The IAEA’s new fusion report says the field is moving from experiments to power plants

Four years ago the International Atomic Energy Agency counted 134 fusion devices that were running, being built or planned. In the 2026 edition of its World Fusion Outlook, released on Nov. 12, the count is 174, and privately funded machines have roughly tripled since 2021. The 128-page report was unveiled at the third meeting of the agency’s World Fusion Energy Group at its Vienna headquarters.

Read next on Morning Overview: The FAA wants $2.39 million from a flight school it says flew a Cessna about 280 times with an unapproved part

The shift in what is being counted matters as much as the total. The fourth edition describes a sector whose center of gravity is moving from laboratory experiments toward integrated facilities, industrial power-plant programs and a supply chain to feed them.

Twenty-six power plants, dated by decade

The Outlook tracks 26 planned fusion power plants. The American Nuclear Society’s Nuclear Newswire lists the breakdown by target decade: eight in the early 2030s (seven private, one public), nine in the late 2030s (eight private, one public), two in the 2040s, one in the 2050s, and six with no date given. The report itself says dozens of plant concepts sit at various stages of development, with completion targets spread from the 2030s to the 2050s.

Those are targets, not commitments, and the private share of the earliest dates is large. Of 63 privately funded facilities in the database, 13, or 21%, are operating. Among 111 public facilities, 85, or 77%, already run. Publicly funded machines remain the majority overall in 2026, according to the report, though the private group is the one growing quickly.

The facility counts carry a caveat that the database itself supplies: a device is counted whether it is operating, under construction or merely planned, so growth from 134 to 174 mixes machines that exist with machines that are intentions. The operating shares above, 77% for public facilities and 21% for private ones, are the better measure of how much of the fleet is real hardware today.

Private money: $21.9 billion committed, $80 billion more wanted

The financing section leans on the investment tracker FusionX Group. According to a summary of the Outlook published by Green Building Africa, private fusion companies have raised $21.9 billion in committed capital, more than 90% of it in the past five years. FusionX estimates that a further $80 billion in private capital could be required over the next 20 years to reach commercial deployment, and argues the industry needs different kinds of investment rather than simply more of it.

Recent flows are brisk. Private fusion companies raised a record $4.5 billion in the 12 months to July 2026, and annual funding could reach $7.6 billion in 2026 at the current pace, the summary says. Companies also spent more than $530 million with suppliers in 2025. The Fusion Industry Association, whose chief executive Andrew Holland took part in the meeting and in drafting and reviewing the publication, notes that its own 2026 global industry survey feeds the investment figures, and that cumulative private funding since 2021 now stands above $14 billion.

AI, data and a century-scale range for electricity

Google DeepMind contributed a section on artificial intelligence, covering faster simulations, better-designed experiments, real-time machine control and new materials. The Outlook’s caution is about data rather than algorithms: building the right data foundations, it says, could prove as important as progress in AI itself. The agency cites its Fusion Data Lake as one effort to make research data easier to reach.

On the long view, the consortium EUROfusion supplied an adapted study of fusion’s eventual share of world electricity. Devdiscourse’s account of the report says the range runs from a few percent to 15-20% by the end of the century, depending on assumptions, and quotes the study’s conclusion that the capacity to build, operate and expand power plants will determine the scale of the eventual contribution. The report also states that the ITER project in southern France is progressing ahead of its 2024 baseline schedule.

The ITER item deserves a closer look because the project has long been the field’s schedule benchmark. The report’s statement that it is running ahead of the 2024 baseline is brief, and the summaries reviewed here give no new date, cost or milestone to go with it, so the claim stands as the agency’s characterization rather than a measured result.

The turn toward implementation was already being claimed a year earlier. In the foreword to the 2025 edition, reported by World Nuclear News, agency Director General Rafael Mariano Grossi wrote that fusion is entering a new phase of real-world implementation and that global private investment had surpassed $10 billion. By the 2026 tally, the cumulative figure is more than double that.

The open question the Outlook leaves is the one its own numbers pose. Twenty-six plants are on paper, and more than half carry target dates in the 2030s, yet no globally harmonized definition of a fusion power plant existed when the 2025 edition appeared, and the 2026 summaries reviewed here do not say whether that has changed.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.