The energy demands of AI are driving investment in futuristic fusion reactors from Big Tech and now Trump Media. Is it viable?
by
Justine Calma
Dec 19, 2025, 6:04 PM UTC
The energy demands of AI are driving investment in futuristic fusion reactors from Big Tech and now Trump Media. Is it viable?
by
Justine Calma
Dec 19, 2025, 6:04 PM UTC
Justine Calma
is a senior science reporter covering energy and the environment with more than a decade of experience. She is also the host of Hell or High Water: When Disaster Hits Home, a podcast from Vox Media and Audible Originals.
Trump Media — yup, the parent company of Truth Social — is the latest entrant in the near century-long race to develop a nuclear fusion power plant. It announced a merger agreement with fusion company TAE Technologies on Thursday, and a bold plan to break ground on the first utility-scale fusion plant some time in 2026.
TAE doesn’t plan to start generating power from its first plant until 2031, which is still an incredibly ambitious timeline. There will likely be a myriad of financial and regulatory issues to contend with along the way, of course. But the scientific and engineering challenges to overcome, which we’ll dive into here, are also enormous.
Fusion is considered the ‘Holy Grail’ of clean energy technologies, and AI companies are salivating over it as a potential source of abundant electricity for data centers. But while a future fusion reactor might one day help solve some of humanity’s headaches, it could just be wishful thinking to expect them to come to the rescue for data centers anytime soon.
Why would Trump Media care about fusion?
With fusion, scientists are trying to replicate the way stars create light and heat — atomic nuclei fuse together, generating a tremendous amount of energy. A fusion power plant could avoid the greenhouse gas emissions from fossil fuels that are causing climate change, as well as the radioactive waste that today’s nuclear fission reactors create by splitting atoms apart to produce energy.
President Donald Trump has made his disdain for clean energy clear by falsely calling climate change a “con job,” ending tax incentives for renewables, and halting federal funding and permitting for solar and wind projects. But he has aligned himself with the tech broligarchy in trying to rush the development of new AI data centers. How to power these energy-hungry facilities has become a major hurdle for the industry, provoking pushback from communities across the US concerned about new data centers potentially driving up electricity rates and pollution.
Big names in tech are funneling money into fusion in the hopes that it might become a silver bullet for everyone’s energy woes. Google and Microsoft have announced agreements to purchase electricity from fusion power plants that other companies plan to complete by the late 2020s or 2030s. Sam Altman, Bill Gates, and Jeff Bezos have also backed startups developing their own fusion technologies.
“Fusion power will be the most dramatic energy breakthrough since the onset of commercial nuclear energy in the 1950s — an innovation that will lower energy prices, boost supply, ensure America’s AI supremacy, and revive America’s manufacturing base, and bolster national defense,” Devin Nunes, Trump Media Chairman and CEO, said on a Thursday investor call.
There are no utility scale nuclear fusion power plants yet, despite the rush of initiatives to design one that might actually work. Success still depends on researchers solving significant scientific and engineering unknowns that they’ve been plugging away at for decades.
How do we know TAE and Trump Media can actually make fusion power plants a reality?
We don’t. As you can imagine, recreating the dynamics of a star on Earth is very difficult. It takes an enormous amount of energy just to kick off a fusion reaction. A stumbling block scientists have faced for decades — and still face — is how to achieve a net energy gain from a fusion reaction.
The first time anyone in the world was able to achieve that net energy gain, often referred to as ‘ignition,’ was in 2022 at the Lawrence Livermore National Laboratory’s National Ignition Facility (NIF). No other group, including TAE, has yet been able to do this with their own technology,.
The 2022 breakthrough at Lawrence Livermore was achieved by firing 192 laser beams at a diamond-encased pellet of fuel. It’s a form of inertial confinement, triggering a fusion reaction by compressing and heating a fuel-filled target. Another competing fusion technology is called a tokamak, which uses magnetic fields to confine plasma and trigger the reaction.
You can think of the reactor TAE is developing as a kind of a hybrid of both strategies, using what’s called field reverse configuration (FRC). It still uses magnetic fields to lock plasma into place, and also shoots beams of fuel directly into the plasma to stabilize it and create the conditions necessary for a reaction to occur.
