Crypto arbitrage: the complete 2026 guide

Right now, somewhere across the exchanges you already have apps for, the same token is trading at two different prices. A coin sits at $1.018 on MEXC and $1.026 on Gate. Buy the cheap side, sell the dear side, keep the gap. That is crypto arbitrage, and unlike a directional bet it does not care whether bitcoin pumps or dumps today - it only cares that price A and price B disagree. This is the foundational guide: what arbitrage actually is, every type that exists, how to read a route that is real instead of a mirage, the risks that quietly eat your edge, and how to start without torching your first deposit. Each type below links to its own deep dive.

What crypto arbitrage is

Crypto arbitrage is profiting from the price difference of the same asset across venues - between exchanges, between blockchains, between spot and futures. The crypto market is split across hundreds of venues, each with its own order book and its own crowd, so one token's price between them constantly drifts. Whoever buys cheaper and sells dearer faster than the rest captures the gap.

Unlike trading, arbitrage is direction-agnostic: it doesn't matter whether bitcoin is up or down, only that one asset's price on venue A differs from venue B. The gap appears because:

• liquidity is fragmented - every exchange and every pool has its own book.
• moving an asset between venues costs time and fees, so arbitrageurs can't flatten the price instantly.
• demand spikes locally - a new listing on Bybit, an airdrop, inflow into a hot chain.

Profit = price gap minus exchange fees, transfer cost and slippage. That subtraction is the entire game. A visible "fat" spread is almost always smaller than it looks, and a lot of the time it is worth exactly nothing. We will come back to why.

Types of crypto arbitrage

There are more flavors than most beginners expect, and they split cleanly by where the trade lives: inside one exchange, across exchanges, between spot and a derivative, or across chains.

Cross-exchange spot (Spot-Spot). Buy spot on one exchange, move it over a network, sell on another. You buy AGLD on Kucoin, withdraw over Arbitrum, sell on OKX. This is the everyday baseline flow, and the best place to learn the mechanics. → spot arbitrage explained and the cross-exchange playbook.

CEX-DEX. A centralized exchange (Binance, Bitget) versus a DEX pool (Uniswap, PancakeSwap). The spreads are often wider because fewer people watch on-chain pools, but you pay for it in gas and on-chain execution risk. → CEX-DEX and DEX dumps.

Futures and funding. A perpetual versus spot (capturing the basis), or perp versus perp across venues (funding-rate divergence), held delta-neutral so price direction nets to zero. No coin ever leaves the exchange, which is why this is the most operationally relaxed family. → futures arbitrage, funding arbitrage and the funding math.

Cross-chain. One token's price gap across blockchains, moved via a bridge or by using an exchange as the bridge. ETH cheaper on Arbitrum than on Base, for example. → cross-chain arbitrage.

Listings. The price gap in the first minutes after trading opens on a new exchange, when one venue has the token and another doesn't yet. One-off windows, roughly 2 to 8 times a month, occasionally double-digit percent but punishing if you mistime them. → new-listing arbitrage.

Pre-market. A token's price in an exchange's pre-market venue versus the expected fair value, before official trading opens.

Triangular. Three pairs inside one exchange (e.g. BTC→ETH→USDT→BTC) that close at a profit. No cross-exchange transfer, purely inside the book, so it lives and dies on fees. → triangular arbitrage.

How to read a tradeable route

Every type reduces to a route - a concrete path: asset + buy leg + sell leg + transfer. The arbitrageur's core skill isn't "spotting a big percent", it's telling an executable route from a phantom. A 7% spread with a closed withdrawal is worth less than a 1.2% spread you can actually clear. Read every route layer by layer.

1. Gross spread (bid − ask) / ask - the ceiling, never the result.
2. Minus fees on both legs (taker roughly 0.1% + 0.1%, so 0.2% gone before you move).
3. Minus transfer - the network, its fee and its time. A slow network means the gap can collapse while your coins are in flight.
4. Check D/W - is withdrawal open on the buy leg and deposit open on the sell leg? Closed on either side = phantom, no matter how pretty the percent.
5. Check depth - a spread for $200 is not a spread for $2000. Thin books quote beautifully and fill terribly.

The full worked example, with the arithmetic, lives in how to read and build arbitrage routes.

The main risks, honestly

• Phantom spreads. Most gaps above 5% aren't tradeable: closed D/W, dead liquidity, a stale price, a network or contract mismatch. The bigger the headline number, the more suspicious you should be. Mitigation: assume a fat spread is fake until the route checks out.
• Closed withdrawal or deposit. The single most common blocker. The spread is right there and you simply cannot move the asset out or in. Mitigation: verify live D/W status before you commit a cent.
• Transfer time. While the asset is in transit the price converges, especially on slow networks and bridges. Mitigation: prefer fast networks, size the trade to the convergence risk.
• Slippage. A thin book eats the spread at real size. Mitigation: read depth, not just the top of book.
• Operational mistakes. Wrong network, missing memo, selling too late. These are the losses that sting most because they are entirely self-inflicted. Mitigation: a fixed checklist, every time, no improvising.

