Right now, on a calm Tuesday with BTC going nowhere, a trader holding a delta-neutral position on Bybit is collecting roughly 0.01% three times a day for doing nothing but waiting. No big call on direction, no transfer between exchanges, no praying for a pump. The price of the coin can rise 8% or fall 8% and the balance barely moves, because one leg gains exactly what the other loses. The only thing dripping in is the funding rate. That is funding arbitrage, and it is one of the few crypto strategies where being right about price is irrelevant to whether you get paid. Below: what funding actually is, how to build the long-spot / short-perp route on real venues, why payout cadence across Binance, Bybit, OKX and Hyperliquid quietly decides your real yield, what basis is, how to compute break-even, and the risks nobody puts in the headline. With an APR worked out in numbers.

What the funding rate is

A perpetual future (perp) is a contract with no expiry. To keep its price from drifting away from spot, the exchange introduces a funding rate - a periodic payment between longs and shorts:

  • When the perp trades above spot (market leans long), funding is positive: longs pay shorts.
  • When the perp trades below spot (market leans short), funding is negative: shorts pay longs.

The rate is a percentage of position notional, transferred at each funding event. In a bull market funding on most assets is persistently positive - there are more longs, and they keep paying shorts. That payment is exactly what you collect by going short, but without directional risk, thanks to a spot hedge. Think of it as the crowd paying you rent for taking the unpopular side of a one-way market.

The delta-neutral setup: long spot + short perp

The base funding-arbitrage route when funding is positive. Say you pick TIA, funding is sitting at +0.012% per 8h on Bybit, and you want to put $10,000 to work per leg:

  1. Short the perp on TIA - for this you receive funding (since longs pay shorts). Open a $10k short on the perp contract.
  2. Long the spot of TIA for the same $10k - this is the hedge: if TIA rises, the perp short is down, but spot is up by exactly as much.

Total delta (price sensitivity) is near zero. TIA pumping to a new high or dumping 20% does not move your capital, because the legs offset each other. And funding keeps accruing on the short leg every period. In effect you turn the funding rate into a near-interest yield on deployed capital, stripped of market risk. You are not betting on TIA, you are renting out your willingness to be short while the longs overpay.

Price of X rises +8% → legs cancel, funding still accrues 0 + − long spot +$800 short perp −$800 net delta ≈ 0 → price P&L = $0 funding income accrues each period →
Price moves cancel between the legs (net delta ≈ 0); only funding accumulates.

Delta-neutral works both ways: under persistently negative funding the setup flips - long the perp (you receive funding as shorts pay longs) + short spot (hedge). In practice persistently negative funding is rare, so the base case is long-spot / short-perp on positive funding.

Why payout cadence decides everything

This is where beginners lose the edge. A 0.01% rate on one exchange is not equal to 0.01% on another, because exchanges pay funding at different cadences:

Exchange Cadence Payouts per 24h
Hyperliquid 1 hour 24
EdgeX 4 hours 6
Binance, Bybit, OKX, MEXC 8 hours 3

The intervals in the table are a snapshot at the time of writing: exchanges change cadence, introduce dynamic funding and hourly windows on individual contracts, so always verify the current interval for your pair on the exchange itself, not from memory. Bitget and Gate, for example, run most contracts on 8h but switch high-volatility ones to 4h or 1h when funding gets extreme.

Here is the trap in one line. A 0.01% rate on Hyperliquid (paid hourly) is 0.24% a day. The same 0.01% on Binance (paid every 8h) is 0.03% a day. That is an 8x difference hiding behind an identical-looking number. Comparing raw on-screen percentages head-on is the single most common funding mistake. You only compare normalized rates correctly. The full breakdown of normalizing to a common window, with the formula and examples, is in funding math. And for the route itself, that cadence asymmetry between venues is often what creates the window: a perp-vs-perp between a high-cadence and a low-cadence venue can yield positive net funding even when both screens show the same percent.

Basis and convergence

Basis is the gap between the perp price and the spot price. Funding is precisely the mechanism that keeps basis near zero: as long as the perp is above spot, positive funding incentivizes shorting the perp and squeezing the basis back. For an arbitrageur, basis matters for two reasons:

  • At entry you lock in the current basis. If the perp is meaningfully above spot, entering short-perp + long-spot earns extra profit on convergence (basis pulls back to zero), on top of the funding itself.
  • At exit the basis should converge back, otherwise closing both legs at an unfavorable basis eats part of the funding income. On liquid assets like BTC, ETH or SOL the basis converges reliably. On thin alts it can diverge unpredictably right when you want out.

A perp-spot route on one exchange that closes through basis convergence is a variety of futures arbitrage. In pure funding arbitrage the basis is secondary, and the main income is the stream of funding payments. If you want the full mechanics of basis and the cash-and-carry trade, that is its own deep dive in futures arbitrage.

