Funding rate arbitrage is the trade behind most of the "delta-neutral" yield in crypto. You buy an asset, short the same amount of its perpetual future, and collect the payments that leveraged buyers make to sellers to keep that contract tied to the spot price. Price moves largely cancel out, and what is left is a stream of funding payments. It is real income with a clear economic source, and it is the engine inside Ethena's USDe, inside institutional basis desks and inside several managed yield strategies, including some of ours.
The tutorials ranking for this term mostly stop at the arithmetic. They explain the two legs, annualise a funding rate and list three risks. What they skip is the part that decides outcomes, what funding has actually paid across market cycles, how a fully hedged position still gets liquidated, and what changes when someone packages the trade for you. That is what this piece covers. Maker Core, one EarnPark strategy, is built around this trade, and later we show what packaging adds and what you give up.
What the funding rate is and why it exists
A perpetual future, or perp, is a futures contract with no expiry date. Dated futures converge on the spot price when they expire, which keeps them honest. A perp never expires, so exchanges need another mechanism to stop its price drifting away from spot. That mechanism is the funding rate, a periodic payment between traders holding long and short positions, as the academic treatment by He, Manela, Ross and von Wachter sets out. When the perp trades above spot, longs pay shorts. When it trades below, shorts pay longs.
Funding intervals vary by venue. Every eight hours is the most common schedule, but some venues settle hourly, and CoinGlass lists each venue at its own settlement interval, eight hours on most and one hour on venues such as Hyperliquid. Before comparing two rates or annualising one, check whether the figure is per eight hours, per hour or already normalised. The rate looks tiny per interval and adds up over a year. On 18 October 2026 CoinGlass showed Bitcoin's open-interest-weighted funding at 0.0071%. If that were an eight-hour rate and stayed there, it would come to roughly 7.8% a year.
Funding has run positive more often than not, at least for Bitcoin and Ether, which is what the funding history Ethena publishes (covered below) shows. One contributing factor is demand for leveraged long exposure, since perps are the cheapest way to get it. Somebody has to take the other side of that demand, and the funding rate is part of what they get paid for doing it.
How funding rate arbitrage works
The classic version is a cash-and-carry trade. You buy one unit of an asset on the spot market and open a short perp position of the same size. If the price rises, the spot holding gains and the short loses by about the same amount. If it falls, the reverse happens. Traders call the position delta-neutral because its value barely responds to price direction. What it does respond to is funding. Every interval the short leg receives the funding payment while the rate is positive, and that payment is the return.
There are variations on the theme. When funding turns negative, the mirror trade, shorting spot with borrowed coins and going long the perp, collects instead, at the cost of borrowing interest. Traders also run perp against perp, long on a venue paying low funding and short on one paying high funding, which is the cross-venue version. And professional desks often stake the spot leg, so a position in ETH or SOL earns staking rewards on one side and funding on the other. Our explainer on liquid staking covers what that staking leg actually yields and costs.
This is a different animal from the arbitrage most people picture. The cross-exchange and triangular trades in our arbitrage explainer chase a price gap that closes in seconds, so they are a contest of speed. Funding rate arbitrage is a carry trade. Nothing has to close for it to pay, and the position can sit for weeks collecting. That makes it far more accessible to ordinary capital, and it is why its risks live in the plumbing, margin, collateral and venues, rather than in execution speed.
How the return is calculated
Funding income per interval is the short position's notional value multiplied by the funding rate. Annualising an eight-hour rate means multiplying by three intervals a day and 365 days, so by 1,095. The figure that matters, though, is the return on all the capital you tie up, not the headline rate.
Take $10,000 and an illustrative funding rate of 0.01% per eight hours. You cannot put the full $10,000 into spot, because the short needs margin. Say $8,000 buys spot and $2,000 sits as margin behind an $8,000 short. Funding pays $8,000 times 0.01%, which is $0.80 per interval, $2.40 a day and about $876 a year. On your $10,000 that is 8.76%, already below the 10.95% you get by annualising the rate itself.
Then costs come out. Opening and closing both legs means four trades, and if they total 0.2% of the $8,000 notional (an illustrative figure, since your fee tier decides it), that is $16, close to a week of funding at this rate. Rotate positions often and fees eat the trade. Funding also refuses to hold still, which is why the daily dollar figure is a better guide than any annualised number taken from a single reading.
