Report contents
Business model
Business model: an owner-operator with no operating plant
Oklo sells electricity, not reactors. It designs small fast-spectrum nuclear plants, which it calls Aurora powerhouses, and intends to build, own, and operate them, selling power under long-term contracts to data center operators, industrial users, and federal sites in the United States. The buyer pays for megawatt-hours, not hardware, which makes Oklo a project developer aiming to become an independent power producer. Type and stage are the whole story here: an unproven, pre-revenue developer whose value rests on one licensed design. The business that matters most is the first unit. Until a plant is financed, built, and delivering power, the announced pipeline is option value.
The customer buys power, and that moves Oklo down the value chain
Selling electricity rather than reactors puts Oklo in the asset-owning position, where the lifetime cash flow sits. A vendor books revenue on delivery and hands construction risk to the buyer. Oklo keeps the plant and the decades of payments, but it must raise project capital, absorb overruns, and wait years for the first invoice. Owning is the better business only if construction costs fall with repetition. Otherwise the company carries utility-style capital intensity without utility-style certainty.
The demand is specific. Data center operators need firm, carbon-free power faster than grid interconnection queues allow, and they will pay for speed and reliability. That is a different buyer from an incumbent utility procuring baseload generation, and it changes what the product has to be: a plant sized to a single large load, deliverable on a committed date. The tradeoff is price against volume. A first-of-a-kind plant with no operating history has little bargaining power, so offtakers can demand concessions, termination rights, or project equity. Announced customer arrangements to date have generally been framework or non-binding agreements rather than executed contracts with disclosed pricing. Interest is visible; pricing power is not yet.
Replication is the investment case, and construction is the weak link
Oklo's margin story depends on the second, fifth, and tenth plant costing less than the first. That is a manufacturing learning curve applied to a construction site, and it is not guaranteed. The company owns the design and the site relationships but depends on suppliers and contractors for the physical build. If unit costs stay flat, Oklo looks like a contractor carrying nuclear liability. If they fall, it owns a scarce, licensed product in a market where few alternatives deliver firm clean power.
Competition sits on two sides. Other advanced reactor developers are chasing the same data center buyers, and gas turbines plus delivered grid power set the price ceiling. Time-to-power decides both races, which is why licensing dates matter more to the commercial story than engineering elegance.
How it gets paid shapes who will fund it
Revenue would arrive as contracted payments for capacity and energy over terms long enough to amortize a plant. That structure is what makes nuclear financeable, because lenders lend against the contract rather than the technology. Oklo therefore needs offtakers with strong balance sheets, and it needs them before construction rather than after. A merchant plant selling into a spot market would carry a completely different risk profile and a higher cost of capital.
Licensing sets the schedule, and the schedule sets the economics
The first Aurora is planned at Idaho National Laboratory, and Oklo's combined license application is with the NRC. A combined license covers construction and operation in a single review, so the regulatory calendar gates revenue by years. Each year of delay raises capitalized overhead, pushes the contract start date out, and gives rivals time to reach the same customers. Oklo does not publish a costed plant design or a priced offtake agreement, so the return on the first unit cannot be tested directly from public disclosure.
Fuel is the second business and the real bottleneck
Most advanced reactors need HALEU (high-assay low-enriched uranium, enriched to between 5% and 20%), and domestic supply is thin. Oklo plans fabrication and recycling capacity so it can fuel its own fleet and eventually sell to others. That is a separate business: capital-intensive, separately licensed, and slow to permit. It could become the strongest part of the franchise if Oklo controls fuel while rivals queue. It could also drain cash before any reactor needs it.
What would change this view: a priced contract and a costed plant
The clearest evidence would be an executed power purchase agreement with a creditworthy buyer at a disclosed price, paired with a plant cost estimate that supports a return above Oklo's cost of capital. An NRC milestone on schedule plus a fixed-price construction contract would carry similar weight. A run of schedule slips, framework agreements that never convert, and capital raises that fund overhead rather than equipment would point the other way.
Questions worth following
- Will Oklo convert announced customer frameworks into binding contracts?
Letters and master agreements with data-center operators cost customers little and obligate no one. A definitive power purchase agreement with a creditworthy buyer is what lets a project be financed and turns pipeline megawatts into dated revenue. Without one, the order book stays an option.
