is the provision that changes the most. It would define “qualified purchaser” for the purposes of section 18(b)(3) of the Securities Act, with the effect that state securities law registration and qualification requirements would be pre-empted in respect of offers and sales of covered investment contracts issued under a Regulation Crypto Assets exemption, and in respect of certain secondary market transactions in those instruments.³

This is the mechanism introduced by the National Securities Markets Improvement Act of 1996 applied to a new class of instrument. Section 18(b)(3) operates on registration and qualification requirements. It does not by its terms displace state antifraud authority, and the proposal should be read on that footing rather than as a general withdrawal of state remedies.³

Where the remedy stops

The point I made about Section 11 survives the release and is sharpened by it. Section 11 of the Securities Act gives a person acquiring a security issued under a registration statement an express claim for a material misstatement or omission in that statement, without requiring proof of scienter.⁴ An exempt offering uses no registration statement, so Section 11 does not apply. Federal antifraud liability remains and requires proof of a different order.

Taken together, the proposal would pre-empt state registration and qualification requirements without supplying investors with a federal remedy equivalent to Section 11. That is a policy choice properly made through notice and comment, and the comment period is where it should be tested. Europe reached comparable ground by legislation. MiCA has applied in full since 30 December 2024 and founds civil liability for defective crypto-asset white paper content under Article 15, on conditions materially different from those of Section 11, though sitting in binding legislation that further legislation would be needed to change.⁵

What has not changed

The durability point stands. A framework built by rulemaking can be amended or rescinded by a later Commission, and any eventual rule would face a court applying its own judgment to disputed questions of statutory authority without Chevron deference. Commissioner Peirce is due to leave in November.⁶ The Digital Asset Market Clarity Act faces a cloture vote on the motion to proceed on 15 September.⁷ Issuing the release answers the question I asked on 18 August about timing. It does not answer the question underneath it, which is whether an administrative instrument can carry the weight Congress has not yet placed in a statute.

Notes

1. SEC, “SEC Proposes New Regulation Crypto Assets”, Press Release 2026-76, 18 August 2026. Regulation Crypto Assets, Release Nos. 33-11434 and 34-106150, File No. S7-2026-27, RIN 3235-AN38, SEC issue date 18 August 2026. The open meeting noticed for 14 August 2026 was cancelled on 13 August 2026 and was not rescheduled.

2. Regulation Crypto Assets, proposing release, summary and overview of proposed rules; Press Release 2026-76.

3. Regulation Crypto Assets, proposing release, Subpart E, defining “qualified purchaser” for the purposes of section 18(b)(3) of the Securities Act of 1933. The qualified purchaser category in section 18(b)(3) was introduced by the National Securities Markets Improvement Act of 1996. See also SEC, Defining the Term “Qualified Purchaser” Under the Securities Act of 1933, December 2001.

4. Securities Act of 1933, s.11, 15 U.S.C. s.77k.

5. Regulation (EU) 2023/1114 (MiCA), Article 15; applicable in full from 30 December 2024.

6. Bloomberg, 21 May 2026; Regent University School of Law announcement. Commissioner Peirce to join the faculty in November 2026.

7. Cloture filed on the motion to proceed to H.R. 3633, the Digital Asset Market Clarity Act, on 8 August 2026; vote scheduled for 15 September 2026.

The SEC proposed Regulation Crypto Assets on 18 August 2026, Release No. 33-11434, five days after it cancelled the open meeting at which the proposal was to be considered.¹ Coverage has concentrated on the two offering exemptions. The provision with the widest federalism consequence is the definition of “qualified purchaser”.

What the Commission proposed

The proposal has three principal exemptive elements. A startup exemption would permit one-time, non-exclusive offerings of covered investment contracts of up to $5 million over a four-year period, subject to public filings at the beginning and end of that period and principles-based narrative disclosure to investors during it. A fundraising exemption modelled in part on Regulation A would operate in two tiers, $20 million in any twelve months under Tier 1 and $75 million under Tier 2, with financial statements required in both cases and audited financial statements for Tier 2, alongside ongoing reporting modelled on Regulation A.² An investment contract safe harbour would treat a covered investment contract as having ceased to exist where the issuer has completed or permanently ceased all essential managerial efforts it represented or promised it would undertake, is not making and does not intend to make any new such representations or promises, and makes a public filing certifying the position with a supporting analysis.

Issuers relying on either exemption remain subject to the antifraud and antimanipulation provisions of the federal securities laws. The comment period runs for 60 days from publication of the proposing release in the Federal Register.

A definitional power used sparingly since 1996

The fourth element is the pre-emption provision. The proposal would add a definition of “qualified purchaser” under the Securities Act, with the consequence that state registration and qualification requirements are pre-empted for offers and sales of covered investment contracts made under the regulation. Secondary market transactions by any person other than an issuer, underwriter or dealer would also be pre-empted, for covered investment contracts initially sold under a Regulation Crypto Assets exemption or another federal exemption, and only for so long as the issuer continues to satisfy the applicable information, filing or periodic reporting requirements.

The authority is section 18(b)(3) of the Securities Act, added by section 102(a) of the National Securities Markets Improvement Act of 1996. It provides that a security is a covered security with respect to offers or sales to “qualified purchasers, as defined by the Commission by rule”, and permits the Commission to define the term differently for different categories of securities, “consistent with the public interest and the protection of investors”.³ Section 18(c)(1) preserves state jurisdiction to investigate and bring enforcement actions for fraud or deceit, so what is displaced is registration, qualification and merit review rather than fraud enforcement.

The Commission has adopted a definition under that subsection twice, each time confined to a single exemptive regime. Rule 256, adopted in the 2015 Regulation A amendments and effective 19 June 2015, provides that for purposes of section 18(b)(3) a qualified purchaser means any person to whom securities are offered or sold in a Tier 2 offering under Regulation A.⁴ Rule 504 of Regulation Crowdfunding, added by the 2020 exempt offering harmonisation rules and effective 15 March 2021, does the same for offerings made under that regulation.⁵ Both are exemption-specific, and the closer precedent is Regulation Crowdfunding, where a rule-based definition sits on top of exemptive treatment the statute already supplies. Two earlier proposals to define the term by reference to investor status were not adopted. Release No. 33-8041 of 19 December 2001 proposed a Rule 146(c) under which a qualified purchaser would mean any accredited investor as defined in Rule 501(a). It drew opposition from state regulators.⁶ Release No. 33-8828 of 3 August 2007 proposed treating “large accredited investors” as qualified purchasers for offers and sales complying with proposed Rule 507. That proposal was also not adopted.⁷

Two pre-emption proposals in the same year

Regulation Crypto Assets is not the only 2026 rulemaking narrowing state registration authority. On 19 May 2026 the Commission proposed Registered Offering Reform, Release No. 33-11418, which would pre-empt state registration and qualification requirements for all registered offerings. Comments on that proposal closed on 27 July 2026.⁸

What practitioners should watch

Chairman Atkins said in his statement that legislation remains indispensable to rules durable enough to survive “a future rogue regulator”, and that the Commission continues to support passage of the CLARITY Act.⁹ The observation applies equally to the exemptions, the safe harbour and the qualified-purchaser definition. Each would remain an agency rule capable of amendment or repeal by a later Commission. The definition is nevertheless distinctive because, while operative, it reallocates regulatory authority between the Commission and the states.

Two practical consequences follow. Secondary market pre-emption is conditional and capable of lapsing. If an issuer stops meeting the applicable reporting conditions, the proposed federal pre-emption would cease to apply to subsequent secondary transactions in that asset, and otherwise applicable state requirements may again apply, which makes this a continuing diligence obligation for trading venues and holders rather than a single check at issuance.

The definition may also attract judicial review. The statutory question is not whether the Commission may define the term differently for different categories of securities, which it expressly may, but whether this definition fits the concept of a qualified purchaser and whether the Commission has adequately explained its investor-protection rationale.

Notes

1. SEC Press Release 2026-76, “SEC Proposes New Regulation Crypto Assets” (18 August 2026). Proposing release File No. S7-2026-27.

2. SEC Fact Sheet, Regulation Crypto Assets, Release No. 33-11434 (18 August 2026).

3. Securities Act of 1933, s.18(b)(3), 15 U.S.C. § 77r(b)(3), as added by Pub. L. 104-290, s.102(a).

4. 17 CFR § 230.256. Adopted in Amendments for Small and Additional Issues Exemptions Under the Securities Act (Regulation A), Release No. 33-9741, effective 19 June 2015.

5. 17 CFR § 227.504, Regulation Crowdfunding Subpart E. Adopted in Release No. 33-10884, effective 15 March 2021.