Leverage changes the arithmetic. Institutional desks borrow stablecoins against collateral, buy more spot with them and short more perps, so the same capital collects funding on a larger notional. That raises the return and also pulls margin mechanics into the centre of the trade, which is where most of the risks below come from.
What funding has actually paid
The best neutral evidence on long-run returns comes from central bank researchers. A BIS working paper studying the futures basis for Bitcoin and Ether found the cash-and-carry return averaged above 10% a year over its study period, much larger than carry in equities, bonds, currencies or commodities, and that it could reach as much as 60% a year. That figure measures the basis on dated futures, the classic version of the trade, rather than perp funding. The authors trace the size of the premium to two forces, smaller investors chasing leveraged upside in boom periods and a relative scarcity of arbitrage capital willing to take the other side. They also stress that running the trade is risky because margins spike and liquidations hit during drawdowns, and that high carry tends to come before price crashes.
The perp research points the same way. He and his co-authors found deviations from no-arbitrage perp prices were larger in crypto than in traditional currency markets, and that a strategy exploiting them produced high Sharpe ratios. They also found those deviations shrinking over time, which is what you would expect as more capital chases the same trade.
Ethena publishes the most detailed funding history from a market participant, so treat it as an interested party's data, though still useful. Its funding risk analysis, puts average annualised BTC and ETH funding between 7.8% and 9% over the three years to 31 December 2024, including the 2022 bear market. Those are gross funding rates, before fees, idle margin and the other costs above, not the net return of a strategy. It also found that 15.9% of days for BTC and 17.5% for ETH had a net negative funding return, and that the longest unbroken negative streak lasted 13 days. Funding is positive on average, swings hard with the market, and turns negative often enough that any honest projection has to include it.
Where the trade breaks
Delta-neutral describes price exposure. It says nothing about the other ways a position loses money, and those other ways tend to arrive at the same moment.
Negative funding
When sentiment turns, longs stop paying and shorts start. A short leg that collected for weeks begins paying out every interval. Negative stretches have historically been short, but nothing guarantees the next one will be, and a position that cannot afford to wait one out closes at a loss.
Basis blowouts
The perp and the spot asset are different instruments trading in different markets, and the gap between them, called the basis, moves. For a long spot and short perp position, the damage comes when the perp pulls further above spot. The short then loses more than the spot leg gains, so the position shows a loss on paper even though the hedge is in place. That tends to happen in a sharp rally, when leveraged buyers crowd into perps. The timing makes it worse. Losses on the short are charged against margin in cash as they happen, while the matching gain on spot stays unrealised, so a rally can drain margin before the spot gain is any use. If you are forced to close at that point and thin order books mean the two legs cannot be unwound at their marked values, the paper loss becomes a real one.
Liquidation of a hedged position
This is the risk the tutorials understate. The short perp needs margin, which is collateral you post with the exchange to back the position. If the price rallies hard, the short loses money and eats into that margin, while the offsetting gain sits in a spot position that may be held on another venue, or on the same venue but not accepted as margin at its full value, or at all. Exchanges also do not count collateral at market price. They apply a discount, called a haircut, and they can change it without notice. A position can be perfectly hedged on paper and still fall below the exchange's maintenance margin, at which point the exchange closes it on its own terms, and a liquidation cascade can push prices further while it does.
Exchange safety systems add a second layer. During the crash of 10 October 2025, when more than $19 billion of leveraged positions were liquidated in about a day, Galaxy Research documented auto-deleveraging (ADL) kicking in on several perp venues, forcibly reducing user positions to cap further losses. In a crash the profitable side is the short, and as Galaxy put it, "ADL trims winners". If ADL reduces or closes the profitable short, the hedger may be left holding the spot asset without its hedge during the crash.
Exchange failure
Every leg of the trade sits on a venue, and that venue is a counterparty you are trusting with your collateral. FTX is the reference case. Its founder was sentenced in 2024 for what the U.S. Attorney described as stealing over $8 billion of customers' money. A hedge does nothing for collateral the exchange itself has lost, and positions on a failed venue, hedged or not, become a claim in an insolvency process. Splitting the legs across venues spreads that risk but means a failure on one side leaves the other side unhedged. Our guide to proof of reserves covers what exchanges now publish about their holdings and what those reports leave out.
