Global data-center electricity demand doubles past 1,000 TWh as AI sites hit 750 MW each
Gartner estimates global data-center electricity demand will exceed 1,000 TWh in 2026 — double the 2023 baseline — with modern AI facilities drawing 100 to 750 megawatts per site. NVIDIA has pledged a $250 billion push behind OpenAI's infrastructure, and a single project could top $500 billion and 800 MW. The power story is now the whole story.
Doubling total data-center demand in three years is the number that reframes the buildout. A jump past 1,000 TWh means AI compute is now a material share of global electricity, and per-site draws of up to 750 megawatts put a single facility on the scale of a mid-sized city's consumption. The grid, not the chip, is the binding constraint.
The capital commitments confirm the scale is real. NVIDIA's $250 billion push behind OpenAI's infrastructure, and projects topping $500 billion delivering hundreds of megawatts, are bets that this demand persists long enough to justify power plants and substations that take years to build. The money is chasing megawatts, not just silicon.
The strategic consequence is that energy access is becoming the competitive moat. Whoever secures power-ready land and grid contracts holds the scarce asset, which is why frontier capital is flowing toward energy deals as hard as toward chips — and why the compute story of 2026 keeps returning to the same ceiling: you can buy the GPUs faster than you can power them.
TechPlusTrends — AI data center power requirements 2026: the grid-to-chip guide → · Al Jazeera — Nvidia plans $250bn push to bolster OpenAI's infrastructure ambitions →