Skip to content

Hyperliquid

11 Jun 202613 min

Key Takeaways

  • New primitive, Same Rails: HIP-4 adds binary outcome contracts natively to Hyperliquid's L1 CLOB — fully collateralized, settling to 0 or 1 USDC, no leverage, no liquidations, unified margin with existing perp positions.
  • The demand is already there: 3.3% of Polymarket users are already on Hyperliquid and drive 12% of its volume. ~$850M/month in outcome flow currently routed off-platform. HIP-4 captures that without acquiring a single new user.
  • Trust is the edge: Polymarket outsources resolution to UMA, a $36M market cap token where voting power is proportional to holdings. Permissionless HIP-4 deployments put the builder’s HYPE stake at risk of slashing, aligning incentives.
  • Repeating the HIP-3 Playbook: At normalized take-rates (3.5bps), the base case projects $48M annualized fees which feed the same HYPE buyback flywheel as perp and spot fees. This is the same path HIP-3 took before the market priced in normalization.

Background

Prediction markets have grown at a 1,144% CAGR over the last two years, now crossing $30B in monthly notional volumes (total face value of all contracts outstanding), a figure roughly 12x since 2023. The sector is effectively a duopoly between Kalshi which holds 54.5% share and Polymarket which holds 29.1%, together accounting for over 75% of all volume. This is a winner-takes-all market, and 2 players currently own it. HIP-4 is Hyperliquid's bid to break that duopoly. The proposal was co-authored in Sep 2025 by Kalshi’s Head of Crypto, John Wang.

What HIP-4 proposes is distinct from both incumbents. Kalshi operates as a fully centralized exchange under CFTC oversight while Polymarket runs a hybrid model using offchain order matching with onchain settlement on Polygon. HIP-4 puts outcome contracts natively on Hyperliquid's own L1 CLOB — fully collateralized, settling to 0 or 1 USDC, no leverage, no liquidations. A trader’s USDC balances, perp positions, and their outcome exposure can all share the same margin, on the same execution layer, with no split custody.

Out of 3M Polymarket users, 3.3% of them are already on Hyperliquid and they drive 12% of Polymarket's volume. These traders currently split their USDC margin across two platforms to do what could be done from a single account. HIP-4 consolidates that flow. With the sector projected to grow at 80% CAGR to 2030, the revenue collected feeds the same Assistance Fund buyback flywheel already proven by HIP-3.

Our March memo on HIP-3 flagged a case for re-rating HYPE on the back of tokenized equity and commodity trading picking up, particularly during Iran war weekends when Hyperliquid became the only venue pricing oil in real time.

HIP-3 fees remain in growth mode today, running at roughly $100-150K daily across $2-3B in volume (~0.67bps) against perps generating $2-3M daily on comparable notional (~3.75bps). The re-rating focused on HIP-3 as a product with proven, durable demand that hasn't realized its fee potential yet, with full normalization also being priced in.

Product-market fit is the most defensible assumption for HIP-4 as the user overlap data already shows latent demand from traders who are effectively doing cross-platform arbitrage today. Fee normalization is probable but not guaranteed because HIP-4 needs enough liquidity depth before pushing fees toward market rates. Buyback contribution follows mechanically once fees scale. HYPE’s re-rating is speculative as it depends on markets pricing in HIP-4's long-run fee contribution before that contribution is fully visible, which is exactly what happened with HIP-3 but isn't guaranteed to repeat.

Each market poses a yes/no question with a fixed expiry. At resolution, “YES” tokens pay out 1 USDC and “NO” tokens pay out nothing, or vice versa. Positions are fully collateralized at open with no funding rates or liquidation risk. Fees are zero across all outcome markets during initial testing.

Outcomexyz was the first builder live on HIP-4 at mainnet launch on May 2, 2026. They run crypto price binary markets with settlements based on HyperCore's mark price at expiry, and real-world event markets covering CPI, Fed decisions, and sports outcomes are deployed and settled by the validator set. Permissionless HIP-4 deployment is not yet live but following HIP-3’s model: external builders can deploy their own markets with their own oracle and resolution logic. Validators can slash their stake in the event of malpractice but the exact specs remain unconfirmed.

The Resolution Problem

HIP-4's market opportunity also lies in how trustworthy the resolution model is. Every prediction market has two layers: trading and resolution. The trading layer typically runs on a CLOB, fills orders, and tracks positions. The resolution layer, on the other hand, handles the settlement of outcomes on a market, often through an oracle. The standard architecture separates these across protocols. For instance, Polymarket handles trading, UMA handles the resolution. That split creates an incentive misalignment that is the core vulnerability of the incumbent model because of the following reasons:

  • Oracle governance capture. UMA is a $36M market cap token. Any market with open interest larger than that makes it economically rational for a whale to buy enough UMA supply to control resolution in their favor.
  • Token-weighted resolution power. Voting weight scales with holdings, so resolution can be bought.
  • Value asymmetry. The economic security of the oracle (UMA market cap) is much smaller than the value at risk across active Polymarket markets.