Stablecoin and staking legs
Many versions of the trade borrow or hold stablecoins and stake the spot leg, which adds a stablecoin's peg risk and the staking protocol's smart contract risk on top of everything above. Our ranking of the safest stablecoins covers the first of those.
The real danger is correlation. In a sharp market-wide fall, funding turns negative, exchanges raise haircuts on volatile collateral, borrowing gets more expensive and order books thin out, all at once. Each of those risks is manageable on its own. Together they can produce a loss larger than their sum.
How Ethena turned the trade into a dollar
Ethena took the cash-and-carry trade and wrapped it in a token. USDe is backed by crypto collateral hedged with short perpetual positions, and the funding those shorts collect, plus staking rewards on the collateral, flows to holders who stake USDe as sUSDe. Our Ethena breakdown walks through its fees and yield history. Ethena says in its documentation that a reserve fund steps in when combined revenue turns negative. Ethena also says it shifts backing toward liquid stablecoins when funding is low or negative.
The supply history shows how cyclical the trade is at scale. DefiLlama's USDe data puts circulating supply at a peak of about $14.8 billion on 4 October 2025 and about $4.7 billion on 18 October 2026. The crash of 10 October 2025 tested the design. According to Galaxy, USDe traded as low as about $0.65 on Binance while minting and redemption kept working at Ethena and it traded near $1 on most other venues. On Binance the low price fed into that exchange's liquidation cascade.
USDe shows both sides of the trade at once. Funding income was real enough to build a multibillion dollar stablecoin alternative on it, and supply contracted by roughly two thirds from that peak. And a product built to be delta-neutral met its worst moment through venue mechanics rather than price.
How EarnPark's strategies use the same trade
Some of EarnPark's strategies are funding rate arbitrage, and we can say exactly which, because our help centre publishes a full description of each one. The Maker Core description gives its type as "delta-neutral interest-rate / funding-rate arbitrage using borrowed funds; liquidity provision (market making)". In plain terms, it borrows against your deposit to run a hedged funding trade and also quotes prices on both sides of the market. On Binance, it says, the pair "typically takes the form of a cash-and-carry (spot-futures) trade", buying an asset such as ETH or SOL with borrowed stablecoins, staking it and shorting the same amount on the derivatives market. It is designed to earn the funding rate on the short leg and the staking yield on the long leg, alongside a market-making book that earns the spread between bids and asks. Per its strategy description as of September 2026, it runs on that single venue, through a portfolio margin account, at a working leverage of 8 to 10 times, with an automated process that cuts positions when leverage climbs. The market-making leg also carries inventory and execution risk. Maker Core carries the Market-Neutral label, and EarnPark's risk labels describe what kind of risk a strategy takes rather than rating how much, so Market-Neutral describes the intended exposure, not a guarantee against losses.
The DeFi strategy is the cross-venue version. In plain terms, it goes long an asset on one venue and short the same asset on another, and earns the difference in interest and funding rates between them. Its description calls it a delta-neutral cross-exchange interest-rate arbitrage, with a single-exchange cash-and-carry variant. Alpha Vault, in the Advanced Yield section, works differently altogether. Our September report describes it as roughly half coin and half stablecoin, the help centre says the position rebalances back toward that split as prices move, and it carries the Market-Exposed risk label.
Three details in those descriptions are worth checking with any provider that packages this trade, us included. You receive a target rate, shown net of the company's margin. The target rate is not a guaranteed return. If the strategy earns more, the company keeps the excess, and if it loses, your return can fall below the target and can become a loss, so in the description's own words your upside is capped at the target rate while your downside is not. Your deposit becomes collateral in an account in the company's name, pooled with other participants in that strategy, and what you hold is an unsecured contractual claim. And no insurance fund, compensation scheme or deposit guarantee applies.
Maker Core gives you access to a managed version of the trade. In return you accept the strategy's terms, the operator and venue risk, and an unsecured claim. Running the trade yourself gives you more control but leaves the monitoring and the risk with you.
