Funding Rates vs Borrow Costs in Crypto Trading

Why these two costs confuse traders

Every leveraged position in crypto carries a carrying cost, the price paid for borrowing other people's money or other people's liquidity to amplify exposure. The trouble is that the same idea shows up under two different names in two different venues, and the way each cost is paid, calculated, and reset is genuinely different.

On a perpetual futures exchange, whether a centralized order-book venue or an on-chain perpetual DEX, the cost of leverage is called the funding rate. On a margin account at a centralized exchange, where a trader borrows USD or a stablecoin to go long, the cost shows up as a borrow APR. Both numbers quote an annual percentage, both move with market conditions, and both can overwhelm a position that is going nowhere on the chart.

The confusion usually shows up when a trader compares the two side by side and assumes they are substitutes. They are not exactly substitutes. Funding is a payment between two sides of a perp market that swings sign depending on whether longs or shorts are crowded. Borrow APR is a one-directional fee the borrower pays to a lender. A trader who runs both at once, which is common when a leverage trade is hedged with a perp, can be paying both layers at the same time.

Risks of comparing the two naively

Before walking through the mechanics, the failure modes are worth naming. Most traders who get hurt on these costs are not blindsided by the headline number; they are blindsided by the hidden layers underneath it.

First, comparing an annualized funding rate of, say, 12% to a CEX borrow APR of 6% ignores that funding can flip negative. In a crowded-short regime, longs collect funding, which means the position earns carry rather than paying it. A borrow loan cannot do that. The asymmetry breaks the simple comparison.

Second, slippage and taker fees are not in either quoted number. A market order that walks the book on a thin perp, or a margin borrow that triggers a liquidity premium on the exchange's lending pool, can cost more than the headline rate.

Third, liquidation thresholds are venue-specific. The same BTC position with the same notional size can survive on one venue and get auto-deleveraged on another, because cross margin, isolated margin, insurance funds, and mark-price methodology each feed into the liquidation engine differently. Carrying cost quotes do not capture this risk at all.

Fourth, counterparty risk differs. A CEX margin borrower is exposed to that exchange's solvency and to its lending-pool rules. A perp trader on a decentralized venue is exposed to smart-contract risk, oracle risk, and the auto-deleveraging auction in the insurance fund. Both cost stacks are not risk-free.

How perpetual funding rates work

A perpetual future tracks an underlying asset's spot price through a funding payment rather than through expiry and settlement. At fixed intervals (every eight hours is the industry standard, though some venues run every one or four hours) the exchange computes a funding rate. If the rate is positive, longs pay shorts. If negative, shorts pay longs. The payment is applied directly to the position; there is no wire transfer, and winners do not need to do anything to collect.

The funding rate itself has two components. The interest-rate component anchors the perp price to the equivalent of a money-market rate, designed to keep the contract tethered to the underlying over time. The premium component captures the gap between the perp's mark price and the spot index at the moment of the funding snapshot. Hot markets push the premium positive; cold or heavily shorted markets push it negative.

A simple way to read funding: if BTC perp trades $50 above spot and the funding window hits, longs pay shorts a small percentage. If BTC perp trades $50 below spot (rare but it happens), shorts pay longs. The rate is usually expressed as an eight-hour figure but is almost always quoted as an annualized number by data terminals.

Funding payments stack on top of trading fees but are independent of them. A trader who opens and closes a position within the same eight-hour window and never holds through a funding timestamp pays no funding at all, only the entry and exit taker fees. A trader who holds through one timestamp pays one period's worth. Holding for a week means paying seven periods, with compounding psychology if the rate climbs.

How CEX borrow costs work

On a centralized exchange, a margin borrower takes a loan of USD, USDT, or USDC (sometimes the base asset, though that is rarer) to fund a long position. The borrow APR is set by the supply of available margin in the exchange's lending pool, and the borrower pays interest on the outstanding balance at each interval, often hourly, sometimes daily.

Unlike funding, the borrow APR is one-directional. The borrower always pays the lender. There is no scenario in which the borrow rate goes negative and the borrower is paid to hold the loan. This is the cleanest mechanical difference between the two costs.

Borrow APR compounds against the borrower. Most exchanges quote an hourly rate that, when annualized, sounds modest (3% to 15% on USDT is common in calm markets), but the daily accrual means the trader's debt grows even on a flat-position day. For BTC and ETH specifically, the borrow rate is usually a function of cross-asset lending demand, not the volatility of the underlying.

A subtler mechanic: CEX margin borrow APR often steps up sharply when utilization in the lending pool crosses a threshold. If the exchange's USDT lending pool is 90% utilized, the borrow APR jumps to discourage further borrowing. This creates a regime where a trader thinks they are paying 6% to lever long, wakes up the next morning to a 25% APR, and is bleeding fast against a flat position. Funding rates do not have this same step behavior on most venues.

Negative funding regimes, explained honestly

Crowded-short environments produce negative funding on perps, meaning shorts pay longs to keep the short open. This often gets framed in trading communities as "free leverage" for the long side. It is not free, and the framing obscures real costs.

