What is Hyperliquid (HYPE)?

---

The Book Anyone Can Read: Why Hyperliquid Exists, and Why HYPE Might Matter

Imagine an exchange that cannot keep a secret.

Every order you place is written into a public ledger. Every cancellation. Every fill. Every liquidation. Nobody in a back office can quietly reroute your trade, invent a price, or borrow the coins you thought were sitting in your account. The trading engine is not a black box in a server room. It is a machine whose rules run in the open, and whose history is a chain that anyone can inspect.

That sounds like a slogan until you remember what the opposite looked like.

In November 2022, FTX collapsed and took with it the pleasant fiction that a popular crypto exchange is just a shop with a sign on the door. Customers had deposited coins and FTX had treated those deposits as a private slush fund. The order book people thought they were trading against was, in important ways, a story the house was telling. When the story broke, the money was gone.

Hyperliquid was built as an argument against that kind of story. Not a moral lecture. A piece of machinery. If the matching engine lives on a blockchain, and the collateral never leaves the trader’s control in the old custodial sense, then the exchange cannot quietly become a casino with other people’s chips. The cost of that honesty is brutal: the chain has to be fast enough that professional traders will actually use it. Most blockchains are not. That is why this project did not start by renting space on Ethereum or Solana. It started by building its own.

The token that later appeared on that chain is called HYPE. The interesting question is not whether HYPE has gone up. Plenty of tokens go up. The interesting question is why an exchange this busy needed a coin at all - and whether the coin is doing real work, or merely collecting applause from the volume it sits on top of.

---

A trader who decided the venue was the problem

The story does not start with a token sale. It starts with a person who already knew how to extract money from markets, and then decided the markets themselves were poorly built.

Jeff Yan grew up in the Bay Area, the child of Chinese immigrants, raised largely by a mother who worked as an accountant. He won gold at the International Physics Olympiad, then studied mathematics and computer science at Harvard. After a stretch in high-frequency trading - including time at Hudson River Trading - he built an anonymous crypto market-making operation called Chameleon Trading, named after an old video-game handle. He started it with savings on the order of ten thousand dollars and, for a few years, ran it as a private money machine from Puerto Rico.

That is not the usual founder myth, and it is worth pausing on what it actually means. Yan was not a marketer who needed an exchange. He was a professional who already had one, in the only sense that matters to a trader: he could trade. What he did not have was a venue he trusted. Centralized books were fast and deep and opaque. Decentralized books, on the chains that existed, were honest and slow. After FTX, the opacity stopped being an inconvenience and started looking like a design flaw with a body count.

He and a small group around Hyperliquid Labs - including a Harvard classmate who goes by iliensinc - looked at every chain they could find and decided none of them could host a real order book. So they wrote one. The team that formed around that decision was tiny by the standards of a company that would later process billions of dollars a day. Public accounts have described a core group on the order of a dozen people, with backgrounds at places such as Citadel, Hudson River Trading, Caltech, and MIT. They took no venture capital. Yan has said the point of that refusal was not romance. If the platform was going to claim it was a neutral venue, it could not begin life owing a boardroom.

Hyperliquid opened near the end of February 2023. The first users were not hedge funds. They were, in the early weeks, people who had never traded a perpetual contract, placing tiny orders and learning leverage in paper competitions. A points program followed that November - weekly scores without a public formula, on purpose, to make life harder for farms that reverse-engineer an airdrop and leave. Market makers who were used to being paid to show up were told, in so many words, to show up or not. The exchange would live or die on whether the book was actually good.

On November 29, 2024, the chain issued its native token. About 310 million HYPE - 31 percent of a one-billion cap - landed in roughly ninety-four thousand wallets that had used the platform. No claim portal. No vesting for those coins. No allocation to venture funds, centralized exchanges, or hired market makers. If you had been there, you woke up owning a piece of the machine.

That distribution is not automatically virtuous. Early users can dump. A team that skipped venture capital can still own a large future slice. But it is a different starting condition from the usual crypto launch, and it is the condition Hyperliquid chose.

---

What “on-chain exchange” actually means

Most decentralized exchanges are not exchanges in the old sense. They are pools.

