Signalview

Hyperliquid Vaults Explained: HLP, User Vaults & Yield

How Hyperliquid vaults work in 2026: the HLP market-making vault vs user-created vaults, how depositors earn from market-making and liquidation PnL, the leader's 10% profit share and 5% skin-in-the-game, the 4-day HLP lockup, realistic APY, and why vaults can and do lose money.

A Hyperliquid vault is an on-chain pool of USDC that runs a trading strategy, splits the profit and loss across everyone who deposited, and settles it all on the exchange's own L1 with no off-chain custodian in the middle. The headline example is HLP — the Hyperliquid Provider vault — a community-owned pool that market-makes and backstops liquidations for the whole exchange, and whose depositors earn a share of the resulting PnL and fees. Alongside it, anyone can create a user vault: deposit 100 USDC, run a strategy, and let others copy your capital into it.

This post covers what a vault actually is, how HLP differs from a user-created vault, exactly how depositors earn (and lose), the leader's own stake and profit share, the deposit lockups, and realistic APY numbers for 2026. The honest through-line: these are trading strategies wearing a yield label, not deposits earning interest. Vaults can and do lose money.

Published July 12, 2026. Figures reflect Hyperliquid's published mechanics and on-chain history as of that date; vault terms and returns change, so confirm against the official docs and the live vault page before depositing.

What a Hyperliquid vault actually is

Mechanically, a vault is a shared account. You deposit USDC and receive vault shares proportional to your deposit divided by the vault's total assets. As the vault's strategy makes or loses money, your shares are worth more or less in USDC — the same accounting a mutual fund uses. When you withdraw, you redeem shares at the vault's current net asset value, and the USDC lands back in your Hyperliquid spot balance. There is no fixed rate, no coupon, and no promise: your return is whatever the underlying trading produced, minus the leader's cut.

Because it lives on Hyperliquid's L1, the whole thing is non-custodial in the sense that positions, PnL and share balances are on-chain and visible in real time — you can watch a vault's equity curve tick rather than trusting a monthly statement. That transparency is the best feature of the design and the reason to judge any vault by its actual on-chain PnL history, not its advertised APY. If you're new to the exchange itself, How to Use Hyperliquid walks through deposits, the interface and getting an account funded before you touch vaults at all.

HLP: the protocol's own vault

HLP (Hyperliquid Provider) is the flagship vault and it is not a normal user vault. It runs several automated market-making strategies, provides USDC liquidity, and — critically — acts as the backstop liquidator for the entire exchange: when a trader gets liquidated and no one else takes the position, HLP absorbs it. It also receives a portion of the platform's trading fees. Depositors into HLP are effectively becoming the house's market-making desk and its liquidation counterparty, earning the spread, the liquidation PnL and a slice of fees in exchange for warehousing risk.

Two things make HLP unusual. First, it charges no profit share — unlike user vaults, there is no leader taking 10% off the top; the strategy is run by the protocol and returns flow to depositors. Second, it is large: HLP has carried on the order of several hundred million dollars in TVL through 2026 (roughly $400M in mid-2026), which is what lets it absorb large liquidations without blowing up. Deposits into HLP lock for 4 days from the moment you deposit; after that, withdrawal requests settle at end-of-day vault NAV. For where those fees come from in the first place, Hyperliquid Fees Explained breaks down the maker/taker schedule that partly feeds HLP.

User vaults: anyone can be the leader

A user vault is created by a vault leader who defines and runs the strategy — funding-rate arbitrage, directional trading, a market-making bot, whatever they choose. It takes a minimum 100 USDC to create one. Other users deposit alongside the leader, and their capital trades the same book, so a user vault is essentially copy-trading with pooled margin: everyone rides the same positions in proportion to their share.

The leader earns a 10% profit share — a flat 10% of profits above the vault's high-water mark. The high-water mark matters: the leader only earns on new profit beyond the vault's previous peak, so if the vault draws down and recovers, they don't get paid twice for the same gains. User vaults carry a shorter 1-day lockup by default (versus HLP's 4 days), though a leader can set a longer one. The obvious risk is that you are trusting a strategy you may not fully see: you can watch the equity curve, but you are handing trading discretion to someone whose track record may be short and whose incentives (10% of upside, none of the downside beyond their own stake) are asymmetric.

Skin in the game: the leader's own stake

Hyperliquid tries to align the leader with depositors by forcing the leader to hold at least 5% of the vault's equity at all times — they cannot withdraw if it would drop their share below 5%. This is real skin in the game: a leader who blows up the vault blows up their own 5% too, and 5% of a large vault is meaningful money. It is the single best structural check on a user vault, and it's worth verifying a leader's stake before depositing.

