Hyperliquid !
The first on-chain order book that traders actually prefer to a centralised exchange, funded with no VC round and distributed almost entirely to users. Extraordinary execution, with validator centralisation as the standing refutation.
Hyperliquid is the strongest single piece of product execution in crypto over the last three years, and the reason is unfashionably simple: it made a decentralised exchange that feels like a centralised one. Not approximately. Not "good enough if you care about self-custody". Traders who have every option available to them — professionals with accounts everywhere — route size through Hyperliquid because the fills are good, the latency is imperceptible and the order book is deep. Every other perpetuals protocol in this archive asked users to accept a worse product for ideological reasons. Hyperliquid declined to make that trade, and it took the market.
The architecture is purpose-built rather than general-purpose, and that is the whole insight. Instead of deploying a matching engine as a smart contract on somebody else's chain and inheriting their block times, the team built a bespoke layer 1 whose consensus — HyperBFT — is tuned for one workload: a fully on-chain central limit order book. Orders, cancels, matches and liquidations are all consensus-level operations, not contract calls competing for blockspace with unrelated activity. The result is sub-second finality, order placement and cancellation that costs nothing, and an order book whose entire state is publicly verifiable. Nobody has to trust a sequencer's word about what filled and in what order.
The economic design is the other half. Hyperliquid held no venture round, sold no allocation to funds, and ran no private sale. In November 2024 it distributed roughly 31% of supply directly to users who had actually traded on the protocol, in what remains the largest genuine user airdrop in the industry's history, with no insider tranche unlocking into it. Fee revenue does not flow to a company; it flows to the assistance fund, which buys HYPE on the open market, and to the HLP vault, which lets ordinary users take the other side of market-making and liquidation flow that on a centralised exchange would be captured by a privileged internal desk. Revenue-to-holders, in production, at meaningful scale, with the numbers publicly visible on chain.
HyperEVM extended the thesis from an application into a platform. A general-purpose EVM environment now runs alongside the order-book core, with the two sharing state, so any deployed contract can read the book and route through it natively. That produces a composability advantage no centralised venue can match: lending markets that liquidate directly into deep on-chain liquidity, structured products that hedge atomically, vaults that execute strategies against a real order book instead of an AMM curve. The ecosystem that has grown on top is early but unusually coherent, because everything built there inherits the one thing Hyperliquid already had — genuine liquidity.
The JELLY incident in March 2025 is the episode that must be addressed, because it is the clearest window into the project's actual governance. A trader engineered a position designed to force the HLP vault into a catastrophic loss on a thin market. Validators voted to delist the market and settle it at a favourable price, protecting the vault. The defence was effective, the fund was preserved, and users lost nothing. It was also, unmistakably, a small set of validators intervening to change the outcome of a live market. The consequences of a real exchange failure would have been worse, and the team's post-incident work on position limits and market listing criteria was substantive. But the event demonstrated that Hyperliquid, at that moment, could be governed by hand. That is not a property a mature settlement layer should have.
Which brings us to the standing objection: validator decentralisation. The set is small, the foundation's stake is influential, and the primary node software has been closed source for much of the protocol's life, which materially limits independent verification and client diversity. Every advantage we have praised — the performance, the tuned consensus, the speed of iteration — is purchased with exactly this. It is a defensible trade for an exchange in its growth phase and an unacceptable one for infrastructure holding billions in user margin indefinitely. The team has been broadening the set and moving toward permissionless validation. That trajectory is the single most important thing to track, and it is the entire reason this review is not a five.
The bridge is the second concentration risk. Assets enter through a bridge secured by the validator set, which means the security of user deposits is not independent of the same small group. The contracts have been audited and the design is conservative, but bridge failures are the industry's most reliable source of nine-figure losses, and a bridge whose trust assumptions match its chain's trust assumptions offers no diversification of risk. We would want to see this hardened and independently reviewed on an ongoing basis.
Tokenomics deserve scrutiny even where they impress. A substantial share of supply remains reserved for future emissions and community rewards, and while the absence of a VC unlock cliff removes the ugliest overhang in the asset class, ongoing distribution is still a supply source that must be matched by growth in fee revenue. The buyback is real and sizeable, funded by actual usage rather than treasury sales, which is the correct structure. The dependency is that perpetuals volume is cyclical: a prolonged decline in speculative activity compresses the revenue that supports the entire flywheel.
Competitively, the position is strong but not unassailable. Centralised exchanges are responding, other chains are building order-book venues, and the technical moat — a fast custom consensus — is replicable by a well-funded team given eighteen months. What is much harder to replicate is the liquidity and the trust earned by a protocol that gave a third of its supply to its users and took nothing from a private round. Liquidity begets liquidity, and Hyperliquid's flywheel has been spinning long enough to be self-sustaining across a full market cycle.
Hyperliquid earns 4.5 out of 5. It has done what almost nobody thought possible — built a decentralised venue that wins on product merit against the best centralised competition — and it did so with a distribution model that treated users as owners rather than exit liquidity. The withheld half point is not a quibble: a small validator set, a closed-source core for much of its history, and one demonstrated instance of manual market intervention are exactly the conditions under which a great exchange becomes a captured one. Open the client, grow the set, and this becomes a five. Strong move: !