What Is Oklo? The Sam Altman Nuclear Startup Explained

• The short answer
• What Oklo actually does
• The Sam Altman and SPAC story
• Where Oklo stands now
• Customer pipeline and data-center deals
• The risks and the skeptic's read

The short answer

Oklo is a U.S. nuclear startup building small "Aurora" microreactors and planning to sell the power they make. The company was chaired until April 2025 by Sam Altman, the chief executive of OpenAI. Oklo is not a big utility. It is a young, pre-commercial firm with a bold idea and a famous backer.

Oklo wants to power the AI boom. Data centers need huge amounts of steady electricity. Oklo says small nuclear reactors can supply it. So far, the company has no operating reactor and no revenue. It has plans, deals on paper, and a lot of investor attention.

That mix of hype and unproven technology is why Oklo is worth a careful look. This page explains what Oklo does, how it went public, where its first reactor stands, and where the real risks sit. It does not tell you what to do with the stock.

What Oklo actually does

Oklo designs small nuclear "powerhouses" called Aurora and plans to own and run them, selling the electricity to customers. This is different from most reactor makers. Oklo does not just want to build and sell reactors. It wants to be the power company too.

The Aurora is a microreactor. Early versions are designed to make roughly 15 to 75 megawatts of electricity, with larger versions planned. That is small next to a normal nuclear plant, which can top 1,000 megawatts. The idea is to place these compact units close to the customers who need power, such as data centers or industrial sites.

Aurora runs on HALEU, or high-assay low-enriched uranium. This fuel is more enriched than the fuel in today's big reactors. Oklo also says its design can run on recycled nuclear waste. In April 2026, Oklo signed a letter of intent with Centrus Energy to supply HALEU for up to five Aurora powerhouses, with deliveries planned to start in 2029.

Oklo's business model is called "build, own, operate." Customers sign long-term power purchase agreements. Oklo builds the reactor, keeps it, and bills for the electricity over many years. If it works, that means steady income for decades. If it does not, Oklo carries the cost and the risk. That model is the whole bet: Oklo sells power, not reactors.

Oklo has also stretched beyond just reactors. The company says it is building a nuclear fuel recycling business and wants to produce radioisotopes used in medicine and industry. In plain terms, it hopes to make money from the fuel as well as the power. That widens the story, but it also spreads a small company across many hard, first-of-a-kind projects at once.

The Sam Altman and SPAC story

Oklo went public in May 2024 by merging with AltC Acquisition Corp, a shell company tied to Sam Altman, and now trades on the NYSE as "OKLO." Jacob DeWitte and Caroline Cochran founded Oklo in 2013. Altman later became chairman and helped take it public. He stepped down as chairman in April 2025 to free Oklo to sign power deals with OpenAI and other AI firms; co-founder Jacob DeWitte became board chair.

A SPAC, or special purpose acquisition company, is a shell firm that raises money and then merges with a real business to bring it to market. AltC Acquisition Corp was that shell. The deal gave Oklo about $306 million in gross proceeds, per company disclosures. Shares began trading on the New York Stock Exchange on May 10, 2024.

The debut was rough. On its first trading day, the stock fell about 54%, CNBC reported. Since then, the share price has swung sharply, both up and down. This page is not investment advice. It explains the company and the risks. It does not tell you to buy or sell any stock, and it names no price target.

The Altman link cuts both ways. It brings attention, money, and credibility. It also raises questions. Altman runs OpenAI, a company that needs enormous power for AI. Critics ask whether the excitement around Oklo has run ahead of its real progress. A famous name is not the same as a working power plant.

Where Oklo stands now

Oklo has no operating reactor and no revenue yet, but its first Aurora is under construction at Idaho National Laboratory, with first power targeted for late 2027. The road to this point was not smooth.

In January 2022, the U.S. Nuclear Regulatory Commission ended its review of Oklo's first Aurora license application and denied it. The NRC dismissed the application "without prejudice." That means Oklo was allowed to try again. The regulator said Oklo had not provided enough safety information.

Oklo changed course. Its current Aurora is moving through a U.S. Department of Energy pilot program, not the standard NRC path. Oklo broke ground at Idaho National Laboratory in September 2025, per the American Nuclear Society. In March 2026, the DOE's Idaho office approved a Nuclear Safety Design Agreement for the plant. In May 2026, the NRC approved a key design-criteria report on an accelerated schedule.

The DOE pilot program is meant to speed up advanced reactors by letting them be built and tested on federal sites. Backers call it a smart shortcut around a slow system. Others worry it moves fast on designs that have not been proven. Either way, it is a different route than the one a normal commercial reactor must take.

These are real steps forward. But a design report and a groundbreaking are not a running reactor. Oklo still must build, fuel, test, and safely operate the plant. For later commercial projects, it will also need full NRC licenses. Nuclear licensing for a new design takes years, and the outcome is not guaranteed.

Customer pipeline and data-center deals

Oklo reports a customer pipeline of around 14 gigawatts, but most of it is non-binding letters of intent, not firm contracts. The demand story centers on AI. Data centers run day and night, so they want power that never stops.

The biggest name is Switch, a data-center developer. Oklo and Switch signed a letter of intent for up to 12 gigawatts of power over several decades, Utility Dive reported. That is a very large number. It is also non-binding, which means either side can walk away.

Oklo also has a deal with Equinix. Equinix made a $25 million pre-payment tied to a 20-year power agreement for up to 500 megawatts, according to Power Engineering. Oklo lists Meta among its interested customers as well.

Here is the key point. A letter of intent is a promise to keep talking, not a signed sale. Oklo has reported zero revenue. The pipeline shows strong interest in nuclear power for AI. It does not yet show cash coming in the door. Big numbers on a slide are not the same as money in the bank.

The risks and the skeptic's read

Oklo is a pre-revenue company selling an unbuilt product, so the main risks are technology, licensing, timing, and hype. None of these are small, and any one of them could delay the whole plan.

First, the technology is unproven at scale. No commercial Aurora has ever run. Fast reactors like Oklo's design have a long and troubled history around the world. Building one on time and on budget is genuinely hard.

Second, licensing risk is real. The NRC rejected Oklo once already. The DOE pilot path is faster, but commercial reactors will still face full NRC review. Delays are common in the nuclear industry, and they cost money.

Third, the cash burns fast. Oklo reported a net loss of $33.1 million in the first quarter of 2026 and no revenue, per its filings. It held about $2.5 billion in cash. That funds years of work, but the clock is ticking toward first power in 2027, and reactors often run late.

Fourth, watch the hype. A famous chairman and an AI-power story can lift a stock far ahead of the physical build. The share price has been very volatile. Skeptics warn that the story may run ahead of the plant. Again, this is not investment advice. It is a reminder to separate the pitch from the progress.

There is also a dilution point worth knowing. To raise cash, Oklo has sold large amounts of new stock, including more than $1 billion through a share-sale program in early 2026. Selling new shares funds the work, but it also splits ownership into more pieces. So even if Oklo succeeds one day, early believers may own a smaller slice than they expect. For a company with no revenue, funding the years ahead is a real and ongoing challenge.

Want the bigger picture? Start with our pillar on nuclear power for data centers, then see how Oklo compares in Oklo vs NuScale. You can also read about the nuclear companies powering AI, who is investing in nuclear for AI, and whether small modular reactors are safe. To see where all this power is headed, explore our data center map and the growing AI backlash.