Nuclear fusion’s moment of truth

FT Big Read on fusion’s capital boom versus scientific and commercial reality: start-ups raised a record $3.8bn in the first eight months of 2026 (vs $3.3bn in all of 2025), with 50+ private companies and backers including Google, Nvidia, Altman, Bezos, Gates and the Trump family’s media group; Pacific Fusion broke ground in Albuquerque and Helion aims to supply Microsoft by 2028. No fusion experiment has yet produced more electricity than its facility consumes — the 2022 NIF “net gain” delivered 3.15MJ of fusion energy while lasers used ~422MJ of electricity. ITER chief scientist Alain Bécoulet says grid connection “will never happen” within 5–10 years and expects fusion in the energy mix only in the last decade of this century; Princeton’s Steven Cowley expects a net-electricity pilot late-2030s and commercial plants around mid-century. Insiders warn of a “Theranos moment” from overpromising on cost/timelines; cost claims (e.g. General Fusion $64–$73/MWh, CFS ~$50/MWh vs ~$71 nuclear / ~$38 utility solar) are contested, and a former General Fusion engineer’s analysis found only ~40% of plant electricity might be saleable to the grid. Technical risks include radiation damage, tritium fuel scarcity and maintenance (UCLA’s Abdou estimates first-gen plants could fail ~every 20 minutes with months-long component swaps).