The reason funding exists at all is to keep the perp price anchored. If nobody were willing to take the other side of a crowded trade, the perp would drift far from spot and arbitrage would break. Funding redistributes a small slice of PnL between the two sides to compensate whoever is providing the stabilizing liquidity. In a negative-funding regime, that stabilizer is the short, paying the long for the privilege of staying short.

Who is actually on the other side when funding is negative? Usually a market maker running a delta-neutral book, or a basis trader who is long spot and short perp simultaneously, collecting the negative funding as income while hedging the directional risk. Institutional desks, designated market makers, and specialized basis funds dominate this trade. Retail traders who happen to be long during a negative-funding window are incidental beneficiaries of a market structure that exists for other reasons.

There are also costs that come bundled with negative funding. Liquidity is thinner on the side being paid, because the economics only work for sophisticated players. Spreads widen when funding is sharply negative. And the rate tends to revert. A trader who opens a long position betting that negative funding will continue is making a directional bet on sentiment, not a cost-free trade.

Liquidation thresholds are not the same across venues

A trader's worst day starts with a liquidation. The price at which a leveraged position is auto-closed depends on the venue's mark-price methodology, maintenance margin requirement, insurance fund depth, and whether the account is cross or isolated.

On a CEX margin position, liquidation is typically driven by the mark price crossing the maintenance margin threshold. The exchange auto-sells the position, sometimes at a loss to the trader and sometimes with the insurance fund covering the gap. The relevant cost question is not just "what is my borrow APR" but "how thick is the maintenance buffer before this triggers."

On a perp DEX, the venue's oracle and mark-price feeds feed a liquidation engine that may execute in an auction rather than directly on the order book. Auto-deleveraging (ADL) is the backstop when the insurance fund cannot absorb a loss. ADL means another trader's position is force-closed against the liquidated trader's losses. ADL is rare but not theoretical, and it tends to show up in the same high-volatility windows when funding rates are also spiking.

The same BTC position with the same notional can therefore survive a 3% drawdown on one venue and get liquidated on another, with no change in the underlying chart. The headline borrow APR or funding rate tells the trader nothing about this margin of safety.

Hidden costs that both quotes leave out

The advertised funding rate and the listed borrow APR are both net of the most expensive line items. A trader who optimizes on these headline numbers is missing several layers.

Taker and maker fees apply on every trade, both on entry and exit. On perps, taker fees of 0.04% to 0.06% per side are typical for retail; on CEX margin spot pairs, fees are lower but still apply. For a position opened and closed inside one funding window, fees can exceed the funding cost.

Slippage on entry and exit, especially on less liquid pairs or during volatile windows, can add another 0.05% to 0.20% per side. Token-margined perps (where BTC is the collateral and the position is also in BTC) carry an additional embedded volatility exposure that USD-margined perps do not.

Funding payments also include a small basis component that reflects the spot-versus-perp gap. In low-volatility regimes that basis component is near zero; in trending markets it can spike. Traders who think they are paying a clean "funding rate" are sometimes paying a stretched-implied spread on top.

CEX margin borrowers may also pay a liquidation fee if the position is force-closed, drawn from the borrower's remaining balance. Some exchanges also charge a borrowing fee on top of the hourly interest accrual, especially when the loan is initiated. These charges are listed in the fee schedule but easy to miss.

How to think about which is cheaper in which regime

Without giving advice, a useful mental model is to match each cost to the condition under which it is cheapest.

Funding is the cheaper carry when rates are negative or neutral and the trader is willing to hold through eight-hour windows. It becomes expensive when the market is hot, longs are crowded, and funding has run positive for weeks. Holding a long through a 0.05% per eight-hour window means about 22% annualized, which is a real cost against a flat position.

Borrow APR is the cleaner choice when the trader wants a flat, predictable cost and is willing to pay a one-directional fee to avoid funding surprises. It is more expensive in low-rate regimes than negative funding, but it does not flip sign. For a long-term leveraged position measured in weeks or months, a stable borrow APR is easier to model and budget against.

Traders often run both layers simultaneously, using a CEX margin loan to fund the spot leg of a basis trade and a perp short to hedge. In that setup, the relevant comparison is the cost of the loan (paid to the exchange) versus the funding payment received (paid by the short side of the perp). The trade only works when the funding received exceeds the borrow paid.

Before opening either, the practical checks are similar: confirm the venue's liquidation buffer, review the fee schedule for hidden charges, and run a worst-case scenario where the rate spikes into the top quartile of the past year. Carrying costs are quiet when the position works and loud when it does not.

Stay ahead of funding and borrow rate moves

Funding rates and borrow APRs shift on a timescale of hours, not days, and they shift independently from the underlying chart. A market that looks quiet on the candle can be paying 30% annualized to stay long. Tracking which regime you are trading through, and how crowded the funding side is, is something most traders do by hand and most regret trying to do that way. Zippfeed surfaces BTC, ETH, HYPE and broader-market perp headlines with sentiment scoring (bullish, neutral, or bearish) and an importance rating, so you can see when funding is flipping or borrow demand is stepping up before the cost eats the trade.