Recently, while taking Professor Xiadong Chen of NTU’s course “Social Impacts of Disruptive Technologies,” I’ve been rethinking nuclear energy.

I used to always think nuclear power is a typical “national-level project”: huge investments, long timelines, and complex approvals—making it difficult for ordinary companies and capital to truly participate.
But the emergence of AI data centers may be changing this logic.
The biggest change is not that nuclear technology suddenly became mature, but that for the first time nuclear power has encountered business buyers that are both capable enough and urgent enough.
In the past, nuclear power mainly addressed national energy security; now AI companies like Microsoft, Google, and Amazon are also proactively looking for energy solutions that can provide long-term, stable, low-carbon power.
So several changes are happening at the same time:
Large nuclear power plants still belong to national-level infrastructure; small modular reactors are moving into commercial validation, and are more suitable for the electricity demand of data centers, industrial parks, and high-energy-consumption enterprises.
Meanwhile, improvements in passive safety technology, modular manufacturing, and regulatory efficiency are also trying to turn nuclear energy from a “one-time super project” into a gradually replicable industrial product.
The value of nuclear energy is not only about electricity generation.
Heating, industrial steam, hydrogen production, seawater desalination, and even stable energy supply for high-energy-consumption manufacturing industries could all become new business scenarios.
Behind this is not just nuclear power plants themselves, but an entire supply chain, including fuel, forgings, instrumentation and control, high-temperature superconducting magnets, and more.
Of course, nuclear energy is not without risks.
Construction timelines, cost overruns, nuclear waste, fuel supply, regulation, and project financing—if anything goes wrong in any link, the project may shift from a long-term asset to a stranded asset; and fusion’s path to true commercialization still needs time.
So my current view on nuclear energy is:
AI hasn’t made nuclear power mature overnight, but it is giving nuclear energy commercial demand, capital support, and real orders that it has never had before.
In the next decade, nuclear power may not be the most “sexy” narrative in the AI industry, but it could become one of the most underrated pieces of infrastructure.
More complete research notes—I’ll continue to share them later.