A step-by-step first trade

Here is a concrete, conservative spot-spot run you can actually execute. Numbers are illustrative but realistic.

1. Pick a small size. Say $300. Big enough that the percent matters, small enough that a mistake is a lesson, not a wound.
2. Find a route on a fast network. Scanner shows a token at $1.018 on MEXC and $1.026 on Gate - a 0.79% gross gap - withdrawable over Polygon. Reject anything routed only over Ethereum at this size, the gas alone can erase the edge.
3. Subtract before you click. Taker 0.1% buy + 0.1% sell = 0.2%. Polygon fee on $300 is pennies, call it 0.05%. Estimated slippage at this size, 0.1%. Net expectation roughly 0.44%, about $1.30. Small, and that is the point - you are buying reps, not a Lamborghini.
4. Confirm D/W is open both ways. Withdrawal open on MEXC, deposit open on Gate, same network on both sides. If either shows ❔ or closed, kill the trade.
5. Buy the cheap leg. Market-buy ~$300 of the token on MEXC.
6. Withdraw immediately, correct network selected, address and memo double-checked. Watch the confirmations.
7. Sell on arrival. The moment it lands on Gate, market-sell. Do not wait for "a little more", the gap is already closing.
8. Write down the real result. Actual fill prices, actual fee, actual time, actual net. Your spreadsheet teaches you more than any guide after ten trades.

Tools - where Finder fits

Watching 20+ exchanges by hand is impossible, and eyeballing a raw percentage is exactly how phantoms get traded. The Finder scanner computes routes with fees and network math already subtracted, shows honest deposit/withdrawal statuses (an honest ❔ when status is unknown, never a fake 🟢), and filters phantoms across 20+ CEXs, DEXs and perp venues. For the delta-neutral families it also tracks live funding rates. The filtering is the point - you only look at routes that already survived it. You can open the whole dashboard here.

How to start practicing

Treat the first month as tuition, not income.

1. Capital on 3+ exchanges. Not every window appears where your funds sit. Pre-fund Binance, Bybit, MEXC, OKX, Gate so you can act on whichever side lights up.
2. Fast networks ready. Solana, BSC, Base, Polygon, Arbitrum for transfers. Avoid Ethereum mainnet by default for small sizes.
3. A monitoring tool, because manual scanning does not scale past a couple of pairs.
4. Exit discipline. Know your take point in advance and do not hold and hope.
5. Scale slowly. Start at $50 to $300 per trade. Move up only once your logged net results are consistently positive after all costs.

How much you can realistically make

It depends on the type. Recurring types - spot-spot, CEX-DEX, funding - are a stream of small 0.5% to 2% net windows, and earnings scale with capital and speed, not luck. One-off types - listings, large cross-chain gaps - are rarer but wider, 2% to 30%+, and carry more risk per attempt. The thing that actually determines your results isn't the size of the headline percent, it is the share of routes you can execute after filtering phantoms, multiplied by your discipline. Honest expectation for the early weeks: small or even negative net while you learn what a real route looks like. That is normal, and it is the part the scam-y "passive income" content leaves out. This is real work with a real chance of losing on a given trade. For a longer reality check, see is crypto arbitrage profitable and how much starting capital you need.

FAQ - crypto arbitrage

What is crypto arbitrage in simple terms?

Buying an asset where it's cheaper and selling where it's dearer, profiting on the gap net of fees and transfer. It doesn't depend on whether the market goes up or down.

Can you still make money on crypto arbitrage in 2026?

Yes, but it isn't "easy money". Liquid windows close in seconds to bots, and most "fat" spreads are phantoms. The money goes to people who quickly filter the executable routes and trade them with discipline.

Which type is easiest to start with?

Cross-exchange spot arbitrage has the clearest mechanics, so most people start there. Funding arbitrage is delta-neutral with lower directional risk. Listings are high-risk and need preparation, so leave them for later.

Do you need a bot for arbitrage?

For liquid assets where the window lives seconds, yes, speed wins and bots win speed. For illiquid and cross-chain windows that last minutes, you can keep up by hand with a good scanner.

How much money do you need to start?

Technically a couple hundred dollars, but fees and transfers eat small trades. Practically, spread $3,000 to $10,000 across several exchanges so you can catch windows wherever they appear.

What's the most common beginner mistake?

Trading the headline percentage without subtracting fees, transfer and slippage, and selling too late once the gap starts closing. Both are fixed by a checklist and a pre-decided exit.

Deep dives by type: spot · cross-exchange · futures · funding · cross-chain · new listings · triangular. Mechanics: how to read a route · funding math · withdrawal windows. Live routes across 20+ exchanges - in the web dashboard.