- Can Oklo build its first Aurora plant on schedule?
Delivery on schedule would show Oklo can convert a licensed design into an operating asset, the step that turns announced capacity into revenue. Slips raise the capital needed before first power and weaken the cost curve the fleet case depends on.
- Can Oklo secure HALEU and license its fuel plant in time?
Fuel is the binding constraint behind the reactor schedule. Domestic HALEU output is at demonstration scale, and Oklo's plan rests on a fabrication facility and a recycling process that each need approval. A fuel gap stops the first core from loading and sets lifetime generating cost.
Financial quality
Financial quality: liquidity raised, returns unproven
Oklo is a pre-revenue reactor developer, not yet an independent power producer. Its September 30, 2026 Form 10-Q describes a company still designing, licensing, building and eventually operating powerhouses and fuel facilities. At that stage the usual quality tests split: the balance sheet is unusually strong, the income statement says almost nothing about earning power, and the only financial question that matters is whether roughly $2.7 billion of liquid resources carries the first units to revenue before another equity sale.
The loss line measures spending, not profitability
Revenue rounds to nil, so a net margin of -2,243% and a gross margin of 41% in the September quarter are arithmetic, not economics. The gross margin comes from a sliver of non-electricity revenue and says nothing about the price Oklo will eventually get for power. Returns are equally uninformative: the 103% ROE reported for FY2023 and the -68% for FY2024 both came out of an equity base that was negative and then tiny, and the -1.2% quarterly ROE now reflects a $3.65 billion equity base absorbing a $40 million loss.
What is measurable is the spending trend. Operating loss went from $50 million in FY2024 to $140 million in FY2025 and $200 million in the first nine months of FY2026. Operating cash use over those nine months was $110 million, roughly half the operating loss, which points to noncash charges, most likely stock-based compensation, carrying the gap. That gap matters for how the loss reads: the cash drain is smaller than the loss suggests, but the compensation is paid in shares, so it returns as dilution.
| USD bn | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 | FY2026 9M | Assessment |
|---|---|---|---|---|---|---|---|
| Revenue | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | Very weak: no product revenue |
| Operating income | -0.01 | -0.01 | -0.02 | -0.05 | -0.14 | -0.20 | Weak: spend widening |
| Net income | -0.01 | -0.01 | -0.03 | -0.07 | -0.11 | -0.12 | Weak |
| Operating cash flow | -0.00 | -0.01 | -0.02 | -0.04 | -0.08 | -0.11 | Weak but contained |
| Free cash flow (reported) | -0.00 | -0.01 | -0.02 | -0.04 | -0.11 | -0.39 | Weak: investment phase |
| Investing cash flow | -0.50 | -0.00 | -0.00 | -0.18 | -0.49 | -1.30 | Portfolio purchases dominate |
| Financing cash flow | 0.51 | 0.01 | 0.02 | 0.30 | 1.26 | 2.25 | Very strong: equity-funded |
| Total assets | 0.01 | 0.01 | 0.01 | 0.28 | 1.53 | 3.77 | Strong, from issuance |
| Total liabilities | 0.01 | 0.04 | 0.05 | 0.03 | 0.05 | 0.12 | Strong: minimal leverage |
| Liabilities/assets (%) | 162.1 | 128.0 | 330.8 | 11.0 | 3.3 | 3.2 | Strong now; shell-era ratios void |
| Cash | 0.01 | 0.01 | 0.01 | 0.10 | 0.79 | 1.55 | Exceptional liquidity |
| Current ratio (x) | 23.9 | 16.5 | 4.4 | 43.5 | 49.1 | 31.2 | Exceptional, financing-derived |
| ROE (%) | -81.9 | 91.9 | 103.0 | -68.0 | -12.2 | -1.2 | Not meaningful |
*Flows are nine-month sums; balance-sheet items are at September 30, 2026. Ratios for FY2026 are the reported third-quarter figures. FY2021–FY2023 reflect the shell-stage entity, with $10 million of assets and negative equity.
