
Important Disclaimer
This content is for educational purposes only and does not constitute financial, investment, or trading advice. Cryptocurrency arbitrage involves significant risks including potential loss of capital. Always conduct your own research and consult with qualified financial advisors before trading.
The Short Answer
Price and funding gaps between decentralised and centralised venues are genuinely wider than gaps between two centralised exchanges. That is not a market failure - it happens because the two venue types compute prices by different mechanisms and attract different traders. Whether any given gap is tradeable is decided by a heavier cost stack: swap fees, gas charged per transaction, price impact against pool depth, withdrawal lag, and the possibility of being front-run.
There Are Two Different DEX/CEX Trades
Most guides pick one and call it "DeFi arbitrage". They are separate strategies with separate economics.
| Spot price gap (pool vs order book) | Perp funding divergence (on-chain vs centralised) | |
|---|---|---|
| What diverges | The asset price on each venue | The funding rate on each venue |
| P&L source | One-time gap capture | Recurring funding payments |
| Does anything move on-chain? | Yes - swaps, often transfers | No - only margin sits on each venue |
| Dominant costs | Gas, swap fee, price impact | Taker fees, funding sign risk |
| Competitive pressure | MEV searchers and bots | Slower - structural, not latency-driven |
| Realistic for manual traders? | Rarely on expensive networks | Yes, with discipline |
Why the Prices Differ at All
A centralised exchange price is the top of an order book: market makers post and cancel quotes continuously, so the price updates even when nobody trades.
An automated market maker computes price algorithmically from pool reserves. In the classic constant-product design, reserves of the two tokens satisfy:
The consequence is that an AMM price only moves when someone trades against the pool, and every trade moves it by an amount proportional to the trade size relative to pool depth. That is price impact, and it is not a fee - it is the mechanism itself.
Two practical implications:
- On deep, heavily traded assets, the two surfaces stay within fractions of a percent. Not enough to trade.
- On thin or newly listed assets, gaps of several percent appear regularly - and those are exactly the pools where your own trade moves the price against you.
A common rule of thumb among on-chain traders is to keep position size to a low single-digit percentage of the pool's total liquidity. Above that, you are trading against your own price impact.
Spot CEX↔DEX: Netting Out a 2% Gap
The gap looks generous. Here is where it goes on a $5,000 trade - illustrative arithmetic, one round trip:
Gross gap (2.00%): +$100.00
CEX taker fee (0.10%): -$5.00
DEX swap fee (0.30%): -$15.00
Price impact / slippage (~0.50%): -$25.00
Gas (two on-chain transactions): -$30.00
CEX withdrawal + network fee: -$12.00
-----------------------------------------------
Net result: +$13.00
Net ROI: 0.26%
A 2% headline gap becomes 0.26%. Change two assumptions and it inverts: on a congested expensive network, gas alone can exceed $100 for a round trip, and a thinner pool can double the price impact. The same trade is then meaningfully negative.
Warning
Gas is a fixed cost, and that is the whole story for small accounts. $30 of gas is 3% of a $1,000 trade and 0.03% of a $100,000 trade. On-chain spot arbitrage does not scale down - it scales up. This is the single biggest reason retail on-chain activity migrated to cheaper networks.
Three more costs that are easy to underestimate:
Withdrawal lag. If the trade requires moving the asset from a centralised venue to a wallet, you are exposed for as long as the withdrawal takes to clear - and the gap you saw was priced before that delay. Traders who do this seriously keep inventory pre-positioned on both sides rather than moving assets per trade. The same logic applies to centralised arbitrage, which we cover in best exchanges for arbitrage.
Front-running. Your intended swap is visible in the public transaction queue before it settles. Searchers monitor that queue and can execute the same trade ahead of you. On competitive pairs this is not an edge case, it is the default outcome for naive execution.
Network mismatch. The asset on the centralised venue must be on the same network as the pool you intend to trade. Getting this wrong means either a bridge - with its own risk and delay - or a lost transfer.
Perp Funding Divergence: The More Accessible Trade
The second DEX/CEX trade never touches a token transfer. You hold opposing perpetual positions on an on-chain venue and a centralised venue and collect the funding difference. If funding mechanics are new to you, start with funding rates explained.
Two things make this different from the spot version.
First, normalise the settlement clocks before comparing anything. Perpetual venues settle funding on different intervals - hourly on some on-chain venues, typically every 8 hours on centralised ones. A printed rate is per interval, so identical-looking numbers can differ by a large multiple in daily terms. Convert both to a common window first; comparing raw on-screen rates is the most common error in cross-venue funding trades.
Second, the divergence is structural, not a latency race. On-chain perpetual venues are accessed from self-custodied wallets and skew towards retail participants with a long bias, while centralised venues carry more professional flow that continuously suppresses the premium. BitMEX's research team describes precisely this asymmetry in their 2026 derivatives report. The practical effect: on-chain funding tends to sit persistently above centralised funding instead of converging within hours the way two centralised venues do.
Persistent is not the same as permanent. Before entering, the same checks apply as for any funding trade - does the rate keep its sign across cycles, how many intervals do entry and exit costs take to recover, and how far can the position move against you first. That checklist is in how to evaluate a funding rate arbitrage trade, and the funding rate calculator does the interval arithmetic.
Compare Funding Across Venues, Normalised
Cross-exchange funding rates and spread statistics across 13+ venues, with settlement intervals accounted for.
