Oklo Rose 22.7% on $1.86 Million of Services Revenue: The AI-Power Trade Now Pays for Milestones, Not Meters

The quarter's cash statements matter more than its income statement, and both point the same way. Oklo used $45.5 million of operating cash and $154.0 million of capital expenditure in the September quarter, for $199.5 million of free cash outflow, and covered it with $402.3 million of stock issuance. Operating expenses of $83.2 million — $41.6 million of research and development and $41.6 million of general and administrative — ran against $1.86 million of revenue.

Interest income of $27.9 million did more to shape the headline loss than anything the company sold, which is the clearest signal that this is a financing story right now, not an operating one.

The mechanism is regulatory sequencing, and it changed three weeks before the print. On Oct. 22 the Nuclear Regulatory Commission finalized a rule amending 10 CFR Parts 50 and 53 so that applicants can lean on analyses, test data and other information from Energy Department- and War Department-authorized demonstration reactors when they apply for a commercial license — 136 unique comments in, and effective 30 days after Federal Register publication, which Oklo's own filing puts at Nov. 25, 2026.

That single change converts everything Oklo has done on the Energy Department side into licensing currency. The Aurora powerhouse under the Department's Reactor Pilot Program, the Groves test reactor's startup authorization and criticality, the preliminary documented safety analysis approved for Aurora-ID on June 11, 2026 — all of it becomes evidence a commercial applicant can reuse. The company no longer needs the NRC's patience as much as it needs a calendar it can point at.

Set the balance sheet against the market value and the arithmetic becomes stark. At the Nov. 6 close, Oklo's market capitalization was about $4.38 billion against $3.13 billion of cash and marketable securities, leaving roughly $1.25 billion of value attached to a business with $3.07 million of trailing revenue. Strip out the cash and the option on the whole pipeline is priced at roughly 400 times sales.

That is not a balance-sheet problem; it is a financing plan. Oklo raised about $2.26 billion of equity in nine months, and the weighted-average diluted share count rose28.8% to 188.6 million. Every quarter of waiting is paid for in stock, and the stock is down about 75% year to date, so the issuance is happening at a discount to where the company was valued twelve months ago.

Meta's January basket is the cleanest illustration of the structure. The company committed to up to 6.6 GW of nuclear capacity: more than 2.1 GW of life extension and uprates from Vistra's Perry and Davis-Besse, 2.8 GW of TerraPower Natrium capacity, and up to 1.2 GW from Oklo in Pike County, possibly online as early as 2030 and fully targeted by 2034. The first two sell electrons in the late 2020s; the third sells a license that does not exist yet.

The market knows the difference and prices it accordingly. Across the chain, the businesses that already sell electrons or the hardware to build them change hands at roughly two to five times trailing sales, while the two listed developers change hands in the hundreds of times.

The facts are narrow and checkable: $1.86 million of revenue, a $41.7 million net loss, $3.13 billion of cash and marketable securities, $402.3 million of stock sold in the quarter, a final NRC rule effective Nov. 25, a forging reservation with deliveries from 2030, a fuel letter of intent with deliveries from 2029, 900 megawatts under interconnection consideration, and a 750 megawatt project withdrawn from the first study cycle with the complaint denied.

The inference is that the market has re-rated this leg on de-risking steps rather than on power, and that this is a policy transmission channel: hyperscaler capital is underwriting federal policy that in turn underwrites developer equity, and the developer equity is the most expensive, most dilutive place to hold that view. The further inference — that any of it converts to cash — is speculation with a 2029 fuel date and a 2028 safety analysis in front of it.

What breaks the case is specific: a license application that does not arrive, fuel that does not move, a queue that does not reopen, and dilution that continues at a quarter a year while the stock sits75% below its highs. What would strengthen it is the opposite: the first application built on Department of Energy data, a definitive fuel agreement with prepayment, and a PJM cluster study that converts900 megawatts into a firm position.