Crypto market times: open all day, busy only part of it
Crypto exchanges never close, so there is no opening bell. But the people and firms who trade are spread across time zones, and they are not awake at the same time. Liquidity, the money waiting in order books, follows them. That makes some hours cheaper to trade than others, especially for large orders. We looked at two things: when trading volume is highest (over nine years of hourly data from Binance) and when a market order actually costs the most (from CostDepth's own order-book samples).
All times below are UTC. New York is 5 hours behind UTC in winter (4 in summer), London 0 (1 in summer), Singapore 8 hours ahead.
Volume by hour of the day
Average quote volume of Binance BTCUSDT spot in each hour, as a share of the daily total. 1-hour candles to 10 Nov 2026. The three busiest hours are highlighted.
Show the numbers
| Hour (UTC) | Last 12 months | Since 2017 |
|---|---|---|
| 00:00 | 3.84% | 4.16% |
| 01:00 | 3.82% | 3.81% |
| 02:00 | 3.67% | 3.58% |
| 03:00 | 3.35% | 3.31% |
| 04:00 | 3.14% | 3.15% |
| 05:00 | 3.26% | 3.15% |
| 06:00 | 3.34% | 3.38% |
| 07:00 | 4.41% | 3.86% |
| 08:00 | 4.21% | 4.03% |
| 09:00 | 3.92% | 3.85% |
| 10:00 | 3.15% | 3.79% |
| 11:00 | 3.21% | 3.79% |
| 12:00 | 4.33% | 4.71% |
| 13:00 | 5.91% | 5.57% |
| 14:00 | 7.44% | 6.55% |
| 15:00 | 7.53% | 6.29% |
| 16:00 | 5.83% | 5.63% |
| 17:00 | 5.02% | 4.69% |
| 18:00 | 4.24% | 4.39% |
| 19:00 | 3.7% | 4.26% |
| 20:00 | 3.3% | 4.13% |
| 21:00 | 2.94% | 3.28% |
| 22:00 | 3.27% | 3.31% |
| 23:00 | 3.19% | 3.34% |
The pattern is clear. Volume builds from about 12:00 UTC and peaks at 15:00 UTC, in the hours after the US stock market opens (14:30 UTC in winter, 13:30 in summer). The 3 busiest hours, 13:00, 14:00, 15:00, carried 20.9% of all daily volume between them. The quietest hours were 04:00, 10:00, 21:00 UTC, with about 3.1% of the day's volume each. A smaller bump appears around 07:00 to 08:00 UTC, when Europe starts work. Over the full history since 2017 the shape is the same, a little flatter: the busiest hour then was 14:00.
| Day | Share of weekly volume |
|---|---|
| Monday | 16.3% |
| Tuesday | 15.5% |
| Wednesday | 15.9% |
| Thursday | 16.9% |
| Friday | 18.2% |
| Saturday | 8.2% |
| Sunday | 9% |
Weekends are much quieter. Saturday and Sunday together carried 17.1% of the week's volume over the last year, about half of what two weekdays carry. Traditional markets are shut, many professional desks are thinner, and books follow.
Price swings by hour
Average (high - low) / open of Binance BTCUSDT 1-hour candles, 17 Aug 2017 to 10 Nov 2026.
Show the numbers
| Hour (UTC) | Average range |
|---|---|
| 00:00 | 1.056% |
| 01:00 | 0.947% |
| 02:00 | 0.875% |
| 03:00 | 0.799% |
| 04:00 | 0.801% |
| 05:00 | 0.787% |
| 06:00 | 0.798% |
| 07:00 | 0.818% |
| 08:00 | 0.893% |
| 09:00 | 0.853% |
| 10:00 | 0.87% |
| 11:00 | 0.876% |
| 12:00 | 1.017% |
| 13:00 | 1.108% |
| 14:00 | 1.206% |
| 15:00 | 1.143% |
| 16:00 | 1.128% |
| 17:00 | 0.988% |
| 18:00 | 0.946% |
| 19:00 | 0.948% |
| 20:00 | 0.997% |
| 21:00 | 0.942% |
| 22:00 | 0.962% |
| 23:00 | 0.915% |
The busiest hours are also the most volatile. The average BTC candle moved 1.21% from low to high at 14:00 UTC and 0.79% at 05:00 UTC. More movement means a bigger chance the price moves between the moment you decide and the moment your order fills, which is a cost of its own even when the book is deep. US economic data is usually released at 13:30 UTC (in winter), which shows in the 13:00 to 15:00 hours.
What a market order cost, hour by hour
Spread + slippage relative to each exchange's own average (1.00 = average), averaged over 14 exchanges and 4 days, 6 Nov 2026, 00:00 UTC to 10 Nov 2026, 03:00 UTC. Highlighted: the costliest hours.
