Wind-to-Data Center Co-Location Sizing Calculator
Co-locating a data center directly with a wind farm is one of the most talked-about strategies in the AI infrastructure boom — but a 200 MW wind farm does not equal 200 MW of continuous data center load. Wind is variable; data centers are not. This calculator compares a wind farm's real annual energy output (nameplate × capacity factor × 8,760 hours) against a data center's annual energy demand, then shows exactly how much of the load the wind farm can cover on an annual basis, the gap that must come from the grid, storage, or backup generation, any surplus (curtailed) wind energy available, and the maximum continuous load the wind farm could fully match. Pair it with our Battery Arbitrage Revenue Calculator to value the storage that covers the gap, and our PPA vs. Ownership Calculator for the traditional wind procurement model.
The rated maximum output of the wind farm under ideal conditions — not the continuous output it actually delivers.
U.S. fleet average is ~33–36%; modern turbines in strong wind regions like Texas can reach 40–45%+.
The continuous megawatts the data center is designed to draw — a hyperscale AI campus, colocation build, or HPC facility.
How close to nameplate the load runs across the year. AI training clusters run near flat-out: 90–98% is typical.
nameplate MW × 8,760 × capacity factor
target load MW × 8,760 × load factor
Annual wind energy output ÷ annual data center energy demand
MAX(0, data center demand − wind energy output) — the shortfall wind cannot cover
annual wind energy ÷ (8,760 × load factor) — the firm-equivalent capacity of the wind farm
Results update live as you type. For planning and field-check estimates — always verify against applicable standards and equipment ratings.
How we calculate this →Real projects aren't claiming 100% wind, and that's the point. Project Hedy pairs 120 MW of data center load with a 200 MW wind farm — but at a realistic 38% capacity factor, that wind farm can only continuously match about 80 MW of that load on an annual basis. The remaining demand has to come from the grid, storage, or backup generation. That's not a flaw in the model, it's the model: co-located data centers are designed to soak up wind when it's abundant, not to claim 100% renewable power every hour.
How wind-to-data center co-location sizing is calculated
The calculation compares a wind farm's real annual energy production against a data center's annual energy demand, then derives the coverage, the gap, any surplus, and the firm-equivalent load the wind farm can match. Four inputs drive everything.
Annual wind energy output (MWh/yr) = wind farm nameplate capacity (MW) × 8,760 hours × capacity factor (% ÷ 100). Nameplate capacity is the maximum output under ideal wind conditions, not what the farm actually delivers. The capacity factor converts that theoretical maximum into real annual production — a 200 MW wind farm at a 38% capacity factor produces 200 × 8,760 × 0.38 = 665,760 MWh per year, the same annual energy a ~76 MW generator running flat-out would produce. The U.S. wind fleet averaged 33.5% in 2023 and 35.9% in 2022 (an all-time high); strong-wind regions like Texas and modern, taller turbines can reach 40–45%+.
Annual data center energy demand (MWh/yr) = target data center load (MW) × 8,760 hours × load factor (% ÷ 100). AI training clusters run near flat-out, so load factors of 90–98% are typical — a 120 MW data center at 95% load factor consumes 120 × 8,760 × 0.95 = 998,640 MWh per year.
Wind energy coverage of annual load (%) = (annual wind energy output ÷ annual data center energy demand) × 100. At the defaults, 665,760 ÷ 998,640 = 66.7% — the wind farm covers about two-thirds of the data center's annual energy, and the rest must come from elsewhere. When coverage exceeds 100%, the wind farm generates more energy annually than the load needs, and the surplus becomes the relevant figure instead of the gap.
Energy gap requiring grid / storage / backup (MWh/yr) = MAX(0, annual data center energy demand − annual wind energy output). This is the shortfall the wind farm cannot cover on an annual basis — it must be served by utility grid power, battery storage, or firm backup generation such as natural gas or fuel cells. At the defaults, the gap is 332,880 MWh/yr. Surplus wind energy available (MWh/yr) = MAX(0, annual wind energy output − annual data center energy demand) is the mirror image: it is non-zero only when coverage exceeds 100%, and it represents the curtailed or stranded-energy opportunity a flexible co-located load can soak up. Only one of the two is ever non-zero at a time.
Maximum continuous data center load this wind farm could fully match (MW) = annual wind energy output ÷ (8,760 × load factor). This is the firm-equivalent capacity of the wind farm — the continuous megawatts it could supply if its annual energy were spread evenly across the year at the data center's load factor. At the defaults, 665,760 ÷ (8,760 × 0.95) = 80.0 MW, meaning a 200 MW / 38% wind farm can fully match about 80 MW of continuous load, well below its nameplate rating. That gap between nameplate and firm-equivalent capacity is the central economic reality of wind-to-data center co-location: the architecture around the wind resource — grid, storage, and firm generation — is what turns variable wind into reliable computing power.
Treat the result as an annual energy planning estimate. It compares total annual production against total annual demand and does not model hourly matching, transmission congestion, battery sizing, or the economics of curtailment — for those, pair it with the Battery Arbitrage Revenue Calculator and the PPA vs. Ownership Calculator.