You deposit two assets into a tank. A formula quotes a price. You swap against the tank. That design - the automated market maker - is one of the cleverest ideas in crypto, and it is the wrong tool for a professional trading floor. A real floor is an order book: a ranked list of people willing to buy and people willing to sell, matched by price and time. Centralized exchanges run that list on private servers because the list changes thousands of times a second. Putting the same list on a typical blockchain is like trying to run a stock exchange through the postal service.

Hyperliquid’s answer is a purpose-built layer-1 chain whose main job is to be that postal service, except the letters arrive in a fraction of a second.

The chain has two rooms that share one roof.

HyperCore is the trading room. It holds the perpetual-futures book and the spot book. Margin, funding rates, liquidations, and fills all live here. The project’s own documentation says mainnet can take on the order of 200,000 orders per second, with median end-to-end latency around two-tenths of a second for a well-placed client. Those numbers will move as the software is tuned. Treat them as a claim about what the live chain is built to do, not as a promise that every wallet on earth will feel like a colocated server.

HyperEVM is the workshop next door. It is an Ethereum-style environment, so developers can write familiar smart contracts, pay gas in HYPE, and - this is the unusual part - reach into HyperCore. A contract can read positions, balances, oracle prices, and vault equity, and it can send certain actions back to the trading engine. The two rooms are not separate chains glued together with a bridge. They share one consensus and one ordered list of transactions.

That consensus is called HyperBFT, a proof-of-stake design in the HotStuff family. Validators who have staked HYPE propose and vote on blocks. More than two-thirds of the stake has to agree before a batch is final. Finality is meant to be immediate in the practical sense that matters to a trader: once the chain says the fill happened, it does not wander off and unhappen.

Around that core sit the products people actually click.

Perpetual futures are the flagship. A perp is a bet on a price that never expires. You can go long Bitcoin without owning Bitcoin, or short it without borrowing it, and the contract pays a periodic funding rate so the bet does not drift too far from the real market. Leverage makes the bet louder. It also makes liquidations a regular event rather than a rare disaster. Hyperliquid’s pitch is that those liquidations, like everything else, happen in public.

Spot markets sit on the same engine, so the chain is not only a casino for derivatives. You can trade the assets themselves.

HLP, the Hyperliquidity Provider vault, is a communal pot that helps make markets. Depositors put up capital. The vault quotes and takes risk. Profits and losses flow back to the pot. This is useful and double-edged. A shared market-maker can deepen the book. It can also become the thing that eats a bad trade when a thin market is shoved around - which is exactly what later happened, and which this article will not skip.

HIP-3 opened the listing desk. Instead of waiting for a core team to bless every new perpetual market, a builder who stakes a large amount of HYPE - 500,000 on mainnet, locked for a minimum stretch after deployment - can launch a book with its own oracle, leverage limits, and settlement rules. That is how markets on equities, commodities, and other things that are not “just another coin” have started to appear. It is also how oracle risk stops being only the protocol’s problem and becomes the deployer’s.

If you have used Binance or Coinbase, the screen will feel familiar: prices, sizes, a chart, a button that says buy or sell. The difference is underneath. On a centralized venue, “your” coins are an entry in a company’s database. On Hyperliquid, the exchange is a program, and the program does not take custody in the FTX sense. That does not make you safe from leverage, bad oracles, or your own fingers. It makes you safe from a particular kind of theft: the kind where the house spends the inventory.

---

Why the chain had to exist before the token did

There is a lazy story in crypto that every product needs a coin, and the coin is the product. Hyperliquid ran for well over a year without one. Volume showed up anyway. That is the first clue that HYPE is not a ticket you buy to enter a ghost town.

The second clue is architectural. A proof-of-stake chain needs something to stake. An EVM needs something to pay gas. A listing market that wants skin in the game needs something expensive to lock. Fee revenue needs somewhere to go that is not a founder’s checking account if the project wants the success of the venue to touch the people who hold the asset. Those are four different jobs. Hyperliquid assigned all four to the same ticker.

That is either elegant or crowded, depending on your taste. It is not accidental.

