Nuclear renaissance: the US, AI and SMRs

While the world’s largest tech corporations are building massive AI data centers and aggressively competing for electricity, a new energy megatrend is taking shape in the United States – one that just a few years ago looked too expensive, too slow, and nearly forgotten. It is the return of nuclear energy, but in a completely new form.

The US Department of Energy (DOE) has made a move that the market sees as a major signal for the entire small modular reactor (SMR) industry. The agency selected Oklo (OKLO) and several other players for discussions on converting weapons-grade plutonium into civilian nuclear fuel.

At first glance, this sounds like science fiction or a late “Terminator” plot: former nuclear weapons material being transformed into fuel for AI data centers. But this is increasingly what the new US energy strategy looks like.

And the key issue is not even plutonium itself, but the problem Washington is trying to solve.

The main SMR problem is not reactors – it is fuel

Over the past two years, the SMR sector has seen a surge of investor interest. Small modular reactors are promoted as the future of nuclear power: compact, scalable, potentially safer, and cheaper than traditional nuclear plants.

The AI boom has given SMRs a second life. The reason is simple: artificial intelligence requires enormous amounts of energy.

Next-generation data centers already consume electricity comparable to small cities. And if AI growth projections hold, global AI infrastructure energy demand could multiply several times by 2030, with some estimates pointing to a tenfold increase in compute capacity.

The problem is that existing power systems are not designed for such a shock.

Solar and wind power are too unstable for continuous data center operations. Gas is politically and environmentally sensitive. Coal is largely off the table in the West. As a result, nuclear energy is once again emerging as one of the few stable baseload power sources for the AI era.

But the market has hit an unexpected bottleneck: fuel.

Most SMR projects rely on HALEU – high-assay low-enriched uranium. It is a next-generation nuclear fuel required for many advanced reactor designs.

And this is where geopolitics enters the picture.

The US suddenly discovered dependency

Historically, a significant share of global HALEU production infrastructure was linked to Russia. After sanctions and deteriorating relations, the US found itself in an uncomfortable position: it can design reactors, but securing fuel supply is far more difficult.

This is why the DOE decision to explore plutonium recycling was interpreted as a strategic pivot.

Effectively, the US is trying to build a temporary “energy bridge” until domestic HALEU production reaches commercial scale. This reduces external dependence while accelerating the SMR industry.

The market reacted instantly

Sector stocks moved sharply higher.

Oklo (OKLO), linked to OpenAI’s Sam Altman, rose about 4.5%. Wedbush maintains a price target of around $110, continuing to bet on long-term sector growth.

NuScale Power (SMR) gained about 6%. Bank of America analysts highlighted its advantage: its reactors can run on conventional uranium, reducing reliance on scarce HALEU and providing a timing edge.

Nano Nuclear Energy (NNE) surged nearly 9%, behaving almost like a meme-stock rally.

This reaction is highly telling.

Investors are starting to treat SMRs not as a distant futuristic concept, but as part of the core infrastructure of the AI economy.

Why Big Tech suddenly loves nuclear power

Just a few years ago, it was hard to imagine tech giants seriously discussing nuclear energy.

Today the picture has completely changed. Microsoft is signing nuclear power agreements. Amazon is investing in energy infrastructure for data centers. Google is actively searching for long-term stable energy sources.

The reason is simple: AI demands massive compute power. Large language models, video generation, data processing, and training all require always-on data centers with enormous energy consumption.

The world is effectively entering a phase where the main constraint on AI growth may no longer be Nvidia chips, but access to electricity. This is why nuclear power is suddenly back in focus.

But risks remain significant

Despite the excitement, the SMR market remains highly speculative. Most companies in the sector still have no meaningful commercial revenue, and many projects are closer to future promises than operational businesses.

Firms continue to burn capital, and reactor deployment timelines are frequently delayed. Nuclear energy is also one of the most heavily regulated industries globally, where licensing delays, safety concerns, or political shifts can quickly change the outlook.

There is also a classic market risk: during megatrend formation, investors often overprice even immature technologies.

But the macro trend is strengthening

Still, the key story is not quarterly earnings of individual companies, but the transformation of global energy infrastructure itself. AI is reshaping not only technology markets but the physical foundation of the economy.

If GPUs and data centers defined the early AI boom, energy is becoming the next layer.

That is why SMRs are increasingly described as the beginning of a new supercycle. Nuclear energy, once associated with the 20th century, is unexpectedly becoming part of the digital economy’s future infrastructure.

Paradoxically, artificial intelligence – the defining technology of the future – may be the force that brings one of the oldest and most controversial energy technologies back into the center of the global economy.