HIP-1 Explained: Hyperliquid's Native Token Standard

Hyperliquid

HIP-1 is Hyperliquid's native token standard, the spot-side equivalent of ERC-20. The difference that matters: a HIP-1 token is not a smart contract. It lives in HyperCore's own state, and the protocol gives it a real order book quoted against USDC the moment it exists. As of August 12, 2026 there are 485 HIP-1 tokens trading across 324 spot pairs.

It is also the least-discussed of Hyperliquid's standards, which is odd, because HIP-3 and HIP-4 both get written about constantly and HIP-1 is the one that every spot balance on the exchange actually runs on. That includes HYPE, PURR, the Unit assets, and the xStocks equity tokens listed in August 2026.

What a Deployer Actually Specifies

Deployment is a sequence, and only the first step is on a clock. The genesis transaction locks in the token's identity:

Two of those deserve a closer look.

Ticker names are not unique. The standard imposes no uniqueness constraint, so nothing at the protocol level stops someone deploying a second token called HYPE. Checking the 485 tokens live today, no collisions exist yet, which is a fact about who has bothered rather than a guarantee. Verify the token index or contract, not the three letters in the interface, before you buy anything unfamiliar.

anchorTokenWei is the interesting one, because it is the seed of everything in the last section of this article. It lets a new token's genesis supply be sprayed across the holders of an existing token, in proportion to their balances, with no claim step. Airdrop as a protocol primitive. Its limitation is that it fires exactly once, at genesis, and never again.

The measurements that come out of these choices are dull but load-bearing. Lot size on a book is 10 ** (weiDecimals - szDecimals), which is why 327 of 485 tokens use weiDecimals of 8 and 343 use szDecimals of 2. Get the combination wrong relative to your supply and the deployment can wedge, at which point the HYPE you paid is gone. Hyperliquid's own documentation tells deployers to rehearse the exact deployment on testnet first.

The Slot Has to Be Won First

You cannot deploy a HIP-1 token by paying a flat fee. Slots are sold through a 31-hour Dutch auction paid in HYPE, opening at twice the last clearing price and decaying linearly to a floor of 500 HYPE. When this article was written the live auction sat at 668 HYPE, down from a 1,000 HYPE start.

The same auction mechanic governs HIP-3 perp market slots, and our Hyperliquid auctions explainer covers the pricing dynamics, historical clearing levels, and where to watch both auctions live. The short version of why it exists: a free listing standard becomes a spam listing standard, and a fixed price is either too high in a bear market or too cheap in a mania. The auction lets the market set it.

HIP-2: Liquidity Without a Market Maker

A brand-new token with no market maker has no book, and a token with no book has no price. HIP-2 hyperliquidity solves this inside the protocol rather than outsourcing it.

The deployer seeds a starting price, a number of levels, an order size, and how many bid levels to fund. From there the protocol runs the strategy itself as part of block transition logic. Each price level sits roughly 0.3% above the last (px_i = round(px_{i-1} * 1.003)), and the ladder refreshes on any block at least three seconds after the previous update. Fills on one side rotate into orders on the other.

There is no operator, no keeper bot, and no off-chain process. The same consensus that matches trades also maintains these orders, which means the strategy cannot be turned off, front-run by its own operator, or quietly abandoned when the team loses interest. Manual market makers coexist with it in the same book and can tighten the spread whenever they want to.

Deployer Fee Shares, and What Deployers Actually Choose

A HIP-1 deployer can keep a slice of the trading fees generated on their token's book, up to half of what traders pay. Two rules constrain it: the share can only be ratcheted down once set, never back up, and anything the deployer declines is burned rather than redirected to the protocol or to liquidity providers.

That second rule makes the setting a genuine choice rather than a formality, and the distribution across all 485 live tokens shows deployers treat it as close to binary:

Nine tokens out of 485 picked a middle value. Everyone else either takes everything available or takes nothing and sends the fees to the burn. If you have ever wondered whether a token's team is monetizing your trades, that number is public and worth checking.

