Nuclear Fusion Timeline, 1920–2040: From Eddington to ITER

Fusion research is declassified and international cooperation begins (unverified)

Nations are reported to have opened their fusion programmes at the Second International Conference on the Peaceful Uses of Atomic Energy. The year before, Britain's ZETA had announced and then retracted a claim of fusion — a period when verifiability itself was at issue. Dates and details still need primary sourcing.

People and institutions: IAEA / UK/US/USSR

UnverifiedDeclassification of fusion research at the Second International Conference on the Peaceful Uses of Atomic Energy (Geneva, September 1958). Not yet confirmed against IAEA or other primary sources.

Neutral beam injection breaks the heating barrier (unverified)

PLT is reported to have reached ion temperatures far above previous tokamaks using neutral beam heating — the point at which the second condition, getting the plasma hot enough, began to look tractable alongside confinement. Figures and year need primary sourcing.

People and institutions: Princeton Plasma Physics Laboratory

UnverifiedIon temperature record on PLT (Princeton Large Torus) under neutral beam heating. Neither the year nor the figures have been confirmed against PPPL primary sources.

ITER is born at the Geneva summit

Gorbachev proposed to Reagan that their countries collaborate on fusion for peaceful purposes. In 1986 Europe, Japan, the USSR and the US agreed on a joint design — the origin of ITER. That this machine begins in diplomacy rather than science shapes the character of every schedule problem that follows.

People and institutions: Mikhail Gorbachev / Ronald Reagan

• ITER History — ITER Organization

JET produces controlled fusion power for the first time

Two shots on a 90% deuterium / 10% tritium mix reached a peak of 1.7 MW, averaging 1 MW over 2 seconds — a ratio to input power of about 0.12. The first experiment that put a measured number on the distance to the Lawson criterion.

• Peak fusion power 1.7 MW
• Q ~0.12

People and institutions: JET / EUROfusion

• JET Deuterium-Tritium Results and their Implications (JET-P(98)70) — EUROfusion / M. Keilhacker

TFTR reaches 10.7 MW on a 50:50 D-T plasma

Extensive 50/50 deuterium-tritium experiments began in December 1993, and in 1994 TFTR produced 10.7 MW — comparable to the demand of roughly 3,000 households. But this is instantaneous fusion power; net of the electricity fed into the machine, the balance is still deeply negative.

• Peak fusion power 10.7 MW

People and institutions: TFTR / Princeton Plasma Physics Laboratory

• Tokamak Fusion Test Reactor — Princeton Plasma Physics Laboratory

JET DTE1 — a record that stands for 25 years

An ELM-free H-mode discharge (pulse 42976) produced 16.1 MW of fusion power at a ratio to input power of 0.62. Just short of scientific breakeven — and that "just short" would go unimproved for more than a quarter of a century.

• Peak fusion power 16.1 MW
• Q (vs input power) 0.62
• Total energy 22.7 MJ

People and institutions: JET / EUROfusion

NoteThe paper gives Qin = 0.62 (Qtot = 0.95±0.17). The figure is commonly quoted as 0.67; cite the source explicitly when writing this up.

• JET Deuterium-Tritium Results and their Implications (JET-P(98)70) — EUROfusion / M. Keilhacker

The ITER Agreement is signed at the Élysée Palace

Representatives of the seven members — Europe, Japan, China, Russia, Korea, India and the US — signed the 34-page agreement in Paris. The site near Cadarache had been chosen in 2005; the ITER Organization was formally established on 24 October 2007.

People and institutions: ITER Organization / EU / Japan / China / Russia / Korea / India / USA

• 10 years ago: the agreement that established ITER — ITER Organization

Wendelstein 7-X first plasma — the stellarator line returns

The stellarator, which had lost out to the tokamak, came back via three-dimensionally optimised superconducting coils. Its inherent advantage for steady-state operation only became buildable once computers could optimise the geometry.

People and institutions: Wendelstein 7-X / Max Planck Institute for Plasma Physics

NoteFirst plasma (helium) is given as 10 December 2015 and hydrogen as 3 February 2016. Dates should be fixed against IPP primary sources.

• Wendelstein 7-X — Max-Planck-Institut für Plasmaphysik

A 20-tesla high-temperature superconducting magnet — the key to shrinking the machine

CFS and MIT reached 20 tesla with a fusion magnet wound from high-temperature superconductor: performance that would otherwise require a device forty times larger. Because fusion performance scales with roughly the fourth power of field strength, this rewrote the assumptions about machine size and capital cost — and marks the point where private capital arrived in earnest.

