Nuclear's Quiet Renaissance Has Crossed the Point of No Return

In September 2023, Microsoft signed a power purchase agreement with Constellation Energy for electricity from the Three Mile Island Unit 1 reactor — the plant that had been closed since 2019 and whose name is synonymous, in the American imagination, with the case against nuclear power. The company did not frame the deal as a climate decision or a values statement. It framed it as an infrastructure decision: it needed reliable, carbon-free baseload power for its data centers, and nuclear was the only source available at the required scale and reliability.

This framing — nuclear as infrastructure rather than ideology — marks the shift that has been underway since approximately 2021 and has now reached a point where its momentum is self-reinforcing.

The Signal

The International Atomic Energy Agency's 2025 update to its global nuclear power projections revised upward its estimate of nuclear capacity additions through 2035 by 34%, driven by new construction commitments in Poland, the Czech Republic, the UK, South Korea, Japan, India, and the United States. The revision was the largest upward adjustment in the IAEA's projection history. More significantly, the drivers of the revision were not primarily climate policy commitments — they were energy security assessments by countries that experienced the geopolitical vulnerability of natural gas dependence in 2022-2023 and concluded that diverse domestic baseload generation was a national security requirement.

The combination of energy security motivation, AI data center power demand, and the mathematical constraint that renewables-only grids cannot provide reliable baseload without storage technology that does not yet exist at the required scale has produced a policy environment that is, for the first time since Three Mile Island, genuinely favorable to nuclear construction across multiple major economies simultaneously.

The Historical Context

Nuclear power has experienced prior moments of apparent revival that did not materialize. The "nuclear renaissance" declared in the early 2000s produced a small number of projects, most of which ran dramatically over budget and over schedule — Vogtle in Georgia, Hinkley Point C in the UK, Flamanville in France. These projects are still being cited by nuclear skeptics as evidence that nuclear is too expensive and too slow to matter in the energy transition.

The structural context for the current revival differs from the 2000s renaissance in ways that the prior-failure argument does not fully account for. The 2000s projects were greenfield deployments in countries that had allowed their nuclear construction expertise to atrophy over 20 years following Chernobyl. They were expensive and slow because they were rebuilding a construction capability that had been systematically abandoned. The countries now at the leading edge of new nuclear construction — South Korea, China, France — have maintained continuous construction programs and have demonstrated learning curves in cost and schedule that the once-off Western projects could not achieve.

The technology landscape has also changed. Small modular reactors, which were speculative in 2005, have now progressed to the point where NuScale's design has received NRC design certification, Rolls-Royce's SMR program has UK government funding, and several designs are in various stages of regulatory review. SMRs do not solve the immediate baseload problem — they will not produce power at scale before 2030 at the earliest — but they represent a technology pathway that reduces the capital concentration risk of traditional large reactors.

The Mechanism

Three forces have compounded to produce the current political consensus for nuclear in countries that had previously moved against it.

Energy security calculations changed fundamentally after Russia's invasion of Ukraine and the subsequent energy crisis in Europe. Countries that had relied on Russian gas for a significant fraction of their electricity generation — Germany being the most extreme example — discovered the political and economic cost of that dependence in a way that abstract energy security arguments had never conveyed. The response has varied: Germany accelerated renewables deployment while closing its last nuclear plants, which most analysts regard as a strategic error it will correct; Poland, the Czech Republic, the Netherlands, and Belgium have moved toward new nuclear construction as part of their energy security response.

AI infrastructure demand has added a new and politically powerful constituency for nuclear power: the technology companies that are building massive data centers and need reliable, carbon-free power at scale. These companies have the financial resources to sign long-term power purchase agreements that reduce the financing risk of nuclear projects, and they have the political influence to advocate for regulatory processes that enable construction timelines.

The renewable energy intermittency problem has become more visible as renewable penetration has increased. Grids with high shares of wind and solar power experience periods of over-supply (requiring curtailment or export) and under-supply (requiring flexible backup generation) that increase with penetration. The storage technology required to eliminate this intermittency at grid scale — specifically, long-duration storage for seasonal imbalances — does not exist at the required cost. Nuclear baseload is the technically mature alternative.

Second-Order Effects

The uranium market implications are already being priced in. Uranium spot prices have increased from $25 per pound in 2020 to $110 in early 2026. The supply side has not responded proportionally — the long lead time for uranium mine development means that supply additions are 7-10 years away. The resulting price trajectory is favorable for uranium producers but will constrain nuclear economics in the 2028-2033 window as construction accelerates.

The geopolitical implications of the nuclear revival are underanalyzed. Russia's Rosatom remains the dominant supplier of nuclear construction services and enriched uranium globally — a position of strategic leverage that the current revival will create strong incentives to reduce. The development of alternative enrichment capacity in the US, UK, and France is accelerating; the reduction of Rosatom's market share is a stated policy objective of the G7. Watch whether nuclear construction contracts signed in the next five years show a measurable shift away from Russian supply chains.

What to Watch

Construction commencement announcements: Watch for final investment decisions and construction commencement at the planned projects in Poland, Czech Republic, and the UK. These decisions, when made, represent committed capital that is difficult to reverse.

SMR regulatory timelines: The NRC's review timeline for SMR designs in active regulatory processes is the best indicator of whether SMRs will contribute to the 2030 grid or only the post-2035 grid.

Uranium enrichment capacity: Watch US and European government announcements on enrichment facility investment. The politics of reducing Rosatom dependence will drive enrichment investment decisions in the next 24 months.

German nuclear policy revision: Germany's decision to close its last operating nuclear plants while building LNG import terminals is widely regarded as strategically incoherent. Watch for political movement toward reversing the closures, particularly as electricity prices remain high.