Nuclear Energy's Comeback: Why the World Is Betting on Fission Again
For most of the last forty years, nuclear energy was a story of retreat. Chernobyl in 1986. Three Mile Island before that. Fukushima in 2011. Each accident hardened public opposition, emboldened anti-nuclear activists, and gave policymakers a reason to phase out existing plants rather than build new ones. Germany shut down its last reactor in 2023. Japan mothballed most of its fleet. The narrative was settled: nuclear was dangerous, expensive, and obsolete.
That narrative is now reversing — faster and more completely than almost anyone predicted.
In 2026, nuclear energy isn't just surviving. It is the fastest-growing area of energy investment in the developed world. Governments that banned new plants are reversing policy. Tech companies are signing 20-year power purchase agreements with nuclear operators. A generation of new reactor designs — smaller, safer, cheaper — is moving from blueprints to construction sites.
The question is why now. And whether the optimism is justified.
The Trigger: AI's Insatiable Appetite for Power
The immediate catalyst for nuclear's revival isn't climate policy or energy security — it's data centres.
Training a large AI model like GPT-5 consumes roughly the same electricity as a small city uses in a month. Running that model at scale — millions of inference requests per day — requires continuous, reliable, carbon-free power. Lots of it.
Solar and wind are intermittent. Batteries can't yet store enough energy cheaply enough to make them reliable as baseload power. Hydropower is geography-dependent. The only carbon-free energy source that runs 24 hours a day, 365 days a year, at industrial scale, regardless of weather, is nuclear.
This is why Microsoft signed a 20-year agreement to restart Unit 1 of Three Mile Island — the plant synonymous with America's worst nuclear accident — specifically to power its AI data centres. Google followed with a deal to purchase power from a fleet of small modular reactors (SMRs) being developed by Kairos Power. Amazon has made similar commitments.
When the world's most valuable tech companies start staking their infrastructure on nuclear, the industry's investment climate changes overnight.
What's Actually Being Built
The nuclear renaissance isn't about the old model — massive, billion-dollar, decade-long construction projects. It's about two new categories.
Small Modular Reactors (SMRs)
SMRs are nuclear reactors with generating capacity of 50–300 MW, compared to 1,000–1,600 MW for conventional plants. The key advantages:
- Factory-built: Components are manufactured at scale in controlled facilities, then assembled on-site. This reduces the construction delays and cost overruns that plagued previous nuclear projects.
- Smaller footprint: Can be sited in locations unsuitable for conventional plants — near industrial facilities, on retired coal plant sites, in remote areas.
- Passive safety systems: Modern SMR designs use physics — gravity, convection, natural circulation — rather than active pump systems to keep the reactor cool in an emergency. They can shut down safely without any operator intervention or external power.
- Faster deployment: Target 5–7 years from permit to power, versus 15–20 for conventional plants.
Companies like NuScale (US), Rolls-Royce (UK), and GE Hitachi are in various stages of regulatory approval and early construction.
Advanced Reactor Designs
Beyond SMRs, several next-generation reactor types are advancing:
Molten Salt Reactors — Use liquid fluoride salts as both fuel and coolant. Cannot melt down in the traditional sense because the fuel is already liquid. Thorium-based variants produce far less long-lived radioactive waste than conventional uranium reactors.
High-Temperature Gas Reactors — Use helium as coolant and graphite as moderator. Operate at higher temperatures, enabling industrial heat applications (steel, cement, chemical manufacturing) that electricity alone can't easily provide.
Fusion — Still experimental, but Commonwealth Fusion Systems and TAE Technologies have made genuine technical progress. Commercial fusion power remains a decade or more away, but it is no longer purely theoretical.
The Government Reversal
The policy shift is as striking as the investment shift.
Japan — After Fukushima, Japan shut down virtually its entire nuclear fleet. By 2023, the government had reversed course, committing to restart existing plants and develop next-generation reactors. By early 2026, over 20 reactors are operating again.
Germany — Shut its last plant in April 2023 in one of the most controversial energy decisions in recent European history. The energy crisis triggered by Russia's invasion of Ukraine — which led to surging gas prices and power shortages — has reopened the debate. A parliamentary review in 2025 concluded the shutdown was a mistake. Restarting the closed plants is technically feasible but politically difficult; the conversation is ongoing.
United Kingdom — Committed to building up to eight new large nuclear plants and fast-tracked planning approval for SMRs. Rolls-Royce's SMR programme has received £210 million in government funding.
United States — The Nuclear Regulatory Commission approved its first SMR design in 2023. The Inflation Reduction Act included significant production tax credits for nuclear power, treating it equally with wind and solar for the first time.
