Is Oklo the Next Great AI Story -- Or Just Nuclear Hype?

Oklo (NYSE: OKLO) has already become an artificial intelligence story. And like all stories, Oklo's is mixed with factual and fictional elements -- the measurable with the speculative -- which can make it challenging to distinguish the genuine opportunity from pure nuclear hoopla. Buzz around the nuclear stock has started to die -- it's fallen 44% in 2026 -- which makes now a good moment to revisit its story and ask: What exactly are investors buying, and is it worth buying today?

Image source: The Motley Fool.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

The story so far: How Oklo became an AI stock

Oklo debuted on the market in 2024. But its history stretches back farther than that. Indeed, we could even go as far back as CEO Jacob DeWitte's graduate work at MIT, where he studied reactor protection systems for sodium-cooled fast reactors, the same broad class of nuclear tech that would eventually underpin Oklo's Aurora powerhouse.

As an AI story, however, a more revealing first chapter would open with something more consequential: Sam Altman.

In 2013, Altman and DeWitte met for the first time at an MIT dinner, and by 2015, the future co-founder and CEO of OpenAI had become an early investor in Oklo. In that same year, Altman became Oklo's chairman and remained in that role until 2025, when he stepped down to allow Oklo to pursue partnerships with AI companies, including OpenAI. By that time, however, the association had already done its work, and Oklo had become an AI-power story.

And what does this AI-power story amount to so far? Nothing commercial, but there have been noteworthy milestones. A list of them would have to include its roughly 18 gigawatt (GW) order book, which includes a non-binding agreement with data center operator Switch to deploy up to 12 GW of nuclear power through 2044. Adding to that is its landmark agreement with Meta Platforms, which involves developing a 1.2 GW nuclear power campus in Pike County, Ohio, to power Meta's AI data centers.

Adding to these is the White House's strong support for novel nuclear technologies -- driven in part by AI's power needs -- which has helped Oklo move its Aurora design through the federal authorization process (Aurora is not yet approved to operate). In 2025, for example, the Trump administration ordered the Department of Energy (DOE) and the Nuclear Regulatory Commission to accelerate testing of advanced reactors and create a DOE pilot program to fast-track construction of these reactors outside laboratories.

Partnerships, non-binding agreements, and federal backing have all been tangible parts of Oklo's story. But they're only the opening act, and frankly, the next act will determine whether Oklo become a nuclear success.

The harder part of Oklo's AI story

Two figures can tell us a lot about Oklo's speculative side: $8 billion (Oklo's market cap) and $1.2 million (its trailing 12-month revenue). Don't be fooled by that latter figure, though: Most of Oklo's revenue has come from businesses it has acquired; none of it is reactor revenue, and Oklo still has no firm date for when its first reactor will operate commercially.

If we dig a little deeper, the picture gets even more complicated.

Data by YCharts

Compare revenue estimates with Oklo's net income. Two fiscal years from now, analysts expect Oklo to generate about $55 million in annual revenue, which is roughly a third of its current $153 million trailing loss. With Oklo's current spending trajectory, that loss will be larger, too. Indeed, Oklo lost about $48.5 million in the second quarter of this year, up from $33 million in the first. That loss is softened by the $23 million in interest and dividend income on its large cash balance.

"But AI companies will buy!" one might argue. "All Oklo needs to do is to certify Aurora, and the sales will go through the roof."

That might be true in some capacity -- the demand for off-grid power is true, at least -- but there's one hitch to that bullish scenario: fuel.

To scale reactors, Oklo will need an adequate supply of nuclear fuel. And unfortunately, the kind of fuel it needs is also the hardest to find. Only one U.S. facility is licensed to make high-assay low-enriched uranium (HALEU), and Oklo isn't the only company trying to get its hand on it.

It is true that investing heavily in fuel recycling could eventually give Aurora another fuel source. But that only accentuates the problem I'm getting at: scaling reactors (and fuel) is expensive business (investment for Oklo's Tennessee fuel center is estimated at $1.7 billion), and Oklo investors may have to wait years before that capex turns into revenue.

Oklo isn't all hype, but the business needs time to grow

I've been an investor in Oklo for a long time. And while I know the short term will be messy -- the losses won't be pretty -- I trust Oklo's management to steer through the uncertain middle. That said, investing in Oklo today means accepting an AI story that is still being written. That may be worth the risk for some, but not everyone should feel compelled to buy the stock today. There's nothing wrong with waiting until the nuclear side of its story gets past the first draft before buying.

Steven Porrello has positions in Nvidia and Oklo. The Motley Fool has positions in and recommends Meta Platforms and Nvidia. The Motley Fool has a disclosure policy.