When the European Commission opened its targeted consultation on the review of the Markets in Crypto-Assets Regulation on 20 May 2026, it put prediction markets and perpetual futures on crypto-assets in the same place. Section 4.4 of the consultation document carries both names in its heading and asks five questions across them, three about prediction markets and two about perpetuals. The window closed on 30 September 2026 at 23:59 CEST. We have read the filings that respondents published themselves, and the interesting thing is not that the two categories were paired. It is that the two halves of one section came back with answers of completely different quality, and the reason has nothing to do with how well anyone writes.
The section names two markets and describes only one
The opening of section 4.4 treats the pairing as a single observation. The Commission writes that one of the most marked developments in crypto markets since MiCA was adopted is the emergence of DLT based prediction markets and the volume of trading activity in perpetual futures. Two phenomena, one sentence, one policy problem.
What follows is a second paragraph, and it describes only prediction markets. In the Commission's own words, these are markets where parties "take a bet on a future development in almost any area, often far removed from financial assets or services". Perpetual futures get no equivalent paragraph. They are introduced by a quantity, the volume of trading activity, and then the questions begin.
That is an uneven starting position and it matters more than it looks. We wrote in our piece on Question 71 that a description becomes a definition when nobody offers a better one. For perpetuals the document offers no description at all, which leaves the field open. For prediction markets it offers one, and the verb in it is bet.
One of the two classification questions lets you answer neither
Question 70 asks whether prediction markets, where they are DLT enabled and facilitated through smart contracts, should be governed by MiFID or MiCA. It offers two boxes: By MiFID, By MiCA. Question 72 asks the same thing of perpetual futures and offers three: Yes by MiFID, Yes by MiCA, and No.
Read those two option sets next to each other and the asymmetry is structural. A respondent who thinks perpetual futures belong under neither EU financial framework has a box for that view. A respondent who thinks the same about prediction markets has to tick one of two frameworks and then write their way out of it in prose, or leave the question alone. Questions 71 and 73, both asking what substantive requirements should apply, come with no answer options at all on either side.
A banking association put a rulebook in one box and a wish list in the other
The Association for Financial Markets in Europe published its response dated 30 September 2026. It answered every question in section 4.4, and the contrast inside its own filing is the cleanest illustration available of what we are describing.
On Question 73, perpetual futures, AFME writes that perpetuals should be subject to the same substantive requirements as other complex leveraged derivatives, "including MiFID II investor protection, MiFIR transparency, transaction reporting, and market abuse rules". Four named bodies of existing law, each with articles, delegated regulations and supervisory practice behind it. A reader can go and look them up.
On Question 71, prediction markets, the same association writes in a single sentence that authorisation and licensing, product governance including a prohibition of unlawful markets, customer protection, and oracle rules should be considered. Twenty two words. Three of those four items exist as recognisable regulatory concepts. The fourth, oracle rules, does not exist as a rulebook anywhere in the European Union, or anywhere else. It is a placeholder for work nobody has done.
This is not sloppiness on AFME's part. It is an accurate reflection of what is available to point at. For leveraged derivatives Europe has thirty years of rules. For resolution of an event contract it has none, so the honest answer is a noun.
One filing wrote the prediction market standard out in full
Global Digital Finance is the exception, and it deserves naming precisely because it breaks the pattern. In its response dated 30 September 2026, GDF answered Question 71 with something close to a draft standard: transparent order books and participant driven price discovery, full collateralisation, segregation of client funds, conflicts controls, appropriateness and market abuse safeguards, published resolution rules, independent data sources and a clear dispute procedure. It adds a line that could have been lifted from our own guidelines: "Operators should not take positions against clients except through disclosed regulated market making."
GDF also tells the Commission where the real line sits, saying that product duration, fee based operator revenue and the operator's economic interest should be considered when distinguishing venue traded event contracts from harmful binary option models. That is a structural test rather than a label, and it is the same instinct behind our second and fourth guidelines.
Then look at what GDF does with the perpetuals half. It merges Questions 72 and 73 into one answer and anchors that answer on numbers that already exist, citing crypto contract for difference parameters including a 2:1 retail leverage limit and account level margin close out at 50 per cent of minimum required initial margin as benchmarks for assessment. The instrument behind those figures is ESMA Decision (EU) 2018/796 of 22 May 2018, the temporary restriction on contracts for differences, which the Hyperliquid Policy Center filing discussed below cites by that number at the same point in its own argument. Even the filing that wrote our half out by hand had a numbered benchmark waiting for the other half.
An energy exchange association answered three questions out of eighty six
Europex, the association of European energy exchanges, which describes itself as a not for profit body with 36 members, filed a response dated 29 September 2026 in the document itself. Its entire submission to an eighty six question consultation covers three questions: Question 1, Question 72 and Question 73.
Questions 69, 70 and 71 are not answered, declined or marked as outside its remit. They are simply not in the document. Europex argues that financial instruments within MiFID scope should stay under MiFID whatever the technology, and extends that to any perpetual futures including commodity and energy related perpetuals referencing TTF, Brent oil and power contracts.
