Big Tech Goes Nuclear
Three Mile Island was shut down in 2019 as an uneconomical relic. By 2024, Microsoft had a deal to restart it. Facing multi-year waits for grid connections and pressure to avoid carbon-heavy power, the hyperscalers went straight to nuclear operators and signed long-term agreements that looked more like utility contracts than tech deals. Microsoft, Amazon, Google, and Meta each anchored gigawatts of capacity, in several cases funding plants that would not otherwise exist. None of them wanted to be in the power business.
For operators dependent on a critical input they don't control, this case sharpens a decision about leverage and timing. It asks what resource is becoming existential to your model while you remain a price-taker, and whether waiting for the market to solve it is simply too slow. The principle the hyperscalers are acting on, and the specific way they reshaped an entire supply landscape, is the lesson held back inside.
Frequently asked questions
Why are big tech companies investing in nuclear power for AI?
Hyperscalers like Microsoft, Amazon, Google, and Meta need vast, reliable, low-carbon electricity for AI data centers, and faced multi-year waits for grid connections. So they went straight to nuclear operators and signed long-term agreements that looked more like utility contracts than tech deals. In several cases they funded plants that would not otherwise exist.
Did Microsoft really restart Three Mile Island for AI?
Yes. Three Mile Island was shut down in 2019 as an uneconomical relic, and by 2024 Microsoft had a deal to restart it to power its AI ambitions. It became one of the clearest signals that AI demand was reshaping the energy landscape. The hyperscalers each anchored gigawatts of nuclear capacity.
Why is it notable that hyperscalers entered the nuclear power business?
None of them wanted to be in the power business, yet the scarcity of firm clean energy forced them to act like utility buyers and even financiers of new capacity. It shows companies vertically integrating into a critical input they could not otherwise secure. The move reshaped an entire supply landscape around AI's energy needs.
What can operators learn from big tech going nuclear?
When a critical input you do not control is becoming existential, waiting for the market to solve it may be too slow, and securing supply yourself becomes a leverage move. The lesson is to name the resource your model depends on and decide whether to act before scarcity bites. CaseBook turns this into a move you apply to your own company, with an AI coach that reads your answer.