So when are we going to have electricity from nuclear fusion?
Most experts The Verge has talked to over the years have been cautious to put a date on when a commercial fusion power plant could actually start powering homes and businesses.
Not only does TAE still have to demonstrate that it can achieve a net energy gain — that gain has to be big enough to make economic sense. The Lawrence Livermore breakthrough, for example, achieved a net energy gain of 1.5 megajoule gain (the experiment yielded 3.15 megajoules of energy compared to the 2.05 megajoules the lasers used to trigger the fusion reaction). A laser fusion power plant would likely need to achieve a gain of 50 to 100. The lab has repeated its experiments to work to get higher gains — reaching a record gain of 4.13 MJ in April of this year.
That’s all to say that there are still a lot of milestones for the industry to reach. The Trump administration released a roadmap in October for the advancement of fusion technologies, and the Department of Energy has set a goal of deploying commercial-scale fusion power to electricity grids by the mid-2030s.
“We’re really at kind of that first-of-kind technology innovation, and it’s kind of hard to assign direct timelines to these things,” says Patrick White, group Lead of fusion energy safety and regulation at the Clean Air Task Force. “As [TAE starts] looking at either building a scientific demonstration machine or going straight to a commercial prototype, those types of machines take a substantial amount of investment. And this [merger] potentially is going to be a pathway for them to receive the capital they need to actually start testing and deploying their fusion technology.”
Commercial fusion reactors would also need robust supply chains for fuel and materials strong enough to withstand the extreme pressure and temperatures necessary for nuclear fusion. In short, designing a working reactor would be an enormous step forward — but then comes more difficult work to build the infrastructure and business around it.
Then how big a deal is this merger?
The biggest impact this merger is likely to have — if it ultimately goes through — is to infuse a lot more cash into TAE’s ambitions.
TAE, founded in 1998, claims that its research has advanced enough to the point that “capital is now becoming our biggest challenge,” TAE CEO Michl Binderbauer said on the investor call.
Trump Media has agreed to shell out $300 million to TAE as part of the transaction. That adds to the more than $1.3 billion in private capital that TAE says it has raised to date from Google (TAE has been partnering with Google since 2014 to incorporate machine learning into its research), Chevron Technology Ventures, Goldman Sachs, and others.
With that money, TAE says it’ll be able to start construction on its first utility-scale fusion plant by the end of 2026, and generate “first power” in 2031. The first plant is supposed to have a capacity of 50 MWe, similar to a fission microreactor. TAE didn’t immediately respond to questions from The Verge about where that facility would be located and how much it would cost. But the company’s already planning to build more fusion plants after that with a capacity reaching up to 500 MWe.
Those plans are also contingent on regulatory approvals. But in a controversial decision by the Nuclear Regulatory Commission that was codified by Congress this year, fusion reactors are regulated as particle accelerators in the US, similar to equipment used in cancer therapies and to sterilize medical equipment. That allows fusion plants to bypass lengthy federal licensing requirements that apply to fission reactors.
TAE has developed five iterations of its fusion reactor design, and said as recently as April that it planned to unveil its sixth, called Copernicus, “before the end of the decade.” From there, it would work toward developing the company’s first prototype power plant, Da Vinci, “in the early 2030s.”
Then in November, TAE announced that it’s now leapfrogging over Copernicus and moving straight to Da Vinci. That’s a result of its latest fusion research reactor, Norm, being small and efficient enough to shave down costs by up to 50 percent, according to TAE.
“Norm is such a tremendous breakthrough that it renders Copernicus unnecessary – saving us considerable time and cost,” Binderbauer said in a November press release.
But it’ll still be resource-intensive for TAE to take the next leap in its reactor design. “As [TAE starts] looking at either building a scientific demonstration machine or going straight to a commercial prototype, those types of machines take a substantial amount of investment,” CATF’s White says. “And this [merger] potentially is going to be a pathway for them to receive the capital they need to actually start testing and deploying their fusion technology.”
Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.
Justine Calma
Environment