---

The token, in plain English

HYPE is the native asset of the Hyperliquid blockchain. The maximum supply is one billion. That ceiling was set at genesis. New coins are not printed without limit the way some proof-of-work or inflationary chains print them. Coins can, however, enter circulation from warehouses that were labeled on day one, and coins can leave circulation when the protocol buys them and treats them as gone. Both directions matter. A cap without a schedule is only half a sentence.

The official split of that billion, published by the Hyper Foundation at genesis, looks like this:

• 38.888% - future emissions and community rewards (388,880,000 unminted at genesis)
• 31.0% - genesis distribution to early users (310,000,000), fully unlocked
• 23.8% - current and future core contributors
• 6.0% - Hyper Foundation budget (60,000,000)
• 0.3% - community grants (3,000,000)
• 0.012% - HIP-2, a sliver for on-chain liquidity (120,000)

Add the user-facing buckets and you get the number the project likes to quote: about 76 percent earmarked for community programs rather than the team. Core contributor tokens were locked for one year after November 29, 2024. The Foundation said most vesting would finish between 2027 and 2028, with some schedules running later. There was no private-investor tranche, which removes one familiar source of unlock drama and leaves another: the team’s own coins, and the large reserve for “future emissions” that does not come with a public, month-by-month calendar the way Bitcoin’s halving calendar does.

Trackers disagree about how many coins are freely circulating at any given moment, because they treat staking, foundation wallets, burned balances, and unclaimed vestings differently. As of early September 2026, widely used figures cluster around the low-to-mid two hundred millions for circulating supply, against a total supply in the mid-nine-hundred millions after burns, and a fully diluted value that still assumes the rest of the cap exists. HYPE was trading in the mid-eighties of dollars, with a circulating market value often quoted near the high teens to low twenties of billions. Those prices will be stale the week after you read them. The structure will not.

So what is the token for?

It pays for computation. On HyperEVM, HYPE is gas. Base fees are burned. Because of the way HyperBFT works, priority fees are burned too, rather than tipped to a block producer the way they are on Ethereum. That is a small, continuous drain on supply whenever the workshop is busy.

It secures the chain. Validators must self-delegate at least 10,000 HYPE, locked for a year, to register. Holders who do not want to run a node can stake to someone who does. Staking rewards come out of that future-emissions bucket. Reported yields have been modest - a couple of percent annually in recent snapshots - which is a feature if you think security should not require double-digit inflation, and a yawn if you came for a savings account.

It is a ticket into the listing business. HIP-3’s 500,000 HYPE bond is not a souvenir. It is supposed to make a careless oracle expensive. Whether 500,000 is the right number will be argued for years. The design idea is clear: if you want to open a new derivatives pit on this chain, you put a pile of the native asset in harm’s way.

It is a governance weight. Validators vote on matters such as delisting a market. That power is not theoretical. It has already been used in anger.

It is the thing the exchange is structurally pushed to buy. This is the mechanism that made HYPE famous among people who do not trade perps. The protocol routes the vast majority of trading fees - figures of 97 to 99 percent are the ones that circulate in research and in the project’s own framing - into an Assistance Fund that purchases HYPE on the open market. In December 2025, validators voted to treat the Fund’s holdings as burned rather than as a rainy-day pile that might be spent. Cumulative buybacks have been measured in the high hundreds of millions to more than a billion dollars, depending on the window you pick. Trailing-twelve-month figures published in 2026 have put fee capture on the order of a billion dollars, with the large majority of that recycled into HYPE.

Read that last paragraph twice, because it is easy to turn into a fairy tale.

You do not own a claim on the exchange’s cash. You own a commodity that the exchange’s rules currently compel the protocol to keep buying, in proportion to how much people trade. That is closer to a corporate buyback than to a dividend, and closer to a policy than to a law of nature. Governance can change the percentage. Volume can fall. A quiet tape means a quiet bid. The flywheel is real when the floor is crowded. It is a parked fan when the floor is empty.

There is a second supply story running in the opposite direction. Contributor tokens began coming due after the one-year cliff. On paper, linear vesting implied a heavy monthly drip. In practice, through 2026, on-chain claims have often been a small fraction of the theoretical maximum - hundreds of thousands of tokens in months when the schedule could have allowed millions. That is a fact about behavior so far, not a covenant. Behavior can change. The 388.88 million reserved for future emissions is a larger long-term question than any single month of team vesting, because “as needed” is not a date.