Free preview, no signup required
Risks On-Chain Adds
These are additional to the usual arbitrage risks, not instead of them.
| Risk | Why it matters here |
|---|---|
| Smart contract failure | Your capital interacts with code; a bug or exploit is a total loss on that leg |
| Oracle / index construction | On-chain perps price off an oracle. If it lags or is manipulated, your hedge is not what you think |
| Gas spikes | Network costs rise precisely during the volatility that creates opportunities |
| Front-running | Public transaction visibility means someone can act on your intent first |
| Bridge exposure | Cross-network movement adds delay and a separate failure surface |
| Self-custody errors | Wrong network or wrong address is unrecoverable; there is no support desk |
| Liquidation mechanics | On-chain liquidation rules differ from centralised ones - read them before sizing leverage |
Tip
A useful filter: if you cannot name the oracle a venue uses to price your position, you are not hedged - you are guessing that two price sources agree.
What Is Actually Tradeable Without Infrastructure
Honest assessment for someone trading manually with modest capital:
- Spot gaps on expensive networks: usually no. Fixed gas plus competitive searchers means the surviving margin is thin and often negative at small size.
- Spot gaps on cheap networks: sometimes. Lower gas restores the maths, but price impact and pre-positioning requirements still apply, and the assets with the widest gaps are the thinnest pools.
- Funding divergence: yes, most accessibly. Nothing moves on-chain, opportunities last hours to days rather than seconds, and the edge is structural rather than latency-based. This is why it is the on-chain trade we would point a manual trader to first.
- MEV-style atomic strategies: no. Those require contract development and private transaction routing. Different discipline, not a scaled-down version of this one.
If you are still choosing between market-neutral approaches generally, spot vs funding arbitrage compares the two simpler options, and is crypto arbitrage profitable sets expectations on realistic returns.
When to Skip the Trade
- Gas exceeds a meaningful share of the expected capture. Set a hard threshold as a percentage of expected profit and respect it.
- The pool is too thin for your size. If your trade would move the price by a large fraction of the gap, the gap is not really there.
- The asset is not on the same network on both sides. A bridge converts an arbitrage into a directional bet for the duration of the transfer.
- The gap is enormous. Very large on-chain gaps usually indicate something broken: disabled withdrawals, a stale price, a honeypot token, or liquidity that will not let you exit.
Information
What we do and do not cover. We track spot and perpetual markets across centralised venues and perp DEXs, including funding rates and spread statistics. We do not quote AMM pool prices, model gas, or execute on-chain transactions. For pool-level swap analysis you will need on-chain tooling in addition to what we provide.
Frequently Asked Questions
Why are DEX and CEX prices different in the first place?
They are produced by different mechanisms. A centralised exchange price comes from an order book where market makers update quotes continuously. An automated market maker computes price algorithmically from pool reserves, so it only moves when someone trades against the pool. Add different participants on each side and the two surfaces drift apart routinely, especially on smaller assets.
Is DEX arbitrage more profitable than CEX arbitrage?
The gross gaps are usually wider on-chain, particularly on low-liquidity assets, but the cost stack is heavier: swap fees, gas charged as a fixed amount per transaction, price impact against pool depth, and the risk of being front-run. Wider gaps do not automatically mean better net returns.
Why is gas such a problem for small DEX arbitrage trades?
Gas is charged per transaction, not as a percentage of your trade. The same $30 of gas is 3% of a $1,000 trade and 0.03% of a $100,000 trade. That makes small on-chain arbitrage structurally unprofitable on expensive networks, which is why most retail on-chain activity moved to cheaper layer-2 networks.
Why are funding rates usually higher on perp DEXs than on centralised exchanges?
The participant mix differs. On-chain perpetual venues are accessed from self-custodied wallets and skew towards retail traders with a structural long bias, while centralised venues carry more professional flow that continuously arbitrages away the premium. On-chain funding therefore tends to sit persistently above centralised funding rather than converging quickly.
Do I need a bot or smart contracts for DEX arbitrage?
For spot price gaps on-chain, competitive execution effectively requires automation and private transaction routing, because searchers monitor the public transaction queue. Funding divergence between a perp DEX and a centralised venue is far more accessible manually, since nothing has to move on-chain except your margin. We compare tooling approaches in bots vs scanners vs manual trading.
What risks does on-chain arbitrage add that centralised arbitrage does not have?
Smart contract and oracle risk, bridge risk when an asset must cross networks, gas spikes that arrive exactly when volatility creates the opportunity, front-running by searchers, and self-custody mistakes such as sending to the wrong network - with no support desk to reverse an error.
Summary
- DEX/CEX gaps are wider for structural reasons: algorithmic pool pricing versus order books, and different participant mixes on each side.
- The spot version carries a heavier cost stack, and gas is a fixed cost - which is why it does not work at small size on expensive networks.
- A 2% headline gap can net well under 0.3%, or go negative when gas or price impact rises.
- The funding divergence version is the more accessible on-chain trade: nothing moves on-chain, and the edge is structural rather than a latency race.
- Normalise settlement intervals before comparing funding rates across venue types, or you will compare numbers that are not the same units.
- On-chain adds contract, oracle, bridge and custody risk. Very large gaps are usually a warning, not an opportunity.
Track Centralised and On-Chain Perp Markets Together
Funding rates, cross-venue spreads and the statistics behind them across 13+ venues. Start on the free tier.
No credit card required