6. Defining the Term “Qualified Purchaser” Under the Securities Act of 1933, Release No. 33-8041 (19 December 2001), File No. S7-23-01. Comment period closed 25 February 2002.

7. Revisions of Limited Offering Exemptions in Regulation D, Release No. 33-8828 (3 August 2007), 72 Fed. Reg. 45116 (10 August 2007), File No. S7-18-07. Comments due 9 October 2007. See pp.6-7 and 12-13 of the release, including n.45.

8. Registered Offering Reform, Release No. 33-11418 (19 May 2026), File No. S7-2026-17. Comments closed 27 July 2026.

9. Paul S. Atkins, “Statement on Regulation Crypto Assets: Fit-for-purpose Exemptions for Crypto Market Innovation” (18 August 2026).

Update, 20 August 2026. The Commission issued Regulation Crypto Assets as a proposed rule on 18 August, later the same day this piece was published, and without rescheduling the cancelled meeting. Three statements below were overtaken by that issuance. The substantive content of the proposal is no longer known only from Chair Atkins’s preview. The proposing release described below as never published has been issued as Release Nos. 33-11434 and 34-106150, File No. S7-2026-27. The $5 million and $75 million figures are no longer illustrative. The analysis of Section 11 and of administrative durability is unaffected by the issuance. I have set out the released text, including the state pre-emption provision in Subpart E, in “Issued Without a Meeting”. The text below is left as published on 18 August.

On 13 August, the SEC cancelled the open meeting scheduled for the following morning, at which its three commissioners were to decide whether to publish proposed rules creating a tailored offering regime for certain investment contracts involving crypto assets.¹ The formal notice gave no reason and no replacement date. What disappeared was not a final rule, but the vote needed to expose a proposal to public scrutiny. Everything known about the proposal’s substantive content comes from what Chair Paul Atkins has previewed, not from a published release, and I keep that distinction in view throughout.

A framework with six-year roots

The framework traces to the Token Safe Harbor Proposal that Commissioner Hester Peirce first floated in 2020 and revised in 2021, which would have given a token network a grace period to decentralise before securities rules applied.² Peirce now chairs the SEC’s Crypto Task Force. Atkins’s preview develops her earlier idea. It rests on the interpretation the SEC and CFTC issued on 17 March 2026, which set out a five-part crypto-asset taxonomy and explained that four categories ordinarily comprise non-security crypto assets, while separately addressing digital securities and investment contracts.³ The rulemaking remains under review at the White House Office of Information and Regulatory Affairs as RIN 3235-AN38. Its continued appearance there as a pending proposed rule suggests postponement rather than withdrawal, although the SEC has announced no new date.

The framework previewed by Atkins contemplated three possible routes. A startup exemption would last up to four years, with an illustrative $5 million aggregate cap and whitepaper-style disclosure in place of a registration statement. A fundraising exemption would carry an illustrative $75 million cap in any 12-month period, supported by principles-based disclosure, a discussion of financial condition and financial statements. An investment-contract safe harbour could apply once the issuer had completed or permanently ceased the essential managerial efforts represented or promised under the investment contract. The figures are illustrative because Atkins offered them as examples. The proposing release that would fix them was never published.

Where the remedy stops

The routes carry a cost that the caps obscure. Section 11 of the Securities Act gives a person acquiring a security issued under a registration statement an express claim for a material misstatement or omission in that statement, without requiring proof of scienter.⁴ An exempt offering carries no such claim. Federal antifraud provisions continue to apply, and other federal or state-law remedies might remain available depending on the facts, but none replicates the no-fault registration-statement claim that anchors investor protection in the registered market. Senators Warren and Van Hollen warned in April that the SEC’s direction risked exemptions that undermine decades of protection. Separately, former SEC Chief Accountant Lynn Turner described the CLARITY Act’s parallel statutory exemption framework as “severely deficient”.⁵ While the vote remains postponed, token issuers have the March interpretation but no tailored offering regime, leaving investors with neither the protections contemplated in Atkins’s preview nor a comprehensive statutory framework.

The European contrast

Europe addressed the same subject through legislation. MiCA, Regulation (EU) 2023/1114, has applied in full since 30 December 2024 and requires many offerors of crypto-assets, and persons seeking their admission to trading, to publish a crypto-asset white paper whose defective content can found civil liability under Article 15.⁶ That liability is not the equivalent of Section 11. A holder must establish that the information was not complete, fair or clear, or was misleading; that reliance affected the decision to acquire, sell or exchange the asset; and that loss resulted. Even so, the obligation sits in binding EU legislation. Changing it would require further legislative action. The US is trying to reach a comparable destination by rulemaking because Congress has not passed a statute, and that route is exposed in two ways. Peirce is due to leave in November for Regent University, which would reduce the Commission to two serving members unless another appointment is made.⁷ Any eventual rule would also face a post-Loper Bright court applying its own judgment to disputed questions of statutory authority, without Chevron deference. Neither point prevents rulemaking. Both show how much of the framework rests on administrative rather than legislative durability, since a later Commission could amend or rescind it through further rulemaking.

The near-term calendar

The weeks ahead will show which way the executive branch leans. The CFTC convenes its first Innovation Advisory Committee on 20 August, under the banner “Crypto’s Regulatory Evolution: From Uncertainty to Clarity”, a commodities regulator occupying ground a securities regulator has just stepped back from.⁸ The Digital Asset Market Clarity Act, meanwhile, is reported to face a cloture test on 15 September. Senator Lummis has warned that failure this year could push comprehensive federal legislation into the next Congress.⁹ For issuers and compliance teams the planning assumption should shift. Do not assume publication on the original timetable, watch the status of RIN 3235-AN38 on Reginfo.gov for the first firm signal and price in a securities framework that, if it arrives, can be revised by a later Commission. A statute-shaped problem is being met with an administrative instrument, and that mismatch is where the exposure lies.

Notes

1. SEC, notice of cancellation of the open meeting of 14 August 2026 (sec.gov); scheduling explanation reported by Reuters, 13 August 2026.

2. SEC Commissioner Hester M. Peirce, Token Safe Harbor Proposal (2020); revised proposal (2021).

3. SEC Press Release 2026-30, “SEC Clarifies the Application of Federal Securities Laws to Crypto Assets”, 17 March 2026 (joint SEC and CFTC interpretation).

4. Securities Act of 1933, s.11 (15 U.S.C. s.77k).

5. Senators Elizabeth Warren and Chris Van Hollen, letter to SEC Chair Paul Atkins, 27 April 2026 (US Senate Committee on Banking, Housing, and Urban Affairs, minority), written response requested by 8 May 2026. Lynn Turner, former SEC Chief Accountant, letter to the Senate Banking and Agriculture Committees, 13 January 2026, describing the CLARITY Act framework as “severely deficient”.

6. Regulation (EU) 2023/1114 (MiCA), Articles 4–15; applicable in full from 30 December 2024.

7. Regent University School of Law faculty announcement (Peirce to join November 2026); Bloomberg, 21 May 2026.

8. CFTC Press Release 9279-26; Federal Register notice 2026-16328 (11 August 2026).

9. Majority Leader John Thune filed cloture on the motion to proceed to the Digital Asset Market Clarity Act (H.R. 3633), first procedural vote set for 15 September 2026 (The Block; American Banker, 8 August 2026). Senator Lummis’s timing warning as reported (Tech Times, August 2026).

At 4:52 a.m. Eastern on Saturday 8 August 2026, at the close of an overnight session, Senate Majority Leader John Thune filed cloture on the motion to proceed to H.R. 3633, the Digital Asset Market Clarity Act.¹ The Senate then left for its summer recess. The cloture vote is set for 2:15 p.m. Eastern on Tuesday 15 September, the day after the chamber returns.² That vote will not pass the Act or even place the Senate formally on the bill. It will invoke cloture on the motion to proceed, limiting further debate on the preliminary question whether the chamber should take the legislation up.

That a statute of this consequence should turn on so preliminary a question captures the state of American digital-asset regulation. In eighteen months the United States has assembled most of a market-structure regime. One part is already law. The larger part has passed the House, cleared two Senate committees and been merged into a single text. It now waits on sixty votes to begin a debate.

Two statutes and a hierarchy

The first piece became law in July 2025. On 18 July the President signed the GENIUS Act, the first comprehensive federal statute devoted to a cryptoasset product, which established a licensing and reserve regime for payment stablecoins.³ The Senate had passed it 68 to 30 in June, the House 308 to 122 in July.⁴ Its provisions are presently scheduled to take effect on 18 January 2027, unless final implementing regulations trigger the alternative commencement date of 120 days after their issuance.⁵

GENIUS is narrow by design. A permitted payment stablecoin must be backed one for one by cash or short-dated government assets, and only a subsidiary of an insured depository institution, a federally qualified nonbank issuer or a state-qualified issuer may issue one, with state supervision confined to issuers below ten billion dollars in circulation. Issuers must publish their reserves monthly and disclose a redemption policy.⁶ The statute governs a single instrument, and governs it tightly.

