I Compared Funding Rates for the Same Coin Across 7 Exchanges. BTC Was Identical. 28 Weren't Even Close.

I've already written about how expensive funding gets on a single exchange — 1,520 pairs, one morning, one pair costing 1,415%/year. This time I asked a different question: does the same coin cost the same to hold no matter which exchange you're on? I pulled this morning's rates for every coin listed on 4 or more of the 7 exchanges RektCalc tracks — 624 of them — and checked the gap between each coin's cheapest and most expensive venue.

Majors move together

This is the part that should be true if arbitrage capital is doing its job, and it mostly is. Across all 7 exchanges checked this morning, the big, liquid coins barely move:

Nine of the ten coins here sit under 10%/year annualized spread between the cheapest and priciest venue — that's the market doing exactly what it's supposed to. BNB is the outlier, and there's a reason: it's Binance's own exchange token, and Binance's funding formula for it doesn't track the same way as how Bybit, Bitget or Gate price it. One coin, one obvious exception, everything else lines up.

The tail is a different market entirely

Once you leave the top 10-15 coins by liquidity, the picture changes fast. Across all 624 coins listed on 4+ exchanges, the median annualized spread was 10.0%/year — basically what the majors show. But the mean was 24.0%/year, and the gap between those two numbers is the tail:

Roughly one in four coins with decent exchange coverage has a real, checkable annualized gap over 20%/year between its cheapest and priciest venue right now. One in eight clears 50%/year. And 28 coins — 4.5% of the sample — clear 100%/year, meaning the difference in what you'd pay to be long on one exchange versus another is bigger than most people's entire annual return target.

The widest gaps this morning

These are all listed on 4+ exchanges, so it's not a single thin order book making up a number:

COTI is the widest gap on the list: Bitget prices it at essentially zero, +0.0013% per 8 hours, while Gate prices the same coin at −0.4344% per 8 hours — a difference of 0.4357 percentage points every 8 hours, annualized out to 477%. Notice that Gate shows up as the "expensive to be long" side on five of these ten coins. One exchange's funding formula, or its specific order book skew, is doing a lot of the work in this list.

What it actually costs to chase the gap

On paper, being long COTI on Bitget and short the same size on Gate looks like free money: you're flat on price, and you collect the 0.4357%-per-8h difference every funding interval. It isn't free, for three reasons.

Fees eat a real chunk of it. Opening both legs and closing both legs later is 4 fills. At a typical 0.05-0.075% taker fee per fill, that's roughly 0.2-0.3% round trip. Capture one funding payment of 0.4357% and you've cleared maybe 0.14-0.24% net after fees — real, but a fraction of the headline number, and you need the position open long enough to collect more than one payment before the math clearly works.

The rate itself doesn't sit still. When I surveyed funding rates in early August, several of the coins near the top of that list had ranges spanning hundreds of percentage points across exchanges within the same morning. A 477%/year gap this morning can be a 50%/year gap by the next funding interval. You're not locking in a year of that rate — you're locking in the next 8 hours, and re-evaluating.

Thin coins mean real slippage. Every coin in that worst-of table is thin enough that its funding rate can sit far from zero in the first place — that's usually a symptom of a lopsided, low-liquidity order book. The same thin book that produces an extreme funding rate also produces worse fill prices when you actually try to open size on both legs at once.

What I actually do with this

I don't trade the tail of this list. The size of the number is exactly why I don't trust it to hold still. What I do use this for is smaller and more boring: before I hold anything longer than a session, I check whether my coin's funding rate looks like a BTC-shaped number or a COTI-shaped number. If it's within the 9-10%/year band the majors sit in, it's not going to surprise me. If it's showing 100%+/year spread between exchanges, that's not a yield I want to farm — it's a sign the coin's order book is thin enough that I'd rather size down or skip the leverage entirely.

If you do want to check the gap for a specific pair before you touch it, run it through the funding arbitrage calculator with real fee inputs, not the headline annualized number. And check live funding rates before you hold anything longer than a day — this morning's snapshot is already a few hours old by the time you read this.

Before you touch a cross-exchange funding gap

1. Majors barely move. BTC, ETH, SOL, XRP, ADA all sat within 9-10%/year annualized spread across 7 exchanges this morning. A wide gap on a major would be the surprise, not the norm.
2. Count the fee legs before trusting the headline spread. 4 fills to open and close both sides typically costs 0.15-0.30% round trip — that comes out of the gap before it's profit.
3. The extreme end of the list is thin-book territory. 28 of 624 coins (4.5%) cleared 100%/year spread this morning — those are also the coins where slippage on entry can outweigh several funding cycles.
4. Re-check before every payment. The rate that made a trade attractive at 06:30 UTC can be a different number by the next 8-hour mark.

→ Funding arbitrage calculator · → Funding calculator · → Live funding rates · → Learn