Your feed this week has a new pitch: a builder vault on Hyperliquid offering 40% APY on a perpetual market you have never traded, for an asset you cannot pronounce. The chart is green, the lockup is short, and the fee is suspiciously low. You have seen this film before. This time the mechanism is genuinely different — and that difference is the point.
These markets are not rogue. They are built on Hyperliquid's own HyperCore stack, deployed under a rule called HIP-3, and backed by someone who staked real HYPE to bring them into existence. That changes the questions you should ask — not whether you should ask them. HIP-3 (Hyperliquid Improvement Proposal 3) lets third-party builders deploy and operate their own perpetual markets on Hyperliquid, choosing the oracle, the margin asset and the market configuration, and collecting a share of the fee flow their market generates. In practice it turns Hyperliquid from one exchange into a platform on which many exchanges can be built.
By the end of this piece you will know what a builder can actually control, how the fees split between the builder and the protocol, and why a builder market and the flagship HLP vault are two completely different trades. Fee numbers stay consistent with the rest of the site: Hyperliquid 0.045% taker / 0.015% maker, versus dYdX 0.05%/0.02% and GMX ~0.05%. Education, not financial advice; every number below has a risk attached to it.
TL;DR. HIP-3 lets any builder who stakes 500,000 HYPE deploy their own perp markets on Hyperliquid, defining the oracle, margin asset and contract specs, with independent margining but the same shared liquidity engine. The builder earns a configurable share of trading fees — per current documentation an added fee scale of 0% to 300% (0–100% in growth mode), with the protocol fee rising to match above 100%; the commonly cited baseline is a roughly even split. On $50M of daily volume at a 0.045% taker rate, that is about $22,500 a day, or ~$8.2M a year, divided between builder and protocol. For a yield seeker, HIP-3 opens new venues and builder-run vaults that can pay more than the proven flagship HLP vault — but with thinner liquidity, a builder-chosen oracle, incentive misalignment and the real possibility that the market is wound down. Read the split, then read the risks.
What HIP-3 actually lets a builder do
Before HIP-3, Hyperliquid listed the markets it wanted to list. If you wanted to trade a perpetual on some long-tail asset, you waited for the protocol to add it. HIP-3 flips that: it hands market creation to anyone willing to post the stake. A deployer locks 500,000 HYPE and can then stand up a perp DEX, define which markets it lists, pick the oracle for each market, set the contract specifications, and choose the margin asset. Per the documentation, each HIP-3 DEX runs with independent margining, which matters because it isolates the builder's markets from the flagship book: a blow-up on a thin builder market does not automatically cascade into the main exchange's margin system.
Most of the plumbing is inherited, and that is the selling point. HIP-3 markets inherit the full HyperCore stack — the high-performance order book, matching and margining engine that make Hyperliquid fast. A builder does not have to build an exchange; it operates a market on an existing one. That is why a small team can launch a perpetual on a niche index, a basket, or a tokenized asset the core protocol would never prioritize.
The trade-off is baked in: the builder controls the parameters that decide how good or how dangerous the market feels. Oracle choice, margin asset, leverage caps and the fee scale are all builder decisions. When you trade a HIP-3 market you are not choosing "Hyperliquid" — you are choosing a specific builder's operating choices. Our Hyperliquid review covers the base protocol's mechanics and its B risk grade.
Where the fees go: a $50M/day worked example
The economics are the reason builders show up at all. A HIP-3 market charges the deployer's fee scale on top of the base protocol fee. Per current documentation, the deployer can configure an additional fee share between 0% and 300% (0% to 100% in growth mode), and if the share exceeds 100% the protocol fee is raised to match the deployer fee. The commonly cited baseline arrangement splits the total fee roughly evenly between builder and protocol. Separately, HIP-3 markets incorporate the usual fee discounts — HYPE staking tiers, referrals and aligned-collateral discounts all still apply.
