How Often Is Funding Paid? Every Venue Runs a Different Clock.

Funding payment frequency is set by each venue, not by any market standard. Hyperliquid settles funding every hour on every market. Most legacy centralized exchanges, including Binance, settle every eight hours at fixed UTC times. Variational’s Omni uses a variable calculation window between one and eight hours depending on the market, matching Bybit’s window where the same market exists on Bybit. There is no universal funding clock in perps, and the same asset routinely settles on three different schedules across three venues.

The interval matters more than it looks, because it changes what the displayed rate means. A venue does not pay you the rate on the screen continuously. It measures the perp’s premium over a window, computes a rate, and pays at the snapshot. Between snapshots, the displayed figure is a projection that updates as the premium moves. On a one-hour venue, the projection has at most an hour to drift before it becomes a payment. On an eight-hour clock, the rate you see at the start of the window can bear little resemblance to the rate that eventually pays.

The venue mechanics are worth knowing precisely. Hyperliquid computes its rate with the standard premium-plus-interest formula over an eight-hour basis, then pays one eighth of that rate every hour, sampling the premium every five seconds, with the rate capped at 4 percent per hour. The hourly cadence makes its funding fast-twitch: dislocations show up quickly and get paid quickly. Variational samples the premium every sixty seconds and averages it over the market’s calculation window using a recency-weighted scheme, with the rate capped at 2 percent per hour, and for a market like BTC with an eight-hour window the hourly rate is zero for seven hours and then computed off the full eight hours of data in the final hour. I went deep on that recency-weighted design in the funding rate with no order book.

Three practical consequences. First, comparing raw rates across venues is meaningless until you normalize for the interval: a 0.01 percent rate paid hourly is a completely different cost of carry than 0.01 percent paid every eight hours, roughly eight times different, and any screener showing the two side by side without normalizing is showing you noise. Second, the interval sets your exposure to drift. The longer the window, the more the rate can move between what you saw at entry and what actually pays, which on a thin market is the difference between a trade and a guess. Third, cross-venue strategies inherit a structural timing mismatch whenever the legs settle on different clocks, and that mismatch, not the spread itself, is usually what kills the trade. That failure mode is the subject of the 1,500% APR that doesn’t exist.

If you want the interval data alongside the rates rather than having to dig it out of each venue’s docs, surfacing exactly that alignment is one of the things my funding page exists to do.