ETH funding rate arbitrage

Why the majors are the hardest funding trade, not the easiest

Funding arbitrage pays for one thing: disagreement between exchanges. You collect the difference between what one venue's longs pay and what another's receive, so the size of the opportunity is the size of the gap. Anything that makes venues agree destroys it.

Ethereum is the market most exposed to everything that produces agreement. It is listed on effectively every venue, carries the deepest order books, and attracts the most competing capital — including market makers running at fee tiers and latency a retail account cannot match. The more participants watching a spread, the faster it closes. By the time a funding difference on ETH is large enough to notice, it has usually been taken.

Costs then finish the job. Every delta-neutral position crosses the cross-exchange spread twice — once opening, once closing — and pays fees on all four fills. That sets a floor the funding income has to clear before the trade earns anything. On thin, crowded spreads the floor is frequently higher than the spread itself, which is a losing trade dressed up as a market-neutral one.

Where the dispersion actually lives

The same forces run in reverse further down the market list. A token quoted on several venues but followed by far less capital can carry genuinely different funding on each, for long enough to be worth trading. That is the universe we publish: 500 markets quoted on at least two venues, 0 of them on all 10 tracked venues. The live funding rates page ranks them by current annualized net spread.

None of which makes those markets easy. They carry thinner liquidity, wider spreads and sharper reversals, and the same execution arithmetic applies. The difference is that the gap is often large enough to survive it. See the execution costs guide for the breakeven math and the risk guide before sizing anything.

Would you ever add ETH?

If Ethereum funding started dispersing across venues in a way that cleared costs, yes. The exclusion is a judgment about current market structure, not a permanent rule — and it is our own judgment, made from running the scanner, rather than a data limitation. We would rather say plainly that a market is not worth trading than publish a number implying it is.