• Field strength 20 T

People and institutions: Commonwealth Fusion Systems / MIT Plasma Science and Fusion Center

• MIT-designed project achieves major advance toward fusion energy — MIT News

Ignition — the Lawson criterion is crossed for the first time

2.05 MJ of laser energy on target yielded 3.15 MJ of fusion energy: output exceeded input, 67 years after Lawson wrote the inequality. But this is a target-level balance and excludes the roughly 300 MJ the laser itself consumes. The gap between scientific ignition and generating electricity became explicit here.

• Output 3.15 MJ
• Laser input 2.05 MJ
• Target gain 1.5

People and institutions: National Ignition Facility / Lawrence Livermore National Laboratory

• Achieving Fusion Ignition — Lawrence Livermore National Laboratory

JT-60SA first plasma — the largest operating tokamak in the world

The Japan–EU superconducting tokamak at Naka reached first plasma. As the machine closest in design to ITER, it works out operating methods ahead of time and feeds back manufacturing lessons — filling the decade-plus before ITER itself runs.

People and institutions: JT-60SA / QST / Fusion for Energy

• Japan and EU mark first plasma from JT-60SA fusion device — World Nuclear News

69.26 MJ in JET's final experiments — and then retirement

0.21 milligrams of fuel produced 69.26 MJ over six seconds, beating 22.7 MJ in 1997 and 59 MJ in 2022. What mattered more than the number was that the conditions were reproduced across multiple pulses. JET ended forty years of operation with this campaign.

• Total energy 69.26 MJ
• Duration 6 s
• Fuel 0.21 mg

People and institutions: JET / EUROfusion / UKAEA

• Breaking New Ground: JET Tokamak's Latest Fusion Energy Record — EUROfusion

ITER pushes its schedule back by a decade

The old "first plasma in 2025" milestone was dropped for a new baseline: start of research operation in 2034 with deuterium-only plasma, full magnetic energy in 2036, and D-T operation in 2039, at an additional cost of around €5 billion. Director-General Barabaschi argues that starting with more systems in place is the more robust path.

• Start of operation 2034
• D-T operation 2039
• Additional cost ~€5 billion

People and institutions: ITER Organization / Pietro Barabaschi

• New baseline to prioritize robust start to exploitation — ITER Organization
• ITER fusion reactor hit by massive decade-long delay and €5bn price hike — Physics World

Tokamak pulses pass 1,000 seconds — first EAST, then WEST

On 20 January 2025 China's EAST held a high-confinement (H-mode) plasma for 1,066 seconds, about 2.6 times its own 2023 record of 403 seconds. A little over three weeks later, on 12 February, France's WEST held a hydrogen plasma for 1,337 seconds (about 22 minutes). Neither used tritium, so almost no fusion power was produced. The contest here is not output but whether the walls and heating systems can survive long operation — the other condition a power plant cannot do without.

• EAST H-mode 1066 s (3.05 GJ injected)
• WEST 1337 s (2.6 GJ injected)

People and institutions: EAST / Institute of Plasma Physics, Chinese Academy of Sciences / WEST / CEA-IRFM

• Overview of recent experimental results on EAST in support of ITER new research plan — 30th IAEA Fusion Energy Conference (2025), overview paper, X. Gong et al.
• WEST sets a new plasma duration record — CEA-IRFM, February 2025

NIF reaches 8.6 MJ — gain quadruples within two and a half years of ignition

8.6 MJ from 2.08 MJ of laser input, the highest yield to date and about 2.7 times the gain of December 2022. Iteration is clearly paying off — though NIF remains a stockpile stewardship facility, not a machine built to generate power.

• Output 8.6 MJ (±0.45)
• Laser input 2.08 MJ
• Target gain ~4.1

People and institutions: National Ignition Facility / Lawrence Livermore National Laboratory

• Achieving Fusion Ignition — Lawrence Livermore National Laboratory

W7-X sets a long-duration triple product record — a different axis entirely

In the OP 2.3 campaign, a high value of the fusion triple product was sustained for 43 seconds, alongside an energy turnover of 1.8 GJ over 360 seconds. This is a contest about duration rather than peak power, and it is where the stellarator's structural advantage finally shows up as a number.

• Triple product sustained 43 s
• Energy turnover 1.8 GJ / 360 s

People and institutions: Wendelstein 7-X / Max Planck Institute for Plasma Physics

• New performance records for Wendelstein 7-X — Max-Planck-Institut für Plasmaphysik

A privately built machine demonstrates measurable D-T fusion for the first time

Helion's Polaris prototype became the first privately developed device to show measurable deuterium-tritium fusion (January 2026, announced 13 February). Plasma temperature reached 150 million °C, up from 100 million on the earlier Trenta prototype. Experts at Sandia National Laboratories and the University of Michigan are reported to have verified the results independently.

