Oklo Stock: The Nuclear Bet That Could Power AI’s Trillion-Dollar Future—Or Burn Through Cash First

Naming your nuclear startup after a location in Gabon where uranium spontaneously achieved fission 1.7 billion years ago has an almost poetic quality. It implies aspiration. Perhaps hubris. Maybe both. The foundation of Oklo Inc., a Santa Clara-based company that designs small modular reactors, is the conviction that nuclear, compact, and quick energy will be the energy of the future, particularly the kind required to satisfy artificial intelligence’s voracious appetite.

Nevertheless, the business doesn’t make any money. In 2025, it reported a $105 million loss. In just six months, its stock has dropped 56%. Even after the numbers were made clear, insider selling caused anxious chatter among retail investors. Its initial application for a license was rejected by the Nuclear Regulatory Commission. No construction has begun. It is anticipated that commercial operations won’t begin until 2028, assuming all goes according to plan.

Investors continue to show up, though. The implied volatility of the stock is close to 94.58, indicating that the market anticipates significant fluctuations. There are currently about 15 gigawatts in the order book. Agreements have been signed by Meta and Switch.

There are rumors that Sam Altman’s OpenAI is considering a power deal. Oklo was selected for the Department of Energy’s Reactor Pilot Program. In Oak Ridge, Tennessee, the company is quietly and tenaciously constructing the first commercial fuel recycling facility of its kind.

The contradiction is difficult to ignore. Oklo is oddly captivating and profoundly troubled. The question is not whether the technology could function, but it most likely could with sufficient funding and time. Whether the business can outlive its own balance sheet is the question.

Jacob and Caroline DeWitte, the company’s founders, are not like other energy executives. They are young, technically proficient, and not scared to present an idea that sounds more like science fiction than practical infrastructure.

Their reactor, the Aurora powerhouse, is built to run on recycled nuclear fuel, generate 15 to 75 megawatts of electricity, and function in locations that the grid cannot reach, such as remote industrial sites, data centers, and military installations. It is a liquid metal-cooled fast reactor that has been tested in different configurations since the 1950s but has never been mass-produced in the US.

The appeal is clear. AI firms are in dire need of dependable, carbon-free baseload power. Data centers cannot operate solely on sporadic wind and solar power. Density and consistency are provided by nuclear. Additionally, you may have a business if you can make the reactors smaller, quicker to deploy, and less reliant on conventional uranium enrichment.

However, the cost of transportation is high. In addition to $450 million in capital expenditures, Oklo anticipates spending between $80 million and $100 million on operating costs in 2026. The rate of cash burn is increasing. In May 2024, the company went public through a SPAC merger, generating $306 million in gross proceeds. The cushion is getting smaller. Quick.

It was not helped by the insider selling. According to early reports, executives and co-founders had sold about $21 million worth of shares. The discrepancy was mostly caused by the way family trusts were reported; the real amount was closer to $10.9 million. The timing, however, drew criticism. Even normal liquidity events appear suspicious when a company is losing money and its stock is sharply declining.

In order to protect executives from allegations of insider trading, Jacob DeWitte and CFO Richard Bealmear sold shares under pre-established Rule 10b5-1 plans. However, investors are aware that plans can be set up and changed. There is some comfort in the fact that insiders still own sizable stakes, mostly through trusts. Not much, though.

The idea that Oklo is just early is what sustains the faith, at least for some. The business isn’t attempting to take on well-established utilities. In a world where AI workloads double every few years, it is attempting to create a completely new type of energy infrastructure, one that does not yet exist but may be crucial. In that case, being first might be more important than making money in 2026.

The regulatory route is still unclear. The NRC’s January 2022 rejection was a request for additional information rather than a categorical denial. Since then, Oklo has finished Phase 1 of the pre-application readiness assessment and is employed at Idaho National Laboratory under Department of Energy supervision. By July 4, 2026, the company hopes to reach criticality, which is the point at which the reactor produces a self-sustaining chain reaction. The deadline is lofty and seems more symbolic than practical. However, ambition is a component of the brand.

A more general change is taking place outside of the nuclear sector. Power purchase agreements are being signed by large tech companies with developers that they previously disregarded. Microsoft is supporting efforts to bring Three Mile Island back to life. Small modular reactors are being purchased by Amazon. Fusion is being investigated by Google. The internet’s power infrastructure is being rebuilt, and nuclear is once again being discussed.

Oklo is in the middle of this transition, neither established nor unimportant. Although it doesn’t have any operational plants, it has partnerships. Orders exist, but there are no deliveries. Although it lacks regulatory approval, it has technology. Additionally, it has money, but not enough to last forever.

A long-term play is seen by some analysts. Some perceive a cash furnace. They could both be correct. In a slow-moving regulatory environment, the company is spending money to create something that doesn’t yet exist for a market that might take years to fully materialize. Quarterly stability cannot be achieved with that. It’s a recipe for either quiet failure or enormous success.

As this develops, it’s easy to draw comparisons between Oklo and Tesla ten years ago—another capital-intensive, pre-profit business that placed all of its bets on a technology that the market wasn’t ready for. Tesla made it through. Many others, however, didn’t. Execution, timing, and whether the world is changing quickly enough to meet the vision halfway are usually what make a difference.

For the time being, Oklo is still a risky bet on the nexus of artificial intelligence and nuclear physics. The technology is genuine. The necessity makes sense. The market is intrigued. However, the route to profitability is costly, time-consuming, and fraught with regulatory obstacles. Investors who wager on Oklo are not purchasing a business. They are investing in the idea that Oklo will still be in business when the future demands what the company is creating.

The validity of that belief is still up for debate. However, it’s difficult to deny that people continue to believe in spite of everything.