Put the pieces together and HYPE is supposed to be three things at once: the fuel and bond of a chain, the security deposit of a marketplace, and the sponge that soaks up the marketplace’s fees. The token needs the exchange. The exchange, strictly speaking, ran without the token. What the token adds is a way to turn the exchange’s cash engine into pressure on a scarce asset, and a way to make validators, builders, and users share one unit of account.

---

Why this token might matter when a thousand other tickets exist

Crypto is full of exchange tokens. Most of them are coupons: hold this, pay a slightly lower fee, receive a governance vote nobody attends, hope the company keeps doing well. A coupon is not nothing. It is also not a reason for a new asset to exist.

HYPE is trying to be a different object.

The first difference is the product underneath. Hyperliquid did not bolt a token onto a thin wrapper around someone else’s chain. It built a matching engine that other on-chain venues had spent years saying was impractical, then let that engine eat a startling share of decentralized perpetual trading. By mid-2026, independent industry tallies were putting Hyperliquid in the front of the perp-DEX pack by open interest - on the order of several to nearly ten billion dollars depending on the month - and often in the neighborhood of a third to two-thirds of decentralized perp open interest among the large venues. Volume share moves around as competitors launch points programs and fade. Open interest is the stickier number. It is capital that chose to sit in these books rather than somewhere else. A slice of that interest has also moved into markets that look like traditional finance - equities, commodities - which is a hint that the chain wants to be more than a crypto-native casino.

The second difference is the fee pipe. Plenty of protocols say they will “share revenue.” Few take almost all of it and spend it, every day, buying their own token in public. Whether that is wise capital allocation is a separate debate - a growing company in the ordinary world might rather hire, or cut fees, or insure the vault. As a design for tying a token to a business, it is unusually direct.

The third difference is the refusal, at birth, to sell the future to a fund. That does not make the team selfless. They still have 23.8 percent. It does mean the usual venture unlock calendar is missing, and that the first giant distribution went to people who had already been trading rather than to people who had written a check.

The fourth difference is the attempt to turn the exchange into a platform. HyperEVM plus HIP-3 is a bet that other people will build books, vaults, and applications against the same liquidity, and that some of those builders will need HYPE to deploy, to pay gas, or to stake. If that bet works, HYPE is not only a claim on one venue’s fees. It is the reserve asset of a small financial city. If that bet fails, HYPE is the coin of a very successful derivatives shop - which is still a business, just a narrower one.

That is why the token can justify its existence in a crowded market. Not because the ticker is clever. Because the thing it meters - an on-chain order book that professionals will actually touch - was scarce, and because the protocol’s cash flows are wired, today, into the ticker itself.

---

What can go wrong

An honest article has to stay in the room after the applause.

The chain is faster than it is open. Hyperliquid’s node software has remained closed source while the team says it will open when the system is stable. Validators run a binary they cannot fully audit. Users can see the resulting state. They cannot independently prove, the way they can with Bitcoin or Ethereum, that the program that produced the state is the program they think it is. That is the most serious decentralization critique that does not depend on any particular scandal. Stake can be distributed. Geography can improve. None of that replaces readable code.

The validator set is still small, and influence is still concentrated. The active set has grown from a handful of nodes to the mid-twenties or low thirties, depending on the snapshot. Foundation-linked validators have at times controlled a large share of stake - reports in 2025 put that share painfully high; later 2026 snapshots showed it falling, in some tallies toward half or lower. Direction matters. So does the residual fact: a vote of “the chain” can still look like a vote of people who built the chain.