But be honest about what 5% does and doesn't do. It caps how much the leader personally loses to a fraction of the pool, while the profit share gives them a full 10% of the upside. A leader running an aggressive strategy is playing with roughly 95% other-people's-money on the downside and keeping a tenth of the wins — an incentive that rewards swinging for the fences. Skin in the game reduces outright rug behavior; it does not make a risky strategy safe.

What the APY really looks like

HLP has historically produced something in the rough range of 10-30% APR, averaging on the order of 15-20% annualized across most windows in its history — but that average hides the important part. The returns are lumpy and event-driven: HLP earns steadily from spreads and fees, then takes sharp drawdowns of 5-12% when it's stuck holding the losing side of a crowded, fast-moving market. There is no single 'HLP APY.' The only reliable way to judge it is the on-chain PnL curve, which shows the smooth-most-of-the-time, occasionally-cliff shape that any market-making book has.

User vault APYs are all over the map and far less trustworthy as advertised, because a short winning streak can imply an absurd annualized number that a single bad week erases. Treat any user-vault APY over a few weeks old with heavy skepticism, and never annualize a two-week track record. Compared with the fixed, protocol-level reward of staking HYPE, vault yield is higher on average but genuinely at-risk of principal — a distinction laid out in Hyperliquid Vaults vs Staking and in the mechanics of the Hyperliquid Staking Guide.

Vaults can and do lose money: the JELLY case

The clearest proof that HLP is not risk-free yield is the JELLY incident of March 2025. An attacker opened fresh accounts, took an over-leveraged short in the low-liquidity JELLYJELLY perp, and then pushed the thin spot price up so the perp's mark price followed. The short was liquidated into HLP as designed — and because the same actor kept driving the price, HLP's unrealized loss on that forced position reached roughly $13.5M, about a 27% temporary drawdown on the vault. Depositor capital was directly on the hook.

It resolved in an uncomfortable way: Hyperliquid validators voted on-chain to delist the JELLYJELLY perpetual and settle it at a favorable price, and HLP ultimately closed the position at a small profit of around $703K rather than a large loss. Depositors were made whole, but only because the protocol intervened — which raised its own decentralization questions. The lesson for a depositor is blunt: HLP's backstop role means it can be handed a toxic position it can't hedge, and a worse or slower response could have crystallized that 27% drawdown as a real loss. Whether you find the intervention reassuring or concerning, it belongs in any honest read of Is Hyperliquid Safe?

Where Signalview fits

Signalview (our product) is not a vault and does not pool your money. Our non-custodial AI agents trade Hyperliquid perps from your own account using scoped agent keys that can place orders but never withdraw, driven by strategies backtested over 18 months and compressed into one score from −100 to +100. A vault takes custody of pooled USDC and a leader trades it; Signalview leaves your funds in your own wallet and your capital is never commingled with anyone else's. They solve different problems — a vault is passive exposure to someone else's book, our agents are automation of a signal you can inspect. If that model interests you, Non-Custodial AI Trading Agents explains the key-scoping in detail.

The honest comparison: a vault requires zero ongoing decisions and spreads risk across a strategy and a leader's skin in the game; our approach keeps you in custody but makes you responsible for choosing and monitoring a strategy. Neither removes market risk. HLP itself proves that pooling capital under professional market-making still took a 27% drawdown in a single event.

Risk note: vaults are trading strategies, not savings accounts — deposits are at-risk principal, HLP has drawn down sharply before, user-vault leaders keep the upside while risking mostly your money, and APY figures are backward-looking and not promises. Nothing here is investment advice.

Frequently asked questions

What is a Hyperliquid vault?
An on-chain pool of USDC that runs a trading strategy and splits the resulting profit and loss across depositors by share. HLP is the protocol's own market-making vault; user vaults are created by individual leaders.
What APY does the HLP vault pay?
There is no fixed APY. HLP has historically averaged roughly 15-20% annualized, ranging about 10-30%, but returns are lumpy and it has taken drawdowns of 5-12% in normal conditions and about 27% during the March 2025 JELLY incident. It is at-risk, not guaranteed.
How much does a Hyperliquid vault leader earn?
User-vault leaders take a 10% profit share on gains above the vault's high-water mark and must keep at least 5% of the vault's equity as their own stake. HLP charges no profit share — it's run by the protocol and returns go to depositors.
How long is my deposit locked in a Hyperliquid vault?
HLP deposits lock for 4 days from deposit. User vaults default to a 1-day lockup, though the leader can set a longer one. After the lockup, withdrawals settle at end-of-day vault NAV back to your spot balance.