There have been a couple of controversies in the past showcasing different attack surfaces:

What’s consistent across all 3 is that disputed markets ultimately escalate to UMA's token-weighted vote regardless of how the dispute originated. The oracle inherits whatever ambiguity the platform layer creates or allows, and whales can then resolve that ambiguity in favor of their positions. This creates the trust gap that HIP-4 capitalizes on.

These 3 instances show the integrity issues with incumbent prediction markets, ranging from direct governance attacks, to ambiguous settlements or outright shifting of goalposts. What connects them is that disputed markets ultimately escalate to UMA's token-weighted vote regardless of how the dispute originated. Even when the failure starts at the platform layer, as it did with the MSTR market, whales can weaponize the ambiguity UMA inherits and resolve it in favor of their positions.

A WSJ investigation on 2 June showed that in most disputed markets, more than half of UMA votes came from the ten largest wallets, at least 60% of active UMA voters could be linked to live Polymarket accounts, and roughly one in five disputes had at least one voter with a direct financial stake in the contract being settled. Polymarket has logged over 1,150 disputed markets in 2026, already past its full-year 2025 total.

Kalshi resolves through Kalshi Klear LLC against CFTC-filed contract terms without any token vote, dispute windows, or traditional governance attack vectors. This structure means the entity writing the market rules also makes every resolution call, presenting centralization risks. So far, there haven't been any controversial settlements, also in part due to their market selection being more conservative to avoid ambiguous resolution situations.

HIP-4 sidesteps the UMA problem by removing the dependence on an external protocol and the susceptibility to economic attacks via its token. It uses the Hyperliquid validator set to settle outcomes by voting based on information from automated newsfeed software. HIP-4 market rules are also encoded at deployment and the validator set has no mechanism to amend them mid-market.

Attacking the resolution layer means corrupting Hyperliquid's validator set directly and coordinating across independent staked validators with slashable HYPE at risk, which is significantly more expensive. External information still enters the system, and the 26 validators still have to interpret ambiguous outcomes just as UMA holders do. The cost of mounting a malicious attack, however, is far higher.

The User Overlap Thesis

Prediction markets are truth-aggregation mechanisms rather than gambling venues. When they work, they are the most efficient way to price the probability of future events and are more accurate than polls, faster than analysts, and continuous in a way that news cycles aren't. That only holds if resolution is trustworthy. Sophisticated traders who size into genuine mispricings get punished for being correct if the underlying settlement layer can be bought. Traders who care most about resolution correctness are the ones who size largest, and those traders are disproportionately already on Hyperliquid.

Prediction markets and perp markets are also more complementary than they appear. Prediction markets price event probabilities. Perps price asset exposure with leverage and continuous P&L. A perp trader running directional BTC exposure into a Fed decision can use an outcome contract to isolate the event risk from the price risk. If a prediction market is too shallow to absorb meaningful size, the same view can be expressed downstream in perps. These use cases have always existed but have never been executable from the same account.

The case for HIP-4 rests on this overlap. On April 21, both Kalshi and Polymarket announced perpetual futures products. However, expanding from prediction markets to perps is a weaker dynamic compared to Hyperliquid's expansion from perps to prediction markets. First, perp traders already use prediction market instruments to hedge macro exposure or arbitrage mispriced events, so outcome contracts are a natural extension of the same directional framework. 3.3% of Polymarket users are already Hyperliquid traders and account for 12% of Polymarket's volume, confirming the overlap and skew toward high-conviction participants.

Second, the learning curve for a perp trader adding outcome contracts is near zero. Margin, collateralization, and position sizing already govern how they operate. A binary outcome contract is a simpler instrument than what they run daily. A casual prediction market user has never operated with funding rates, liquidation curves, and active margin management. These are separate risk disciplines.

Third, HIP-4 removes the only reason native Hyperliquid traders routed outcome bets off-platform. That 12% of Polymarket volume is capturable on HIP-4 without acquiring a single new user.

Beyond familiarity, many Hyperliquid traders run custom infrastructure on top of the platform (bots, alerts, portfolio trackers, API integrations, etc.) that creates switching costs. Kalshi and Polymarket are building perp products to retain users who are already partially elsewhere. HIP-4 brings the product to users who shouldn’t need to leave.

