---
title: 'Hyperliquid funding rate history: how often is it negative?'
url: https://comparand.xyz/blog/hyperliquid-funding-rates-data
description: 'Measured hourly Hyperliquid funding for 24 coins: how often it was negative, typical rates, yearly averages, and how much it changed by month.'
updated: '2026-10-21'
category: Data
---

# Hyperliquid funding rate history: how often is it negative?

Measured hourly Hyperliquid funding for 24 coins: how often it was negative, typical rates, yearly averages, and how much it changed by month.

On Hyperliquid, BTC funding was negative in 15.37% of hours from April 2025 through September 2026 (13,152 hourly records), and SOL in 30.47%. Across the 24 coins we measured, the share ran from 0.79% to 58.59%, and it changed a lot by month: SOL was negative in 85.27% of hours in February 2026 but only 7.66% in July 2025. BTC's typical hour paid 1/8 of a basis point from longs to shorts. How often funding is negative does not tell you the total cost, because payment sizes vary.

For what funding is and who pays whom, see [What is a funding rate?](https://comparand.xyz/blog/crypto-funding-rate-explained).

## What does negative funding mean?

A perpetual, or perp, is a contract that lets you trade on a coin's price without owning the coin. It has no expiry date. Funding is a regular payment between traders on opposite sides of that contract, which keeps the perp price near the coin's price. A long position gains when prices rise; a short position gains when prices fall.

When funding is positive, longs pay shorts. When funding is negative, shorts pay longs. Hyperliquid settles funding every hour.

We show rates in basis points (bp). One basis point is 0.01%, so 1 bp on a $10,000 position is $1. In the tables, bp/hour means basis points per hour. Many coins sat at exactly 1/8 bp per hour for much of the time, which is the same as 1 bp every 8 hours. That is why so many medians below match. On a $10,000 position, 1/8 bp is 12.5 cents an hour, or $3 a day.

Funding is separate from the price move. Receiving funding does not mean a position made money overall, because the price can move against it at the same time.

*Diagram: Show positive funding flowing from longs to shorts and negative funding flowing from shorts to longs, with price changes shown as a separate part of the result.*
## How often was funding negative for each coin?

These are Hyperliquid hourly funding records from April 2025 through September 2026, in UTC (the shared world clock). Each coin has 13,152 records, one for every hour in that period.

The **median** is the middle rate after sorting from lowest to highest. The **p95** is the rate that 95% of hours were at or below, so it shows the high end. Our summary does not include the single lowest or highest rate. The last column is the average hourly rate stretched over a year (see below).

| Hyperliquid coin | Hours negative | Median, bp/hour | p95, bp/hour | Average, yearly scale |
| --- | ---: | ---: | ---: | ---: |
| BTC | 15.37% | 1/8 | 0.16 | 7.93% |
| ETH | 16.99% | 1/8 | 0.18 | 7.47% |
| SOL | 30.47% | 0.11 | 0.14 | 2.92% |
| HYPE | 7.76% | 1/8 | 0.52 | 13.34% |
| AAVE | 6.94% | 1/8 | 0.23 | 11.16% |
| ACE | 31.85% | 1/8 | 1/8 | -70.01% |
| ANIME | 40.83% | 1/8 | 1/8 | -55.57% |
| BERA | 58.59% | -0.15 | 1/8 | -56.02% |

A positive yearly figure means longs paid shorts on average. For BTC, the average was positive in 17 of 18 months; the only negative month was April 2026 (-0.85%), when funding was negative in 47.36% of hours.

Of the 24 coins, 11 had a negative full-period average. But be careful with that count: all 24 coins had a full set of hours, so our "most hours" rule tied and the alphabet broke the tie. The 20 picked coins all start with A or B, plus BTC, ETH, SOL and HYPE. This group does not represent Hyperliquid as a whole.

## Can funding be mostly positive but average negative?

Yes. ACE was negative in fewer than a third of hours, and its median was positive. Yet its average on a yearly scale was -70.01%. ANIME showed the same thing: a positive median and a -55.57% average.

That happens when the negative rates are much larger than the positive ones. Counting negative hours ignores how big each rate was, and the median ignores the extremes. SOL and ACE had similar negative-hour shares, yet SOL's average was positive and ACE's was strongly negative. You cannot turn a share of negative hours into a payment total.

## Does the answer change by month?

Yes, even for the same coin. Here is SOL on Hyperliquid in four months. Each row covers every hour of that month.

| SOL month, UTC | Hours | Hours negative | Median, bp/hour | Average, yearly scale |
| --- | ---: | ---: | ---: | ---: |
| July 2025 | 744 | 7.66% | 1/8 | 19.85% |
| February 2026 | 672 | 85.27% | -0.17 | -17.62% |
| March 2026 | 744 | 72.45% | -0.09 | -9.74% |
| September 2026 | 720 | 15.28% | 1/8 | 7.16% |

SOL's funding went from mostly positive to mostly negative and back. Its full-period average hides those swings. BTC also had its most negative stretch in early 2026: 31.70% of hours in February, 40.32% in March and 47.36% in April.

