Oklo Inc. OKLO

Key data

The business

Oklo designs the Aurora, a small fast-fission "powerhouse" that starts around 15 megawatts and scales up, cooled by liquid metal and descended from an old Idaho test reactor. The twist that defines the company is the sales model: it does not sell reactors, it plans to build and own them and sell the electricity under multi-year power contracts, the way an independent power producer does. Its customers so far are letters of intent and framework agreements, weighted toward data-center operators desperate for round-the-clock power and a military base in Alaska. In June 2026 Oklo bought three engineering and fuel-services firms, ARMEC, Precision Manufacturing Engineering Solutions, and Creative Engineers, which is where the first $1.2 million of revenue comes from. The moat, if it earns one, is a granted regulatory license plus secured fuel and a site, none of which it holds in commercial form yet. Everything rests on a first plant that does not exist.

The numbers

There is no earnings story here, only a burn story and a funding story, and both are the whole point. The last five quarters:

The loss roughly doubled in a year as Oklo hired, engineered, and absorbed acquisitions. The full-year picture shows the same slope, steeper each year:

Notice the loss per share barely widened even as the loss doubled, because the share count ballooned. That dilution is the second story, and it funded the first. Here is the cash against the burn:

Cash went from $90 million at the end of Q1 2025 to $1.64 billion at the end of Q2 2026, a fifteen-month rise funded by selling stock into a rising market, and that is why the company is funded for years of operating losses. The catch hides in what the cash has not yet done: capital spending was $33 million for all of 2025, a rounding error against what a fleet of reactors costs. The money is raised, the construction has barely begun, and the gap between those two facts is the timeline to first power. What this memo believes that the price does not is plain: the burn and the build are still ahead, not behind, and the single print that settles the story is not an earnings number at all but the regulator's decision on the first plant.

Management

Co-founder and chief executive Jacob DeWitte and finance chief Craig Bealmear were sellers, not buyers: about $60 million sold across 49 filings in twelve months, zero purchases, DeWitte's stock leaving in steady monthly slices that look automatic but whose plan status is not disclosed in the filings I can see. There is no profit to judge pay against, so compensation runs on stock, which feeds the same dilution the share count already shows. The team's proven record is capital-raising, and at that they have been effective, turning a hot story and a rising stock into a war chest. That is a real skill for a company at this stage. It is not the same skill as building reactors on budget, which remains untested.

How it fails or surprises you

The license does not come, or comes late. The regulator turned Oklo's first application away in 2022 for thin detail, and the company has not yet cleared a resubmission. If approval slips past 2028 or arrives narrowed, the whole revenue timeline moves right and the cash burns against nothing. Watch the licensing docket for the first Aurora.

Dilution never stops. The $1.6 billion does not build a fleet; the first commercial plants cost billions more, and with no earnings that money comes from selling more stock, likely below the old highs. Watch shares outstanding each quarter against capital spending. Rising shares with flat capex is the bad combination.

Data-center demand converts (right tail). Hyperscalers are signing for firm power years out and paying up for it. If even a slice of Oklo's gigawatt-scale interest list becomes binding contracts and the first plant runs, a build-own-operate fleet throws off decades of contracted cash the market prices at almost nothing today. Watch for the first binding power contract with a construction start date attached.

Closing thoughts

Nothing in the next four quarters settles this. It is a survival-and-permission story: whether the cash lasts to a licensed, operating first plant, with everything else noise until then. The near-term tail is fatter on the downside, delay and dilution, because those are the default paths for a pre-revenue builder, and the market has already marked the stock down almost 80% from its high to say so. The right tail is genuinely large and genuinely unpriced, but it pays out on no schedule you can underwrite. What is at risk if licensing breaks is most of the equity; what the upside is worth is a power producer with a multi-decade contracted book, and only one of those is knowable in advance.

The bet is still that Oklo builds small nuclear power plants, owns them, and sells the electricity to data centers and the military. It breaks if the regulator says no or if the cost of building forces endless share issuance, and it wins if the first plant clears its license and runs. The one pair that tells you which way it is going, quarter by quarter, is the licensing milestone on the first Aurora set against capital spending rising into that cash balance. Everything else on this page is a way of waiting for those two numbers.