The CLARITY Act is the wider project. The House passed it on 17 July 2025 by 294 to 134, with more than seventy Democrats joining the Republican majority, the broadest cross-party margin any digital-asset bill has drawn in a chamber of Congress.⁷ Where GENIUS governs one instrument, CLARITY attempts to allocate the principal categories of digital asset between federal regimes. It sets out a three-part statutory taxonomy. Digital commodities fall to the Commodity Futures Trading Commission. Assets sold pursuant to an investment contract fall to the Securities and Exchange Commission, though the operative distinction is between the asset and the transaction in which it is sold rather than a fixed property of the token. Permitted payment stablecoins sit in a third category, governed by GENIUS and outside both the securities and the commodity regimes.⁸

The analysis distinguishes where necessary between the House-passed text, the Senate Banking substitute and the proposed merged Senate substitute released in July. The three pursue the same jurisdictional settlement, but their definitions and transition mechanisms are not identical, and the version senators would consider is the proposed merged substitute rather than the bill the House passed.⁹

The taxonomy is the point of the exercise. For most of a decade the classification of a token turned on the Howey test, the investment-contract standard the Supreme Court devised in 1946 for orange groves and citrus-service contracts, applied to crypto case by case through enforcement.¹⁰

Regulation by enforcement

To understand why an industry lobbied for its own regulation, look at the decade that preceded it. Without a statute, classification was worked out in court. In SEC v. Ripple Labs Judge Analisa Torres held in July 2023 that institutional sales of the XRP token were unregistered securities offerings, while programmatic sales to anonymous buyers on exchanges were not, a distinction that turned on what each class of purchaser could have expected rather than on any property of the token itself.¹¹ The ruling satisfied few and bound no other court. Parallel actions against the major exchanges produced further fragments of doctrine, none of them a system.

The ground then shifted beneath the agencies. In Loper Bright Enterprises v. Raimondo the Supreme Court ended mandatory Chevron deference to an agency’s reasonable construction of an ambiguous statute.¹² The decision did not displace Howey, and courts considering SEC crypto cases were already required to decide independently whether the facts satisfied the investment-contract test. Its broader significance is nevertheless structural. An agency interpretation cannot supply the durable jurisdictional settlement that Congress has declined to enact.

The decentralisation pivot

The line between the two contested categories is decentralisation. In the House-passed text a digital commodity is an asset whose value is intrinsically linked to the use of its blockchain, excluding securities, derivatives and stablecoins.¹³ The Senate Banking substitute approaches the securities side through a different construct, the ancillary asset, a token sold with or as part of an investment contract but carrying its own disclosure, disposition and certification machinery rather than a simple relabelling of the House category.¹⁴ The regulatory treatment of transactions involving the same token can change over its life.

The mechanism for that movement, in the House-passed text, is the mature blockchain system. A network qualifies as mature if it is functional, composed of open-source code, governed by pre-established and transparent rules and not subject to the control of any single person or group, including through the holding of twenty per cent or more of the tokens.¹⁵ A token initially sold in a transaction constituting an investment contract may later trade as a digital commodity once the associated network satisfies the applicable decentralisation or maturity test.

The proposed Senate substitute uses a different certification mechanism. It begins with a rebuttable presumption that a network token is an ancillary asset. The token’s originator (or, subject to additional due-diligence conditions, a digital-asset intermediary) may submit evidence certifying that the token is not an ancillary asset. The certification becomes effective if the SEC does not object within sixty days, although the Commission may deny it on finding that the token remains an ancillary asset or carries a disqualifying financial right.¹⁶ The Senate mechanism therefore does not ask simply whether a network is mature. It asks whether the asset remains dependent on the entrepreneurial or managerial efforts that justify ancillary-asset treatment. The burden of initiative still moves towards the promoter or intermediary, but the proposition asserted and the statutory presumption differ materially from those in the House bill.

What the Act would build

Around the taxonomy the Act constructs a market-conduct regime. The proposed substitute creates a notice-of-intent-to-register regime for digital commodity exchanges, brokers and dealers while the CFTC completes its permanent rules. Qualifying firms would be treated as registered for specified purposes while complying with customer-asset, disclosure, financial-responsibility and supervisory requirements. Existing listings could continue until 180 days after the permanent registration rules take effect, subject to the Act’s conditions and delisting machinery.¹⁷

The grant to the CFTC is more novel than it appears. The Commission has long policed manipulation and fraud in commodity spot markets, but it has never held comprehensive registration and supervisory authority over a spot market as the SEC holds it over securities. CLARITY would give it that authority over digital commodity trading for the first time, and would have to fund the supervisory apparatus the role demands.¹⁸ On the securities side, issuers of ancillary assets would make tailored disclosures to the SEC rather than file full registration statements, a lighter regime calibrated to tokens rather than equities. The Act does not abolish securities regulation of crypto. It rescales it.

The text also draws a line around software itself. Its decentralised-finance provisions, carried in part by the Blockchain Regulatory Certainty Act, shield non-custodial software developers from money-transmitter obligations and Bank Secrecy Act requirements, and a separate exclusion exempts validators and open-source publishers from registration.¹⁹ The distinction is between those who hold customer assets and those who write or run code. The first are regulated intermediaries. The second, on the Act’s logic, are not intermediaries at all. That line will bear a great deal of weight, because much of what the market calls decentralised finance sits close to it.

The regulators moved first

While the bill has waited, the agencies have not. On 11 March 2026 SEC Chairman Paul Atkins and CFTC Chairman Michael Selig signed a memorandum of understanding to coordinate on matters of shared concern, the product of a joint initiative the two chairs called Project Crypto.²⁰ The memorandum created a Joint Harmonization Initiative across six workstreams, spanning product definitions, clearing and margin, dual-registration friction, a fit-for-purpose framework for crypto assets, regulatory reporting and cross-market surveillance and enforcement.²¹

Six days later, on 17 March, the two agencies issued a sixty-eight-page joint interpretation of how the federal securities laws apply to certain crypto assets and transactions, announced at the DC Blockchain Summit. It set out a five-part classification of digital commodities, digital collectibles, digital tools, stablecoins and digital securities, and identified sixteen major assets as digital commodities on the basis of their characteristics at the date of the release, among them Bitcoin, Ether, Solana and XRP. The CFTC stated that it would administer the Commodity Exchange Act consistently with the interpretation.²² The regulators have begun to draw the line the legislature has not yet enacted.

The durability of that administrative settlement is the open question. The interpretation does not bind a court and, after Loper Bright, cannot claim mandatory Chevron deference.²³ What the agencies have built by agreement, a later administration can unbuild by agreement. Only a statute can make the basic allocation materially less dependent on the personnel leading the two commissions. That is the argument for passing CLARITY even now that the regulators have acted, and it is the argument its supporters have found hardest to carry against a crowded Senate calendar.

What changes on enactment

For practitioners the practical effect is a change in default posture. A token originator today operates against the risk that the SEC will characterise transactions involving its token as investment-contract transactions after the fact, a risk priced into every launch and listing. Under the proposed substitute, an originator, or in specified circumstances a digital-asset intermediary, can seek to displace the presumption that a network token is an ancillary asset by certifying that the statutory criteria are no longer satisfied. The SEC must object within the statutory period or the certification becomes effective. Exchanges gain a notice-of-intent-to-register status under which they can operate lawfully while the CFTC completes its rules, rather than trading in the shadow of an enforcement theory. Custody, customer-asset segregation and disclosure move from a patchwork of state licensing requirements and agency positions towards a more coherent federal market-conduct regime.

The regime changes the character of legal risk rather than eliminating it. The contested question becomes whether a network token remains an ancillary asset, including whether continuing entrepreneurial or managerial efforts support that treatment, and the certification that answers it is an assertion the issuer must be prepared to defend. An unsuccessful or inaccurate certification provides no safe harbour. The originator or intermediary must defend it through the SEC process on the statutory timetable. A denial constitutes final agency action subject to judicial review. The Act replaces the diffuse uncertainty of classification by enforcement with the sharper, and more manageable, uncertainty of a defined test applied to particular facts.

