Oklo reported second-quarter revenue of $1.2 million against a consensus of $83,800 — a beat of roughly fourteen times on a number so small it barely registers as revenue. The loss per share was $0.28 against $0.16 expected, and the net loss doubled to $48.5 million from $24.7 million.
The stock opened about 7% higher, extended through the session, and finished up roughly 15%, closing the week above $48. It remains down about 45% year to date.
This is what a milestone-driven asset looks like, and it is worth understanding on its own terms rather than through an ordinary earnings lens.
The quarter
| Metric | Q2 2026 | Expected |
|---|---|---|
| Revenue | $1.2M | $83,800 |
| Loss per share | −$0.28 | −$0.16 |
| Net loss | $48.5M | $24.7M a year earlier |
| Liquidity | ~$3B | — |
Why the loss did not matter and the reactor did
Oklo's Groves Isotope Test Reactor in Texas achieved first criticality — a self-sustaining chain reaction — less than a year after construction began.
For a company whose entire valuation rests on the proposition that it can build fast-fission plants faster and cheaper than the industry consensus believes possible, this is the only category of news that carries information. Revenue of $1.2 million tells you almost nothing about a business that intends to sell power by the gigawatt. A reactor going critical inside twelve months tells you the execution claim is not purely theoretical.
The doubled loss is affordable in a specific and checkable sense: roughly $3 billion of liquidity against this burn rate is many years of runway. The company is not solvency-constrained; it is timeline-constrained.
The honest framing of the risk
Our team's view is that Oklo is not an earnings story and should not be sized like one. It is a sequence of binary technical and regulatory events, each of which either happens on schedule or does not.
That has three consequences an investor should price:
Volatility is the product, not a defect. A stock down 45% year to date that adds 15% on a single milestone is behaving exactly as designed. Position sizing, not conviction, is what determines whether that is survivable.
Cash does not equal certainty. Three billion dollars removes the risk of dying before the thesis is tested. It does not shorten the licensing timeline, which is the actual bottleneck for commercial deployment in the United States, and which no balance sheet can accelerate.
Being chosen is not the same as being paid. Oklo's selection for a US programme supporting reactors for AI infrastructure is a genuine catalyst and a genuine validation. It is also, at this stage, an agreement to try. Converting programme selection into contracted, delivered megawatts is the step where most of the industry's history has stalled.
What the next print has to show
- The next milestone landing on the stated timeline. The company's whole differentiator is schedule.
- Revenue that is contractually tied to deployment, not test services — the difference between a research operation and an energy business.
- Burn scaling with construction, not overhead. A widening loss driven by building reactors is investment. A widening loss driven by corporate cost is dilution waiting to happen.
The reactor is real. The business model is still a forecast.