Two mechanics are worth flagging. Growth mode cuts protocol fees and rebates by roughly 90%, letting a new builder market keep more of what it charges while it bootstraps volume. And because the deployer's fee sits on top of a roughly doubled HIP-3 fee base, the headline taker rate on a builder market can be meaningfully higher than the flagship 0.045%. The example below holds the taker rate at the standard 0.045% so the arithmetic stays comparable to the rest of the site.
| Line item | Calculation | Figure |
|---|---|---|
| Daily taker volume | assumed | $50,000,000 |
| Taker fee rate | base standard | 0.045% |
| Gross daily fee | $50M × 0.045% | $22,500 |
| Builder share (est.) | ~50% baseline | $11,250 / day |
| Protocol share (est.) | ~50% baseline | $11,250 / day |
| Annualized gross fee | $22,500 × 365 | ≈ $8,212,500 |
| Annualized builder take (est.) | $11,250 × 365 | ≈ $4,106,000 |
Read the table as an estimate: the split depends on the fee scale the deployer chooses, and a builder can dial its share well above the even baseline, up to 300%, at which point the protocol fee rises to match. Growth mode changes the math again. But the mental model holds — the fee stream is the product, and the builder's cut is why these markets exist. If you provide liquidity to a builder vault, that vault's return is a slice of this fee stream, so volume and fee scale are the two variables that decide whether the advertised APY is real. Track funding and fee drag with the funding-rate calculator, and read the mechanics in our perpetual funding-rate math guide.
A builder market vs the flagship HLP vault
Here is the comparison that decides whether HIP-3 belongs in your portfolio. Both are ways to be "the house" on Hyperliquid, but almost nothing else about them is the same. The flagship HLP vault is the protocol's own market-maker and liquidation backstop. A builder market is a separate venue with its own oracle, its own book and its own owner.
| Dimension | HIP-3 builder market | Flagship HLP vault |
|---|---|---|
| Fee take | Builder keeps a configurable share (0–300%, 0–100% growth mode); protocol fee rises to match above 100% | No performance, management or deposit/withdrawal fees charged to depositors |
| Liquidity source | The builder's own vaults, LPs and market makers on that market | Pooled USDC across the flagship order book, plus USDC supply interest |
| Oracle | Chosen by the deployer — may be custom to the market | Protocol-defined price feeds on the core Hyperliquid markets |
| Who bears the risk | Builder and its LPs; deployer's staked HYPE and market design at risk | HLP depositors — adverse selection, drawdown months, 4-day lockup, platform risk |
| Track record | Newer, market-by-market; quality varies widely | ~450% since inception, low-20s% trailing CAGR as of Q1 2026 |
The row that should stop you is the oracle row. On HLP you take exposure to the core protocol's price feeds and governance. On a builder market you are trusting a specific deployer's choice of oracle — the very input liquidations are computed against. If the oracle is wrong, thin or manipulable, the loss lands on whoever is on the other side of the liquidated position. For context on where HIP-3 sits against the incumbents, see our dYdX vs GMX vs Hyperliquid comparison.
The risks that do not show up on the leaderboard
Every one of the four risks below is a structural consequence of permissionless market creation, not an edge case. Take them seriously and HIP-3 is a legitimate frontier; ignore them and you are the liquidity.
- Fragmented liquidity and wider spreads. A new market starts thin. Thin books mean wider bid-ask spreads and brutal slippage on size, and the spread you pay is exactly the spread some LP is earning. If a builder market advertises high LP yield, ask what it costs a taker to cross — the two numbers are the same trade viewed from opposite sides.
- Builder quality and incentive misalignment. The same party that runs the vault you might deposit into also sets the fee scale and can operate its own market-making. A builder's incentives are not always aligned with its users'. Staking HYPE is a real skin-in-the-game signal, but it is a signal, not a guarantee of competence or good faith.
- Custom oracles and manipulation surface. When the deployer picks the oracle, a poorly designed feed becomes an attack vector. Illiquid or non-standard assets are the classic soft target: if the price can be moved cheaply, positions can be liquidated cheaply. Hyperliquid has already shown, on its core book, that governance may intervene to settle a market at fair value when an attack unfolds — a user-protective move that is also a reminder the rules are not purely mechanical.
- A builder market can be wound down. This is the risk yield seekers most often forget. A market can be delisted or closed by the builder or by governance. When that happens, your position may be settled at a protocol-defined price rather than the exit you would have chosen, and your vault may be forced to redeem. An APY on a market that might not exist in six months is not the same asset as an APY on a venue with a multi-year record.
Notice how these compound: a thin builder market with a custom oracle, run by an unproven deployer, is a different risk object than a deep market with a battle-tested feed. The honest framing is that HIP-3 gives you access to venues and yields that did not exist before, and prices that access in exactly the currency you would expect — liquidity, oracle quality and longevity. The yield calculator is where you turn an advertised rate into a net figure after fees, funding and spreads.
