Can Fusion Power AI? From NIF's Ignition to Tech Giants' Power-Purchase Bets

How did NIF break through fusion ignition, from 2022's first shot to 2025's gain record?

The National Ignition Facility (NIF) achieved fusion ignition for the first time on December 5, 2022, when a target produced 3.15 megajoules of fusion energy from 2.05 megajoules of laser energy delivered to it, a target gain of roughly 1.5CITE:E1. NIF has since repeated and improved on that result: on April 7, 2025, the facility recorded its highest target gain to date, 4.13, producing 8.6 megajoules of fusion energy from 2.08 megajoules of laser inputCITE:E2.

Does a target gain of 4.13 mean NIF is generating electricity?

No — NIF's gain figures measure fusion output against only the laser energy striking the target, not the total electricity a power plant would need to produce and deliverCITE:E6. Both the 2022 ignition shot and the 2025 record are science-facility milestones, not output from a power plant, since the gain calculation excludes the driver laser's own enormous electricity draw and does not account for a plant-wide energy balanceCITE:E1CITE:E2. The distance between a laser target achieving ignition and a facility that reliably feeds a power grid remains a large engineering gap rather than a solved problemCITE:E6.

Why has fusion become an investment hotspot, and how fast is capital moving?

The Fusion Industry Association (FIA) reports cumulative private and public investment in the fusion industry has reached about $9.77 billion across 53 companies, with roughly $2.64 billion of that added in the twelve months ending July 2025CITE:E3.

How are tech giants betting on fusion power for AI, and what did Microsoft and Google actually commit to?

Microsoft and Google have each signed power-purchase agreements with fusion startups to help meet future data-center electricity demandCITE:E4CITE:E5. Microsoft signed the world's first fusion power-purchase agreement with Helion Energy in May 2023, targeting grid delivery by 2028 with output ramping to at least 50 megawatts within a year of start, with Constellation acting as power marketerCITE:E4. Google's commitment is larger: in June 2025 it agreed to purchase 200 megawatts from Commonwealth Fusion Systems' (CFS) first ARC commercial plant in Virginia, a facility designed for 400 megawatts of net output, targeting grid connection in the early 2030s — the largest corporate fusion power-purchase agreement to dateCITE:E5.

Can the 2028–2030s targets actually be met?

As of 2026, no fusion device anywhere has delivered net electricity to a power grid, leaving Helion's 2028 target and CFS's early-2030s target as goals rather than completed projectsCITE:E6. Those two dates — 2028 for Helion and the early 2030s for CFS's ARC plant — remain the earliest commercial fusion supply targets on record, set against a science base where NIF's own gain metric still excludes the laser system's power consumption and a plant-scale energy balanceCITE:E1CITE:E2CITE:E6.

What this means

Three timelines now run in parallel: NIF's ignition science, which moved from a 1.5 gain in 2022 to a 4.13 gain in 2025CITE:E1CITE:E2; the capital build-up, with $9.77 billion invested across 53 companies and $2.64 billion added in just the past yearCITE:E3; and the corporate offtake commitments — Microsoft's 50 MW by 2028 and Google's 200 MW by the early 2030sCITE:E4CITE:E5. None of these three lines has yet produced a fusion plant that has put electricity on a gridCITE:E6, so the record laser gain, the investment total, and the purchase agreements are three measures of momentum toward a still-unmet outcome.