Withdrawal terms differ by strategy, and a funding strategy is not automatically liquid. Maker Core pays daily, and withdrawals go through a bonding period, which means a withdrawal can take up to 30 days to arrive after you request it, with the rate still accruing during that wait. USDT and USDC DeFi carry the Cross-Venue label and allow withdrawal requests at any time, and the help centre says USDT and USDC withdrawals are processed instantly. Alpha Vault pays out monthly and processes withdrawals monthly. Current rates and terms sit on each strategy page, the USDT page shows the range across all of them, and our standalone risk statement covers platform-level risks. Whichever strategy you pick, the published rate is the base every depositor gets holding zero PARK, the platform's own token, as our page on token gating explains.
Running it yourself or holding a packaged version
Running the trade yourself gives you control and complete visibility. You choose the venues and monitor reported funding payments, keep the proceeds after fees and costs, and close when you choose, subject to venue liquidity and execution. In exchange you carry every risk above personally, need to watch margin around the clock in volatile weeks, and pay whatever fees a retail account pays.
A packaged version, whether a token like USDe or a managed strategy, delegates day-to-day management to an operator. Whether that improves execution or reduces costs depends on the product's fees, scale and terms, and it adds reliance on the operator and its venues. The price is a layer of trust. You rely on the operator's risk model, venues and candour, and depending on the structure you may give up part of the upside. The questions worth asking are the same everywhere. Where exactly does the yield come from, which venues hold the collateral, how much leverage sits behind it, who keeps the excess in a good month, who absorbs a bad one, and how quickly can you leave? Look for written explanations of the strategy, its risks and how it performed in both strong and weak periods. Those disclosures give you something to assess, but they do not remove the risks. A provider that answers with "proprietary algorithms" gives you nothing to assess.
Funding rate arbitrage is one of the few crypto yields with a clear economic source, leveraged demand paying someone to take the other side. The return is real and cyclical. The hedge reduces simple directional price exposure, while basis moves, margin pressure, venue failures and correlated risks can still cause losses. That makes it worth understanding even if you never place the trade yourself.
If you would rather hold the trade than run it, review each strategy's risk label and withdrawal terms in the app, and if the structure fits, Start earning there.
FAQ
Is funding rate arbitrage risk free?
Funding rate arbitrage is not risk free, even though the hedge removes most direct price exposure. Funding can turn negative, the perp can pull further above spot and put the position at a loss on paper, the short leg can be liquidated if margin runs short, and the exchange holding the position can fail. The BIS research on crypto carry links the trade's high returns to spikes in margins and liquidations during drawdowns.
How much can you earn from funding rate arbitrage?
How much you can earn from funding rate arbitrage depends on the period, the venues and your costs, and the best evidence measures gross carry rather than net returns. A BIS study of the futures basis for Bitcoin and Ether, the dated futures version of the trade, found carry averaging above 10% a year over its study period, with peaks up to 60%. For perp funding itself, Ethena's data puts average annualised BTC and ETH funding at 7.8% to 9% over the three years to 31 December 2024, which is gross funding rather than the net return of a strategy. Fees, capital held as margin and negative periods all come out of those figures.
What happens when the funding rate turns negative?
When the funding rate turns negative, shorts pay longs, so the short leg of a cash-and-carry trade starts paying out every interval instead of collecting. Traders either hold through it and expect funding to recover, close the position, or flip to the reverse trade by borrowing the asset to short spot. In the funding history Ethena publishes, the longest unbroken negative streak lasted 13 days.
Can a delta-neutral position be liquidated?
A delta-neutral position can be liquidated, because the exchange margins it on its own rules rather than on the logic of the hedge. A sharp rally drains margin on the short perp while the matching gain sits in spot, and exchanges count collateral at a discounted value they can change. If the account falls below maintenance margin, the exchange closes positions regardless of how well hedged they are.
Is funding rate arbitrage the same as a cash-and-carry trade?
Funding rate arbitrage is the perpetual futures version of the cash-and-carry trade. Classic cash-and-carry pairs spot with a dated future and earns the basis as the contract converges at expiry, while the perp version earns periodic funding payments because a perp never expires. The risks are similar, though a perp position has no built-in end date.
Do you need leverage for funding rate arbitrage?
Funding rate arbitrage does not require leverage, since a fully paid spot holding against a modestly margined short works. Without leverage the return on total capital sits below the headline funding rate, because part of the capital is held as margin. Professional desks borrow to scale the position, which raises returns and brings liquidation risk closer.