Solar-Powered Bitcoin Mining: Size the Load Before Buying Panels or Batteries
2026-10-16 10:15

Solar-powered Bitcoin mining starts with the ASICs' electrical load and intended operating hours. Calculate the total site load, compare it with local hourly solar output, and then size any grid or battery support needed to cover the gaps. For example, a miner drawing 3.645 kW uses 87.48 kWh over 24 hours before external cooling and other site loads are added.

Bitcoin mining hardware draws a relatively steady, high-power load while operating, while solar photovoltaic (PV) generation varies by hour, weather, season, and array orientation. Panel nameplate capacity or an annual kilowatt-hour total alone cannot establish whether the system can run the intended miners. Hourly modeling provides an initial energy assessment; stable electrical supply and, where applicable, sufficient battery recharge must also be checked.

Define the ASIC load

The first step is identifying which ASIC model will run, in which power mode, and what else draws electricity at the site. Manufacturer specification sheets state a "power on wall" figure, which is the miner's typical electrical draw under stated test conditions, separate from its hashrate and its J/TH efficiency rating.

As an air-cooled example, BITMAIN's ANTMINER S21 XP is specified at a typical 270 TH/s hashrate, 3,645 W power on wall, and 13.5 J/TH under the manufacturer's stated conditions. Over 24 hours at that power draw, the miner alone consumes 3.645 kW × 24 h = 87.48 kWh. Over a six-hour scheduled run, the same miner uses 3.645 kW × 6 h = 21.87 kWh. These are ASIC-only figures: they include the unit's own fans and power-supply losses, but exclude external ventilation, networking equipment, and losses elsewhere in the site's power system. Do not add the ASIC's internal power-supply losses again to its power-on-wall figure. See the S21 XP Product Manual.

Hydro-cooled ASICs add further load outside the miner itself. BITMAIN's S23 Hyd. is specified at a typical 580 TH/s and 5,510 W power on wall at a 35°C inlet coolant temperature, giving roughly 9.5 J/TH under those stated conditions. The unit requires a 380–415 V three-phase electrical supply with a maximum input current of 12 A, plus a coolant loop with its own flow and temperature requirements. Circulation pumps and dry coolers draw power separately from the ASIC's own rated figure and must be added to the site load unless the manufacturer's specification already states otherwise. See the S23 Hyd. Product Manual.

A practical site-load estimate for planning purposes adds ASIC power, cooling-system power, networking and monitoring equipment, and any other site loads such as lighting or controls. This is a planning sum, not a named industry metric, and its accuracy depends on using the manufacturer's stated test conditions and on including every piece of equipment the miners actually require to run.

Map generation by hour

Power and energy are not interchangeable: power, in kilowatts, describes what is being drawn at a given moment, while energy, in kilowatt-hours, describes consumption over a period of time. Solar generation should be modeled the same way the load is modeled, hour by hour, rather than as a single annual total.

Tools such as PVWatts estimate PV output using location-specific solar-resource data and can produce hourly or monthly results rather than only an annual figure. PVWatts' default system-loss input is 14%, with adjustable categories such as soiling, shading, wiring, and equipment age; this default should not be treated as a universal real-world loss figure for every installation. Inverter conversion efficiency is modeled separately from that loss input. See the PVWatts output documentation and system-loss guidance.

Compare generation and demand at the same electrical boundary. When using PVWatts' AC output, do not deduct the modeled inverter or system losses again. Add only losses between that output boundary and the loads that have not already been included. Battery charging and discharging losses should likewise be counted once along their respective paths.

An array that produces sufficient energy over a full year can still generate little or no power for several hours after sunset or during extended cloud cover. Comparing hourly solar output against hourly load identifies likely energy surpluses and shortfalls, but an hourly average can hide shorter power drops. A system intended to operate without the grid must also maintain the voltage and frequency the miners require and respond to rapid changes in generation. Conventional grid-following inverters depend on an external grid reference; independent operation requires equipment designed for that purpose. See the U.S. Department of Energy's inverter guidance.

Grid-tied vs battery scenarios

Once the hourly generation profile is mapped against the load, three broad configurations become relevant, and each shifts cost and risk differently.

A grid-tied design uses solar generation when it is available and draws electricity from the grid to cover any shortfall, effectively purchasing power for the hours solar cannot supply. This avoids battery capital cost but means part of the operation's total electricity cost remains tied to grid tariffs and their own seasonal or time-of-use variation.

A battery-backed design stores surplus solar energy and discharges it during low-generation periods. It can support continuous mining without a grid connection only if generation, storage, and the independent power system can meet demand throughout the modeled conditions. Two separate battery ratings matter: energy capacity in kWh and continuous discharge power in kW.

For an initial estimate of a single discharge period:

Required nominal battery capacity (kWh) ≈ AC energy the battery must supply (kWh) ÷ (usable fraction of nominal capacity × discharge-path efficiency)

The usable fraction is the planned operating window between the upper and lower state-of-charge limits. The calculation assumes the battery begins at the upper limit and uses efficiencies expressed as decimals. If a supplier quotes usable capacity for that same window, do not apply the usable fraction again. Confirm whether its quoted energy is measured at the battery or AC output. This is a preliminary estimate, before allowances for temperature, aging, and reserve capacity. See SAM's battery sizing guidance.

A simple overnight example

Assume one S21 XP runs at a constant 3.645 kW, with no external loads included, and the battery must supply all power for 12 hours. For illustration, assume an 80% usable capacity window and 90% efficiency from stored battery energy to the AC load:

  • Overnight AC energy: 3.645 kW × 12 h = 43.74 kWh.
  • Nominal battery capacity: 43.74 ÷ (0.80 × 0.90) = 60.75 kWh.
  • Required continuous AC output: at least 3.645 kW for the miner alone.