There is nothing wrong with this. An energy exchange association answering the two questions that touch its members and skipping the rest is exactly how targeted consultations are supposed to work. It is worth recording only because it shows the perpetuals half of section 4.4 drawing in an established exchange constituency with an existing position, while the prediction markets half drew nobody from the equivalent position on our side.
The filing from the side of the market that builds both
The Hyperliquid Policy Center filed its response on 30 September 2026 through the Commission's online questionnaire. It describes itself in that document as "an independent research and advocacy organisation dedicated to a clear, regulated path for the evolution of onchain markets, including those available on Hyperliquid". It is an advocacy body and not a venue, which is worth stating plainly before drawing anything from it.
HPC answered all five questions in section 4.4. Its answer to Question 69 runs to roughly 195 words and is generous about prediction markets, calling the informational opportunity principal and the hedging of real world events a legitimate risk management function rather than a speculative novelty. Its answer to Question 70 runs to roughly 373 words, picks MiFID, and urges the Commission to "resist any approach which treats DLT-enabled prediction markets as a single category".
Its answer to Question 71, the question asking what substantive requirements should apply to prediction market service providers, is one line: "Please see our response to Question 70 above." Eight words. HPC says why, and says it in the open, which is to its credit: "Prediction markets are not the principal focus of this response, and we have therefore not set out detailed views on how such contracts are best calibrated, though HPC would be pleased to assist the Commission on those questions."
The answer to Question 73 is a different document. By our count it runs past 3,300 words. It opens by defining a perpetual future as a standardised, cash settled derivative with no predetermined expiry date, sets out the funding mechanism, and then proposes requirements with the specificity a regulator can actually implement. The funding methodology should be published in full, the interval defined and applied consistently, and the reference price derived from more than one independent input so that no single source can determine it. Maintenance margin requirements and liquidation thresholds should be published in advance with notice of change. The filing closes that passage with a five item disclosure list covering funding methodology, the speed and irreversibility of automated close out, the effect of leverage, the derivation of the reference price and its limitations, and the risk that a settlement asset deviates from its reference currency.
Read that list again and notice what it is. It is a venue disclosure standard. It is the thing our second guideline asks for, written by somebody else, for the other product, and filed in Brussels inside the same section as ours.
The two primitives sit next to each other in one set of documentation
The pairing in section 4.4 is not an administrative accident, and the clearest evidence of that is technical rather than legal. The Hyperliquid documentation lists its improvement proposals in sequence. HIP-3, builder deployed perpetuals, lets a deployer define a market including its oracle and contract specifications, set oracle prices and leverage limits, settle the market when needed, and take a configurable share of fees, against a mainnet staking requirement of 500,000 HYPE held for a minimum of 183 days. HIP-4, outcome markets, defines "fully collateralized contracts that settle within a fixed range" with no leverage and no liquidations.
So the Commission's hypothesis is correct at the level of infrastructure. One stack, two primitives, adjacent proposal numbers. And note the detail that should interest anybody writing a standard: the HIP-4 primitive is already fully collateralised and unleveraged, which is two of the three conditions GDF told the Commission should govern retail access to event contracts, the third being appropriate conduct and market integrity protections. Part of the answer to one of the hardest questions in the European file was sitting in public documentation while the file was open. None of the published filings we read mentions it, and the words outcome market do not appear anywhere in the Hyperliquid Policy Center submission.
What an eight word answer costs the category
None of the respondents above did anything improper, and none of them owes our category a submission. The point is narrower and it is about us. The deadline that matters is in the regulation itself. Article 140(1) of Regulation (EU) 2023/1114, quoted in full in footnote 4 of the consultation document, requires the Commission to present a report on the application of the regulation to the European Parliament and the Council by 30 June 2027, accompanied where appropriate by a legislative proposal. As of 29 October 2026 the Commission's consultation page lists the consultation as closed and publishes no responses, no summary and no feedback statement, so the public record consists of what respondents chose to publish themselves.
On that record, perpetual futures enter the European file with a named instrument definition, a published funding methodology requirement, margin and liquidation disclosure with notice of change, a multi source reference price, a five item disclosure list, and an existing decision from 2018 to calibrate against. Prediction markets enter it with one trade body's paragraph, one banking association's twenty four words, and a cross reference.
That gap is not a regulatory injustice. It is a description deficit, and it is the kind this initiative exists to close. A category that does not describe itself gets described, and it gets described in the vocabulary of whoever held the pen. In section 4.4 that vocabulary is already on the page, and the word in it is bet.
The practical lesson for any venue reading this is not to lobby harder next time. It is that the documents a consultation rewards are documents you should already have, because they are documents your users should already have: a written instrument definition, published resolution rules with the data sources named, a dispute procedure, a statement of where the operator's revenue comes from and what it does not depend on, and a change log. Hyperliquid's perpetuals filing worked in Brussels because the underlying facts were written down before anyone asked. The next consultation will reward the same thing, and the one after that will not wait for us.