Hyperliquid Hired Washington

On February 18, 2026, the Hyper Foundation announced that it had funded a new 501(c)(4) advocacy organization in Washington. The vehicle is called the Hyperliquid Policy Center. The funding was one million HYPE tokens, worth roughly $29 million at the time. The CEO is Jake Chervinsky, formerly Chief Legal Officer at Variant and Chief Policy Officer at the Blockchain Association. The policy counsel came from Sullivan & Cromwell. The policy director came from Variant.

The launch announcement is written in the careful prose of someone who has read every comment letter the CFTC has ever received. It talks about "the next generation of American markets," about derivatives as "an American story of continuous innovation," about how decentralized markets can deliver on the policy objectives of the Commodity Exchange Act through code rather than through institutions. It is a good document. It is also not really what HPC is for.

HPC is for not getting indicted.

What the Last Cycle Taught the Founders

The reason any perp venue needs a Washington office in 2026 is that the US enforcement record on derivatives is unambiguous, and that record was set by a CEX. In 2020, the CFTC and the Department of Justice charged the founders of BitMEX — a centralized perpetuals exchange — with operating an unregistered derivatives venue and willfully violating the Bank Secrecy Act. The CFTC eventually extracted a $100 million civil penalty from the corporate entities. The three founders pleaded guilty to BSA violations. Arthur Hayes was sentenced to six months of home detention and a $10 million fine. He was later pardoned. The pardon does not undo the precedent.

BitMEX is the precedent because it is the case where founders learned what their personal exposure actually was. The architecture of the venue did not matter to the indictment. What mattered was that American users could reach the product, that the product was a derivatives contract, and that no US registration existed. The Seychelles entity was not a shield. The IP-level geo-block was not a shield, because it was not enforced strictly enough. The disclaimer was not a shield. The only thing that has ever protected a derivatives venue accepting American counterparties is a CFTC registration. Centralized perp exchanges took the BitMEX file and either built compliance shops or stopped serving the US. Decentralized perp exchanges read the same file and reached different conclusions, because their architecture made some of those compliance shops impossible to build in the first place.

What Perp DEXs Did Differently

Decentralized perp venues responded to the same precedent with a different toolkit. They had no FCM to register. They had no central order book to license. They had no single legal entity sitting in front of the matching engine. What they did have was infrastructure they could rearrange.

The first move was the sequencer. Several of the largest perp DEXs of the previous generation rebuilt or relocated their order-matching infrastructure so that no US person operated it. The second move was a chain migration. At least one major venue restructured its entire L1 architecture, leaving an Ethereum L2 in favor of an app-specific chain explicitly to thin out the US legal nexus. The third move was the front-end. The Cloudflare-level IP block became standard. So did the wallet-level geo-check. So did the legal entity sitting in a Cayman or BVI structure with no US officers.

The fourth move was quieter. Teams stopped naming themselves publicly. Founders disappeared from podcasts. Job postings dropped any reference to the protocol's headquarters because there was no headquarters to refer to. This was not paranoia. It was the rational response to a regulatory environment in which a named, US-resident protocol founder running a venue serving US users was assumed to be one subpoena away from a problem.

Hyperliquid grew up inside this constraint, and shows it. Most of its volume is non-US. Its public-facing front-end blocks American users. The core team is deliberately quiet about geography and personal identity. None of this is decorative. It is the cost of running a derivatives venue with American counterparties watching from the sidelines, and it is the reason a $29 million policy organization is the next logical line item.

Why the Law Did Not Fit

The reason American users are excluded from decentralized perpetuals is not that the CFTC dislikes them. It is that the law for derivatives markets in the United States is built on a single load-bearing assumption: that there is a centralized intermediary in the middle. The Commodity Exchange Act assumes a Designated Contract Market or a Swap Execution Facility. It assumes a Futures Commission Merchant holding margin. It assumes someone whose license can be revoked, whose books can be examined, whose officers can be subpoenaed. Decentralized perpetuals have none of these things. The order book is on a chain. The risk engine is in a contract. The collateral sits with the user. There is no FCM to register, because there is no entity holding the funds.