These months show that funding changes, not a calendar rule. We did not test whether the month predicts funding, or whether funding predicts the next price move.

*Diagram: Show SOL's share of negative funding hours on Hyperliquid in every month from April 2025 to September 2026, with February 2026 the highest.*
## What does an annualized funding rate actually mean?

Annualized means put on a yearly scale. We multiply the average hourly rate by 8,760 (the hours in a year) and write it as a percent, with no compounding.

So BTC's 7.93% does not mean anyone earned 7.93%. It means: if the average rate from these 18 months had lasted a full year, a short of a fixed size would have received about that share of its size in funding, before anything else. A real result also depends on which side you held, when, the price move, and costs. In our other studies we assume a 0.05% fee to enter and 0.05% to exit, or 0.1% total (an assumption, not a quote for every account). Funding figures leave fees out, along with the spread (the gap between buy and sell prices) and the money you must keep in the account.

## How we measured this

We used Hyperliquid's hourly funding settlement records for 24 coins from April 1, 2025 to October 1, 2026, grouped by the UTC hour of each record. We picked the 20 coins with the most recorded hours (all tied, so alphabetical order decided) and added BTC, ETH, SOL and HYPE. Rates stored per 8 hours were divided by 8 to get hourly rates. Missing hours were left out, never filled. We did not test any trading rule, and the coin group was chosen by data, not by size or trading volume.

## What should you check before using funding history?

First, check when each rate was actually known. Each coin had two delayed records in July 2025, about 1 minute 51 seconds and 14 minutes 48 seconds after the hour. A test that uses a rate before its recorded time uses information from the future, as our [guide to look-ahead bias](https://comparand.xyz/blog/look-ahead-bias-backtest-example) explains.

Second, complete records do not prove a market was active. AI on Hyperliquid had a zero median, zero p95 and zero average in every month from May 2025 through September 2026. Those zeros do not show that you could have traded it.

## How can you use this in your own research?

If you backtest a perp idea that holds a position for hours or days, add the funding for the exact hours held, from the same coin's records. Do not assume one fixed rate. For SOL on Hyperliquid, the monthly average on a yearly scale ran from 19.85% in July 2025 to -17.62% in February 2026, and it flipped sign, so the side that paid changed too. These are past records, not a forecast of future funding. The same archive also records forced closes; see [what happened after liquidation cascades](https://comparand.xyz/blog/after-crypto-liquidation-cascades).

## Try this in Comparand

Ask your connected AI assistant: "Check Hyperliquid coverage, then show SOL's funding for February 2026 in UTC."

- hl_coverage (free) lists the archive days, the coins, and which days are complete.
- hl_funding returns SOL's funding rates for your chosen dates. It is included on plans; with pay-as-you-go credits it costs $0.005 per call.

Then ask: "Show BTC's funding payments on Binance for the same dates."

- venue_funding returns that perp's recorded funding settlements on the other exchange. It is free, with up to 400 days per call.
- Keep the time unit in mind when you compare: Hyperliquid pays every hour, so put both on the same scale first.

See the [docs](https://comparand.xyz/docs) for access and our [crypto data API comparison](https://comparand.xyz/blog/crypto-market-data-apis-compared) for other data choices. You can [try 3 coins free](https://comparand.xyz/try) to explore past cases, not a forecast.

## The short version

- BTC funding on Hyperliquid was negative in 15.37% of hours from April 2025 to September 2026; SOL in 30.47%.
- A typical BTC hour paid 1/8 bp from longs to shorts; the average on a yearly scale was 7.93%.
- A positive median can sit beside a negative average, as ACE (-70.01%) shows.
- Funding changes by month: SOL went from 7.66% negative hours in July 2025 to 85.27% in February 2026.
- A yearly funding figure is not a trading return; it leaves out price moves and costs.

## Questions people ask

**How often were Hyperliquid funding rates negative?**

It depended on the coin and month. Across 13,152 hourly funding records per coin on Hyperliquid from April 2025 through September 2026, BTC funding was negative in 15.37% of hours and SOL in 30.47%. Among the 24 coins we measured, the share ran from 0.79% to 58.59%.

**Who pays when funding is negative?**

When funding is negative, traders with short positions pay traders with long positions. When it is positive, longs pay shorts. The amount depends on the rate and the size of the position.

**What is the typical Hyperliquid funding rate for BTC?**

In our April 2025 to September 2026 records, BTC's median hourly rate was 1/8 of a basis point, the same as 1 basis point every 8 hours. Its average rate put on a yearly scale was 7.93%, paid by longs to shorts.

**Does a positive median funding rate mean a positive average?**

No. On Hyperliquid, ACE had a positive median rate and was negative in only 31.85% of hours, yet its average rate on a yearly scale was -70.01%. A few large negative rates outweighed many small positive ones.

**Is annualized funding a return I could have earned?**

No. It stretches an observed average hourly rate over a year. It leaves out price changes, trading fees, the spread and changes in position size.