The politics of a motion to proceed

The Act’s difficulty is not conceptual. It is political, and it is concentrated in provisions that have little to do with token categorisation. There is a paradox in the timing. The March interpretation, by giving the market much of the certainty it sought, has drained some of the urgency from the legislative effort, and an industry that has what it needs administratively lobbies less hard for the statute that would make it permanent. The Senate Agriculture Committee advanced its portion in January 2026, and the Senate Banking Committee reported its version by 15 to 9 on 14 May.²⁴ On 22 July, Senate Republicans released a proposed merged substitute running to 616 pages, joining the Banking market-structure framework to the Agriculture commodity provisions and adding new titles on government ethics and law enforcement.²⁵

The ethics title is the obstacle. Senate Democrats have conditioned their support on a conflict-of-interest rule restraining officials, the President included, from personal crypto interests. The Senate Banking Committee’s Democratic minority staff calculates, from the President’s 2025 financial disclosure, that his crypto ventures generated more than 1.4 billion dollars in income that year.²⁶ The merged text answered with a provision that bars officials from issuing or sponsoring digital assets but permits them to invest, and that sunsets in 2029, so that it expires with the current presidential term. Seven of the Senate’s most pro-crypto Democrats rejected it in a joint statement.²⁷ A second front concerns deposit flight. Banks have warned that stablecoin reward programmes could draw potentially substantial deposits out of the banking system, and at least one Republican senator has withheld support until the point is addressed.²⁸

The arithmetic follows from the shape of the chamber. Cloture requires sixty votes. Republicans hold fifty-three seats. If every Republican votes to proceed, at least seven Democrats or independents must join, and the three unresolved questions of ethics, illicit finance and the folding-in of the Agriculture text all remain open going into the vote.²⁹ Galaxy Research, an industry analyst, has cut its estimate of the odds that the Act becomes law this year from fifty to thirty per cent, a projection rather than a settled fact.³⁰ The calendar is unforgiving. The Senate returns on 14 September, and attention turns quickly to the November midterms.

The view from London and Brussels

The contrast with Europe and the United Kingdom sharpens the design question. The European Union built its market structure, the Markets in Crypto-Assets Regulation, as a single passportable regime under which one authorisation lets a firm serve all twenty-seven member states, and the final EU-wide transitional period expired on 1 July 2026, though individual member states could shorten or decline the grandfathering they offered.³¹ The United Kingdom has taken a third path. Parliament made the Financial Services and Markets Act 2000 (Cryptoassets) Regulations 2026 on 4 February 2026, the Financial Conduct Authority gateway opens on 30 September 2026 and closes on 28 February 2027 ahead of a mandatory regime expected to begin on 25 October 2027, and the Bank of England took oversight of systemic sterling stablecoins on a standard backing of up to seventy per cent short-dated government debt and at least thirty per cent unremunerated central-bank deposits, with a systemic-at-launch step-up under which a new issuer may initially hold up to ninety-five per cent in eligible government debt.³²

Each jurisdiction has answered the same question differently. Europe created one rulebook and one licence. Britain divided the work between a conduct regulator and a financial-stability regulator within a single statutory perimeter. The United States has split authority between two established agencies with a long history of rivalry, joined by a decision tree. The American design keeps the market expertise and the enforcement records of the SEC and the CFTC. It also imports the boundary disputes a single regime avoids by definition.

The classification problem, relocated

What CLARITY resolves is the turf question. After enactment there would be a statutory answer to which agency supervises which asset, and a defined path by which an asset passes from one to the other. That is more than a decade of enforcement produced, and it should not be understated.

What the Act does not resolve is the classification problem itself. Howey asked whether the buyers of an instrument were led to expect profits from the efforts of others, and left the answer to be found in the facts. The House bill asks whether a blockchain has become sufficiently mature and decentralised. The proposed Senate substitute asks a related but different question, whether a network token remains an ancillary asset because its value still depends on an originator’s entrepreneurial or managerial efforts. In both versions the proponent may initiate a certification process, but the presumption, the evidence required and the status being certified are not the same. Whether either certification model proves more determinate than Howey, or merely relocates its factual ambiguity, is the question the legislation leaves open. The experiment cannot begin until the Senate decides, on 15 September, whether to take the Act up at all.

Notes

1. Senate Daily Press floor log, Friday 7 August 2026 (recording the 4:52 a.m. cloture filing on the motion to proceed to H.R. 3633), dailypress.senate.gov; reported in Zack Abrams, “Majority Leader Thune files cloture on Clarity Act, setting up Sept. 15 Senate vote,” The Block, 8 August 2026. Available at: The Block, 8 Aug 2026; Senate Daily Press.

2. Cloture vote on the motion to proceed scheduled for 2:15 p.m. Eastern, Tuesday 15 September 2026, the day after the Senate reconvenes on Monday 14 September. U.S. Senate Daily Press floor log, 7 August 2026 (adjournment to 3 p.m. Monday 14 September); U.S. Senate floor activity records; The Block, 8 August 2026. Available at: Senate floor activity.

3. GENIUS Act, Pub. L. No. 119-27, 139 Stat. 419, signed 18 July 2025. “Fact Sheet: President Donald J. Trump Signs GENIUS Act into Law,” The White House, 18 July 2025; Sullivan & Cromwell LLP, “GENIUS Act Enacted,” July 2025. Available at: GovInfo (Pub. L. 119-27); White House fact sheet; Sullivan & Cromwell.

4. GENIUS Act, S. 1582, 119th Congress; Senate passed 68–30 on 17 June 2025, House passed 308–122 on 17 July 2025. Congress.gov, S. 1582. Available at: Congress.gov, S. 1582; Covington & Burling.

5. GENIUS Act commencement: presently scheduled for 18 January 2027 (eighteen months after enactment), unless final implementing regulations issued by the primary federal payment-stablecoin regulators trigger the alternative date of 120 days after issuance. As at 14 August 2026 no such final regulations had been issued. Congressional Research Service, IN12553; FDIC, proposed rule, GENIUS Act Requirements and Standards for FDIC-Supervised Permitted Payment Stablecoin Issuers and Insured Depository Institutions, 91 Fed. Reg. 18534, 10 April 2026. Available at: CRS IN12553; FDIC proposed rule.

6. Congressional Research Service, “Stablecoin Legislation: An Overview of S. 1582, GENIUS Act of 2025,” IN12553; FDIC, proposed rule, “GENIUS Act Requirements and Standards for FDIC-Supervised Permitted Payment Stablecoin Issuers and Insured Depository Institutions,” 91 Fed. Reg. 18534, 10 April 2026. Available at: CRS IN12553; FDIC proposed rule.

7. H.R. 3633 passed the House on 17 July 2025 by 294–134, Roll Call 199; more than seventy Democrats voted in favour. U.S. House Clerk, Roll Call 199 (2025). Available at: House Clerk, Roll Call 199.

8. H.R. 3633 (as passed by the House). On the asset/transaction distinction, see the SEC–CFTC joint interpretation, SEC Release Nos. 33-11412 and 34-105020 (17 March 2026), and Congressional Research Service, IN12583. Available at: SEC Release Nos. 33-11412 and 34-105020; CRS IN12583.

9. Three iterations are in play: the House-passed text, H.R. 3633 (17 July 2025); the Senate Banking Committee substitute (market-structure discussion draft, May 2026, banking.senate.gov); and the proposed merged Senate substitute of 22 July 2026, drafted as an amendment in the nature of a substitute intended to be proposed to H.R. 3633 (the Lummis substitute, lummis.senate.gov), not yet offered or adopted on the floor. Propositions in the notes below identify the governing version where the texts diverge. Available at: Senate Banking substitute; Lummis substitute.

10. SEC v. W. J. Howey Co., 328 U.S. 293 (1946).

11. SEC v. Ripple Labs, Inc., No. 20-cv-10832 (S.D.N.Y. 13 July 2023) (Torres, J.), holding institutional sales of XRP to be unregistered securities offerings while programmatic exchange sales were not.

12. Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024), overruling Chevron U.S.A. Inc. v. Natural Resources Defense Council, 467 U.S. 837 (1984). The decision ended mandatory deference to reasonable agency constructions of ambiguous statutes; it did not displace the Howey investment-contract test, which courts apply independently.

13. House-passed text, H.R. 3633. The digital commodity definition turns on value intrinsically linked to the use of the blockchain and excludes securities, derivatives and stablecoins. Available at: H.R. 3633 text.

14. Senate Banking Committee substitute (May 2026 discussion draft, banking.senate.gov). The ancillary-asset construct carries its own disclosure, disposition and certification provisions and is not a mere renaming of the House investment-contract-asset category. Available at: Senate Banking substitute.

15. House-passed text, H.R. 3633. The four-part maturity test requires a functional network, open-source code, pre-established transparent rules and the absence of unilateral control, including through holdings of twenty per cent or more of the tokens. Available at: H.R. 3633 text.

