9 min read
Hyperliquid funding carry means shorting a Hyperliquid perpetual, holding the same coin on spot, and collecting the hourly funding that longs pay shorts. In Huginai’s 60-day paper test (2026-05-10 to 2026-07-08) it returned 3.94% to 5.56% net APR after taker fees and an assumed 5 bps of slippage on every leg: HYPE 5.56%, NEAR 5.24%, LDO 3.94%. It is the only strategy this desk still runs; the directional signal book closed on 2026-07-08 at a 40% win rate against the 57.4% it needed. Below: the setup, the numbers, what kills the trade, what is still unmeasured, and who is shut out by the venues’ own terms.
The short version: a real, small, capacity-bound edge of about 4–6% a year on paper, capped near $50–60k per leg, with the tail risk concentrated in which coins you pick. The live summary sits on the front page next to the four strategies that closed.
A funding carry trade, also called funding-rate arbitrage or a delta-neutral funding strategy, holds equal and opposite positions (long spot, short perpetual) so price moves cancel and the position earns the funding payment. What remains is funding income minus the cost of opening and closing four legs.
Hyperliquid’s own funding documentation sets the mechanics:
When a perp trades close to spot, its funding settles near that interest component. We call it the funding floor, and the strategy is built to harvest it on coins where the floor holds for weeks at a time.
The front page keeps the carry test next to the four books that closed, each with the number that closed it. The demo dashboard runs on synthetic data behind a permanent DEMO banner, no account needed.
The test covered three coins over a 60-day window of hourly Hyperliquid funding from public endpoints, costed with taker fees plus 5 bps of assumed slippage on every leg and every side, amortized over a 30-day hold.
After all modelled costs, the three coins returned between 3.94% and 5.56% net APR over the 60-day window, with 96.7% to 100% of days positive.
| Coin | Net APR | Positive days | Spot hedge | 4.5% gate |
|---|---|---|---|---|
| HYPE | 5.56% | 96.7% | Hyperliquid spot, single venue | Clears |
| NEAR | 5.24% | 100% | External spot, cross-venue | Clears |
| LDO | 3.94% | 96.7% | External spot, cross-venue | Below the gate |
HYPE leads on cost: its hedge stays on Hyperliquid, where the spot-leg taker fee in our model is 0.07%, against 0.10% for the Binance spot hedges of NEAR and LDO. On one venue the two legs also share margin. LDO fails the 4.5% gate on its own expectation, so it no longer qualifies for capital even in principle.
The 60-day result above is the measured test. To check whether the floor still holds, we pulled the latest 60 days of hourly funding from Hyperliquid’s public API on 28 September 2026. These are gross funding rates on the perp side, before any cost, and they are a condition check on the market, separate from the carry result.
| Coin | Gross, 60d annualized | Positive hours, 60d | Trailing 7-day sum |
|---|---|---|---|
| HYPE | 10.57% | 90.6% | 22.7 bps |
| NEAR | 17.63% | 98.8% | 52.7 bps |
| LDO | 11.65% | 93.9% | 42.9 bps |
| BTC | 8.54% | 91.0% | 20.6 bps |
| ETH | 9.15% | 93.8% | 23.6 bps |
All three basket coins sit well above the 7 bps kill threshold as of 28 September. Gross rate and net APR move apart on the cross-venue legs, which pay more in fees and slippage; that gap is exactly what the cost model charges for.
Three things end a funding carry: funding turning negative on the coin you hold, position size outgrowing the order books, and the costs of rotating between coins. Our risk register, measured on 231 coins over 90 days ending 2026-07-10, puts numbers on each.
The defence is coin selection plus a hard exit. The basket holds only coins with at least 90% positive days, and the 7-day kill rule closes a position when its trailing 7-day funding sum drops below 7 bps. HYPE, NEAR and LDO each kept a positive worst 7-day sum across the 90-day sample and never flipped sign between its two halves.
Four inputs to the result are assumptions or single observations, and each one could move the net APR.
Hyperliquid’s interface is closed to Restricted Persons, and its Terms of Use (last updated 15 June 2026, section 1.6) define them to include persons who reside in, are located in, or are incorporated in the United States or Ontario, Canada. The same terms prohibit using VPNs or proxies to conceal location.
The spot side has its own limits. Binance’s HYPE listing announcement bars residents of Canada, the United States and several other regions from the new spot pairs. For a reader in Ontario, both legs of this trade are closed by the venues’ own terms. That is one more reason Huginai keeps this strategy on paper. Check your own jurisdiction first; this article is research, and nothing in it is investment or legal advice.
It is a delta-neutral trade: short the Hyperliquid perpetual and hold the same coin on spot, so price moves cancel and the position collects the hourly funding that longs pay shorts when the rate is positive.
In Huginai’s paper test, 3.94% to 5.56% net APR over 60 days (2026-05-10 to 2026-07-08), after taker fees and 5 bps assumed slippage per leg. HYPE 5.56%, NEAR 5.24%, LDO 3.94%. Past paper results do not predict future returns.
Every hour. The interest-rate component is 0.00125% per hour, and funding is capped at 4% per hour, according to Hyperliquid’s documentation.
The short perp pays funding to the longs. In our 90-day sample, 31 of the 125 coins that started net positive, about one in four, flipped to net negative in the next 45 days. The kill rule closes the pair when the trailing 7-day funding sum falls below 7 bps.
Hyperliquid’s Terms of Use name persons in Ontario, Canada and in the United States as Restricted Persons, and Binance excludes Canadian residents from its new HYPE spot pairs. Check the rules that apply to you with a qualified advisor.
The front page shows the live paper test with its window, cost basis and kill rule, and it changes when the measurements change. The demo dashboard shows how the desk reads funding, on synthetic data.