France — Never fully abandoned nuclear (it generates around 70% of French electricity) and announced plans to build six new EPR2 reactors in 2022, with more under consideration.
India and China — Both have aggressive nuclear expansion programmes. China is building more nuclear reactors than the rest of the world combined — over 20 under construction as of 2025, with plans for dozens more.
The Case For Nuclear
The revival rests on several converging arguments that have strengthened significantly over the past five years.
Carbon emissions are the lowest of any energy source. Life-cycle analysis — accounting for construction, fuel, and decommissioning — puts nuclear's carbon footprint at 4–12 grams of CO₂ equivalent per kilowatt-hour. This is lower than solar (20–50g) and wind (7–15g), and dramatically lower than gas (490g) or coal (820g).
It is safer than people believe. Deaths per unit of energy produced: nuclear causes fewer deaths than any other energy source, including solar (which has rooftop installation fatalities). Chernobyl was a uniquely flawed Soviet-era design operated in ways modern reactors cannot replicate. Fukushima killed one person directly from radiation.
Energy density is extraordinary. A kilogram of uranium fuel contains roughly 3 million times more energy than a kilogram of coal. The fuel cost and land use of nuclear are minimal compared to the energy produced.
Reliability. Nuclear plants operate at 90%+ capacity factors — meaning they produce near their maximum output almost all the time. Solar averages 25%, onshore wind around 35%. For grid stability, this difference is enormous.
Waste is smaller than imagined. All the nuclear waste ever produced by US commercial reactors — over 60 years — would fit within a single football field stacked 10 metres high. It is solid, containable, and does not disperse into the environment. Managing it is an engineering problem, not an unsolvable one.
The Case Against (Honestly)
The pro-nuclear case is strong. But the concerns that drove the previous retreat aren't imaginary.
Cost overruns are real. The last two large nuclear plants built in the US — Vogtle Units 3 and 4 in Georgia — cost over $35 billion, more than double the original estimate, and took 7 years longer than planned. The story of Western nuclear construction in the 21st century has been consistent: massive delays, staggering cost overruns. SMRs promise to change this, but no SMR has been built at commercial scale in the West yet. The promises are real; the proof is not.
The waste problem is managed but not solved. Current storage is safe and stable. But permanent geological repositories for high-level waste don't exist at scale in most countries. The US's Yucca Mountain repository — planned for decades — remains blocked by politics. "We can store it safely for now" is not the same as "we have solved the waste problem."
Proliferation risk. Nuclear technology and weapons technology share some of the same knowledge base. Expanding nuclear power globally, to countries with less robust regulatory frameworks, increases the risk of weapons proliferation. This is a genuine strategic concern, not a scare tactic.
Public acceptance. Support for nuclear has grown significantly in polls — particularly among younger people who prioritise climate over accident risk. But in many countries, a nuclear plant announced today would face years of legal challenges. Social licence is easier to support in a survey than to maintain through an actual planning process.
What It Means for India
India is one of the most energy-hungry countries in the world, with electricity demand projected to triple by 2050. The country currently gets less than 3% of its electricity from nuclear — the lowest share among major nuclear nations.
India has a domestic thorium-based nuclear programme — thorium is a more abundant fuel than uranium, and India has massive reserves. The three-stage programme, developed over decades, is approaching its second stage. If it succeeds at scale, India could have a nuclear fuel source that is domestic, abundant, and produces less long-lived waste.
For a country simultaneously trying to industrialise, electrify rural areas, decarbonise, and reduce energy import dependence, nuclear's combination of reliability and low carbon emissions is compelling. The constraint has been capital, regulatory speed, and public acceptance — all of which are being actively worked on.
The Bigger Picture
Nuclear energy's revival says something important about how technology and society actually change.
For thirty years, the story seemed settled. Nuclear was over. The accident at Fukushima was supposed to be the final confirmation. But the energy crisis, the climate crisis, and the AI power demand crisis arrived simultaneously — and suddenly every drawback of nuclear looked smaller relative to the problem it solves, and every alternative looked more limited than it had appeared.
This pattern — the return of a discarded technology when circumstances change — is more common than it seems. Nuclear is not unique in being written off prematurely.
The honest summary: nuclear energy is not a silver bullet. It is expensive and slow to build. It produces waste that requires careful long-term management. It carries proliferation risks that require serious governance.
It is also the only carbon-free energy source that runs at industrial scale, continuously, regardless of weather. In a world where we need to power civilisation without cooking the planet, that matters enormously.
The world isn't betting on nuclear because it's perfect. It's betting on nuclear because the alternatives have limits that are becoming impossible to ignore.
Published by Publixly. Follow for weekly analysis on technology, energy, and the forces shaping the world.