Hyperliquid, a decentralized derivatives exchange, plans to let outside parties create prediction markets on its platform without needing approval for each one. The mechanism, an extension of its HIP-4 outcome markets, will launch on testnet before moving to mainnet, with no dates announced.
Prediction markets let people trade contracts that pay out based on whether a stated event occurs. A contract on an election, an inflation print or a sports result trades between zero and one dollar, and the price is readable as the market's implied probability. Hyperliquid launched outcome markets in May and they attracted around $100 million of volume in the first month.
What is actually being proposed
The design tries to be permissionless in listing while retaining control over structure.
Validators approve standardized templates that specify how a category of market must be defined. Individual deployers then create markets within those templates, without needing per-market approval.
To do so, a deployer must stake 500,000 HYPE, locked for six months, and must settle all outstanding markets before unstaking. With HYPE trading around $60 in mid-July, that is roughly $30m of capital committed per deployer.
In return, deployers may charge fees of up to 50% on their markets, and receive an initial allocation to run 100 outcomes, equivalent to 200 outcome tokens. Settled markets free up allocation for reuse, and an auction for additional allocation is planned.
The part that carries the risk
The critical detail is who settles the market.
Deployers are responsible for defining and settling individual markets in line with the template's criteria. Validators approve templates and can penalize non-compliance: the stake can be partially or wholly slashed by validator vote for a poorly defined market, an incorrect settlement, or a failure to settle within one week of the scheduled end.
So the party that profits from a market's fees is also the party that decides its outcome, with the check applied after the fact. Resolution has always been the hard problem in prediction markets, because the money at stake is often larger than the ambiguity is small, and real-world events rarely resolve as cleanly as their contract wording assumes.
The economic argument for this design is that $30m locked for six months makes dishonest settlement irrational: you would lose far more than any single market could earn. That reasoning holds for ordinary markets. It holds less well if a deployer could list a market large enough to make the stake worth forfeiting, which is precisely the scenario the slashing rule needs to cover.
The counter-design used elsewhere is to separate resolution from the listing party entirely, whether through an external oracle, as Polymarket does, or through a regulated exchange's own rulebook, as Kalshi does. Hyperliquid's approach concentrates both functions in the deployer and relies on collateral to align them.
The regulatory backdrop is unusually live
Prediction markets are in the middle of a jurisdictional fight in the United States, and the timing matters.
The Commodity Futures Trading Commission regulates event contracts and has asserted exclusive jurisdiction over them. It has taken legal action against a series of states, among them Arizona, Connecticut, Illinois, New York, New Mexico, Minnesota, Rhode Island and Wisconsin, over state efforts to treat these contracts as gambling. Kalshi followed the CFTC in suing Minnesota over a law criminalizing prediction markets, and courts have split: a federal court in Tennessee sided with Kalshi in February 2026, while a Massachusetts court rejected the same argument.
The CFTC's Enforcement Division has also issued an advisory on prediction markets. Regulated venues operate under constraints on what may be listed, with contracts tied to violence and similar categories off limits.
A permissionless system raises that question directly. If anyone meeting a collateral requirement can list a market within an approved template, the template becomes the entire compliance perimeter. Hyperliquid has not published rules or technical controls addressing categories that regulated venues prohibit, and we could not verify that such controls exist.
The jurisdictional question is genuinely unresolved rather than merely unanswered. Decentralized protocols without a licensed operating entity are the case US regulators have found hardest to reach, and this proposal will test it.
Scale, and why it matters
Hyperliquid is not a marginal venue. It is among the largest decentralized perpetual futures exchanges by activity, with open interest reported in the region of $11bn in mid-July.
That scale is what makes this more than a product note. A permissionless outcome-market layer attached to a venue of that size, resolved by the parties who list the markets and collateralized in a token whose price moves with the platform's own fortunes, is a genuinely new structure. The case in its favor is that it lets markets form on questions nobody thought to ask permission for, which is where prediction markets add most informational value. The case against is that it makes resolution integrity a function of collateral value, and collateral value is not constant.
Nothing is live yet. Testnet comes first, and the details that matter most, meaning template scope and the handling of disputed settlements, have not been published.