Break-even: what's subtracted from funding

Funding is a gross income, and like any spread, costs are subtracted from it:

  • Entry and exit taker fees. You open two legs and close two - that's four trades, ~0.1% each. Round trip easily adds up to 0.2-0.4% of notional. To recoup them, funding must accrue over several periods.
  • Negative funding in a bad hour. Funding isn't guaranteed: the rate floats and can turn against you on individual periods. The income is the sum of payments over the holding time, not a fixed figure.
  • Cost of frozen capital. Both legs reserve collateral and spot. Efficiency depends on the margin mode - see cross vs isolated margin.

So you hold the route not for one period but until accumulated funding has confidently covered the entry-exit fees. Re-opening the position too often kills the return through fees. The arithmetic of how many days you must hold to clear those four trades is the whole game, and we work it next.

Example: working out a funding route's APR

Take TIA again, with a persistently positive funding of 0.01% per 8-hour period on Bybit, a liquid venue. Notional per leg is $10,000 (short perp $10k + long spot $10k, ~$20k capital deployed in total).

$1,095 / yr funding → what the APR actually is APR / $10k notional 10.95% APR / $20k capital 5.5% round-trip fees −$40 annualized yield → break-even: $40 fees / $3.00 day ≈ 14 days held
The same $1,095 funding is ~10.95% on notional but only ~5.5% on deployed capital.
Route: SHORT $10k perp + LONG $10k spot (delta-neutral)
Capital deployed: ~$20,000
Funding: +0.01% per 8h → 3 payouts per day

Daily funding:   3 × 0.01% × $10,000 = $3.00 / day
Annual funding:  $3.00 × 365 = $1,095 / year

── APR on short notional ($10k) ──
$1,095 / $10,000 = ~10.95% per year

── APR on deployed capital (~$20k) ──
$1,095 / $20,000 = ~5.5% per year

── Minus entry/exit costs (one cycle) ──
4 trades × 0.1% × $10k = $40 (one-off)
Recouped in: $40 / $3.00 ≈ 14 days of holding

So "10% APR" sounds attractive, but on deployed capital it's ~5.5%, and the entry-exit fees take almost two weeks of holding to recoup. The route makes sense with stable funding and a sufficient horizon. With a rate hovering around zero, or with a short hold, costs eat the income. At a bull-market peak rates can be several times higher (0.05-0.1%+ per period), and then the APR flies into the tens of percent, but that's a temporary window, not the norm.

Step-by-step: opening your first funding route

Here is the exact sequence, with figures, so you can run it once on small size before scaling. Assume $1,000 per leg to start, not $10k, while you learn the mechanics.

  1. Find a positive, stable rate. In the funding dashboard, sort by normalized funding and look for a coin paying steadily positive over the last few days, not a one-off spike. Say SUI on Bitget at +0.015% per 8h. Avoid anything that flipped sign in the last 24h.
  2. Check both legs are liquid. The spot pair (SUI/USDT) and the perp must each hold $1k without you walking the book. Thin alts move the price against you on entry and again on exit.
  3. Confirm spot withdrawal is open if spot and perp sit on different venues. If you trade both on Bitget, no transfer is needed and this risk disappears, which is why same-venue routes are the cleaner beginner choice.
  4. Size the legs equal. Long $1,000 of SUI spot. Short $1,000 of SUI perp. Equal notional is what makes delta zero. Eyeballing it leaves you partly directional.
  5. Set the perp on cross or portfolio margin with a collateral buffer, so a sharp pump cannot liquidate the short leg. This is the single most common way the trade blows up - see cross vs isolated margin.
  6. Record your break-even. Four taker fees at ~0.1% on $1,000 is ~$4. At +0.015% × 3 per day that is $0.45/day, so you need ~9 days of holding just to clear fees. Write that number down before you enter.
  7. Hold and monitor the normalized rate. Collect funding each period. The moment the normalized rate turns negative and stays there, close both legs together, not one at a time.
  8. Close both legs in the same minute. Buy back the perp short, sell the spot long. Closing them apart re-introduces the directional risk you spent the whole trade avoiding.

Beginner tip: run steps 1-8 on one same-venue coin with $200-500 first. The goal of the first route is not profit, it is proving you can open, hold, and close delta-neutral without fumbling a leg.

How to start practicing

Funding arbitrage rewards patience over reflexes, which makes it a good first "real" arbitrage. A sane ramp:

  • Week 1-2, paper and observe. Watch the funding dashboard daily. Notice which coins pay steadily positive and which whip between signs. Get a feel for what "stable funding" looks like versus a spike that is gone by tomorrow.
  • Week 3-4, one tiny live route. $200-500 per leg, same venue, liquid coin. Open it, hold a week, close it. Measure your actual net after the four fees against what you projected. The gap between projection and reality is your real lesson.
  • Month 2+, scale what worked. Only after a few clean cycles, raise size to $1k-5k per leg and consider cross-venue routes where the cadence asymmetry adds yield. Never deploy capital you cannot leave parked for weeks.