• Plasma temperature 150 million °C

People and institutions: Helion Energy / Polaris

• Helion Achieves New Industry-First Fusion Energy Milestones — Helion Energy

The US proposes to regulate fusion machines outside the reactor framework

The US Nuclear Regulatory Commission published a proposed rule that would license fusion machines not as fission reactors but under the same framework used for radioactive materials in medicine and industry (10 CFR Part 30). It is the regulatory system acknowledging a physical difference: there is no chain reaction, and switching the machine off stops the reaction. Open questions put out for comment include reporting thresholds for the large amounts of tritium fuel and how to dispose of parts made radioactive by neutrons. By law the final rule must be in place by the end of 2027.

People and institutions: U.S. Nuclear Regulatory Commission

• Regulatory Framework for Fusion Machines (proposed rule) — Federal Register 91(38), 9476 (26 February 2026), U.S. Nuclear Regulatory Commission
• NRC Kickstarts Process for Creating Regulatory Framework for Fusion Machines — U.S. Nuclear Regulatory Commission, press release No. 26-023, 26 February 2026

Japan's Q-DEMO — an ITER-sized prototype aiming to generate electricity in the 2030s

Japan's National Institutes for Quantum Science and Technology (QST) presented the conceptual design of Q-DEMO, a prototype reactor with an ITER-sized plasma, to a Cabinet Office task force. A machine larger than ITER would cost more time and money, so the size was deliberately matched to ITER, and building to phase 1 was estimated at about 2 trillion yen on the basis of ITER component manufacturing. Phase 1 (proposed to start in 2038) runs for minutes and aims only to show that the electricity generated roughly covers the plant's own use. Making more fuel than it burns comes in phase 2, and a net output of about 100 MW only in phase 3.

• Construction cost (to phase 1) about ¥2 trillion
• Fusion power (phase 1) about 500 MW
• Tritium breeding ratio target (phase 2) 1.05

People and institutions: QST (量子科学技術研究開発機構) / Cabinet Office, Government of Japan (内閣府)

• ITERサイズ原型炉(Q-DEMO)の概念設計について (Conceptual design of the ITER-sized prototype reactor Q-DEMO) — QST, 6th Task Force on Social Implementation of Fusion Energy, Cabinet Office, Material 2-1 (16 March 2026)
• 「Q-DEMOの概念設計」報告書(概要) (Q-DEMO conceptual design report, summary) — QST, same task force, Material 2-2 (16 March 2026)

Can a reactor make its own fuel? A facility to measure breeding blankets

Every power-plant design that burns deuterium and tritium assumes a "blanket" around the plasma in which neutrons strike lithium and breed the tritium fuel. Yet no one has confirmed this in a neutron environment like that of real fusion. With US Department of Energy funding, Kyoto Fusioneering and Oak Ridge National Laboratory will build UNITY-3, which uses an accelerator to fire 14 MeV neutrons into blanket test pieces and measure how much tritium is produced. A 2023 review notes that computed breeding ratios still carry uncertainties of 5–15%, so whether a reactor can keep itself fuelled is, for now, an answer on paper.

• Target completion end of 2028
• Uncertainty in computed breeding ratio 5–15%

People and institutions: Kyoto Fusioneering / Oak Ridge National Laboratory / U.S. Department of Energy

• ORNL, Kyoto Fusioneering to develop fusion technology test facility in East Tennessee — Oak Ridge National Laboratory, 11 August 2026
• Review of commercial nuclear fusion projects — Front. Energy Res. 11, 1157394 (2023), S. Meschini et al.

Every large component of ITER's core machine is now on site

The last vacuum vessel sector, built in Europe, arrived on 2 October, putting every major component of the core machine — superconducting coils, vacuum vessel and cryostat — on site. The vessel is assembled by welding together nine sectors, six of which were in place in the tokamak pit at the time. In-vessel components such as the divertor and blanket are still to come, and the 2024 plan of starting operation in 2034 and D-T operation in 2039 is unchanged. It marks the point where the challenge shifts from making the parts to putting them together and running them.

• Vacuum vessel sector modules 6 of 9 installed (Oct 2026)

People and institutions: ITER Organization / Fusion for Energy

• A remarkable chapter draws to a close — ITER Organization, ITER Newsline, 5 October 2026
• Two-thirds of ITER tokamak core now in place — ITER Organization, press release, 29 July 2026