Hyperliquid Improvement Proposal 4 (HIP-4) is the designation used for two distinct proposals aimed at expanding the trading capabilities of the Hyperliquid decentralized exchange. The most recent proposal, announced by the Hyperliquid team on February 2, 2026, focuses on introducing "Outcome Trading" through a new primitive called "Outcome Contracts" to enable prediction markets and options-like instruments ^{[1] [2]}. An earlier proposal from September 2025, also named HIP-4, was submitted by third-party contributors and focused on a product called "Event Perpetuals" to solve challenges related to deploying prediction markets on the platform ^{[3] [4]}.

HIP-4: Outcome Trading (February 2026 Proposal)

On February 2, 2026, the Hyperliquid team announced its own version of HIP-4, introducing a new primitive called "Outcome Trading" to the protocol ^[2]. This proposal aims to support a wider range of financial products beyond the platform's core offering of perpetual futures. The development was initiated in response to what the team described as "extensive user demand" for both prediction markets and options-style derivatives ^[1].

Key Features and Technology

The core of this proposal is the "Outcome Contract," a new general-purpose primitive designed to be built on Hyperliquid's core infrastructure, HyperCore ^[2].

Key characteristics of Outcome Contracts include:

• Full Collateralization: Unlike leveraged perpetuals, these contracts are fully collateralized, meaning the maximum potential loss is the capital committed upfront. This design completely removes the risk of liquidations for traders of these instruments ^[1].
• Fixed Settlement Range: Each contract settles at a fixed price within a predetermined range, making them suitable for binary or bounded outcomes ^[2].
• Dated Markets: The proposal introduces contracts with specific expiration dates, a feature distinct from the platform's existing perpetual offerings ^[1].
• Non-Linear Payoffs: Outcome Contracts allow for more complex profit-and-loss structures compared to the linear payoffs of perpetual futures ^[1].
• Compositionality: The primitive is designed to be compatible with other Hyperliquid features, such as portfolio margin and the HyperEVM, allowing developers to build novel applications ^[2].

The primary initial applications for Outcome Contracts are prediction markets and bounded, options-like derivatives. The canonical markets are planned to be denominated in USDH ^[2].

Proposed Implementation Roadmap

The rollout of Outcome Trading is planned in multiple phases to ensure stability and incorporate community feedback ^[1]. As of the announcement on February 2, 2026, the feature was actively being developed and tested on the Hyperliquid testnet ^[2].

The planned phases are:

1. Phase 1: Testnet Launch: Initial deployment on a testnet for final technical development and testing.
2. Phase 2: Mainnet Launch with Curated Markets: Following a successful testnet phase, the feature will launch on the mainnet with a limited set of markets curated by the Hyperliquid team and settled using objective data sources ^{[1] [2]}.
3. Phase 3: Potential Permissionless Listing: Depending on the success of the initial phases, the infrastructure may be opened up for permissionless deployment. This would allow any user to create their own outcome-based markets, potentially under a similar model to Hyperliquid's permissionless perpetual markets, which require staking 500,000 HYPE tokens ^[1].

Market Impact

The announcement of HIP-4 for Outcome Trading on February 2, 2026, was followed by a positive reaction in the market for Hyperliquid's native token, HYPE. The token's price increased by approximately 10% in the 24 hours following the news. This came during a period of already strong momentum for the token, which had appreciated by 33% in the prior month, driven by high activity in its existing permissionless markets for assets like cryptocurrencies, stocks, and gold ^[1].

Previous HIP-4 Proposal: Event Perpetuals (September 2025)

Prior to the team's official 2026 announcement, another proposal also designated HIP-4 was introduced on September 16, 2025 ^[4]. Titled "HIP-4: Event Perpetuals," this proposal was authored by four contributors, including John Wang, the Head of Crypto at the CFTC-regulated prediction market Kalshi, and the research entity Bedlam Research ^{[3] [4]}.

Rationale and Problem Addressed

This version of HIP-4 was created as an extension of HIP-3, which had enabled builders to deploy their own perpetuals markets. The authors identified that HIP-3's mechanics were unsuitable for binary outcome prediction markets. The key limitations were ^[4]:

• Slow Price Adjustment: The existing infrastructure restricted price changes to 1% per tick, making it too slow to settle a market to a definitive 0 or 1 outcome. This "asymmetric resolution problem" could take nearly an hour for a market to decay to zero.
• Arbitrage Vulnerabilities: The slow settlement time created significant arbitrage opportunities for traders who had information about a resolved event before the market price could catch up.
• Oracle Dependency: The system's reliance on a continuous oracle price feed was not practical for events that resolve instantaneously, such as the outcome of a sports game or an election ^[3].

Technical Design of Event Perpetuals

To solve these issues, the proposal introduced "Event Perpetuals," a specialized derivative designed for binary prediction markets ^[4].

Market Launch

A key innovation was a 15-minute opening auction to establish an initial market price. The system would algorithmically determine a single clearing price that maximized trading volume, minimized the buy/sell imbalance, and, as a final tie-breaker, was closest to 0.50. All initial orders would execute at this single price before the market transitioned to continuous trading ^[4].

Trading and Settlement

• Oracle-Free Trading: After the opening auction, trading would occur freely on the order book without a continuous oracle feed, with prices representing the market-implied probability of the event ^[3].
• No Funding Rate: Unlike standard perpetuals, these markets would not have a funding mechanism ^[4].
• Binary Settlement: Upon the event's conclusion, a resolution oracle would post the final outcome (0 or 1), trading would be halted, and all positions would be settled instantly. The proposal included an optional dispute window before final settlement ^[3].
• Margin: The proposal specified 1x isolated margin, requiring traders to fully collateralize their positions based on their maximum potential loss ^[4].

Deployment and Governance Requirements

Under this proposal, builders could deploy their own Event Perpetual markets by staking 1,000,000 HYPE tokens ^[3]. Builders would be responsible for defining the event terms and resolution criteria and would be granted the ability to set and collect additional trading fees of up to 50% on their deployed markets ^{[3] [4]}.

Underlying Technology: HyperCore

Both HIP-4 proposals are designed to be implemented on HyperCore, the specialized Layer 1 blockchain that powers the Hyperliquid exchange ^[5]. HyperCore is designed for high-performance decentralized finance, with its state split between the core matching engine and a general-purpose, EVM-compatible environment known as the HyperEVM. Unlike many exchanges, it does not use an off-chain order book, instead processing all transactions through its on-chain consensus mechanism ^[5].

Architecture and Consensus

Hyperliquid's L1 uses a Proof of Stake (PoS) consensus algorithm named HyperBFT, which is a variant of HotStuff consensus. Validators produce blocks in proportion to their staked native tokens, and the mechanism is optimized for low end-to-end latency ^[5].

Performance and Scalability

According to documentation updated in early 2025, the Hyperliquid mainnet demonstrated significant performance capabilities:

• Latency: For a geographically co-located client, the median end-to-end latency (from request sent to response received) was 0.2 seconds, with the 99th percentile at 0.9 seconds.
• Throughput: The chain could process approximately 200,000 orders per second. The documentation noted that the primary performance bottleneck was the execution logic, while the underlying consensus and networking layers were capable of scaling to "millions of orders per second" with future optimizations.

These technical specifications form the foundation upon which features like Outcome Trading and Event Perpetuals are built ^[5].