Risk note. Builder-deployed markets may use custom oracles, carry thin liquidity and be wound down at short notice. Higher advertised yields on new venues are compensation for exactly those risks, not free return. Nothing here is financial advice — size any builder-market position as money you can afford to lose.
How to size a HIP-3 position
The practical answer: HIP-3 belongs in the speculative sleeve of a portfolio, not the core. Keep the majority of any Hyperliquid exposure in venues with a track record, and treat builder markets as a satellite bet on a specific deployer and a specific oracle. Before you trade or provide liquidity, check three things: who the deployer is and how much HYPE is actually staked, what oracle the market uses and whether it is a custom feed, and how deep the book is relative to your size. If any answer is vague, the yield is not mispriced — it is unpriced risk.
For anyone who wants the builder exposure without managing individual markets, a builder-run vault that pools liquidity across its own HIP-3 markets is the cleaner route: you outsource the market-making but still own the builder risk. Read its fee terms like you would a fund's, because the vault keeps a slice of the same fee stream shown above.
Sources and further reading
- Hyperliquid docs — HIP-3 builder-deployed perpetuals — the deployer fee share, growth mode and market specification rules.
- Hyperliquid — HIP-3 announcement — the 500,000 HYPE stake, deployer-defined markets and independent margining.
- DeFiLlama — live perp DEX volume and TVL for context on venue liquidity.
- ethereum.org — Oracles — why oracle design and price-feed integrity are the core attack surface for any derivatives market.
Frequently asked questions
What is HIP-3 on Hyperliquid?
HIP-3 (Hyperliquid Improvement Proposal 3) lets third-party builders deploy their own perpetual markets on Hyperliquid. A deployer stakes 500,000 HYPE, then defines the markets, the oracle used for each, the contract specifications and the margin asset. Each HIP-3 DEX runs with independent margining and inherits the HyperCore order book and matching engine, while the builder operates the market and takes a configurable share of its fees.
How are HIP-3 trading fees split between the builder and the protocol?
A HIP-3 market charges the deployer's fee scale on top of the base protocol fee. Per current documentation the deployer can configure an additional fee share between 0% and 300%, or 0% to 100% in growth mode; if the share exceeds 100%, the protocol fee is raised to match the deployer fee. In the commonly cited baseline arrangement the total fee is split roughly evenly, so the builder and protocol each earn about half. Growth mode reduces protocol fees and rebates by roughly 90% to help a new market bootstrap.
Can I earn yield from HIP-3 builder markets?
Yes, but the yield is builder-run, not the flagship HLP vault. You can trade these markets, or supply liquidity to builder-operated vaults and LP pools that quote on them. Returns can be higher than proven venues because a new market is less efficient and pays the builder's spread, but that edge is unproven, liquidity is thinner, and the builder can change parameters or wind the market down. Treat builder-market yield as a newer, higher-risk sleeve rather than a replacement for HLP.
What are the main risks of a HIP-3 builder-deployed market?
Four stand out. Fragmented liquidity means wide spreads and painful slippage on thin builder markets. Builder quality varies, and a builder's incentives may favor its own vault over users. A custom oracle chosen by the deployer is a new manipulation surface. And a market can be wound down or delisted, so positions may be settled at a protocol-defined price rather than the exit you wanted. None of these are hypothetical; they are the structural cost of permissionless market creation.
Is HIP-3 the same as the HLP vault?
No. HLP is Hyperliquid's flagship community vault: you deposit USDC and the protocol itself market-makes the main order book and absorbs liquidations. HIP-3 is a framework that lets outside builders stand up their own separate perpetual markets with their own oracles, margin assets and fee scales. HLP is one venue run by the protocol; HIP-3 is many venues run by builders. The risk, liquidity source and who earns the fee are different in each case.
Model the fee drag before you trade
Put your position size, entry, exit and holding period into the DifiCalc funding-rate calculator to see what funding and fees cost you on any builder or flagship market.
Open the Funding-Rate CalculatorRelated: weigh the incumbents in dYdX vs GMX vs Hyperliquid; read the Hyperliquid review and the HLP vault breakdown; understand perpetual funding-rate math; then run the numbers in the funding-rate calculator and the DeFi yield calculator.