Skip to main content

Hyperliquid HIP-4 Goes Live: How a Zero-Fee Order Book Just Flipped the Prediction Market Wars

· 10 min read
Dora Noda
Dora Noda
Software Engineer

On May 2, 2026, a small line in Hyperliquid's release notes quietly redrew the map of a $24 billion industry. HIP-4 — the long-awaited "outcome markets" upgrade — went live on mainnet with a single Bitcoin binary contract: would BTC close above $78,213 on May 3? Within hours, the order book was deep, the spreads were tight, and traders were opening positions for free.

Free. Zero fees to open. Fees only when you close, burn, or settle.

That single design decision is the most aggressive shot fired in prediction markets since Polymarket beat Augur on UX in 2020 and Kalshi beat Polymarket on regulation in 2024. It is also a direct attack on the only two platforms that matter today — Kalshi, freshly valued at $22 billion, and Polymarket, sitting at $15 billion. And it lands in the middle of a 96-hour news cycle that has rewritten what "legitimate" prediction markets are allowed to look like.

The Setup: Two Giants, One Wildcard, One Very Bad Week​

To understand why HIP-4's timing matters, you have to understand what the rest of the industry was doing the same week it launched.

Prediction markets had a record-breaking April 2026. Total taker volume across the industry hit $8.6 billion, with Kalshi printing $5.42 billion to Polymarket's $1.99 billion — the first month Kalshi clearly overtook Polymarket on volume. Year-to-date, the gap is even wider: Kalshi has cleared $37.49 billion in 2026, against Polymarket's $29.23 billion. The two platforms now control between 85% and 95% of all prediction market volume on the planet.

But the same month brought a regulatory storm.

On April 22, Kalshi suspended and fined one Senate candidate and two House candidates for insider trading on their own campaigns. On April 25, the U.S. Department of Justice unsealed a criminal indictment against Master Sergeant Gannon Van Dyke, who allegedly used classified information about a U.S. military operation in Venezuela to make roughly $400,000 trading on Polymarket. On April 30, the U.S. Senate unanimously passed a rule barring senators from trading prediction markets at all — effective immediately.

Both incumbents responded with hastily rolled-out integrity policies: Kalshi's technological guardrails preemptively block politicians, athletes, and employees from trading their own contracts; Polymarket's "updated market integrity rules" defined three categories of forbidden insider trading conduct.

It was, in short, the worst possible week for a "trust us" centralized model. And it was the perfect possible week for a permissionless on-chain venue to launch.

What HIP-4 Actually Is​

Strip away the marketing and HIP-4 is engineering, not narrative.

Each outcome market is a pair of binary tokens — typically YES and NO — that float between 0.001 and 0.999. The price is the implied probability. At settlement, one side converts to one USDH (Hyperliquid's native stablecoin) and the other to zero. Positions are fully collateralized; there is no liquidation risk because there is nothing to liquidate.

What makes this different from Polymarket's AMM-based architecture is that HIP-4 lives natively inside Hypercore, the Hyperliquid L1's matching engine. That means outcome markets share the same order types, the same approximately 200,000-orders-per-second throughput, and — critically — the same margin account as a trader's perpetual futures and spot holdings. A trader hedging an event-risk position against a BTC perp does it in one wallet, with portfolio margin, on the same book.

This is the architecture Polymarket cannot ship without rebuilding from scratch, and it is the architecture Kalshi structurally cannot ship at all because Kalshi is a CFTC-regulated centralized intermediary.

The fee model is where the knife twists. Polymarket charges 2% taker fees. Kalshi captures spread through a centralized clearinghouse. HIP-4 charges nothing to open. Fees only kick in on close, burn, or settlement — meaning short-duration traders, high-frequency event arbitrageurs, and anyone with a directional view on a specific outcome can build a position with no entry tax.

For market makers, the implication is even larger: the cost of providing liquidity at the open of a new market is, by design, zero.

Why Token-Economics Is the Third Axis​

Polymarket and Kalshi compete on UX and regulation. HIP-4 introduces a third axis: token-economic alignment.

Hyperliquid uses approximately 97% of its protocol revenue to buy back and burn HYPE tokens. Every fee paid by a prediction market trader on HIP-4 — even just the closing fee — flows back into the same buyback engine that has made HYPE the largest non-Bitcoin position in Maelstrom, Arthur Hayes's family office.

