TerraPower Has Gates' $1B Backing-But the Next Oklo Trade Is Still Indirect

TerraPower has the story, but investors still have to play it indirectly

TerraPower is becoming one of nuclear's most compelling AI-power stories. Bill Gates has poured more than $1 billion into the company, and earlier this year the NRC granted the first construction permit for a commercial non-light-water reactor. That moves TerraPower well beyond theory.

The catch is simple: you can't buy privately held TerraPower. Retail investors cannot buy the company directly, and the Kemmerer project is not expected to produce first power until around 2030. That keeps the theme alive and active, but it also means any stock exposure right now is indirect.

That is why the real investor question is no longer whether advanced nuclear is real. It is how to participate while the market keeps testing which public companies benefit most from the same AI-driven power demand.

The March permit moved TerraPower from concept to execution

In March, the NRC authorized the first commercial reactor construction approval in nearly a decade, and the first construction permit for a commercial non-light-water reactor in more than 40 years. For the advanced-nuclear theme, that is a real milestone. The discussion is now less about whether the technology could work and more about whether it can be built on schedule.

Why the permit matters

TerraPower received the permit for Kemmerer Unit 1, the first commercial-scale advanced nuclear plant to get such approval. That shifts the story from design promise to project execution.

The reactor design also helps explain the demand case. The Natrium system is a 345-megawatt electric sodium-cooled fast reactor paired with a molten salt storage system that can boost output to 500 megawatts when needed. For data-center developers and grids facing tighter power constraints, that combination of firm, high-density power and added flexibility is central to the thesis.

What still has to be proven

The permit strengthened the story, but it did not settle it. TerraPower has said it plans to start construction in the coming weeks, and first power is still years away. Over the next 12 to 24 months, the key questions are whether construction starts on time, whether costs stay controlled, and whether the Bechtel partnership supports reliable execution.

The investable angle is the broader AI-power and nuclear trade

Since TerraPower itself is not publicly tradable, the more useful public-market approach is to look at who benefits if demand for AI-era power keeps rising. TerraPower helps validate the theme, but the tradeable exposure is still indirect.

Meta shows the demand is becoming real

Meta is working with TerraPower and Oklo while also buying power from Vistra, and it wants those new reactors up and running between 2030 and 2032. That timeline matters because it suggests hyperscalers are already planning around firm, round-the-clock power supply.

Funding and fleet language are improving the narrative

TerraPower also closed a $650 million fundraise which included investment from Nvidia. That helps reinforce the idea that compute, power, and grid constraints are increasingly being viewed as linked bottlenecks.

Management is also describing Kemmerer Unit 1 as a commercial blueprint for future Natrium deployments. One plant alone does not create a fleet, but it does make the repeatability argument easier for investors to take seriously.

What to watch instead of a direct TerraPower stock

Because there is no direct share class, the clearer signals are operational and commercial, not purely narrative:

• Oklo and peers: more customer announcements, funding updates, and evidence that they can share the same hyperscaler power window.
• Off-take deals: power agreements and data-center partnerships matter more than design slides because they de-risk demand.
• Supply-chain spill-over: transformers, rotating equipment, containment, and niche engineering firms may see demand before reactor vendors fully monetize it.
• Liquidity changes: any IPO, SPAC, or project-level financing structure would make TerraPower exposure more direct and more tradable.

For now, the theme looks stronger, but the access is still indirect. The better approach is to focus on milestones-construction starts, customer wins, supplier commitments, and any path to public-market liquidity-rather than on legend-building alone.