When perp funding runs persistently positive, shorts get paid simply for holding. Funding arbitrage pairs that paid short with an equal long exposure elsewhere, cancelling price risk while the funding payments keep arriving. It's the closest thing crypto has to a money-market trade — which is exactly why its returns compress when everyone piles in, and why the risks are subtler than the pitch.
Structure 1: cash-and-carry
The classic. Buy spot, short the perp on the same asset, same size. Price moves cancel: if BTC rallies 10%, the spot leg gains what the short loses. What remains is the funding stream on the short — positive funding means the crowd of longs is paying you.
The math: at a sustained +0.01%/8h rate, the short collects ≈11% annualized on notional. During hot markets, rates on popular assets have paid several times that for weeks at a stretch. Against that income: two sets of trading fees, spread costs on both legs, and (on CEXs) the capital efficiency hit of collateralizing both sides.
Structure 2: cross-venue spreads
When the same asset funds at different rates on two venues — common with newly listed markets, and visible in real time on our funding rates page — you can go long on the cheap-funding venue and short on the expensive one. No spot needed; both legs are perps. The spread between the venues' rates is your carry.
This is where venue mechanics matter. Hyperliquid settles hourly and lists new markets fast, so it's frequently one leg of these trades. Lighter is attractive for the frequently-rebalanced leg because its zero fees make re-truing the hedge free. A CEX like Bybit often anchors the other side with deep books. New-listing windows are the fattest opportunities: early funding on a fresh perp can be violently mispriced for days.
What returns are realistic?
Honest ranges, not promises: in calm markets, single-digit to low-teens annualized on deployed capital before costs — comparable to what it is, a carry trade. In frothy periods, materially more for weeks at a time. Compressing toward zero as capital floods in — which it now does fast. Anyone quoting fixed "guaranteed" yields from funding arb is describing a product, not the trade. Model any specific setup in the funding calculator: plug in the rate spread and see the daily/annualized carry against your size.
The failure modes (read twice)
- Funding flips. Your income stream can invert with sentiment. Cross-venue spreads close; carry trades go negative. You need exit criteria, not just entry criteria.
- Leg liquidation. The perp short in a cash-and-carry has liquidation risk on violent rallies if margined too thin — the classic blowup is being right on the structure and liquidated on one leg anyway. Size margin for the wick, not the average; check distances with the liquidation calculator.
- Execution slippage on entry/exit. The edge is often tens of basis points; sloppy fills across two venues can eat the first month of carry. Fees matter too — this is a strategy where a maker vs taker habit decides viability.
- Venue risk. Two venues means two custody/solvency exposures. Weight accordingly.
- Borrow/transfer friction. Moving collateral between venues mid-trade is slow exactly when you need it fast.
Who this trade suits
Funding arb rewards operational discipline over market opinion: monitoring, sizing, rebalancing, cost control. It suits traders who want yield-shaped returns from crypto infrastructure without directional exposure — and it punishes anyone who treats "delta-neutral" as "risk-free". Start small, run one structure end-to-end including the exit, and only then consider size.