Exchange Fees Can Turn a Winning AI Into a Losing One
Fees decide more than most traders think. We tested one AI crypto portfolio, the BTC & ETH Cycle Trader, from 94 monthly start dates since 2018 at seven fee levels. At a 0.10% fee per trade it returned a median of 33.6% a year; at 0.90% it returned 17.1%, below the 24.1% from simply holding the same coins. It ties holding at a fee near 0.55%. Exchange fees range from 0% on Binance.US to 0.90% at Coinbase's US entry tier, so the exchange you pick can matter as much as the AI. Botsfolio reads your account's real fee, uses limit orders where they save money, and is tuned for the fees on your exchange.
Most bot comparisons argue about strategies and subscriptions. The fee your exchange charges on every trade rarely comes up. It should. A strategy that trades regularly pays that fee over and over, and the total can be the difference between beating the market and trailing it.
The test: one AI portfolio, seven fee levels
We took one portfolio, the BTC & ETH Cycle Trader, and ran it from every monthly start date between January 2018 and the present, each held to October 2026. Then we recomputed every run at different fees per trade, from almost nothing to the highest common retail rate:
The thin line marks holding BTC and ETH from the same start dates: 24.1% a year. Green bars beat it, red bars trail it.
The same trades, the same coins, the same start dates. Only the fee changed. Three numbers sum it up:
- At 0.10% per trade the portfolio returned a median of 33.6% a year and beat holding from all 94 start dates.
- Near 0.55% it ties holding.
- At 0.90% it returned 17.1% a year, about 7 points a year behind simply holding.
In the same test it beat holding from every start date at fees up to about 0.27%, and from 90% of them up to about 0.39%.
What exchanges actually charge
Most exchanges charge two rates. A maker order waits in the order book and is usually cheaper. A taker order fills right away against someone else's order and costs more. Your real cost per trade lands somewhere between the two. Here are the four exchanges Botsfolio AI can trade on:
| Exchange | How you connect | Trading fee (maker / taker) | Smallest order | Where |
|---|---|---|---|---|
| Coinbase | Sign in with Coinbase. No keys to copy. | US: 0.50% / 0.90% at entry, lower with volume. EU, UK, Canada: 0.25% / 0.50%. Most other countries: 0.09% / 0.10%. | $1 | Most countries. Trades in USD in the US, USDC elsewhere. |
| Kraken | API key that can trade, with withdrawals off. | 0.40% / 0.80% at entry, lower with 30-day volume or assets held. | $5 | Not in India and a few other countries; in the US, not Maine or New York. |
| Binance | API key that can trade, with withdrawals off. | 0.10% / 0.10% (25% less when paid in BNB). | $5 | Not in the US, Canada, the Netherlands, the UK, Singapore or Japan. |
| Binance.US | API key that can trade, with withdrawals off. | 0% / 0.02%. | $10 | US only, and not every state. |
Rates also fall as you trade more. On Coinbase in the US, the entry tier is 0.50% maker and 0.90% taker; from $10,000 of trading in 30 days it drops to 0.35% and 0.75%, and from $50,000 to 0.15% and 0.30%. Kraken counts either your 30-day volume or the assets you hold there. Binance.US charges 0% maker and 0.02% taker with no volume needed.
Fee tiers are set by volume, so small accounts usually sit at the entry tier. That is exactly where the gap between exchanges is widest. Someone trading $2,000 on Coinbase in the US pays several times what the same trades cost on Binance.US or on Coinbase outside the US.
How to pay less
- Pick the exchange with fees in mind. Where it's available to you, a low-fee exchange is the biggest single saving.
- Use limit orders where the gap is large. On exchanges where maker and taker rates differ a lot, waiting in the book saves real money.
- Trade less when fees are high. A strategy that suits a 0.10% fee may need to trade less often at 0.75%.
- Watch your tier. Volume or assets held can move you to a cheaper tier.
Botsfolio does the parts it can for you. When you connect, it reads your account's own fee tier. It uses limit orders where they save money, and it is tuned for the fees on your exchange. Before you start, it shows the estimated monthly cost. Our own fee comes on top of the exchange's and is small: here is what it comes to by account size.
| Amount the AI trades | Quiet month | Typical month | Busy month | Typical year |
|---|---|---|---|---|
| $1,000 | $0.44 | $1.52 | $2.32 | $18.24 (1.8%) |
| $5,000 | $2.20 | $7.59 | $11.61 | $91.08 (1.8%) |
| $10,000 | $4.40 | $15.18 | $23.22 | $182 (1.8%) |
| $50,000 | $22.00 | $55.52 | $79.65 | $666 (1.3%) |
The full cost picture, including bot subscriptions, is in how much it costs to let AI trade crypto.
A backtest on past prices with the fee charged on every trade. Results are medians across 94 monthly start dates since January 2018, each held to October 2026. It shows how fees change the outcome of the same trades, not what any portfolio will return in future.
FAQ
Watch Botsfolio AI trade live
Botsfolio AI paper trades on the live market, and every entry, stop and exit is public.
- BTC & ETH Cycle Trader+24.2% vs -13.8% buy & hold in 2026 · biggest drop 8.7% vs 46.5%
- BTC & ETH Trend Rider+3.9% vs -13.8% buy & hold in 2026 · biggest drop 6.0% vs 46.5%
- BTC & ETH Steady Trader+2.6% vs -13.8% buy & hold in 2026 · biggest drop 3.2% vs 46.5%
- Bitcoin Cycle Trader+3.2% vs -8.2% buy & hold in 2026 · biggest drop 4.8% vs 39.6%
- Crypto Cycle Trader+27.9% vs +8.7% buy & hold in 2026 · biggest drop 12.4% vs 45.5%
- Altcoin Trend Rider+31.7% vs +12.9% buy & hold in 2026 · biggest drop 13.4% vs 50.2%
- Memecoin Trend Rider-6.3% vs -36.6% buy & hold in 2026 · biggest drop 10.5% vs 63.2%
How to read these numbers: the buy and hold comparison, biggest drop, and how the portfolios are tested.
See everything Botsfolio AI trades live
Paper trading on real prices. Not investment advice.
Educational analysis, not financial advice. Fee figures as published by each exchange in October 2026 and subject to change. Backtests are tests on past prices. Trading crypto can lose money. Past performance does not predict future results.