For most of the last decade, the biggest energy story in tech was renewable energy procurement — wind and solar PPAs signed by the gigawatt, corporate sustainability pledges, RE100 commitments. In 2026, that story has a new chapter, and it looks nothing like the last one. As of this year, every major AI hyperscaler — Microsoft, Google, Amazon, and Meta — has signed at least one nuclear power agreement specifically to secure electricity for AI data centers. Together, these four companies have committed to more than 13 separate deals totaling close to 10 gigawatts of nuclear capacity, an amount that would have sounded implausible from a tech company's balance sheet just three years ago.
The specifics of each deal reveal just how far companies are willing to go to lock in power. Microsoft's $16 billion, 20-year agreement restarts the former Three Mile Island Unit 1 — now rebranded the Crane Clean Energy Center — with commercial operation now targeted for the second half of 2027, a year ahead of the original schedule, after a Federal Energy Regulatory Commission transmission waiver cleared the last major grid obstacle in mid-2026. Google signed what industry trackers describe as the first-ever corporate agreement to develop a fleet of small modular reactors in the US, committing to up to 500 MW across six to seven units with Kairos Power, with the first reactor targeted for 2030. Amazon has taken a two-track approach: expanding its existing offtake agreement with Talen Energy to nearly 2 GW from the Susquehanna plant in Pennsylvania through 2042, while separately leading a $700 million investment round in X-energy to develop up to twelve gas-cooled Xe-100 SMR units. Meta has gone furthest of all, with agreements covering as much as 6.6 GW spread across TerraPower, Oklo, Vistra, and Constellation — a portfolio approach that hedges across nearly every major reactor technology currently in development.
The logic behind this shift is less about climate branding and more about a hard operational constraint that's reshaped how hyperscalers think about site selection entirely. US grid interconnection queues have swollen past 2,600 GW of waiting projects, with average wait times now stretching toward five years and withdrawal rates as high as 80% for projects that simply give up. Wholesale power prices near hyperscale data center clusters have reportedly surged by as much as 267% in some regions, and global data center electricity demand is projected to double by 2027, with AI workloads responsible for the majority of that growth. Against that backdrop, restarting an existing nuclear plant — with its transmission rights, cooling infrastructure, and grid interconnection already built — offers a materially faster path to reliable, always-on power than waiting in line for a new renewable interconnection or building transmission from scratch. Small modular reactors, for their part, promise even more flexibility: factory-fabricated units that can theoretically be sited directly alongside a data center campus, sidestepping transmission bottlenecks altogether.
Whether that promise holds up on the timelines these companies are counting on is the real open question. SMR unit economics still haven't proven out at commercial scale in the West — NuScale's flagship US project was cancelled in 2023 after cost overruns outpaced what its utility customers were willing to absorb, and most credible industry analyses suggest SMR costs won't hit their mature-market targets until roughly 10 GW of cumulative capacity has actually been built and operated somewhere in the world. Fuel supply is its own constraint: several of the advanced reactor designs backed by Google and Amazon require HALEU fuel enriched between 5% and 20%, and Russia currently dominates the commercial enrichment capacity for that fuel grade, meaning a secure Western supply chain still has to be built essentially from scratch. What's clear is that Big Tech isn't waiting for those questions to be fully resolved before writing checks — the sheer scale of committed capital suggests these companies view guaranteed access to firm, carbon-free power as a competitive necessity for the AI race, worth underwriting risk that utilities alone have historically been reluctant to take on. For the broader energy sector, that's arguably the more consequential shift than any single reactor restart: for the first time in decades, private capital outside the utility industry is willing to absorb first-of-a-kind nuclear risk at scale, and that alone could reshape how fast next-generation nuclear technology actually reaches commercial maturity.
#NuclearEnergy #SMR #AIDataCenters #EnergyStrategy #CleanEnergy #BigTech #EnergyTransition #SynergySolutions