Why Meta and Google are betting billions on Nuclear

Is a Nuclear Renaissance finally here? In our latest update, we look at how the insatiable energy appetite of AI is catalyzing a new era of fusion and fission.

• Fusion Hits the Gas: From Commonwealth Fusion’s massive magnet breakthrough to Google’s $1 billion commitment, the "future" of energy is arriving faster than expected.
• Big Tech as a Utility: We break down Meta’s massive new deals with Oklo and TerraPower, and why Amazon is targeting 5,000 MW of nuclear capacity by 2039.
• The Reality of Scale: It’s not all smooth sailing. We examine the four massive hurdles—from NRC regulatory bottlenecks to "NIMBY" pushback—that could still relegate these technologies to a niche market.

Chapters

• 0:00 Setting The Energy Agenda
• 0:13 Offshore Wind Legal Whiplash
• 0:35 Fusion Claims And Timelines
• 1:44 Commonwealth’s Tech, Partners, And Deals
• 3:09 Turning To Gen 4 Fission
• 4:15 Hyperscalers Fuel A Nuclear Push
• 5:16 Costs, Scale, And Hard Truths

Transcript

SPEAKER_00: 0:04

Hi, this is your energy story for the third week of January 2026. We've spent a great deal of time this past week thinking about the energy sources of tomorrow. Offshore wind might fit into that category, but the Trump administration hasn't finished kicking that dog yet. Although just this morning, yet another federal judge, it was announced, lifted a stop work order on the$5 billion New England project owned by Orsted, and Dominion announced it would ask a federal judge to do the same thing for its$11 billion project off of Virginia. Today, though, let's look briefly at the prospects for both fusion and fish and generating technologies, and let's start with the fusion company TAA, which announced in mid-December that it was merging with the Trump Media Group of all things in an all-stock transaction valued at more than six billion dollars. The press release stated that by 2026, the combined company plans to site and begin construction on the first utility-scale fusion power plant at 50 megawatts. On its website, TAE notes the fact that it has achieved proof of science with the ability to sustain plasma confinement with temperatures exceeding 158 million degrees Fahrenheit. That's not too bad. And inhales other technological breakthroughs that may help to create a more cost-effective approach and shorten the road to commerciality. TE also says it's on track to deliver hydrogen boron fusion to the grid by the early 2030s. And on January 6th, it announced it's begun the site selection process for a 20-plus acre site with proximity to a primary distribution grid and metropolitan hub. Its research reactor recently delivered beyond its initial expectations. But from the outside looking in, if I had to guess or bet who'd get the fusion first, I'd have to turn my sites to Commonwealth Fusion. They've got more critical pieces of the puzzle apparently assembled, and they just announced the manufacture of the first magnet necessary to create the container that holds the fusion reaction. In their case, attempts of 180 million degrees Fahrenheit, which is hotter than the core of the sun, even at night. It also has the critically important manufacturing capability to make more of those magnets. Commonwealth is currently building a demonstration plant in Massachusetts, and it also just announced that it's enlisting AI chip manufacturer Nvidia and industrial powerhouse Siemens to develop digital representations of its machines, compiling data to make simulations that can help accelerate the design and make improvements along the way. And finally, the all-important piece is the off taker. Google committed last June to purchase 200 megawatts of power from the first fusion plant in Chesterfield County, Virginia. And importantly, Google's been an investor in Commonwealth Fusion since 2021. It announced last June it's increasing its stake by an undisclosed amount, giving Google the option to purchase even more power from additional plants. And on the heels of that, Commonwealth announced an additional deal with Italian energy company Any worth a billion dollars. So fusion may just be a very real thing sometime in the visible future. Now let's turn to fission, since there's a constant drumbeat of news coming from the modular nuclear sector. This week, Meta announced deals with two of the so-called Generation 4 modular nuclear companies, Oclo and Terrapower. Gen 4 nukes aren't your typical 700 to 1,000 megawatt large nukes. Instead, they're designed to be smaller, more fuel efficient, fail-safe in the case of a plant failure or an event such as a tsunami or earthquake, and with fuel from which it's more difficult to make nuclear weapons. Numerous Gen 4 companies are active these days, and the optimism surrounding them is quite high. Take publicly