This is what Hayes is pointing at when he calls a $150 HYPE target. His thesis isn't a multiple of trading fees. It's a bet that prediction markets become the third revenue vertical — alongside perps and spot — that pushes Hyperliquid's annualized revenue back to the $1.4 billion mark it briefly touched last August. Polymarket has no comparable token-economic loop because POL has no fee-revenue exposure. Kalshi has no token at all.

When Hyperliquid's ~$9.57 billion in perpetuals open interest sits in the same wallet as binary BTC contracts that pay into the same buyback, every category of trader — directional, hedging, arbitrage — becomes a structural buyer of HYPE. That is the loop neither competitor can copy.

The Strange Kalshi Partnership​

There is an unusual wrinkle in this story: HIP-4 was co-authored by John Wang, the head of crypto at Kalshi.

In March 2026, Hyperliquid and Kalshi announced a partnership to develop on-chain prediction markets together. The optics looked like a classic "incumbent defends by co-opting the disruptor" play — Kalshi gets distribution onto a permissionless chain without canibalizing its CFTC-regulated business; Hyperliquid gets the credibility and contract design experience of the volume leader.

In practice, the partnership creates a strange equilibrium. Kalshi is the only one of the three players that genuinely cannot be displaced by HIP-4 — its institutional flow is glued to its CFTC license, and large allocators are not moving to a permissionless venue regardless of fee. Polymarket, on the other hand, sits in the awkward middle: a non-US-regulated AMM venue whose entire competitive moat (UX + crypto-native users) is exactly what Hyperliquid is now competing for directly.

If HIP-4 takes 30% market share within six months of mainnet, the volume comes from Polymarket, not Kalshi. The Kalshi partnership essentially picks the target.

What Has To Be True for HIP-4 To Win​

Prediction market history is unkind to challengers. Augur had the first-mover advantage and the better technology in 2020. Polymarket won by being usable. Polymarket had product-market fit on the 2024 U.S. election and Kalshi won by being licensed. Both losers had reasons to win that didn't matter once the actual fight began.

For Hyperliquid to repeat the cycle in 2026, three things need to happen:

Liquidity has to migrate, not duplicate. Polymarket's edge is that its books are thick on long-tail political and cultural events — exactly the markets where it has 678,342 unique April users to Kalshi's much smaller user base. HIP-4 launching with a recurring daily BTC binary is a clever cold-start because it draws on Hyperliquid's existing trader base, but the harder problem is convincing event-market users to leave Polymarket's familiar UI for an order book.

The category expansion has to land. Hyperliquid has signaled politics, sports, macro releases, crypto, and entertainment as next categories. Each one is a different liquidity bootstrap problem. Politics drags in regulatory complexity. Sports collides with state-by-state US gambling law. Macro is the easiest fit for an order book and the smallest TAM.

Regulatory pressure on the incumbents has to keep tightening. The April insider trading bans were self-inflicted, but the deeper problem is that centralized prediction market platforms have a list of names — every trader, every IP, every account — and that list is now subject to subpoena. Permissionless markets do not. As enforcement intensifies, the gap between "legal but surveilled" and "permissionless and pseudonymous" widens, and HIP-4 sits squarely on the latter side of that line.

If all three happen, the prediction market industry of late 2026 looks like a three-way split: Kalshi keeps institutional flow, Hyperliquid takes crypto-native event traders, and Polymarket gets squeezed in the middle. If only one or two land, HIP-4 stays niche.

The Real Question Isn't Whether HIP-4 Wins​

The interesting question is not who captures the next $10 billion of prediction market volume. It is what happens to the architecture of the industry once a credible permissionless option exists at zero open-fee.

For five years, the prediction market debate has been UX versus regulation. HIP-4 introduces a third option: build it as a primitive inside an existing high-throughput trading venue, collateralize it natively, and tax it only at exit. That design borrows nothing from Augur, nothing from Polymarket, and nothing from Kalshi. It borrows from CME — and turns the page on what a prediction market is supposed to feel like.

The industry was already reshaping itself around insider trading bans, ETF wrappers, and Senate rules. HIP-4 just accelerated the part nobody was watching: the part where the marginal trader stops choosing between Polymarket's AMM and Kalshi's clearinghouse, and starts choosing whether to stay in TradFi at all.

May 2, 2026 will be remembered as the day that choice got cheaper.

BlockEden.xyz provides enterprise-grade RPC infrastructure for builders deploying on Hyperliquid, Solana, Sui, Aptos, and 25+ other chains. If you're building event-driven applications, prediction-market integrations, or trading infrastructure, explore our API marketplace for production-ready endpoints designed for high-throughput trading workloads.

Sources​

Share on Twitter