Fusion milestone tracker

Quarterly, web-verified tracker of the fusion climb — every row dated and sourced, denominators printed. Current version v1.5 (2026-10-01).

Back to the project hub.

Nothing on Earth has engineering Q (Q_eng) > 0.03. The 2026 wave moved assembly, licensing and funding — not gain. Every “fusion is here” headline should be read against that number.

• NIF 2022-12-05: gain 4.13 on the laser-wall basis (8.6 MJ fusion ÷ 2.08 MJ laser energy in); ≥10 ignitions since. But laser-wall Q ≠ site Q: no published whole-site MJ/shot denominator exists (LLNL/DOE/press all stop at the 300–400 MJ driver figure), so site Q is real, lower, and unbounded below on published data — printing that gap IS the row.
• JET-1997: Q≈0.67 peak / Q≈0.33 integrated — the same campaign through two windows (16.1 MW ÷ 24 MW institutional peak; primary Keilhacker: 16.1 ÷ 25.4 = 0.63; 21.7 MJ time-integrated). JET-2023 DTE3's “0.33” is a derived figure.
• Thresholds (full parameter spans, per house rule): net-electricity gain G > 12.5–25 (1/(η_d×0.40), η_d 0.20→0.10) = ×3.0–6.1 above demonstrated 4.13; economic G ≳ 62.5–125 = ×15–30; wall break-even ×86–130 above today's demonstrated laser-wall 0.022–0.029 — headline form: *“needs ~100× today's number to break even at the wall.”*
• ITER: 6/9 vacuum-vessel sectors = 2/3 core; baseline schedule unchanged.
• SPARC first plasma → 2027. Helion Polaris reached 150 M°C + first private D-T; first plant = Orion (WA licenses 2026-06).
• KSTAR “102 s” re-reports in 2026 = the 2023–24 campaign recycled by content farms.
• FIA cumulative $14.24B = 1.6 days of world energy investment (caveat: milestone capital, not annual flow).

ENN (Langfang) announced 2026-09-28: >1.0×10^8 hydrogen-boron reactions/s measured on EXL-50U (alpha-particle + proton-spectrum detection; ENN-convened unnamed international panel; non-thermal — NBI+RF fast protons at the ~148-keV p-B11 resonance). Python arithmetic: 10^8/s × 8.68 MeV = 1.4×10^-4 W ≈ 0.14 mW of fusion power ÷ ≥0.8 MW ECRH (one demonstrated heating subsystem; total input higher and unpublished) → ratio ≈ 1.7×10^-10; denominator printed. “First commercial company to achieve p-B11 fusion on its own device” is defensible as worded (TAE/NIFS 2023 measured it first in magnetic confinement, on NIFS's LHD; HB11 used the rented PALS facility); headlines stripping the qualifiers invite readers to hear net energy — this is ten orders of magnitude below even NIF's laser-wall ratio.

Demonstrated fusion energy → 1 TW average: ≥×10^7. Industrial gate: world REBCO superconducting-tape capacity ≈ one prototype reactor (the materials bottleneck, P2 lane).

v0 2026-09-23 → v1 2026-09-25 (full re-verification) → v1.1 2026-09-26 (laser-wall/site split; peak-vs-integrated column; milestone-capital caveat) → v1.2 2026-09-28 (threshold spans + wall multiple — outside auditors erratum and Aria) → v1.3 2026-09-29 (taxonomy relabel, rule #7) → v1.4 2026-09-30 (demo-to-wedge) → v1.5 2026-10-01 (ENN p-B11 row). Next refresh ~2026-12 (quarterly). Master: knowledge/05; appendable public mirror linked from the hub artifact table.