| name | why-hyperliquid |
| description | Why build on Hyperliquid. HyperBFT consensus, native orderbook, speed, AI agent angle, honest tradeoffs. Use when someone asks "should I build on Hyperliquid?", "why not Ethereum?", or when an agent needs to understand what makes Hyperliquid unique. |
Why Hyperliquid
What You Probably Got Wrong
"Hyperliquid is just a fast Ethereum L2." No. Hyperliquid is a purpose-built L1 with HyperBFT consensus and a native orderbook. HyperEVM is ONE component — it's an EVM environment that runs alongside HyperCore (the native exchange engine). They share finality. They are NOT the same thing.
"MEV sandwich attacks happen here too." They don't. HyperCore processes orders through HyperBFT consensus — there's no public mempool, no reordering by validators. Orders arrive, match, settle. No frontrunning. No sandwich. This is structurally different from any EVM chain.
"The native token is ETH." It's HYPE. HYPE is the gas token on HyperEVM, the staking token for HyperBFT validators, and the governance token. ETH has no role on Hyperliquid.
"I'll just use standard Ethereum tooling." HyperEVM works with standard EVM tools (wagmi, viem, Foundry). But HyperCore does NOT. HyperCore has its own API (/info, /exchange), its own token standards (HIP-1, HIP-2, HIP-3), and its own signing scheme (EIP-712 typed data over the exchange).
The Architecture in One Sentence
Hyperliquid is a purpose-built L1 where HyperCore (native orderbook, perps, spot) and HyperEVM (Solidity contracts) share atomic finality via HyperBFT consensus.
HyperBFT Consensus
HyperBFT is a BFT (Byzantine Fault Tolerant) consensus algorithm optimized for throughput and finality. Key properties:
- Sub-second finality: Transactions are final in ~1 block, not probabilistically safe after 6+ blocks like PoW chains
- No public mempool: Orders and transactions don't sit in a visible pool where bots can frontrun them
- ~100,000 orders/sec throughput: Native orderbook capacity that no EVM chain can match
- Validator set: Permissioned set of validators staking HYPE; slashing for equivocation
- Not PoW, not standard PoS: HyperBFT is a custom algorithm, not Ethereum's Gasper or Tendermint
What this means for builders:
- Finality is fast and hard. No need to wait multiple block confirmations.
- MEV extraction (sandwich attacks, frontrunning) is structurally prevented on HyperCore.
- Order execution is deterministic and fair. First-come, first-served at the protocol level.
HyperCore — The Native Exchange
HyperCore is the core of Hyperliquid. It's not an EVM contract. It's built into the L1 itself.
What lives in HyperCore:
- Perpetual futures: all perp positions, funding rates, liquidations, cross-margin
- Spot orderbook: native spot trading for HIP-1 tokens
- Vault system: protocol-native vaults (like HLP — the Hyperliquid Liquidity Provider vault)
- HIP-1/HIP-2 token standards: native token issuance and automated market-making
- Clearing: margining, settlement, liquidations
You interact with HyperCore via:
POST /info — read-only queries (no auth)POST /exchange — signed actions (orders, cancels, transfers)- WebSocket at
wss://api.hyperliquid.xyz/ws — real-time streams
HyperCore is what makes Hyperliquid unique. No other chain has a native perp DEX running at L1 consensus speed with no MEV.
HyperEVM — The EVM Layer
HyperEVM is a standard EVM environment running alongside HyperCore. It:
- Supports all standard EVM tooling (Foundry, Hardhat, wagmi, viem, ethers.js)
- Has Chain ID 999 (mainnet), 998 (testnet)
- Uses HYPE as the native gas token (not ETH)
- Supports Cancun-era opcodes (no blobs — Hyperliquid doesn't need data availability blobs)
- Burns priority fees (not paid to validators)
- Block time: ~1-2 seconds
- Has a system precompile at
0x2222222222222222222222222222222222222222 — send HYPE here to bridge from HyperCore to HyperEVM
HyperEVM and HyperCore share finality. A HyperEVM transaction and a HyperCore order that execute in the same block are final together — no cross-chain bridge risk between them.
Why Build on Hyperliquid
For Trading/DeFi Applications
- Native orderbook: No DEX approximation. Real limit orders, real price discovery, native CLOB.
- No MEV: Your users' trades execute at the price they expect. No sandwich attacks.
- Speed: Sub-second order acknowledgment. HyperCore doesn't have 12-second block times.
- Capital efficiency: Cross-margin across perps. Unified collateral. No fragmented liquidity across pools.
For AI Agents
- Programmatic trading: The
/exchange API is designed for algorithmic trading. Order placement is deterministic.
- No gas surprises: HyperCore order placement doesn't require gas — only EVM transactions do. Agents can place thousands of orders cheaply.
- Real-time data: WebSocket subscriptions give sub-second price, order, and fill data.
- Agent wallet pattern: Main wallet approves an agent subaccount key with spending limits. Agents never touch the main key.
For Token Launches
- HIP-1 tokens: Native spot tokens with built-in orderbook integration
- HIP-2 hyperliquidity: Protocol-level automated market-making — no need to bootstrap a Uniswap pool
- Bonding curve: x*y=k until graduation threshold, then full orderbook listing — predictable launch mechanics
Honest Tradeoffs
| What Hyperliquid Does Well | What It Doesn't |
|---|
| High-speed, high-volume trading | Ecosystem as deep as Ethereum/Arbitrum |
| Native orderbook with no MEV | Not EVM-only — HyperCore has its own API surface |
| Built-in token launch rails | Developer tooling still maturing vs Ethereum |
| Sub-second finality | Less composability surface (smaller protocol ecosystem) |
| Low gas costs on HyperEVM | HyperCore API learning curve |
| AI-agent-friendly trading API | Smaller DeFi liquidity than Arbitrum |
When NOT to choose Hyperliquid:
- You need deep DeFi liquidity composability (Aave, Curve, Balancer) → Arbitrum
- You need the largest EVM developer ecosystem → Ethereum mainnet
- You're building a simple token with no trading ambitions → Base or Ethereum
- You need ZK proofs or privacy → zkSync or Scroll
When TO choose Hyperliquid:
- Trading application, perp protocol, or high-frequency anything
- Token launch with built-in market-making
- AI agent that needs programmatic market access
- Any application where MEV-free execution matters
Current Network Stats (Early 2026)
- HyperCore throughput: ~100,000 orders/sec (measured)
- HyperEVM block time: ~1-2 seconds
- Finality: Sub-second (HyperBFT)
- HyperEVM Chain ID: 999 (mainnet), 998 (testnet)
- Native gas token: HYPE
- HYPE price: Volatile — verify on CoinGecko or Hyperliquid's own UI
Resources