This is the spot analogue of the HIP-3 deployer fee scale, which since August 2026 has been a continuous value from 0.1 to 3 set per asset. Spot and perps use the same underlying formula but different ranges, and the full fee guide has the numbers for both.

The Parts Nobody Mentions Until They Bite

Dust gets swept daily. At 00:00 UTC the protocol collects balances smaller than one lot size and worth a dollar or less, market sells them together, and redistributes the USDC proportionally to the people who were dusted. Balances too small to sell are burned. The sweep skips tokens with one-sided books, or when total dust exceeds 3,000 USDC (10,000 for PURR). What you get back can be less than the mid-price implied, because the aggregate sale eats slippage.

165 of 485 tokens are linked to HyperEVM. Linking assigns the token a system address, and an ERC-20 transfer to that address moves the balance to HyperCore, with the reverse also true. This is the mechanism behind Unit's uBTC and uETH, behind bridged stablecoins, and behind the wrapped equity tokens. It is also the reason the next section comes with a caveat.

Supply can fall but never rise. maxSupply is fixed at genesis. Inflation is not a feature a HIP-1 token can add later.

What's Proposed Next: scaleWei

In August 2026, Hyperliquid founder Jeff Yan posted a HIP-1 extension in Discord, describing it as a response to builder feedback and inviting more. It is a single deployer-controlled function:

scaleWei { token, totalWei, referenceToken, systemAddress }

Here is the announcement in full, since the details matter and most of the commentary has been secondhand:

Source: Jeff Yan (jeff_hl) in the Hyperliquid Discord, August 2026, used under fair use for educational purposes

The function atomically transfers totalWei of token out of systemAddress and distributes it across every holder of referenceToken, in proportion to their balances, rounding down and excluding the system address itself.

Read the signature twice and the scope becomes clear. It is anchorTokenWei freed from genesis and made repeatable.

When token and referenceToken are different, you are paying holders of one asset in another asset. That is a dividend. It is also an airdrop, a revenue share, a staking distribution, or a rebate, depending on what you call it in your announcement.

When they are the same, the token is scaling against itself, which is a redenomination. A positive totalWei is a stock split. And because the proposal explicitly allows totalWei to be negative, the reverse works too: a reverse split. In that case open orders are canceled and replaced at the new effective scale, rounded down to szDecimals, so the book survives the operation instead of being left quoting stale prices. Call it repeatedly on a small scale and you have a rebasing token.

None of these require the user to do anything. No claim page, no gas, no deadline, no forgotten allocation sitting in a contract for two years.

The Caveat Everyone Skipped

The announcement says plainly that no equivalent atomic functionality exists on the EVM, and that a linked contract may need custom logic to apply the same operation to HyperEVM balances.

That matters more than it sounds. A third of live HIP-1 tokens are EVM-linked. If a token's supply is split between HyperCore balances and HyperEVM balances, and half of it is sitting in a lending market or an LP position on the EVM side, then a scaleWei call updates one half cleanly and leaves the other half to whatever the token's contract was built to handle. "No claiming required" is true on HyperCore. It is a design problem everywhere else.

Worth noting too: the tokens come out of a systemAddress the deployer funds, either the Core-to-EVM system address or a treasury address that signs off. This is redistribution of something that already exists, not minting. And because the split rounds down, the arithmetic leaves dust behind.

Why Equity Tokens Are the Obvious Use Case

Speculation about tokenized stocks followed the announcement within hours, and the logic holds up. Real shares do things tokens historically could not: they split, they reverse split, and they pay dividends. Every one of those is a balance-level operation applied to all holders at once, which is exactly the shape of scaleWei.

Hyperliquid now hosts ten share-backed equity tokens on spot. Those instruments will eventually face a corporate action, and the alternatives today are a claim contract or a manual redeployment. Whether this proposal was written with them in mind is not something the announcement says, and the function is general enough to be useful without them.

How to Read a HIP-1 Token Before You Buy It

The standard gives you more public information than most chains do. Use it.