16. Proposed Lummis substitute, §10103(c)(5), pp. 29–36. A network token is rebuttably presumed to be an ancillary asset unless its originator, or a qualifying digital-asset intermediary, certifies with reasonable supporting evidence that it is not. The certification becomes effective upon SEC non-objection or, absent a timely rebuttal, sixty days after submission. The SEC may deny the certification on finding that the asset is an ancillary asset or includes a disqualifying financial right. This mechanism differs from the mature-blockchain certification in §205 of H.R. 3633 as passed by the House. A denial constitutes final agency action reviewable under applicable law. Available at: Lummis substitute.

17. Proposed Lummis substitute: a notice-of-intent-to-register regime for digital commodity exchanges, brokers and dealers. The CFTC must establish the filing process within thirty days of enactment, and existing listings may continue until 180 days after the permanent registration rules take effect, subject to the Act’s conditions and delisting machinery. The four-year provisional-registration sunset appears in H.R. 3633 as passed by the House, not in the proposed substitute. Available at: Lummis substitute.

18. H.R. 3633; Congressional Research Service, IN12583. The Act extends the CFTC’s remit beyond its traditional derivatives and anti-manipulation jurisdiction to registration and supervision of digital commodity spot markets. Available at: CRS IN12583.

19. Proposed merged Senate substitute of 22 July 2026 (Lummis substitute), incorporating the Blockchain Regulatory Certainty Act and a decentralised-finance exclusion for validators and open-source publishers; see also “What is in the merged CLARITY Act text, and what changed,” Cryptonews, 22 July 2026. Available at: Lummis substitute; Cryptonews.

20. SEC–CFTC Memorandum of Understanding, 11 March 2026, signed by SEC Chairman Paul Atkins and CFTC Chairman Michael Selig; Norton Rose Fulbright and Global Fintech & Digital Assets Blog client analyses, March 2026. Available at: Norton Rose Fulbright; Global Fintech & Digital Assets Blog.

21. The Joint Harmonization Initiative spans six workstreams and is co-led by Robert Teply for the SEC and Meghan Tente for the CFTC. Global Fintech & Digital Assets Blog, March 2026. Available at: Global Fintech & Digital Assets Blog.

22. SEC Release Nos. 33-11412 and 34-105020 (17 March 2026). The release identifies sixteen assets as examples of digital commodities on the basis of their characteristics at the date of the release: APT, AVAX, BTC, BCH, ADA, LINK, DOGE, ETH, HBAR, LTC, DOT, SHIB, SOL, XLM, XTZ and XRP. The CFTC has stated it will administer the Commodity Exchange Act consistently with the interpretation. Available at: SEC Release Nos. 33-11412 and 34-105020; Jenner & Block; Ropes & Gray.

23. The interpretation does not bind a court and, after Loper Bright (note 12), cannot claim mandatory Chevron deference; it may be revised by a later Commission.

24. Senate Committee on Banking, Housing, and Urban Affairs, committee action of 14 May 2026 (reported 15–9); Senate Committee on Agriculture, Nutrition, and Forestry action, January 2026. See also CoinDesk, 14 May 2026. Available at: CoinDesk, 14 May 2026; Elliptic.

25. Proposed merged Senate substitute released 22 July 2026 (616 pages), drafted as an amendment in the nature of a substitute (Lummis substitute), adding government-ethics and law-enforcement titles; “New Clarity Act emerges that’s a start on the final draft,” CoinDesk, 22 July 2026. Available at: CoinDesk, 22 July 2026.

26. Senate Committee on Banking, Housing, and Urban Affairs (Democratic minority), materials on the President’s crypto income and requests for updated financial disclosures, 2026, banking.senate.gov/newsroom/minority. The $1.4 billion figure is the minority staff’s calculation from the President’s 2025 financial disclosure and is a contested political estimate. Available at: Senate Banking minority; TechTimes.

27. Joint statement of seven pro-crypto Senate Democrats opposing the July ethics provision, August 2026; reported in “Crypto Democrats Are Ghosting the Industry,” The American Prospect, 10 August 2026. Available at: The American Prospect.

28. Bank-industry submissions warning that third-party stablecoin reward programmes could draw substantial deposits from insured institutions; Senator Josh Hawley has conditioned support on the point. The Block, August 2026. Available at: The Block, 8 Aug 2026.

29. U.S. Senate Standing Rule XXII (cloture requires three-fifths of senators duly chosen and sworn, ordinarily sixty votes); Republicans hold fifty-three seats. The outstanding issues are ethics, illicit finance and integration of the Senate Agriculture Committee text. The Block, 8 August 2026. Available at: The Block, 8 Aug 2026.

30. Galaxy Research probability estimate, revised from fifty to thirty per cent in mid-2026, reported by The Block. This is one analyst’s projection, not an independently verifiable fact. Available at: The Block, 8 Aug 2026.

31. Markets in Crypto-Assets Regulation (EU) 2023/1114. The final EU-wide transitional period expired on 1 July 2026, though member states could shorten or decline the grandfathering they offered. ESMA, “Statement on the end of transitional periods under MiCA,” 2026. Available at: ESMA statement.

32. Financial Services and Markets Act 2000 (Cryptoassets) Regulations 2026, SI 2026/102, made 4 February 2026, legislation.gov.uk. FCA authorisation gateway 30 September 2026 to 28 February 2027, mandatory regime expected 25 October 2027 (FCA press release, 2026). Bank of England, policy statement on sterling-denominated systemic stablecoins, June 2026 (standard backing up to 70% short-dated government debt and at least 30% unremunerated central-bank deposits, with a systemic-at-launch step-up to 95% eligible government debt). Available at: SI 2026/102; FCA press release; Bank of England policy statement.

References

Legislation and instruments

Digital Asset Market Clarity Act, H.R. 3633, 119th Congress (House-passed 17 July 2025; Senate Banking substitute, May 2026; proposed merged Senate substitute (Lummis substitute), 22 July 2026).

GENIUS Act, Pub. L. No. 119-27, 139 Stat. 419 (2025).

Markets in Crypto-Assets Regulation (EU) 2023/1114.

Financial Services and Markets Act 2000 (Cryptoassets) Regulations 2026, SI 2026/102 (made 4 February 2026).

Case law

SEC v. W. J. Howey Co., 328 U.S. 293 (1946).

SEC v. Ripple Labs, Inc., No. 20-cv-10832 (S.D.N.Y. 2023).

Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024).

Regulatory and official sources

SEC and CFTC, Application of the Federal Securities Laws to Certain Types of Crypto Assets and Certain Transactions Involving Crypto Assets, SEC Release Nos. 33-11412 and 34-105020, 17 March 2026.

SEC–CFTC Memorandum of Understanding and Joint Harmonization Initiative, 11 March 2026.

U.S. Senate Daily Press floor log, 7 August 2026, and Senate floor activity records.

Senate Committee on Banking, Housing, and Urban Affairs (majority and minority) and Senate Committee on Agriculture, Nutrition, and Forestry records, 2026.

FDIC proposed stablecoin rule, 91 Fed. Reg. 18534, 10 April 2026; Congressional Research Service, IN12583 and IN12553.

Financial Conduct Authority cryptoasset regime materials, 2026; Bank of England policy statement on systemic sterling stablecoins, June 2026; ESMA statement on MiCA transitional periods, 2026.

In January 2025, the Texas Attorney General commenced proceedings against Allstate Corporation and its data subsidiary Arity International LLC under the Texas Data Privacy and Security Act. The complaint alleged that Arity had collected trillions of miles of location data from approximately 45 million consumers through software trackers embedded in third-party mobile applications, without adequate disclosure or consent. That data was then sold to insurers for use in premium pricing. A parallel class action survived a motion to dismiss in March 2026.¹ The proceedings represent the first enforcement action under the TDPSA and a further major regulatory challenge to the telematics data supply chain now used in algorithmic insurance pricing across the United States.

Telematics data is only one input among many. Insurers and their data vendors increasingly draw on credit-based information, postcode-level demographic data, telematics, app-derived behavioural signals and other external consumer data to inform underwriting and pricing. The result is a pricing architecture in which traditional actuarial variables coexist with machine learning models trained on consumer data of a breadth and granularity that existing regulatory frameworks were not designed to address. Regulators, courts and legislatures are responding across multiple jurisdictions, but they are doing so unevenly, with different legal traditions and different conceptions of what constitutes unlawful discrimination in insurance underwriting.

Insurance is becoming both a deployer of AI and a market for the liabilities AI creates. That dual role is producing a regulatory structure in which underwriting, discrimination law, privacy enforcement, product liability and insurance coverage are beginning to converge.

