Hyperliquid funding rates and the USDC that changed hands at each hourly settlement, one row per perp market per hour. We sum HyperCore's per-account funding payments for each market and settlement block, so a week is 168 rows per market instead of millions of account rows. Each row has the rate, the USDC paid and received, open interest in base units, and how many accounts were long and short. HIP-3 markets from the xyz, io, para and mkts dexes sit next to the validator-run perps.
Most buyers sum Funding_Paid_USDC by Market_CoinRaw over a week and sort it. In the week of 9 to 15 November 2026 that came to about $25.2 million, 62% of it on HIP-3 markets. For per-account payments, buy Hyperliquid Funding instead; this table joins it on Block_Number and Market_CoinRaw.
Use cases
Rank Hyperliquid markets by funding paid in a week
Funding rate history per market for carry and basis backtests
Compare funding on HIP-3 stock perps (xyz:NVDA) with plain crypto perps
Track open interest and long/short account counts hour by hour
Table
14 columns · free sample for the table
hyperliquid_funding_hourly
14 columns
One row per perp market per hourly funding settlement: rate, accounts long and short, open interest and USDC paid and received.
hyperliquid/hyperliquid-funding-rates/hyperliquid_funding_hourly<YYYY-MM-DD>.parquetone file per UTC day
Send this dataset’s full schema to an assistant and ask it anything. It reads the plain-text brief ↗ first, so the answer comes from the real column list rather than a guess.
Opens a new chat with the question filled in. Nothing about you is sent to Bitquery, and the assistant sees only the public brief.
Delivery
Parquet files in a Requester Pays S3 bucket or a Cloudflare R2 bucket, chosen at checkout. On S3 you download with your own AWS account and AWS bills the transfer; on R2 you download with a read-only API key and pay no egress. Layout:
# hyperliquid_funding_hourly: one file per UTC day
hyperliquid/hyperliquid-funding-rates/hyperliquid_funding_hourly<YYYY-MM-DD>.parquet
Download a free sampleNo email needed.
Pay nowName, work email and licence check.
Pay the invoiceBy card or bank transfer.
Download from S3 or R2The team prepares the export and emails the bucket path, plus a read-only API key for R2 orders.
Need new data after your purchase?
Kafka Streams sends new Hyperliquid data in real time, so your copy can stay current after the files end. It is sold on an Enterprise plan.
One table, hyperliquid_funding_hourly, with 14 columns. Each row is one perp market at one hourly funding settlement: the hourly rate, accounts long and short, open interest, and USDC paid and received. Plain perps and HIP-3 perps from the xyz, io, para and mkts dexes are both in it.
How big is it?
About 8,300 rows a day: roughly 350 markets times 24 settlements, so around 250,000 rows for 30 days. One day is about 370 KB of ZSTD Parquet, so the whole window is around 11 MB. These are estimates from our 14 November 2026 sample; the delivery manifest gives exact figures.
Which Hyperliquid markets pay the most funding?
Sum Funding_Paid_USDC grouped by Market_CoinRaw over the days you want and sort descending. Weekly totals across all markets ran from $19.3 million to $25.5 million in the four full weeks from 19 October to 15 November 2026.
Which way does the money flow?
A positive Funding_Rate means longs pay shorts, a negative one means shorts pay longs. Funding_Paid_USDC is always what the paying side paid, as a positive number, so the rate's sign tells you who paid.
How are the numbers produced and checked?
We group HyperCore's per-account funding rows by settlement block and market. Every market has one rate per settlement, and long sizes equal short sizes in every market we checked. Over 3 to 9 November 2026 the paid side came to $26,923,169 and the received side to $26,923,175, a $6 rounding gap, across 168 hourly settlements.
What is left out?
Per-account payments, which are in Hyperliquid Funding. HIP-3 dexes that margin in USDH, USDE or USDT0 (km, flx, vntl, hyna, cash) are not in our HyperCore feed yet, so their markets are missing. Spot and prediction markets pay no funding.
What limits should I know about?
Open_Interest is in base units of the market, not USD; multiply by a price from Hyperliquid Native Perps or Hyperliquid Trades for a dollar figure. The rate is hourly, so multiply by 24 times 365 for a yearly figure. Block_Time is epoch nanoseconds.
How do I load it?
With DuckDB: SELECT Market_CoinRaw, sum(Funding_Paid_USDC) FROM read_parquet('hyperliquid_funding_hourly/*.parquet') GROUP BY 1 ORDER BY 2 DESC. With pandas: pd.read_parquet('hyperliquid_funding_hourly/').
How far back does it go?
30 days. Our HyperCore history is a rolling 30-day window, so a purchase covers the 30 days up to the day before we prepare the files. There is no longer window to buy.
How is it delivered?
As Parquet files, one per UTC day, compressed with ZSTD, in a Requester Pays S3 bucket. Download costs are billed to your AWS account.