Funding rate arbitrage between Hyperliquid and Aster

Funding rate arbitrage means holding a long perp on the venue where funding is lower and an equal short on the venue where it is higher, in the same coin. Price moves on the two legs roughly cancel, and you collect the difference in funding. Between Hyperliquid and Aster, the spread you see on a screen is real, but it is usually short-lived and smaller than the cost of opening and closing four trades. In Prow's tests on past prices, on 25 coins over 2025 and 2026, with Hyperliquid, Aster and Lighter prices, trading it the way a bot can (one leg as a limit order, the other at market, with a liquidation guard) earned about 0.3–1.4% a year on the capital after costs, depending on the test run. That is less than holding cash.

This guide shows how the trade works, a worked example with real fees, and why the net result is so much smaller than the headline spread. For the other trades an arbitrage bot can run, and the rule it uses to enter, see how crypto arbitrage bots work. The single-venue version, spot against a perp on Hyperliquid, is in the cash and carry guide.

How the trade works

Suppose BTC funding on Aster is higher than on Hyperliquid:

1. Open a short on Aster. While funding is positive there, shorts receive it.
2. Open a long of the same size on Hyperliquid. You pay Hyperliquid's (lower) funding.
3. If BTC rises 5%, the long gains about what the short loses. If it falls, the reverse. Your net price exposure is close to zero.
4. What remains is Aster funding received minus Hyperliquid funding paid, minus fees.
5. Close both legs when the spread closes or flips.

The positions are a pair, each side a leg. The legs must be the same size in coin terms, opened as close together in time as possible, and kept away from liquidation on both venues.

Compare the rates in the same unit first

Hyperliquid pays funding every hour. Aster pays every 8 hours on most markets, and every 4 or 1 hour on some. A raw comparison of the two numbers is meaningless until you convert both to a rate per hour: divide Aster's rate by its interval in hours. The funding rates guide explains why Hyperliquid's rate sits at 0.00125% per hour most of the time.

The interval also matters for timing. On an 8-hour Aster market you are paid only at the settlement times, so a pair opened just after a settlement waits up to 8 hours for its first Aster payment while paying Hyperliquid every hour.

A worked example

BTC funding: Hyperliquid at the baseline 0.00125% per hour, Aster at 0.03% per 8 hours, which is 0.00375% per hour. The spread is 0.0025% per hour, or 21.9% a year if it lasted. You open a $10,000 short on Aster and a $10,000 long on Hyperliquid, each at 2x.

What you earn while the spread lasts

What it costs to open and close

Taker fees, base tier, on each $10,000 leg:

At $6 a day, the venue fees alone take almost 3 days to earn back, before slippage. Funding spreads between these two venues usually close within days. If this one lasts 2 days, the pair earns $12 and pays $17. If it lasts a week at the same level, it earns $42 and keeps about $25, on $10,000 of capital tied up across two venues. That is the best case, and spreads this wide are not the common case.

A limit order on Aster pays 0% maker, which cuts the fees a lot. The catch is in the next section.

Why the net result is so much smaller than the spread

Fees on four trades. Every pair pays to open and close two positions. A spread has to last long enough to cover that before it earns anything.

Spreads don't last. A high funding rate attracts traders who push it back down. Most wide spreads shrink within hours to days, and some flip sign, so the pair starts paying instead of earning.

The cheap way to trade it can't stay hedged. With both legs as limit orders, Prow's tests earned 3–7% a year. But while one limit order waits to fill, the other leg is open on its own and fully exposed to the price. Filling one leg with a limit order and hedging the other at market is the compromise, and that is the version that earned 0.3–1.4% a year.

Limit orders fill at the wrong moment. A resting order tends to fill just before the price moves against it. Each such fill costs a little, and it adds up across many clips.

One leg can be liquidated alone. Each leg has its own margin on its own venue. A sharp move can push the losing leg toward liquidation while the winning leg's profit sits on the other venue, where it can't be used as margin. Lower leverage per leg helps. It also means more capital for the same size. The delta neutral risks guide shows how far the price must move at 1x, 2x and 3x.

USDC versus USDT. Hyperliquid settles in USDC and Aster's main perps in USDT. The two coins trade slightly apart, so a "gap" between the venues' prices is mostly that difference, not something that closes. Your profit and losses also land in two different stablecoins.

Money sits idle. Capital waits on both venues for a spread worth taking, earning nothing in between.

If you do it anyway

• Size each leg so a large move (say 20–30%) does not bring either leg near liquidation, and set a stop on each venue.
• Count the full round trip before you enter: four trades, both venues' fees, slippage, and from Jan 1, 2027 the Prow fee if you trade through Prow.
• Decide in advance when you exit: a time limit, a minimum spread, or funding flipping for a couple of settlements in a row.
• Keep margin free on both venues. You can't move it between them instantly.
• Watch both legs. A pair is only market-neutral while both sides are open and the same size.

Prow has an arbitrage bot that runs this trade on Hyperliquid and Aster. It opens a pair only when the spread pays for both legs' costs within your horizon (72 hours by default) and still clears your entry threshold (8% a year after costs, by default), builds each pair in small clips, hedges each fill at market, and keeps a stop on each venue plus a mirror order that closes a leg if the other leg's stop fires. In our tests on past prices, the bot with its default settings, on 24 coins over the last 180 days, lost about 0.9% a year: it entered few pairs, almost all in HYPE and ZEC. Read Arbitrage bot before you start one. Bots can't trade on Lighter.

Frequently asked questions

Is funding rate arbitrage risk-free?

No. It removes most of the price exposure while both legs are open, but one leg can be stopped or liquidated while the other stays open, limit orders can fill just before the price moves, and fees can cost more than the spread pays. In Prow's tests it earned less than cash after costs.

How much does funding arbitrage between Hyperliquid and Aster earn?

In Prow's tests on past prices, on 25 coins over 2025 and 2026, with Hyperliquid, Aster and Lighter prices, about 0.3–1.4% a year on the capital after costs across test runs, trading one leg with a limit order and hedging the other at market. Keeping both legs as limit orders earned 3–7% a year in the same tests, but a trader can't stay hedged that way.

Why is the price of BTC different on Hyperliquid and Aster?

Hyperliquid prices in USDC and Aster's main perps in USDT, and the two stablecoins trade slightly apart. That steady difference is not a gap that closes, so it is not an arbitrage on its own.

How long does a funding spread need to last to break even?

Divide the round-trip cost by the net funding per day. In the example above, $17 of venue fees against $6 a day of net funding is almost 3 days, before slippage and before the Prow fee that applies from Jan 1, 2027.

Sources

Checked October 5, 2026:

• Hyperliquid docs: Funding, Fees
• Aster docs: Funding rate, Fees
• Prow docs: Fees, Venues and Best fill, Liquidation
• Prow data: Live funding rates and gaps between venues, Perp DEX fees compared