Algorithmic Pricing and Proxy Discrimination

The central risk in algorithmic insurance pricing is proxy discrimination. An insurer may exclude race, ethnicity or disability from its pricing model and still produce discriminatory outcomes if the model relies on variables that correlate with those protected characteristics. Credit-based insurance scores, postcode-level data, telematics records and app-derived behavioural signals can each serve as proxies for characteristics that insurers are prohibited from using directly. The difficulty is that the correlation is often invisible to the insurer itself. A model optimised for predictive accuracy may learn to weight variables that are statistically associated with protected characteristics without any explicit instruction to do so.

The mechanism is well documented in adjacent sectors. In 2024, SafeRent Solutions agreed to a class-action settlement of approximately $2.3 million in Louis v SafeRent Solutions, after plaintiffs alleged that its tenant-screening algorithm disproportionately harmed Black and Hispanic rental applicants using housing vouchers. The Department of Justice was not the claimant, although it had earlier filed a statement of interest on the application of the Fair Housing Act to algorithmic screening.² In July 2025, the Massachusetts Attorney General secured a $2.5 million settlement from Earnest Operations LLC after an investigation found that the company’s algorithmic student lending model incorporated variables that served as proxies for race, resulting in less favourable terms for borrowers of colour.³

The pattern is now visible. Algorithmic discrimination claims are succeeding where claimants or regulators can show that facially neutral variables produce disparate outcomes along the lines of protected characteristics. The analytical framework is transferable to insurance. The same data sources that created liability in housing and lending are now embedded in insurance pricing models across personal lines. No insurance commissioner has yet imposed a fine specifically for algorithmic pricing discrimination. The enforcement gap is narrowing. The question is when the first action arrives, not whether it will.

The US Regulatory Response

Three regulatory developments define the current US position.

Colorado was the first state to legislate directly on algorithmic discrimination in insurance. Senate Bill 21-169, enacted in 2021, prohibited insurers from using external consumer data sources, algorithms or predictive models in a manner that unfairly discriminates on protected grounds. Regulation 10-1-1 initially focused on life insurers. Amendments adopted in August 2025 extended the governance and risk-management framework to private passenger automobile insurers and health benefit plan insurers, with line-specific compliance timing.⁴ The compliance obligation is granular. Insurers must identify the external data sources and algorithms used in underwriting, test for disparate impact across protected characteristics and document the results. Separately, Colorado’s broader AI regime has changed again. SB 24-205 was first deferred and then superseded by SB 26-189, signed in May 2026, which recasts the state’s framework around automated decision-making technology and creates obligations beginning on 1 January 2027. That broader regime will intersect with the insurance-specific requirements, but the operative structure is no longer the original SB 24-205 high-risk AI framework.

The New York Department of Financial Services issued Circular Letter No. 7 in 2024, setting out supervisory expectations for insurers using external consumer data and information sources, algorithms and predictive models in underwriting and pricing. The Circular Letter requires multi-step proxy-discrimination testing using a disproportionate-impact, justification and less-discriminatory-alternative methodology, documented governance, contractual controls over vendors and detailed adverse-decision explanations. It also creates a specific 15-day notice obligation where an applicant cannot be underwritten through an External Consumer Data and Information Sources/Algorithmic and Innovative Systems (ECDIS/AIS) process and must instead proceed through a non-ECDIS/AIS process.⁵ The DFS approach is notable for its specificity. It prescribes the testing methodology by which compliance is to be demonstrated and extends the governance obligation to the insurer’s supply chain. A carrier that purchases pricing data from a vendor such as Arity is expected to hold contractual rights sufficient to audit the vendor’s data collection and modelling practices.

Through the state-regulatory coordination process, the National Association of Insurance Commissioners published a Model Bulletin on the Use of Artificial Intelligence Systems by Insurers on 4 December 2023. The Model Bulletin sets supervisory expectations that insurers maintain a written AI programme, conduct impact assessments and ensure human oversight of algorithmic decisions. More than 24 states have adopted or issued guidance based on the Model Bulletin.⁶ A 12-state pilot examination tool, designed to standardise supervisory review of insurer AI systems, commenced in March 2026 and is expected to run until September 2026, with formal adoption anticipated in the autumn. Once adopted, the examination tool is likely to become the baseline standard against which state regulators assess insurer AI compliance.

The US regulatory framework is therefore layered. State-level legislation, supervisory circulars and model standards adopted through the National Association of Insurance Commissioners (NAIC) process operate concurrently, with varying degrees of enforcement rigour and different definitions of key terms. An insurer writing business across multiple states faces a patchwork of obligations that differ in scope, in the specificity of testing requirements and in the consequences of non-compliance.

The EU Framework

The European Union has taken a different approach, embedding insurance AI within a horizontal regulatory architecture rather than addressing it through sector-specific legislation.

The EU AI Act classifies AI systems used for risk assessment and pricing in relation to natural persons in life and health insurance as high-risk under Annex III, point 5(c).⁷ High-risk classification triggers conformity assessments, risk management systems, data governance requirements, transparency obligations and human oversight. Under Regulation (EU) 2026/1744, the Digital Omnibus on AI, stand-alone Annex III high-risk AI obligations will apply from 2 December 2027, while obligations for high-risk AI systems embedded in regulated products will apply from 2 August 2028. Insurers operating in the EU therefore have a defined implementation window, but the obligations themselves remain substantial and require significant investment in documentation, testing and governance infrastructure. An insurer that uses a third-party AI model for pricing will need to satisfy itself that the model meets the conformity requirements, even if the insurer did not develop the model.

The AI Liability Directive, which would have established a harmonised framework for civil liability arising from AI systems, was withdrawn in October 2025. Its withdrawal leaves the liability position fragmented across national tort regimes. The revised Product Liability Directive (2024/2853) partially fills the gap. The Directive extends the definition of “product” to include software and AI systems, and it introduces a concept of defect that applies to self-learning systems whose behaviour changes after placing on the market.⁸ The revised Product Liability Directive is relevant where an AI system causes damage falling within the Directive’s scope, including defects in software or self-learning systems. It is less likely to serve as the principal civil-liability route for discriminatory insurance pricing, which will more naturally arise through the AI Act, sectoral insurance supervision, equality law, consumer protection, data protection and national tort regimes. Member states must transpose the Directive by December 2026.

The European Insurance and Occupational Pensions Authority published an Opinion on AI Governance in August 2025, setting out supervisory expectations for national competent authorities.⁹ The EIOPA Opinion adopts a risk-based and proportionate approach, emphasising that the intensity of supervisory scrutiny should reflect the materiality of the AI system’s impact on policyholders. It does not create directly binding obligations on insurers but signals the direction of supervisory engagement across the single market.

The EU position therefore combines ex ante classification under the AI Act, product liability exposure under the revised Directive and supervisory expectations through EIOPA. The withdrawal of the AI Liability Directive creates a gap in the harmonised civil liability framework that national courts and legislatures may address differently across member states.

The UK Position

The United Kingdom has no AI-specific insurance regulation. The Financial Conduct Authority has relied on existing conduct rules, principally the General Insurance Pricing Practices rules (PS21/5), which require insurers to offer renewal prices no higher than equivalent new business prices, and the Consumer Duty (introduced in July 2023), which imposes a general obligation to deliver good outcomes for retail customers.¹⁰

The Consumer Duty is broad enough in principle to capture algorithmic pricing discrimination. If an insurer’s pricing model produces systematically worse outcomes for customers who share a protected characteristic, there is a credible argument that the insurer has failed to deliver the good outcomes the Duty requires. The FCA has acknowledged that bias in algorithmic decision-making is a “live issue” for the insurance sector. The Treasury Committee has called on the FCA to produce specific guidance on AI in financial services by the end of 2026.

The UK approach is reactive rather than prescriptive. The legal tools to challenge algorithmic discrimination in insurance pricing exist within the current framework. What is absent is the specific regulatory guidance that would tell insurers how to test for proxy discrimination, what governance structures to maintain and what disclosures to make to affected consumers. The contrast with the New York DFS Circular Letter is instructive. An insurer operating in New York has a prescribed testing methodology, defined vendor audit obligations and a specific 15-day notice obligation where an applicant cannot be underwritten through an ECDIS/AIS process. An insurer operating in the United Kingdom has a general duty to produce good outcomes and an acknowledgement from the regulator that bias is a live issue. The gap between the two positions is one that regulated firms and their advisers must manage without regulatory guidance.

AI in Healthcare Coverage Decisions

The application of AI to coverage determinations in health insurance raises distinct concerns that go beyond pricing.

