Arbitrage Keeps Crypto Prices Together—At a Cost

Arbitrage keeps exchange prices aligned by rewarding traders who buy an asset where it is cheap and sell it where it is dear; Uniswap v4’s January 31, 2025 launch made the fees around that correction programmable. The basic force did not change. What changed was a pool builder’s ability to alter how much the correcting trade costs as market conditions move.

How does crypto arbitrage align exchange prices?

Crypto arbitrage aligns prices when a trader executes offsetting trades against a gap larger than all expected costs. If ETH is offered at $100 on one venue and bid at $101 on another, an arbitrageur can buy on the first and sell on the second. Those orders lift the cheap price and press down the expensive one until the remaining spread no longer covers trading fees, network fees, slippage and execution risk.

In an automated market maker, the trade changes the pool’s reserves and therefore its quoted price. A 50/50 constant-product pool begins with equal values of two assets, an arrangement explained in this SyncSwap Classic pool primer. When outside markets move first, the pool becomes stale; arbitrageurs trade against it and reset its reserve ratio. The oracle is not doing the alignment. Inventory-moving trades are.

• Liquidity providers supply the token inventory and bear price risk.
• Arbitrageurs supply transactions, capital and execution judgment; they keep any spread left after costs.
• Swappers create order flow and accept the pool’s price impact.
• Fee recipients are normally LPs, with protocol or hook recipients included only when a pool’s rules specify them.

Who pays for arbitrage and who earns the fees?

The arbitrageur pays the explicit costs, but liquidity providers can bear the economic cost of selling to a better-informed counterparty. When a reference price jumps, the arbitrageur buys the pool’s underpriced asset before passive liquidity can be repriced. The LP collects a swap fee yet may finish worse off than a continuously updated market maker would. The arbitrage profit and the LP’s loss are not identical after gas, fees and competition, but they come from the same stale quote.

This is why a narrower spread is not free infrastructure. Searchers and block builders may also capture value through transaction ordering, while users benefit indirectly from quotes that track the wider market. Builders must decide which side to subsidize: low fees invite faster correction and more flow; high fees protect inventory but allow larger price gaps to persist.

What did Uniswap v4 change for arbitrage?

Uniswap v4 let pool creators replace a short menu of preset fee tiers with arbitrary fees and hook-controlled dynamic fees. Earlier Uniswap designs embedded more pool behavior in the core protocol: v2 used a fixed 0.30% LP fee, while v3 offered fixed tiers. In v4, a hook can raise fees during volatility, lower them in quiet markets or add logic around a swap.

That flexibility enables experiments aimed at reducing losses to fast arbitrage, but it adds code and model risk. A fee that reacts too slowly still leaks value; one that reacts too aggressively can repel the traders who restore the price. Permissionless deployment also means “v4 pool” does not describe one risk policy. Each hook and fee rule needs separate review.

Does arbitrage guarantee one crypto price?

No: arbitrage creates a price band, not a single universal price. Transfers can be delayed, chains can fragment liquidity, centralized venues can halt withdrawals, and thin pools can impose more slippage than a visible gap will pay.

Our view is that arbitrage remains the practical synchronization layer for crypto markets, but programmable fees make that layer negotiable rather than automatic. The useful test is whether a design narrows executable spreads after all costs while improving LP returns. Public launch materials establish what hooks can do; they do not establish which fee policies win across volatile cycles. That outcome remains an empirical unknown.