traded Oklahoma. Its stock rose from a share price of$23 a year ago to$102 today, and it kissed$175 along the way. Meta recently announced a deal contemplating a plant in Chicago with Oklahoma, the first of many additional projects totaling up to 2,200 megawatts. With the first coming online as early as 2030. Oklahoma has a site use permit for an initial trial reactor and has also ordered fuel for that first machine from Idaho National Laboratories. Oklahoma also submitted a licensed application for both construction and operation of its first plant, and it broke ground for that in September. Meta's second deal with Bill Gates backed Terapower, while not specifying a site, includes up to eight of Terapower's reactors, joined with energy storage systems for load falling capabilities. Two of those plants may be commissioned as early as 2032, with the enterprise eventually scaling to be as large as 2,000 megawatts of nuclear generation paired with 1200 megawatts of storage. If a nuclear renaissance is to occur, it will be the insatiable energy appetites of the hyperscale data centers that likely catalyzes it. Last August, Google, Kairos Power, and Tennessee Valley Authority announced a power purchase agreement between Kairos Power and TVA for up to 50 megawatts in that case that would support Google data centers in Tennessee and Alabama. That makes TVA the first U.S. utility to ink a power purchase agreement with a Gen 4 reactor. Kairos and Google have a separate agreement for 500 megawatts of nuclear capacity by 2035, with the first of those units scheduled for operation by 2030. Meanwhile, Amazon and Xenergy are moving ahead as well. Last August, they announced a relationship with Korea Hydro and Nuclear Power and Dusan entrability to accelerate the deployment of modular reactors in the U.S. with an eventual target of 5,000 megawatts across the country by 2039. In December, XENergy also inked a binding agreement for the manufacture of its main power system steel components for 16 of those 80 megawatt modular reactors. Amazon has also selected its first sites, having signed an agreement with Energy Northwest. There are other modular nuclear companies besides the ones I mentioned, such as NewScale and Holtec, but they all do face similar challenges. First, they have to get regulatory design approved by the Nuclear Regulatory Commission, and only NewScale's done that so far. Then they have to build and trial their first reactors. Even as they start construction of the plants, they will eventually build more machines. To get there, they'll need to fill their order books, because selling one or two of these isn't going to get to the scale needed to drive down costs, especially in today's world where anything related to power is soaring in price. Plus, you have to attract more investment capital. Scalar manufacturing can help reverse the price trends because you can spread capital costs across more units. And a learning curve can kick in if you get volumes high enough. But let's not kid ourselves about that. This isn't like solar panels or battery cells where one creates a virtuous circle during the manufacture of millions or billions of cells or panels. In those instances, one had massive improvements in the underlying technology, think about sunlight conversion efficiencies for solar or lifecycle gains for batteries. Combined with supply chain efficiencies all the way up the ladder, added to manufacturing improvements. The numbers we're talking about here for modular nukes, they won't get you there. Then there's a new workforce that needs to be developed and NIMBY issues to be overcome. If solar and wind are facing increasing pushback, one shouldn't expect anything different with small nuclear plants in the backyard. Could there be a moonshot movement? Perhaps. But also only if a number of these companies perish on the road. The industry won't achieve scale with all of them operating. And the cost issue ultimately will be paramount here as it always is in the electric energy space. New Scale serves as a cautionary example. It came out of the box with an agreement pre-COVID to serve the Utah Association of Municipal Power Systems at$58 a megawatt hour. Post-COVID, that number soared to$89 a megawatt hour, and all the potential off takers headed quickly for the exit ramp. So the fortunes of these new industries, fusion and fission, will be based both on what they can achieve as well as developments in the market around them. Gas prices, storage technologies, the evolving conversation around carbon, loan guarantees and government subsidies, all these are elements of the soup that will eventually determine whether these industries survive and thrive or are relegated to a small niche in the power game of tomorrow. Well, thanks for watching, and we'll see you again soon.