That only addresses one discharge period. If charging from the AC side into stored battery energy is 95% efficient, replacing the energy used overnight takes 43.74 ÷ (0.90 × 0.95) ≈ 51.16 kWh of AC charging energy. Running the miner for the other 12 hours also takes 43.74 kWh, so the simplified daily requirement is about 94.90 kWh of solar AC output delivered at usable times.

These efficiencies and the 12-hour split are illustrative assumptions, not equipment specifications or a solar forecast. The array must serve the miner while supplying the charging power, within the battery's charge limits. Actual site loads, losses outside the stated boundaries, operating margins, and weather can raise the requirement. A location-specific hourly model is still needed to select array capacity and panel count.

A hybrid design combines batteries with grid or generator support for longer shortfalls. Choosing among these configurations depends on the site's solar profile, purchased electricity costs, battery and inverter cost, and tolerance for scheduled downtime.

Model intermittent operation

For any hour where solar generation falls short of the load, the shortfall must be supplied by battery discharge, grid power, a generator, or by reducing the mining load through scheduled shutdown. Model successive hours with the battery's state of charge carried forward, accounting for charging and discharging losses and power limits. Do not assume it starts full every evening: daytime surplus must be sufficient to recharge it while also serving the operating miners. SAM's battery results illustrate the separate tracking of stored charge, charging energy, and discharge output.

Use a full year of hourly generation and load estimates to capture seasonal differences, and examine extended low-generation periods if continuous operation is the goal. Record unmet load or curtailed mining rather than treating a depleted battery as an available source. For operation without grid or battery support, shorter-term fluctuations also require compatible power controls and shutdown or restart behavior; an hourly energy balance alone cannot demonstrate uninterrupted operation.

Temperature also affects the ASIC side of this model. BITMAIN's S21 XP performance curve shows listed hashrate declining from 270 TH/s to 252 TH/s, and listed efficiency rising from 13.5 J/TH to 15.3 J/TH, as ambient temperature rises from the manufacturer's baseline condition to 45°C. Sites with limited cooling capacity or high ambient temperatures during peak solar hours should use the manufacturer's stated performance at the expected operating temperature rather than its baseline specification. See the S21 XP performance curves.

Because BTC mining earnings depend on network difficulty, the miner's realized hashrate, and transaction-fee conditions, the energy model and the earnings model should be kept separate. Use explicit assumptions for these changing inputs rather than extending a current network snapshot into a long-term forecast. BTC price affects the fiat value of earnings, not BTC output by itself.

After installation, comparing ASIC dashboard readings, measured site power, pool-reported worker status and estimated hashrate, and credited BTC earnings helps confirm that the intermittent-operation model matches field results. ViaBTC's BTC Mining documentation describes mining setup and checking operation status and earnings through the Workers and Earnings pages. Its rejection-rate guidance can help interpret rejected-share issues. These pool-side figures remain distinct from local ASIC telemetry and from meter readings at the site.

Compare total project cost

A solar mining project's capital cost typically includes the ASICs, the PV array and its mounting and wiring, inverters, any battery system, switchgear sized for the site's peak power and voltage requirements, and in hydro-cooled designs, the circulation and heat-rejection equipment. Operating costs include any grid or generator electricity purchased to cover shortfalls, maintenance, and pool fees associated with the chosen payout method.

Compare configurations over the same measurement period and against the same target operating hours, using consistent cooling and temperature assumptions. Report the operating hours each design can actually deliver. If one configuration requires more shutdowns, show those lost mining hours alongside its lower costs; when hashrate also differs, compare realized hashrate over operating time. For multi-year comparisons, include explicit assumptions for battery capacity degradation and any replacement costs within the evaluation period. SAM's battery results documentation distinguishes calendar and cycling degradation.

A grid-tied design with no batteries generally has lower capital cost but ongoing exposure to grid electricity prices for the hours solar cannot cover. A battery-backed design raises capital cost in exchange for reduced reliance on purchased electricity, with the actual reduction depending on the site's solar profile, charging opportunities, and the battery's usable energy and discharge power.

Because BTC price, network difficulty, and transaction-fee levels change over time, and because solar generation varies by season and weather, a total-cost comparison built on a single day's figures should be treated as a scenario rather than a fixed outcome. The electrical load and generation model provides the inputs a cost comparison needs; it does not by itself determine whether a given configuration will be profitable at a future date.

FAQ

How many solar panels are needed to run one ASIC miner?

This depends on the miner's power-on-wall draw, external site loads, intended operating hours, and local solar output. An S21 XP drawing 3.645 kW uses 87.48 kWh over 24 hours before external loads and storage losses. Use the site's hourly generation profile to size an array that serves the load and any battery recharge; then convert the selected array capacity into a panel count using the chosen panel's rating.

Can Bitcoin mining run on solar power alone without batteries or a grid connection?

It can in a system designed for independent operation, while solar output is sufficient and stable enough to support the load. This requires compatible power-conversion equipment and controls that maintain the required voltage and frequency and handle generation drops, including stopping the miners when necessary. Ordinary grid-following solar equipment and sufficient hourly energy alone do not establish that capability. See the DOE's inverter guidance.

Does a higher-efficiency ASIC reduce how much solar capacity is needed?

A lower J/TH figure reduces power draw for a given hashrate under the manufacturer's stated conditions, which lowers the load to be sized for, but it does not remove the need to model hourly generation against that load, including any separate cooling or networking equipment.

Is Bitcoin mining electricity cost the same whether power comes from solar, grid, or a generator?

No; each source has a different cost basis—panel and battery capital cost versus purchased grid tariffs versus fuel and generator costs—so total electricity cost depends on the mix of sources used to cover the load across all operating hours, not on any single source in isolation.

References