Fusion Energy Race: From Lab to Grid 2026–2035

ITER slipped to 2034. The money went elsewhere.
Private fusion funding has 8× since 2021, but the physics has not kept pace — and the only state with a published commercial deadline is China.

John Lawson's 1957 criterion — the density × temperature × confinement-time threshold a plasma needs to cross for a self-sustaining fusion reaction — has been the universal yardstick for 67 years. For 65 of them it was the goalpost nobody cleared. In December 2022 Lawrence Livermore's NIF fired shot N221204 and produced 3.15 MJ of fusion for 2.05 MJ of laser drive. That is scientific breakeven — more fusion energy out of the capsule than laser energy into it — and nothing more. NIF draws ~300 MJ from the grid to fire 2 MJ of laser light, so at the plug the ratio is still worse than 1:100.

ITER — the €27B+ state-led megatokamak at Cadarache, France, designed to be the first reactor-scale burning-plasma device — published a new baseline in July 2024: first plasma 2034, deuterium-tritium operations 2039, plus another €5B on top of the €22B already spent. Original first-plasma target was 2016. That slippage is the industry's central political fact. Capital stopped waiting.

The Fusion Industry Association counted $2.64B of new investment in the 12 months to July 2025 and cumulative private funding at ~$15B by September, roughly 8× the 2021 baseline. Commonwealth Fusion installed the first of SPARC's 18 HTS magnets — high-temperature superconductors that allow stronger fields in a smaller reactor — in January, and now targets first plasma in 2027, with a 400 MWe ARC plant near Richmond, Virginia in the early 2030s. Helion broke ground on a 50 MW pulsed field-reversed-configuration plant (a compact, non-tokamak geometry) in Malaga, Washington, and claims it will sell fusion electricity to Microsoft in 2028. Most plasma physicists do not believe the 2028 number, and Helion's performance data is largely self-reported.

China is running a different book. In January 2025, EAST held a 100 M°C burning plasma — a plasma hot enough to heat itself from its own fusion reactions, not from external beams — for 1,066 seconds, three orders of magnitude longer than JET's best. Six months later, CNNC spun out "China Fusion Energy Company" from its research base — the same corporate pivot that preceded the 60-reactor civilian fission buildout now underway. In March 2026, the corporation's chief fusion scientist published a calendared commercialisation roadmap: burning plasma on HL-3 by 2027, engineering reactor by 2035, commercial demo reactor by 2045.

That roadmap is less a physics claim than a supply-chain one. A February 2026 Cryogenics review documents what is actually being built: high-current-density superconducting cable, radiation-resistant magnet joints, multi-stage cryogenic thermal interception. This is the industrial base for CFEDR, China's demonstration reactor — and the thing Western startups do not have. Until a private reactor actually sells watts, the $15B cumulative number is a story about expectations, not output.