The AI buildout is changing the demand curve for electricity as rapidly as it is sucking up capital from Wall Street.
Stated bluntly, data centers require large amounts of power, but unlike residential or ordinary commercial demand, their load is constant. High-performance chips need uninterrupted electricity around the clock.
That makes reliability as important as total generating capacity.
Wind and solar can add supply, but they are intermittent. Natural gas can meet demand, but it adds fuel-price exposure and emissions.
Nuclear power fits the data center requirement because it provides high-capacity, 24/7 baseload electricity with very high operating reliability.
At first glance, it would appear China has been kickin’ the world’s tuckus in new nuclear energy production:
You may even be alarmed that the U.S. is not on this list of new reactors being built globally.
But if you were, you’d be missing the most interesting distinction between the U.S. and Chinese approaches to bringing AI systems online. Russia, for that matter, too.
In the U.S., hyperscalers are moving toward nuclear power through several channels at once. Microsoft (MSFT) signed a 20-year agreement with Constellation Energy (CEG) to restart a shuttered reactor at Three Mile Island, now renamed the Crane Clean Energy Center.
Amazon (AMZN) bought a 1,600-acre data center campus next to the Susquehanna nuclear plant in Pennsylvania, a behind-the-meter approach that gives it direct access to power instead of waiting for grid upgrades.
Alphabet (GOOGL) signed a framework agreement with Kairos Power for up to 500 megawatts of small modular reactor capacity by 2035.
These deals show nuclear is no longer a “utility” but an integral part of the commercial energy strategy for AI.
If you were to evaluate the industrial production of energy, you’d find nuclear has some similarities with oil and gas, but the differences are more important for investors.
Both are energy sectors tied to national security, heavy capital investment, regulation and geopolitics. But oil projects can move from investment to production in a few years, while nuclear plants can take more than a decade to plan, permit and connect to the grid.
Oil and gas wells deplete quickly, while reactors can operate for 60 to 80 years. Oil demand is more cyclical and tied to transport, industrial activity and consumer behavior.
Nuclear fuel demand is steadier because reactors operate continuously and are refueled on long schedules. A small quantity of uranium contains an enormous amount of energy, but it cannot be used directly after mining.
It must be converted, enriched and fabricated into fuel assemblies before it can enter a reactor.
That processing chain is where the strategic competition begins.
China and Russia use state-directed nuclear models. China is building reactors domestically to secure power for industry, AI, manufacturing and national security.
Russia exports nuclear power through bundled agreements that can include construction, financing, fuel supply, technical support and long-term service.
In that model, the buyer does not simply purchase a power plant. It enters a long-term relationship with the vendor country. Because many reactors require specific fuel assemblies and service arrangements, the fuel relationship can last for decades. This creates predictable uranium demand and gives the exporting country geopolitical leverage.
The Western response is what’s known as The “Sapporo Bloc” Strategy.
The United States, Canada, France, Japan and the United Kingdom are trying to create a nuclear fuel and supply-chain system outside Russian and Chinese control.
Like the development of AI applications, the Sapporo model is less centralized than the Chinese or Russian approach. It relies on allied capacity rather than on a single state-directed machine.
Canada contributes uranium. France contributes fuel-cycle capacity and nuclear operating experience. Japan and the U.K. contribute industrial and strategic depth.
The United States contributes capital markets, advanced reactor development, federal support and rising power demand from AI.
For uranium investors, the result is a second source of demand beyond China’s reactor buildout: Western fuel security, domestic enrichment, HALEU production, strategic inventories and long-term supply agreements tied to the electricity needs of the AI economy.
In the Grey Swan Model Portfolio and several of the Special Situation Reports in the archives on the website, we’ve been following the development of next-generation nuclear reactors. Small modular reactors, SMRs.
What matters while the industry gets organized? Uranium itself. We’re looking at a uranium play that fits in with today’s push toward more, not less, nuclear power.
And following the Sapporo strategy to locate the demand for uranium… and who’s producing it.
In today’s Grey Swan Pro, Andrew recommends one player strategically positioned to benefit from the secular rise in uranium demand and prices in the years ahead.
~ Addison