March 26, 2025, made that fact impossible to ignore. A trader built a leveraged structure around a thin memecoin perpetual, JELLY. Price was shoved around on outside venues. Liquidations dumped toxic risk into HLP. Validators then delisted the market and force-settled positions at a chosen price. The vault was spared a much larger hole. Ordinary users, other than flagged addresses, were made whole by the Foundation. Critics - including people with no reason to be kind to centralized exchanges - said the quiet part out loud: this was an emergency brake pulled by a small set of operators, not an autonomous market absorbing a loss. They were not wrong about the description. Defenders were not wrong that an unmanaged blowup could have been worse. Both things can be true. A venue that will rewrite a price when the house is on fire is safer for depositors in that hour and less neutral forever after. Hyperliquid later pushed delistings toward a clearer on-chain vote. The scar is still part of the product.

Oracles do not become honest because the book is on-chain. HIP-3 pushed listing risk out to deployers. That is healthier than pretending the core team can babysit every market. It does not abolish the problem. When a third-party feed prints a bizarre but “valid” number, liquidations still fire. In 2026, a major HIP-3 deployer reimbursed traders after an anomalous equity print. Generosity is not a protocol guarantee.

The economic engine is a volume engine. Fees come overwhelmingly from perpetuals. Perpetuals thrive on speculation, leverage, and the presence of someone on the other side of the bet. A long quiet market, a successful competitor, a regulatory clamp, or a simple change in trader fashion shrinks the Assistance Fund’s bid. HYPE does not have a separate industrial buyer. It has traders.

Supply is not finished with holders. Contributor vesting runs for years. The emissions reserve is large and only partly scheduled in public. Team restraint so far is comforting and unenforceable. A capped supply can still feel inflationary if the warehouse door opens faster than the burn.

The buyback is a policy. It is an unusually aggressive policy, and it has been reinforced by a burn vote. It is not a legal right attached to the token. People who treat HYPE as if it were preferred equity in an exchange are reading a metaphor as a contract.

Users can still lose everything the ordinary way. Self-custody does not save you from 20x leverage. A transparent liquidation is still a liquidation. The chain can be honest and the trade can be stupid.

Regulators have not agreed to the premise. A fast, non-custodial derivatives venue with no KYC at the protocol layer is a triumph of design and a magnet for scrutiny. The legal wrapping around perpetual futures is messy in the United States and in plenty of other places. A product can be technically elegant and still become difficult to touch from the countries where the capital lives.

Competition is not theoretical. Other perp venues have already taken turns stealing headline volume with incentives. Hyperliquid’s lead in open interest is a moat only for as long as the book stays better. Books can be copied. Trust is slower to copy. Trust is also slower to repair.

None of these are reasons the project is fake. They are the bill that comes with trying to put a trading floor on a chain without waiting for the chain to become a trading floor on its own.

---

What would have to go right

For HYPE to be more than a successful season, several ordinary and several strange things have to keep being true at once.

The book has to remain a place professionals want to stand. That means latency, depth, uptime, and the unglamorous work of not surprising market makers.

The fee pipe has to keep pointing at the token, or some equally hard-edged successor, without being quietly rerouted into a slush fund when growth gets expensive.

The emissions reserve has to be spent like a treasury, not like a faucet.

The validator set has to keep spreading, and the node software has to become something a skeptical engineer can read. Until that happens, “decentralized exchange” will always arrive with quotation marks in a serious conversation.

HIP-3 and HyperEVM have to produce businesses that are not just more ways to lever memecoins - or, if they do produce those, they have to produce them without handing the chain a new JELLY every quarter.

And the legal weather has to stay survivable. Code does not repeal securities law, commodities law, or the patience of large countries.

If those pieces hold, HYPE is one of the few tokens in circulation that is wired into a real, high-velocity financial machine rather than into a story about a machine that might exist later. The exchange would still matter if the token vanished tomorrow. The token would not matter if the exchange did. That asymmetry is healthy. It is also the test.

Hyperliquid is trying to prove that the oldest object in finance - a list of bids and offers - can live in public without becoming a toy. HYPE is the asset that chain uses to pay its guards, fuel its contracts, bond its new markets, and recycle its take. The distinctive claim is not that trading is exciting. Trading was exciting on FTX. The distinctive claim is that this time the book is on the wall, the fees have somewhere honest to go, and the people who showed up early were handed the keys instead of a waitlist.

Whether that is a new kind of exchange or a very fast shop with extra steps will be decided the same way every exchange is decided: by whether the book is still there, and still fair, when someone tries to break it.