Crypto withdrawal fees are set per network, not per exchange, and the gap between networks for the same asset can be enormous. Fiat withdrawals add bank and intermediary charges, and sometimes a minimum that makes small withdrawals uneconomic.
Withdrawal behaviour is also the clearest test of a venue. Fees are published; processing time, documentation requests and limits are not, which is why we time real withdrawals ourselves.
The mistake it causes: judging a venue on trading fees alone, then paying it all back on the way out.
Often the largest cost on a small account
Crypto withdrawal fees are set per network and charged flat per transaction, which means they scale inversely with size. Moving $50 can cost proportionally more than moving $5,000, and nobody compares exchanges on this number.
| Amount | Flat fee $1 | Flat fee $8 |
|---|
| $50 | 2.0% | 16.0% |
| $500 | 0.2% | 1.6% |
| $5,000 | 0.02% | 0.16% |
The top-right cell is the one that matters. A trader saving 0.05% on trading fees and paying 16% to get the money out has optimised the wrong number by a factor of several hundred.
Fiat withdrawals have their own layer
Bank transfers add correspondent bank charges that the venue does not control and does not disclose, plus a minimum that can make small withdrawals uneconomic. A broker stating "no withdrawal fees on our side" is being accurate and incomplete.
Speed is the part nobody publishes
Fees are on a page. Processing time, documentation requests and limits are not, and they are what you actually care about the first time you need your money. This is why we time real withdrawals from funded accounts instead of quoting policy: XM bank transfer in roughly ten minutes, XM and HFM crypto in two to three hours, each timed on a recorded date.
What to do with this
- Batch withdrawals rather than taking small amounts frequently.
- Compare networks before withdrawing; the same asset can differ by an order of magnitude.
- Test-withdraw a small amount from any new venue before trusting it with size.
