A full-size air-cooled ASIC miner draws about 3 to 4 kW and runs at that level around the clock. A 3,500 W machine uses 84 kWh per day, 2,520 kWh in a 30-day month and 30,660 kWh in a year. Hydro-cooled machines draw roughly 5 to 11 kW each, and small home units draw a fraction of a full-size machine. Because a miner never idles, its energy use is simply its power multiplied by the hours it runs.
Key points
- kWh per day = power (kW) × 24. kWh per year = power (kW) × 8,760.
- Published power is measured at the wall and is a typical value. Allow for the published tolerance, commonly ±5% to ±10%.
- A site uses more than the sum of its miners. Fans, pumps and transformer losses add an overhead.
- All of the electricity becomes heat: 1 kW ≈ 3,412 BTU/h.
- Measure real consumption with a metered PDU or an energy meter, not from the specification sheet.
The calculation
energy (kWh) = power (kW) × hours
per day = kW × 24
per 30 days = kW × 720
per year = kW × 8,760
To convert watts to kilowatts, divide by 1,000. Multiply the result by your price per kWh for the cost: 84 kWh at an illustrative $0.06 per kWh is $5.04 per day. Electricity Cost and Bitcoin Mining has full cost tables.
Energy use by size of machine
These machines are illustrative round-number examples, not specific models.
| Illustrative machine | Power | kWh per day | kWh per 30 days | kWh per year | Heat output |
|---|---|---|---|---|---|
| Small home unit | 100 W | 2.4 | 72 | 876 | 341 BTU/h |
| Heater-style home unit | 800 W | 19.2 | 576 | 7,008 | 2,730 BTU/h |
| Full-size air-cooled | 3,500 W | 84 | 2,520 | 30,660 | 11,942 BTU/h |
| Hydro-cooled | 5,500 W | 132 | 3,960 | 48,180 | 18,766 BTU/h |
| Large hydro-cooled | 8,500 W | 204 | 6,120 | 74,460 | 29,002 BTU/h |
| Rack-format hydro | 11,000 W | 264 | 7,920 | 96,360 | 37,532 BTU/h |
Manufacturer-published examples
The hashrate and power below are manufacturer-published figures. The energy columns are calculated from the published power, assuming continuous operation.
| Model | Cooling | Published hashrate | Published power | kWh per day | kWh per year |
|---|---|---|---|---|---|
| Antminer S19k Pro | Air | 120 TH/s | 2,760 W | 66.24 | 24,178 |
| Antminer S21 | Air | 200 TH/s | 3,500 W | 84.00 | 30,660 |
| WhatsMiner M60S+ (200T) | Air | 200 TH/s | 3,400 W | 81.60 | 29,784 |
| Antminer S23 | Air | 305 TH/s | 3,355 W | 80.52 | 29,390 |
| SEALMINER A3 Pro Air | Air | 290 TH/s | 3,625 W | 87.00 | 31,755 |
| Antminer S23 Hyd. | Hydro | 580 TH/s | 5,510 W | 132.24 | 48,268 |
| WhatsMiner M63S++ (464T) | Hydro | 464 TH/s | 7,192 W | 172.61 | 63,002 |
| SEALMINER A4 Ultra Hydro | Hydro | 886 TH/s | 8,372.7 W | 200.94 | 73,345 |
| Antminer U3S23H | Hydro, rack format | 1,160 TH/s | 11,020 W | 264.48 | 96,535 |
Two things stand out. The air-cooled machines all sit in a narrow band of roughly 2.8 to 3.6 kW regardless of generation, because a single-phase circuit and air cooling limit how much power one unit can take. Newer generations use that power to produce more hashrate: the Antminer S23 delivers 305 TH/s on 3,355 W, while the Antminer S19k Pro delivers 120 TH/s on 2,760 W. And hydro machines draw more per unit because each one carries far more hashrate. Efficiency, not total power, is what to compare. See Understanding J/TH.
Rated power, wall power and tolerance
Wall power is what the machine draws from the supply, including the conversion loss in its power supply, the control board and the fans. This is the figure your meter records, and it is the figure manufacturers publish. MicroBT labels it power on wall.