Where Finder fits

Finder reads live funding rates across 20+ exchanges and normalizes them to a common window, so you compare a Hyperliquid hourly rate and a Binance 8h rate as apples to apples instead of being fooled by the raw screen number. The funding dashboard ranks coins by real normalized funding and shows the payout cadence per contract. The spread scanner shows the basis between perp and spot, plus honest deposit/withdrawal statuses so a cross-venue route does not strand your spot leg. Open the scanner and sort by funding to see what is actually paying right now.

Realistic expectations

Funding arbitrage is not a money printer, and anyone selling it as one is skipping the math you just read. On deployed capital, a normal positive-funding regime is single-digit to low-double-digit APR, comparable to lending, with the entry-exit fees clawing back the first couple of weeks. The headline "10% APR" routinely halves once you count the spot leg in the denominator and pay four taker fees. The tens-of-percent numbers you see screenshotted are bull-market peaks that last days, not a baseline.

Your first weeks will probably be flat or slightly negative while you learn to size legs evenly, pick cross margin, and close both sides cleanly. That is normal and it is the point. The traders who do well here are not the ones chasing the highest screen rate, they are the ones who hold stable routes patiently, never let the short leg liquidate, and treat funding as a steady yield rather than a jackpot. If that sounds boring, it is, and boring is exactly why it survives bear markets that wipe out directional traders.

Risks of funding arbitrage

  • Hedge-leg liquidation. If the perp short is on thin isolated collateral, a sharp pump liquidates it - and you're left with a naked spot long, the delta-neutrality broken. Solved by margin mode (cross / portfolio) and a collateral buffer - see the margin breakdown.
  • Funding sign flip. The rate can reverse: you expected to receive and started paying. You need to monitor and close the route when the normalized rate turns negative.
  • Basis divergence at exit. On illiquid assets, closing both legs at an unfavorable basis eats part of the income.
  • Exchange counterparty risk. Capital sits on a derivatives venue, and the exchange's own risk doesn't go away.

This is not investment advice. Funding arbitrage uses leverage on the perp leg and carries liquidation risk under insufficient collateral or sharp volatility. The funding rate is not guaranteed and can flip sign. Real return depends on the normalized rate, entry-exit fees, the margin mode, and basis convergence.

FAQ - funding arbitrage

What is funding arbitrage in simple terms?

It is holding long spot and short perp on the same coin at the same size, so price moves cancel out, while you collect the funding rate that longs pay shorts each period. You are paid for being delta-neutral, not for guessing direction.

How much can you realistically earn from funding arbitrage?

In a normal positive-funding regime, expect single-digit to low-double-digit APR on deployed capital after fees, similar to lending. Bull-market peaks can push individual coins to tens of percent annualized, but those windows last days, not months, and the first couple of weeks of any route go to recouping entry-exit fees.

Why is the same funding rate worth more on one exchange than another?

Because exchanges pay at different cadences. A 0.01% rate paid hourly on Hyperliquid is 0.24% a day, while the same 0.01% paid every 8h on Binance is 0.03% a day - an 8x gap. Always normalize to a common window before comparing, which the funding dashboard does for you.

Do I need to transfer coins between exchanges for funding arbitrage?

No, and that is a big advantage. If you run both legs on one venue (spot SUI and perp SUI on Bitget, for example), there is no on-chain transfer and no withdrawal-window risk. Cross-venue routes can add yield through cadence asymmetry but reintroduce transfer and D/W risk.

What is the biggest risk in funding arbitrage?

Liquidation of the short perp leg. If it sits on thin isolated margin and the coin pumps hard, the short gets liquidated and you are left holding a naked, fully directional spot long. Use cross or portfolio margin with a collateral buffer to prevent it. See cross vs isolated margin.

Can funding go negative and cost me money?

Yes. The rate floats and can flip sign, so on a bad period you pay instead of receive. The fix is to monitor the normalized rate and close both legs together once funding turns and stays negative, rather than hoping it reverts.

Where do I see live funding rates across exchanges?

In the funding dashboard: Finder pulls funding from 20+ venues, normalizes to a common window, and shows the cadence per contract plus the perp-spot basis, so you compare real yields instead of raw screen percentages.

Related: the pillar Crypto arbitrage guide, normalizing rates of different cadence in funding math, the basis mechanics in futures arbitrage, reading any route in spot arbitrage, collateral modes and liquidation risk in cross vs isolated margin. Common failures - crypto arbitrage mistakes. Live funding rates and spreads across 20+ exchanges - in the funding dashboard and the spread scanner.