Show the numbers
| Hour (UTC) | $1k | $100k | $1M |
|---|---|---|---|
| 00:00 | 0.96 | 1.17 | 1.07 |
| 01:00 | 1.17 | 1.17 | 1.06 |
| 02:00 | 1.00 | 1.02 | 1.01 |
| 03:00 | 0.80 | 0.99 | 0.97 |
| 04:00 | 0.81 | 0.92 | 0.94 |
| 05:00 | 0.78 | 0.82 | 0.96 |
| 06:00 | 0.83 | 0.87 | 0.96 |
| 07:00 | 0.84 | 0.88 | 0.95 |
| 08:00 | 0.93 | 0.92 | 0.93 |
| 09:00 | 0.81 | 0.91 | 0.94 |
| 10:00 | 0.76 | 0.82 | 0.93 |
| 11:00 | 0.81 | 0.78 | 0.91 |
| 12:00 | 0.90 | 0.93 | 0.93 |
| 13:00 | 0.85 | 0.90 | 0.97 |
| 14:00 | 1.06 | 1.04 | 1.03 |
| 15:00 | 1.23 | 1.02 | 1.04 |
| 16:00 | 0.92 | 0.95 | 0.97 |
| 17:00 | 0.87 | 0.91 | 0.97 |
| 18:00 | 0.80 | 0.83 | 0.94 |
| 19:00 | 0.88 | 0.82 | 0.97 |
| 20:00 | 0.87 | 0.97 | 0.99 |
| 21:00 | 2.79 | 1.72 | 1.19 |
| 22:00 | 1.28 | 1.27 | 1.25 |
| 23:00 | 1.03 | 1.24 | 1.09 |
In our 4 days of order-book data, the eight cheapest hours for a $100,000 BTC futures order were 05:00, 06:00, 07:00, 10:00, 11:00, 13:00, 18:00, 19:00 UTC. The Asian night (01:00 to 07:00 UTC) averaged 0.95 and the US session (13:00 to 19:00 UTC) 0.92, against 1.00 for an average hour. So the busy, fast US hours were not the expensive ones for a market order: deep books made up for the faster prices.
The costliest hour was 21:00 UTC, at 1.72 times average. That figure comes mostly from one evening: on Sunday 8 Nov 2026 the cost at that hour was 4.24 times the exchange's average, while on the other days it was 0.62, 0.86, 1.15. Sunday evening UTC is when the weekend is at its thinnest, just before Asian markets and US futures reopen. Apart from that evening, the weekend was not expensive in our window: across all weekend hours a $1M order cost 0.90 times average, and weekday hours 1.10. Lower weekend volume did not, on its own, mean wider books.
For small orders the hour matters very little: a $1,000 order pays about half the spread, and on BTC that is one price tick at almost every hour. The differences above are for large orders. Four days is a short sample, and the picture will firm up as CostDepth's hourly history grows.
So when is the best time to trade crypto?
- For large orders, the busiest hours (weekday European and US hours, roughly 07:00 to 20:00 UTC) have the most volume, and in our data they were not more expensive to trade.
- Be careful late on Sunday (UTC). In our window that was the one time a large order cost several times its normal amount.
- Expect the biggest price swings around the US open and US data releases (13:00 to 16:00 UTC in winter). Deep books, fast prices.
- For small orders, the time of day barely changes the cost. The fee matters far more: see our cheapest exchange study.
- Holding a futures position over a funding time costs (or pays) funding, whatever the hour. See what holding a perp costs.
Every exchange page on CostDepth has an hour-by-hour chart of its own cost over the last 24 hours and 7 days (for example Binance). This is data about past hours, not a forecast or advice.
Method
Volume and price range: Binance BTCUSDT spot 1-hour candles from 17 Aug 2017 to 10 Nov 2026 (80,800 hours), from Binance's public market-data API. For each hour of the day (UTC) we took the average quote volume (in USDT) and the average of (high - low) / open. "Last 12 months" uses only the most recent 365 days.
Trading cost: CostDepth's stored hourly averages of the spread and slippage of BTC USDT perpetual market orders ($1k, $100k, $1M), 6 Nov 2026, 00:00 UTC to 10 Nov 2026, 03:00 UTC. To compare exchanges with very different books, we divided each exchange's hourly cost by its own average and then averaged across exchanges for each hour of the day: 1.00 is a normal hour, 1.20 is 20% more expensive. We left out exchanges whose $1M cost hardly moves at all (their books show very large resting orders at the best price, so their cost cannot change with the hour). Included: Binance, Bybit, OKX, Bitget, KuCoin, BingX, HTX, WEEX, Gate.io, BloFin, WhiteBIT, Hyperliquid, Lighter, Variational.
The cost data covers only 4 days, so each hour of the day has about four observations per exchange, one of them on a weekend. One unusual evening can move an hour's figure a lot, which is why we show the day-by-day values for the costliest hour. Volume is a far longer record (more than nine years), but volume is not cost: it shows when people trade, which is usually when books are deepest.
The numbers behind every chart and table are in data.json (free to reuse with a credit to CostDepth). This is fee and cost data, not financial advice.