In Estate of Lokken v UnitedHealth Group Inc., proceedings before the United States District Court for the District of Minnesota, the plaintiffs alleged that UnitedHealth used an algorithm called nH Predict to deny post-acute care benefits to Medicare Advantage members. The plaintiffs allege that internal data showed the algorithm’s denial recommendations were overturned on appeal at a rate of approximately 90 per cent, and that UnitedHealth continued to use the system despite knowledge of this error rate. UnitedHealth disputes the allegations. The Court ordered discovery in March 2026, permitting the plaintiffs to obtain documents concerning the development, design, creation, approval, implementation, use and oversight of nH Predict, while not requiring production of the data, rules, source code or medical guidelines on which the system was based.¹¹

The case is significant for two reasons. First, it moves the algorithmic accountability inquiry from pricing to coverage, an area where the consequences of an erroneous decision are immediate and potentially irreversible for the policyholder denied treatment. Second, the discovery order opens the development, deployment and oversight of the algorithm to judicial scrutiny in a way that ordinary regulatory examination has not yet achieved. If the litigation proceeds to trial, it is likely to produce the most detailed judicial analysis of an insurer’s AI system to date.

California enacted SB 1120 in 2024, effective from January 2025, requiring health plans and insurers using AI, algorithms or software tools in utilisation review to preserve human clinical decision-making, use patient-specific clinical information and prevent discriminatory application. The statute also requires disclosure of AI use in utilisation-review policies and procedures. It goes beyond transparency: it imposes substantive constraints on how AI may be used in coverage determinations, including a prohibition on denying care based solely on algorithmic output.

Insuring AI

As insurers adopt AI in their own operations, a parallel market is developing to insure the liabilities that AI systems create for their deployers across all sectors.

Armilla AI, in partnership with Chaucer, launched an AI liability insurance product in April 2025. Munich Re markets aiSure, a suite of AI performance insurance products. The Artificial Intelligence Underwriting Company (AIUC) was reported in 2026 to offer up to $50 million of product-liability coverage for AI agent failures, backed by Beazley paper. Testudo, a managing general agent specialising in AI liability, launched in January 2026.¹² These new entrants are building products for a risk category that the traditional insurance market has been slow to address.

Standard commercial general liability and professional indemnity policies were not drafted with algorithmic decision-making in mind. Exclusions for technology errors, data processing failures and software performance are common and may operate to deny coverage for losses caused by AI systems. The traditional market response has been twofold. First, absolute AI exclusions are appearing in professional indemnity and technology errors and omissions policies, removing any ambiguity about whether existing wordings respond to AI-related claims. Second, insurers are eliminating “silent AI” exposure by expressly addressing AI in policy wordings, ensuring that coverage is either affirmatively granted and priced or explicitly excluded.

The market is splitting. Specialist providers are writing affirmative AI liability coverage at rates that reflect the novelty and uncertainty of the risk. Traditional insurers are restricting or excluding coverage for the same risk. Organisations that identify the gap and purchase specialist cover will have protection. Those that assume their existing policies respond to AI-related claims may discover at the point of loss that they do not. The timing is acute. Regulatory obligations are increasing, litigation risk is rising and the insurance market that would ordinarily absorb those risks is simultaneously repricing and restricting.

Strategic Outlook

The regulatory framework for AI in insurance is fragmented across jurisdictions, between sectoral and horizontal regulation and between legislative rules and supervisory expectations. That fragmentation creates compliance cost for multinational insurers and uncertainty for consumers whose rights depend on the jurisdiction in which their policy is written.

Three consequences follow.

First, proxy discrimination is likely to be the primary litigation vector for the next several years. The analytical framework is established. Enforcement precedents from housing and lending are accumulating. The data sources that create proxy discrimination risk are becoming more central to insurance pricing models. The absence of an insurance-specific enforcement action to date reflects regulatory caution and the difficulty of demonstrating algorithmic causation, not regulatory acceptance of discriminatory outcomes.

Second, the structural gap between the US and EU approaches is unlikely to close. The EU has opted for ex ante classification and conformity assessment under the AI Act. The US is developing a mixture of state-level legislation, supervisory guidance and enforcement actions through litigation. The UK has adopted neither approach in a meaningful form. Each has limitations. Ex ante classification imposes compliance cost regardless of whether a particular system creates risk. State-level variation creates regulatory arbitrage opportunities. Reliance on existing conduct rules leaves regulated firms without clear standards and exposes them to enforcement actions based on rules that were drafted before the technology they are applied to existed.

Third, the AI liability insurance market is developing faster than the regulatory framework it is responding to. If coverage contraction in the traditional market accelerates, organisations deploying AI in insurance and adjacent sectors may face a period in which meaningful coverage is available only from specialist providers at premium rates that reflect the absence of actuarial data on AI-related losses. The convergence of rising regulatory obligations, increasing litigation risk and tightening insurance coverage creates a compounding exposure that boards and risk committees should be addressing now.

The period between mid-2026 and the end of 2027, when the NAIC examination tool is expected to reach formal adoption and stand-alone Annex III obligations under the EU AI Act begin to apply under the Digital Omnibus timetable, is likely to define the structural terms on which AI in insurance operates for the following decade. The regulatory choices made in that window will determine whether algorithmic underwriting remains a competitive advantage available to insurers with robust governance or becomes a source of systemic liability for those without it.

Notes

¹ Texas v Allstate Corp. and Arity International LLC, proceedings commenced January 2025 in the State of Texas. The complaint alleged violations of the Texas Data Privacy and Security Act (TDPSA). A parallel class action survived a motion to dismiss in March 2026.

² Louis v SafeRent Solutions LLC, class-action settlement 2024, approximately $2.3 million. The plaintiffs alleged that SafeRent’s tenant-screening algorithm disproportionately harmed Black and Hispanic rental applicants using housing vouchers. The Department of Justice filed a statement of interest on the application of the Fair Housing Act to algorithmic screening but was not the claimant.

³ Massachusetts AG v Earnest Operations LLC, settlement July 2025, $2.5 million. The Massachusetts Attorney General’s investigation found that Earnest’s algorithmic student lending model used variables that served as proxies for race.

⁴ Colorado SB 21-169 (2021); Regulation 10-1-1 amended August 2025 to extend the governance and risk-management framework to private passenger automobile insurers and health benefit plan insurers, with line-specific compliance timing. Colorado SB 26-189, signed May 2026, superseded the earlier SB 24-205 framework and creates broader automated decision-making technology obligations from 1 January 2027.

⁵ New York Department of Financial Services, Circular Letter No. 7 (2024), on the use of artificial intelligence systems and external consumer data and information sources in insurance underwriting and pricing.

⁶ National Association of Insurance Commissioners, Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, 4 December 2023. A 12-state pilot examination tool commenced March 2026, with formal adoption expected autumn 2026.

⁷ Regulation (EU) 2024/1689 (the AI Act), Annex III, point 5(c). Regulation (EU) 2026/1744, the Digital Omnibus on AI, entered into force on 27 July 2026 and provides that stand-alone high-risk AI obligations apply from 2 December 2027, while high-risk AI systems embedded in regulated products apply from 2 August 2028. The AI Liability Directive was formally withdrawn in October 2025.

⁸ Directive 2024/2853 on liability for defective products (revised Product Liability Directive). Member state transposition deadline December 2026. The Directive extends the definition of “product” to include software and AI systems and introduces a concept of self-learning defect extending beyond placing on the market.

⁹ European Insurance and Occupational Pensions Authority, Opinion on AI Governance, August 2025.

¹⁰ FCA, PS21/5, General Insurance Pricing Practices (2021). The Consumer Duty was introduced in July 2023. The FCA has acknowledged algorithmic bias as a “live issue”. The Treasury Committee has called for AI-specific guidance by the end of 2026.

¹¹ Estate of Lokken v UnitedHealth Group Inc. (D. Minn.). The complaint alleged that UnitedHealth’s nH Predict algorithm denied post-acute care to Medicare Advantage members with an approximately 90 per cent error rate on appeal. Discovery was ordered in March 2026.

¹² Armilla AI, in partnership with Chaucer, launched an AI liability insurance product in April 2025. Munich Re markets aiSure as an AI performance insurance suite. The Artificial Intelligence Underwriting Company (AIUC) was reported in 2026 to offer up to $50 million of product-liability coverage for AI agent failures, backed by Beazley paper. Testudo, a specialist AI liability MGA, launched in January 2026.

A class action now before the Southern District of New York gathers two questions that the law of digital assets has not yet answered. The first is whether a project can be held to account for advertising machine autonomy it did not possess. The second is whether a decentralised autonomous organisation is a defendant that can be served, held to answer and made to pay. Pikabea v. Walters puts both in issue at once. The procedural skirmishing in its opening months shows why neither is straightforward.¹

The complaint

The AI16Z token launched on the Solana blockchain in October 2024. It was marketed around an autonomous artificial intelligence agent said to make its own investment decisions and to manage an on-chain treasury. It also borrowed the cachet of Andreessen Horowitz, the venture firm whose a16z branding the project name evoked without any affiliation.² By early January 2025 the token carried a market capitalisation of about $2.6 billion.³ It then collapsed. The token was later rebranded ELIZAOS.

