Key takeaways
- HIP-4 launched on Hyperliquid mainnet on 2 May 2026, adding binary and multi-outcome event contracts to an existing deep-liquidity perpetuals venue rather than launching a standalone prediction market.
- Builder markets require a slashable stake in HYPE, so a validator posting a faulty outcome loses capital. That is the structural difference from UMA's bonded but unslashed token vote.
- The design matters because a $750 bond was enough to propose a resolution on a Polymarket contract carrying up to $242M of exposure, adjudicated by a token worth around $95M.
- Day-one share was modest at roughly 0.7% of category volume, despite a loud launch claim from an interested party about the first Bitcoin outcome market.
- Stratium founder Wasim Fukase, building on HIP-4, reports roughly $400M in cumulative volume, with World Cup markets driving the spike and daily Bitcoin contracts carrying the recurring flow.
- Liquidity rather than demand is the binding constraint, and no venue has yet proved that shared liquidity with perps solves it.
Why a perps exchange entering this category is worth attention
Most new prediction markets launch as standalone venues and then spend their first year buying liquidity. Hyperliquid did something different in May 2026, bolting event contracts onto an order book that already had deep flow from perpetual futures. That approach tests two questions at once: whether shared liquidity travels across product types, and whether a staked, slashable resolution layer holds up better than a token vote. Both questions come out of the microstructure and integrity chapters of the Hodl Up Prediction Markets Report 2026, which included an interview with an operator building inside HIP-4. For where Hyperliquid sits among the other venues, see our guide to every prediction market platform in 2026.
What HIP-4 is, mechanically
HIP-4 is an onchain central limit order book for binary and multi-outcome contracts, unified with Hyperliquid's perpetuals rather than separated from them. Custody is self-custodial on Hyperliquid L1 and HyperEVM. Settlement uses USDH. The fee structure charges nothing to open, with fees applied on close and settle.
The resolution design is the part worth understanding. Builder markets require a slashable stake in HYPE, so the party running an outcome has capital at risk against getting it wrong. Wasim Fukase described the mechanism directly when we asked what he would change about resolution across the industry.
"For hyperliquid itself there is a mechanism where if there's any kind of fault in terms of the oracle side, you get slashed in terms of your stake from a validator itself."
Running an outcome therefore costs something to get wrong, which is the opposite of the incentive structure in a system where a small bond buys the right to propose an answer.
Three live resolution designs, compared
The 2025 to 2026 record now runs three approaches in parallel. That is the natural experiment the sector needed, and it is also the clearest way to see what HIP-4 is trying to fix.
| Kalshi | Polymarket via UMA | Hyperliquid HIP-4 | |
|---|---|---|---|
| Who resolves | Exchange under a regulated rulebook | UMA token holders through the Data Verification Mechanism | Staked deployer with authorised oracle feeds |
| Cost to challenge or propose | Regulated dispute process | Bond of roughly $750 to propose | Slashable stake in HYPE |
| Conflict controls | KYC, surveillance, voters cannot be traders | No identity check, nothing stops a voter holding a position | Economic penalty for a faulty outcome |
| Known failure | Mention markets under CFTC review | Zelensky market flipped after nine days of disputes | Not yet stress-tested at scale |
| Accountability | Regulator and courts | Token governance | Stake at risk |
The Zelensky case is the reference point. A market on whether Volodymyr Zelensky would wear a suit before July drew somewhere between $160M and $242M in volume, resolved initially to yes, then flipped to no after nine days of disputes routed through UMA's token-holder vote, even though the BBC, the New York Post and roughly forty outlets had described his NATO-summit outfit as a suit. A Wall Street Journal investigation found a UMA voter had held a position in a market they voted on. Recent academic work on Polymarket oracle adjudication examines the same weakness.
None of the three designs has yet been tested by a contested nine-figure market on its own rails, because the Zelensky case tested only the second. That is a real limit on how much credit HIP-4 deserves today.
What the volume actually shows
The launch was loud. The chief executive of Hyperion DeFi, a publicly traded HYPE accumulator and therefore an interested party, claimed via DL News that the first Bitcoin outcome market did roughly three times the combined volume of the equivalent Polymarket and Kalshi markets. Day one nonetheless took only about 0.7% of prediction-market share, and the months since have been modest.
Fukase's numbers from inside the ecosystem are more useful than the launch claim. Stratium has done roughly $400M in cumulative volume, the World Cup final on 19 July drove the largest single spike, and five or six of his top ten markets were tournament-related. Outside those events, the recurring volume is daily Bitcoin up-or-down contracts, which is a different business from event forecasting and closer to a short-dated options product.
He is candid about what is holding it back.
"Right now the liquidity is the issue, and this is being honest about it."
Limitless chief executive CJ Hetherington told Bernstein analysts that the missing momentum behind Hyperliquid's push suggests either that building real liquidity in this category is harder than commonly assumed, or that volume on rival platforms is more inflated than it appears. Both readings are live, and the wash-trading research on Polymarket gives the second one more weight than it would otherwise carry. We cover why headline figures deserve caution in Nobody Knows How Big Prediction Markets Are and in our Kalshi vs Polymarket comparison.
What to watch next
Three signals decide whether the design matters beyond its engineering. Whether a contested high-value market ever tests the slashing mechanism in public. Whether recurring volume broadens past daily Bitcoin contracts into events with genuine forecasting value. And whether shared liquidity with perps proves durable rather than a one-time launch benefit.
Fukase also reports that no Asian jurisdiction has yet approved prediction markets, and that his own is among the first to issue an operator a no-objection certificate, with an invitation to help draft the framework there.
The full conversation with Wasim Fukase covers the builder-code economics and the Asian regulatory picture in more depth. The oracle chapter of the Prediction Markets Report 2026 sets out the UMA comparison in full, and our report summary covers the integrity record around it.
None of this is investment advice.