HIP-4 fees are zero today, which is the same position HIP-3 was in when the market started pricing in normalized fees. Every outcome contract opened on HIP-4 settles in USDH and routes fees through the Assistance Fund into HYPE buybacks which is the same flywheel for perp and spot fees.

Sizing the Opportunity

May 2026 prediction market volumes ran $30B, with Polymarket holding a 22.7% share. Since 12% of Polymarket volumes are driven by Hyperliquid users, that equates to ~$850M/month in outcome contract flow being routed off-platform. We model two volume sources for HIP-4:

  1. Polymarket migrators, which are defined as the subset of Hyperliquid traders currently routing outcome bets through Polymarket. That cohort drives 12% of Polymarket's volume, or $850M/month We’ll use a sensitivity of 10-70% of that cohort.
  2. Organic HIP-4 adoption from existing Hyperliquid traders incorporating outcome contracts into their existing margin accounts, modeled at 3%, 5%, 7%, and 9% of Hyperliquid's $219B monthly platform volume.

The fee matrix has two rates:

  1. 0.67bps as the growth-mode floor, consistent with where HIP-3 sits today.
  2. 3.5bps as the normalized ceiling, which is Hyperliquid's disclosed platform-wide blended take rate.

At growth-mode rates, the output is marginal across all three cases, under $1.5M/month even in the bull scenario. At 3.5bps normalized, the base case generates $3.98M/month or ~$48M annualized), and the bull case hits $7.1M/month (or $85M annualized).

For context, HIP-3 currently generates $3M/month on $45B in volume at 0.67bps. At normalized rates of 3.5bps, HIP-3 volumes would produce $15.75M/month.

Much of the HIP-3 re-rating is already behind us, and the story that Hyperliquid is more than a crypto perp exchange is largely priced in. That doesn't invalidate the thesis. A platform that generates cash and compounds it through buybacks can, over time, detach from the broader crypto risk-on/risk-off cycle.

HIP-4 is the next growth leg the market hasn't priced yet due to its nascency.

HIP-4 expands the TAM beyond crypto-natives. The perp-to-PM user flow already exists, as the overlap data confirms. The next layer is event-driven traders — macro participants who want to express a Fed view, geopolitical outcome, or sports result. These users have never had a venue where those positions share margin with a BTC perp. That cohort doesn't exist on Hyperliquid yet and HIP-4 is the product that brings them in.

Invalidations

HIP-4 risks staying a crypto-native product if canonical event markets fail to build durable open interest beyond the launch window. CPI releases, Fed decisions, and World Cup outcomes are upcoming high volume events to watch. If those markets don't retain volume after the novelty fades, HIP-4 doesn't expand Hyperliquid's addressable market.

The thesis also weakens if the Polymarket overlap cohort doesn't consolidate onto HIP-4 once it's live either because incumbent liquidity on Polymarket is too thick to abandon, or because capital efficiency doesn't actually change trader preference.

The resolution model carries its own tail risk. One high profile canonical market resolving incorrectly damages trust before the category matures in the same way a single UMA incident sets the tone for Polymarket's settlement integrity.

Conclusion

The trade hinges on whether HIP-4 unlocks a new pocket of demand for sophisticated traders consolidating cross-instrument macro books onto one platform, or whether it remains a crypto-native product that repackages the same assets. The user overlap data shows this cohort exists. Capital efficiency via portfolio margin creates hard incentives. If volumes stay sticky across event types beyond BTC price binaries and the system exits growth mode, then HYPE's next leg will come from Hyperliquid becoming the unified margin layer where traders run their whole book.

References

Dune - @datadashboards

Dune - @filarm, HIP-4 Volume Data

Hyperliquid — HIP-4 Outcome Markets Documentation

Outcome.xyz — Crypto Price Binary Markets

Polymarket — Ukraine Mineral Deal Market

Polymarket — Will Zelenskyy Wear a Suit Before July

X/@0xFleck — Hyperliquid/Polymarket User Overlap Data

X/@j0hnwang — HIP-4 Proposal

X/@willo2_Poly — MSTR Resolution Dispute

Disclaimers

Alea Research has never had a commercial relationship with Hyperliquid and this report was not paid for or commissioned in any way.

Members of the Alea Research team, including those directly involved in the analysis above, may have positions in the tokens discussed.

This content is provided for educational purposes only and does not constitute financial or investment advice. You should do your own research and only invest what you can afford to lose. Alea Research is a research platform and not an investment or financial advisor.

Alea Research · 11 Jun 2026 · 13 min · 9 sections

Back to top