Gorka Pikabea, a buyer resident in Spain, commenced proceedings in April 2026 on behalf of a proposed class of purchasers between 24 October 2024 and 20 April 2026. The complaint identifies at least 3,945 wallet addresses said to have sustained losses.⁴ Its central factual allegation is arresting in its simplicity. The autonomous agent, it says, was operated by hand. The machine agency that gave the token its narrative was, on the plaintiff’s account, a person.

The pleaded causes of action are reported as deceptive acts and false advertising under New York General Business Law, unfair competition and false advertising under California law, negligent misrepresentation and unjust enrichment.⁵ The named defendants are Eliza Labs, Inc., its founder Shaw Walters, Sebastian Quinn-Watson, the ai16z DAO itself and fifty unnamed persons. The relief sought includes damages, restitution and disgorgement. Every allegation remains unproven. A motion to dismiss was made in June 2026.⁶

AI-washing as a pleaded wrong

The choice of causes of action is the first point of interest. The plaintiff has not led with the federal securities laws. The wrong is framed as deception. The project said the product did something it did not do. That framing places the case in a lineage that predates the token boom and reaches well beyond it.

In March 2024 the Securities and Exchange Commission brought its first enforcement actions for what it called AI-washing, settling charges against two investment advisers, Delphia (USA) Inc. and Global Predictions Inc., for false and misleading statements about their use of artificial intelligence. Delphia had claimed to use an algorithm to analyse client data that it had not built. Global Predictions had described itself as the first regulated AI financial adviser. The firms paid $400,000 in total penalties without admitting or denying the findings.⁷ The regulatory point was that the misrepresentation of artificial intelligence is an old wrong wearing new clothes. A false statement about capability is actionable whether the capability claimed is human or machine.

Pikabea takes that logic into private litigation and into the crypto market. If the pleaded facts are made out, the deception is not in the token mechanics or the price action. It is in the representation of autonomy. That reframing carries consequences for proof and for reach. Consumer-protection and false-advertising claims turn on statements and their effect on the audience and do not carry the definitional battles over what is a security that have occupied the crypto docket for a decade. A claimant who can plead a clear misstatement of capability sidesteps the harder question of characterisation and gets to causation and loss.

The DAO as defendant

The second question is the one the court reached first and the more novel. The plaintiff named the ai16z DAO as a defendant in its own right, alongside the identifiable people behind it. That raises a problem the law of unincorporated collectives has been circling for several years. Who accepts service for an on-chain organisation with no registered office, no directors and no agent. At the initial pretrial conference the court questioned why the founder had not been served after a single attempt in San Francisco, noted that a co-defendant resides in Australia and heard the position, advanced for the defence, that the DAO could not be sued at all.⁸

The last submission is not fanciful, but it runs against the recent authorities. Two lines of cases now bear on whether a DAO is a defendant with legal capacity.

The first treats a DAO as an unincorporated association. In CFTC v. Ooki DAO the Commodity Futures Trading Commission obtained a default judgment against a DAO that had ignored the proceedings. Judge Orrick held that Ooki DAO was an unincorporated association subject to suit under the Commodity Exchange Act, entered judgment for $643,542, imposed trading and registration bans and ordered the associated website content removed.⁹ The ruling’s importance lay in a single holding, that the collective was a person the statute could reach.

The second line is more consequential for members and runs through the general partnership. In Sarcuni v. bZx DAO the Southern District of California declined to dismiss a negligence claim brought by users who had lost about $1.7 million in a protocol hack, finding that the plaintiffs had sufficiently alleged that bZx DAO was a general partnership under the California Corporations Code and that its members were the partners.¹⁰ A general partnership carries joint and several liability. On that theory the liabilities of the collective become the personal exposure of those who compose it.

The general partnership problem

The general partnership theory was sharpened a year later. In Samuels v. Lido DAO, a claim that the LDO token was an unregistered security, Judge Chhabria held that Lido DAO could be a general partnership and, more pointedly, that certain institutional participants were plausibly its general partners. The court distinguished passive holding from active governance. Paradigm Operations, Andreessen Horowitz and Dragonfly, which held large blocks of governance tokens and had held themselves out as active in the organisation, were plausibly partners. Another investor, Robot Ventures, was dismissed for want of sufficient allegations of active participation.¹¹ The general counsel of a16z crypto called the decision a serious blow to decentralised governance, on the view that any participation, down to a forum post, might fix a member with the liabilities of the whole.¹²

That distinction is the legal core of the coming fight and disciplines the loose language that surrounds these cases. Liability does not attach to token holders as such. It attaches to members or active participants, to those who exercise or hold out that they exercise control. The question in each case is not who owned a token but who ran the enterprise. For ai16z, where the allegation is precisely that a human hand directed what was sold as an autonomous system, the general partnership analysis and the AI-washing analysis converge on the same small group of people.

What it means

Three consequences follow for anyone building or funding in this space.

For AI-token projects the exposure is not confined to the collapse of the token. It sits in the marketing. A project that advertises autonomy it does not have invites a deception claim that may be easier to plead than a securities case because it does not first require the claimant to win the token-classification fight. The Commission has already shown that a regulator will treat the misdescription of artificial intelligence as a discrete wrong. The gap between what a system is said to do and what it does is now a legal surface.

For DAOs the case tests whether the structure absorbs a claim or merely makes the defendants harder to find. The recent authorities suggest the latter. A court that regards a DAO as an unincorporated association may let it be sued. A court that regards it as a general partnership may look through it to the participants who directed it. Decentralisation complicates service. It does not, on the current cases, dissolve liability.

For founders and insiders the practical lesson is that the collective form is not obviously a shield. The claimant strategy these cases invite is straightforward. Name the reachable humans, plead the general partnership to widen the net and let the difficulty of serving a decentralised entity fall on the defence rather than the claim.

The open question

What remains unresolved is whether the merits will be reached at all. The motion to dismiss will test whether AI-washing dressed as consumer protection states a claim. The DAO-capacity issue may arise separately through service, representation and any later challenge to the DAO’s amenability to suit. But the case may turn on service and settlement before it turns on doctrine. If the DAO cannot be served or represented, the claim falls back onto the identifiable people the court can reach. The fork the court has not resolved is whether naming the DAO widens liability to its members or active participants or simply exposes how little of a decentralised collective the law can, in practice, touch.

Code on Trial is a weekly briefing on AI, crypto and digital disputes.

Notes

1 Pikabea v. Walters et al, No. 1:26-cv-03238 (S.D.N.Y.), before Judge Rakoff. Docket via Justia and PacerMonitor. All allegations unproven.

2 Cointelegraph (22 May 2026); secondary crypto reporting on the launch and marketing of the AI16Z token.

3 Cointelegraph, 22 May 2026 (peak market capitalisation of about $2.6 billion, January 2025).

4 Complaint as summarised in secondary reporting (Claim Depot; KuCoin). The class period and the figure of at least 3,945 wallet addresses should be confirmed against the complaint.

5 Pleaded counts per secondary summaries of the complaint: New York General Business Law ss 349 and 350; California Unfair Competition Law and False Advertising Law; negligent misrepresentation; unjust enrichment.

6 Docket: motion to dismiss made 18 June 2026; opposition due 3 July; replies due 13 July 2026; telephonic pretrial conference listed 31 July 2026.

7 SEC, Press Release 2024-36 (18 March 2024); settled orders In re Delphia (USA) Inc. and In re Global Predictions Inc.; combined penalties of $400,000.

8 Inner City Press, report of the 19 May 2026 conference before Judge Rakoff; Cointelegraph, 22 May 2026.

9 CFTC v. Ooki DAO, No. 3:22-cv-05416-WHO (N.D. Cal.), default judgment 8 June 2023 (Orrick, J.); judgment for $643,542 with trading and registration bans and removal of website content.

10 Sarcuni v. bZx DAO, No. 3:22-cv-00618, 664 F. Supp. 3d 1100 (S.D. Cal. 2023), order of 27 March 2023 (partial denial of motions to dismiss).

11 Samuels v. Lido DAO, No. 3:23-cv-06492 (N.D. Cal.), order of 18 November 2024 (Chhabria, J.); Securities Act 1933 s 12(a)(1) claim; general partnership finding as to certain participants.

12 Miles Jennings, General Counsel, a16z crypto, public commentary on the Lido decision (reported by Decrypt, November 2024).