The result is not that decentralized derivatives are illegal in any direct sense. The result is that there is no lane for them to be legal in. They cannot register, because the registration categories assume an intermediary. They cannot operate without registration, because the law requires it. The architecture and the rulebook do not speak the same language. American builders facing this gap have had three options: build something that is not a perp, geo-block American users, or go to prison. Most picked option two.

What Changed in 2025 and 2026

The reason HPC exists right now, in this specific year, is that the gap has started to move. The CLARITY Act passed the House in July 2025 by a vote of 294 to 134. The bill gives the CFTC exclusive jurisdiction over digital commodity spot markets and creates new registration categories — Digital Commodity Exchange, Digital Commodity Broker, Digital Commodity Dealer — that are designed to fit crypto venues rather than legacy futures floors. Section 309 carves out non-controlling blockchain developers from being treated as regulated intermediaries. The CFTC and SEC announced a joint initiative called Project Crypto built around three pillars: regulatory clarity, inter-agency coordination, and support for permissionless innovation. In March 2026, CFTC Chair Mike Selig publicly said the agency was weeks away from a framework for onshore crypto perpetual futures.

This is the window HPC was built to walk through. The existence of a draft framework is not the same as a working pathway for a non-custodial protocol. Someone has to be in the room when the line is drawn between "decentralized enough to qualify for the carve-out" and "centralized enough to need a license." Someone has to be the one explaining what a funding rate is, why a 24/7 perpetual contract is not a 19th-century futures contract that needs a clearinghouse, why a margin check that runs every block is more conservative than one that runs at end-of-day. If no one is in that room, the rules will be drawn by people whose mental model of a derivative was last updated in 2010.

The Honest Read

There is a reason the HPC announcement uses the phrase "Hyperliquid is poised to become the house of all finance" rather than something more modest. Lobbying organizations write in this register because they have to. The audience is congressional staff and agency lawyers, not crypto Twitter, and the goal is to convince that audience that the protocol is a legitimate financial market deserving of a defined legal status. The grandeur is part of the function.

But underneath the grandeur the actual ask is narrow and specific. HPC wants a regulatory definition that recognizes deterministic, non-discretionary code as capable of performing the supervisory functions that an intermediary used to perform. It wants a carve-out for protocols where there is no central operator to register. It wants the CFTC's eventual perp pilot to include a path for non-custodial venues, not just a path for Coinbase Derivatives and CME. If those three things land in the final rule, American users will be able to trade on Hyperliquid legally for the first time, and the founders will not have to think about the BitMEX file every time someone in their Discord posts a New York IP.

That is what $29 million of HYPE is buying. Not a mission. A jurisdiction.

Why It Matters for the Rest of the Stack

A non-custodial trading terminal sitting on top of Hyperliquid does not write the rules and does not need to. But the rules determine which users it can serve. If HPC succeeds, a US-based trader using a self-custodied wallet to interact with Hyperliquid through a third-party front-end will be doing something the law has a name for, instead of something the law has no opinion about because the question was never asked. The difference between those two states is the difference between a market that has 10 million addressable users and a market that has 10 million addressable users plus the largest pool of retail risk capital in the world.

This is the part that gets lost in the cynical reading. Yes, HPC is a lobbying operation. Yes, it exists because the founders watched what happened to BitMEX and decided they did not want to be the next case. But the thing being lobbied for is not a special favor. It is the absence of a structural exclusion. The current state — where a self-custodied American cannot legally use the most liquid onchain perp venue in the world while being free to use any number of higher-risk centralized alternatives — is not a considered policy outcome. It is a gap left by a statute written before the technology existed.

Closing that gap takes someone whose full-time job is to be in the room. Hyperliquid hired the people who can do that. The rest of the ecosystem benefits whether or not it ever sends a comment letter of its own. That is the part of this story worth paying attention to. The press release is written for Washington. The actual outcome is written for everyone who wants to trade on a venue they do not have to trust.