Reference conditions. Published power is a typical value at a stated temperature. Bitmain states the Antminer S21 figure at 25 °C inlet air and the Antminer S23 Hyd. figure at 35 °C inlet coolant.
Tolerance. Bitmain publishes ±5% on power for the Antminer S21, so a unit within specification draws between 3,325 W and 3,675 W. That is 79.8 to 88.2 kWh per day. MicroBT's datasheet for the WhatsMiner M60S+ gives power on wall as 3,230 W to 3,570 W, ±10%, as a range for the model. Budget energy and size circuits for the top of the range.
Power supply rating. The maximum output printed on a power supply is its capacity, not what the miner draws. Do not use it to estimate energy.
Actual draw also moves with operation:
- Hotter intake air makes the fans work harder and raises consumption.
- Low-power and high-performance modes change power and hashrate together.
- A machine running with a failed hashboard draws less and hashes less.
- Third-party firmware and tuning can move power in either direction.
Cooling and site overhead
The miner's power is not the whole bill. Behind a site's meter there are also ventilation fans, pumps and dry coolers for hydro systems, transformer and cable losses, network equipment and lighting. The ratio of total site energy to miner energy is often called PUE (power usage effectiveness).
| Overhead (illustrative) | Site power for 1,000 kW of miners | Extra energy per day |
|---|---|---|
| 3% | 1,030 kW | 720 kWh |
| 5% | 1,050 kW | 1,200 kWh |
| 10% | 1,100 kW | 2,400 kWh |
These percentages are illustrative, not benchmarks for any cooling method. Two points apply generally:
- Hydro and immersion machines have no on-board fans. Their published power excludes the pumps and heat-rejection equipment the site must run, so compare cooling methods on total site power.
- Compressor-based air conditioning is a costly way to cool miners, because it must remove every kilowatt the machines put into the room. Sites normally reject heat with outside air or with dry coolers.
At home, the overhead is whatever you add to move the heat out: usually an exhaust fan of modest power.
Heat output
Essentially all the electrical power a miner draws leaves it as heat. The conversion is 1 kW ≈ 3,412 BTU/h, so:
heat (BTU/h) = power (kW) × 3,412
A 3,500 W miner produces 3.5 × 3,412 = 11,942 BTU/h. One ton of cooling is 12,000 BTU/h, so each full-size air-cooled machine is close to one ton of heat load. Use this figure to size ventilation, or to judge how much space a miner can heat if you intend to reuse the warmth.
Compared with a familiar load
A common portable electric heater is rated at about 1,500 W. One 3,500 W miner is the equivalent of about 2.3 of them.
The comparison understates the difference in energy. A heater on a thermostat cycles on and off, and most appliances run for a few hours a day. A miner runs at full power for all 24. A 1,500 W heater would use 36 kWh if it ran for a full day without stopping; the miner uses 84 kWh every day.
How to measure actual draw
- Miner dashboard. The web interface reports hashrate and often a power estimate. Treat the power figure as an indication, not a meter reading.
- Metered PDU. A PDU with per-outlet or per-phase metering shows current and power for each machine or group while it runs.
- Energy meter on the circuit. A kWh meter fitted to the circuit by an electrician records consumption over time and is the most reliable basis for cost calculations and hosting bills.
- Clamp meter. Clamped around one insulated conductor of the supply cable, it reads current. Multiply by the measured voltage for an approximate power: 14.6 A × 240 V ≈ 3,500 W at PF ≈ 1. Take readings only where conductors are accessible without opening panels or enclosures. Anything inside a panel is a job for a qualified electrician.
Plug-in watt meters made for standard 120 V, 15 A household outlets are not rated for a 240 V miner circuit. Do not use them.
Record energy over at least 24 hours and convert it to average power: kWh ÷ hours × 1,000 gives watts. Divide that by the average hashrate in TH/s that your pool reports for the same period. The result is the machine's real efficiency in J/TH, which you can compare with the published figure.