Funding Rate

A perpetual swap has no expiry, so there is nothing to force its price back to spot the way settlement forces a quarterly future. Funding is the substitute. Every 8 hours, whichever side is in surplus pays the other a fee proportional to how far the perpetual has drifted from the index — longs pay shorts when the perp trades rich, shorts pay longs when it trades cheap. That payment is the tether. This chart plots the volume-weighted 8-hour funding across the major perpetual venues as a daily close: rust fill above the baseline marks settlements where longs paid, slate fill below marks the reverse, and a muted price line sits behind for cycle context.

Today’s reading is +0.0018% — Neutral. The 7-day rolling mean is +0.0017%; the 30-day mean is +0.0037%; the series spans 1,000 daily observations. The right axis is native percent per 8 hours, capped symmetrically at the 99th percentile of absolute funding so a single intraday flush spike does not flatten the rest of the history.

The canonical perpetual-funding formula across the major venues is identical in shape, with only minor variations in clamp magnitude and settlement cadence. Provenance is on the data sources page; the formula is reproduced verbatim in the methodology:

F = P + clamp(I − P, −0.05%, +0.05%)

where F is the funding rate per 8-hour interval, P is the time-weighted average premium of the perpetual price over the index price across the window, and I is a fixed interest baseline of +0.01% per 8 hours on USD-collateralised contracts. The clamp bounds the (interest minus premium) differential to plus or minus five basis points; an outer cap then bounds F itself. The decomposition that matters for reading the chart: in flat markets, with the perp tracking spot tightly, the premium is zero and funding settles to the interest baseline alone — a structural +0.01% positive bias the chart cannot escape.

Two consequences. First, a “neutral” reading is not zero — it is centred on the interest baseline; anything within a few basis points of +0.01% per 8h is the formula at rest. Second, deeply negative readings are mechanically harder to produce than deeply positive ones: the perp has to trade persistently below spot for the (interest minus premium) term to drag the rate negative through the clamp. Sustained negative-funding stretches are concentrated around cycle-bottom panic windows for exactly that reason. The 2025 Q3 derivatives report from the perpetual’s original venue put it cleanly: The perpetual swap formula has a built-in interest component, forcing rates to cluster around 0.01% (positive bias).

Three settlements per day at 00:00, 08:00 and 16:00 UTC; we aggregate the three into one daily close so the series is comparable to the rest of the site. For the intraday funding spikes that fire during a flush, the upstream venue’s live feed is the right surface; this chart is the nightly summary.

Above +0.03% per 8h, longs are paying meaningful yield to stay in — an annualised carry north of +32.9%. That is expensive, and any reversal can snowball as leveraged longs unwind to escape the cost. Below −0.02%, shorts are paying longs — the regime that historically brackets squeeze setups, when bearish positioning crowds into a shrinking pool and a small spot rally cascades into short cover. The middle band, around the +0.01% structural baseline, is the formula at rest. The table below states what each side is implicitly earning or paying, annualised.

Six anchors since the start of our daily series make the regime rotation visible — the 2024 pre-halving euphoria, the August yen-carry unwind, the post-election leg, the January ATH window, and the most recent close. Each row pins a cycle-anchor date; the funding-rate cell and the spot price are pulled from the same daily snapshot powering the chart above, taking the nearest prior close where an anchor falls on a gap.

The framing matters more than any single reading. Arthur Hayes, who co-founded the venue that launched the original Bitcoin perpetual swap, framed the mechanism plainly in his 2021 essay on perp carry: the perpetual swap exchanges a funding rate — an interest income — between longs and shorts every 8 hours. The unit of analysis is the payment, not the sentiment. A single cash-and-carry desk that shorts the perp against spot to harvest the premium can move funding without expressing any directional view at all; the chart records the carry, not the conviction.

Hayes returned to the framing in his 2025 retrospective with a worked example: if the perp traded at an average 1% premium to spot over the last eight hours, if you held a position at the funding period timestamp, longs pay 1% to shorts. The perp’s premium drives the payment, full stop. Reading funding as “people are bullish” or “people are bearish” mistakes the formula for a poll. Read it instead as: at this rate, this is what it costs to be on the crowded side of the book.

The shape of the funding distribution shifted after November 2022. Pre-FTX, extreme positive readings — multiple settlements per cycle north of +0.10% per 8h — were a regular feature; the basis-trade ceiling, where cash-and-carry desks short the perp into spot until the premium compresses, was thin and often vanished in panic windows. Post-FTX, that ceiling has thickened. The 2025 Q3 derivatives report framed the structure: the large pool of undeployed capital acts as a ceiling for funding rates – preventing it from staying high for long. Extreme positive funding still happens, but it now persists for hours, not days.

The corresponding floor — sustained negative funding — behaves differently. The post-FTX market-structure literature documented that a venue’s disappearance can collapse the basis market that normally clears the perp’s premium against spot; in those windows, both extremes show up at once on different venues, and a volume-weighted aggregate looks artificially neutral. The 30-day rolling mean on this chart smooths through that distortion; the daily print still records the micro-regime.

The regulated leg is invisible (post-Jan-2024 spot-ETF era). Quarterly cash-settled Bitcoin futures and the spot-ETF creation/redemption flow have no funding mechanism — the CME leg settles on contract expiry, and ETF AP arbitrage clears against NAV, not against a perp premium. Since the US spot ETFs launched in January 2024, an increasing share of marginal price action runs through books that never touch this chart. Perp funding can sit flat while the tape moves on ETF flow, because the leg moving it is not here. The open-interest stack is where the regulated leg shows up.

Venue dispersion during dislocations (Nov 2022, Aug 2024). The post-FTX market-structure analysis documented that a week after the collapse, we noticed that global crypto liquidity had halved — a window where FTX’s funding feed simply stopped publishing while panic shorts crowded surviving books. The 5 August 2024 yen-carry unwind produced the same fingerprint in miniature: funding gapped hard negative on some venues and stayed near the floor on others within the same day. A volume-weighted aggregate across those windows averages a stale cell with live ones and misreads dispersion as neutrality. We drop affected days from the source feed where we can; readings during venue-discontinuity events still deserve the sceptical eye.

The structural floor anchors readings positive. The formula’s +0.01%-per-8h interest baseline biases the chart toward positive prints even in flat markets. “Funding is slightly positive, longs must be euphoric” misreads the formula clearing its throat. Always interpret the print against the +0.01% floor, not against zero — which is also why a mildly negative current read like +0.0018% is more informative than its size suggests: it means the premium term actually went negative, against the design’s built-in upward pull.

Treat extremes as setups, not directional triggers. Sustained funding above +0.05% per 8h is annualised carry north of +54.8% — a real cost the long side is paying to stay in. That cost has to be earned back through spot appreciation, and historically it has not been. Pair it with open interest for scale — rising OI plus persistent positive funding is the canonical leverage-flush precursor — and with the liquidation heatmap for the geometry of where the unwind sits.

The contrarian read lives on the other tail. Sustained negative funding is rarer and harder for the formula to produce, so when it persists it carries information: the perp is trading below spot for a reason, usually crowded shorts at a cycle-bottom panic. Confirm against a low long/short ratio before treating it as a squeeze setup — funding alone tells you carry has flipped, not who is about to be forced to cover.