Solvency is not the issue; funding dependency is
Liabilities of $120 million against $3.77 billion of assets, and a 31x current ratio, are not the marks of a fortress. They are the marks of a company that has sold stock. Financing inflows were $2.25 billion in the first nine months of 2026, and equity rose from $1.48 billion at December 31, 2025 to $3.65 billion at September 30. Cash stood at $1.55 billion, with about $1.17 billion more in non-cash current assets, mostly short-term investments. Interest expense is nil and total liabilities are mostly working capital. The register shows 195,289,593 Class A shares outstanding at November 4, 2026; that is the price of the liquidity.
Investing outflows of $1.30 billion over nine months look alarming until they are decomposed. The gap between operating cash flow (-$110 million) and reported free cash flow (-$390 million) implies about $280 million of capital spending, so a large share of the remaining outflow went into the securities portfolio sitting in current assets. Real project spending is therefore a fraction of the headline investing number, but it is accelerating: roughly $30 million in the March quarter, $90 million in June, $160 million in September.
| USD m | Revenue | Net income | Op. cash flow | Capex (implied) | Net PP&E | Total assets | Liabilities | Cash | Current ratio | Assessment | Trend |
|---|---|---|---|---|---|---|---|---|---|---|---|
| FY2025 Q1 | ≈0 | -10 | -10 | ≈0 | n.d. | 300 | 30 | 90 | 36.2 | Thin but debt-free | Starting point |
| FY2025 Q2 | ≈0 | -30 | -20 | ≈0 | n.d. | 730 | 30 | 230 | 71.3 | Funded by issuance | Cash up sharply |
| FY2025 Q3 | ≈0 | -30 | -20 | ≈0 | n.d. | 1,250 | 40 | 410 | 67.5 | Ample liquidity | Assets climbing |
| FY2025 Q4 | ≈0 | -40 | -30 | 30 | n.d. | 1,530 | 50 | 790 | 49.1 | Ample, pre-revenue | Capex begins |
| FY2026 Q1 | ≈0 | -30 | -20 | 30 | n.d. | 2,700 | 70 | 1,590 | 59.9 | Strong after $1.2bn raise | Cash jumps |
| FY2026 Q2 | ≈0 | -50 | -50 | 90 | n.d. | 3,360 | 80 | 1,650 | 48.5 | Strong; burn rising | Capex accelerating |
| FY2026 Q3 | ≈0 | -40 | -40 | 160 | n.d. | 3,770 | 120 | 1,550 | 31.2 | Strong; spending steps up | Cash dips |
FY2024 was disclosed on an annual basis only, so seven standalone quarters are shown. Capex is derived as the difference between operating cash flow and reported free cash flow and is approximate.
Capital spending with no return to test
Capital intensity without revenue produces no ROIC, no asset turnover and no capex-to-sales ratio worth quoting. What can be checked is whether the money is landing in construction in progress and whether the licensing calendar holds, because a license delay stretches the same spending over more quarters. On recent rates, about $110 million of operating cash use over nine months plus a capex pace that reached $160 million in the September quarter alone, the liquid balance covers roughly three to five years. The range is wide because first-of-a-kind construction is where estimates usually break, and project-level capital costs for the first unit are not disclosed with enough precision to narrow it.
Three things worth checking
- The split inside investing. Purchases of property and equipment should be separated from purchases of marketable securities. If much of the $280 million implied capex is capitalizable development work rather than steel and concrete, the physical asset base is thinner than the cash outflow implies.
- The licensing docket. A combined license application for the first Aurora unit is the gate between spending and revenue. A schedule slip changes the runway arithmetic more than a change in quarterly operating burn does.
- Issuance cadence. With $2.25 billion raised in nine months, the question is not whether Oklo can fund itself today but how much of the next construction phase is sold in advance. Watch the share count and any at-the-market capacity against the capex curve.
Debt is not the constraint. Revenue is, and until a powerhouse sells power, the funding model stays the same.
Questions worth following
- Can Oklo fund its first powerhouses without another equity raise?
The answer sets dilution and solvency risk for the next three years. Oklo has no revenue and rising capital spending, so its $2.7 billion of cash and investments either covers construction to first power or forces another stock sale at whatever price the market offers.
- Can Oklo build its first Aurora plant on schedule?
Delivery on schedule would show Oklo can convert a licensed design into an operating asset, the step that turns announced capacity into revenue. Slips raise the capital needed before first power and weaken the cost curve the fleet case depends on.
- Does stock compensation explain most of Oklo's reported loss?
The split between cash and noncash expenses determines whether the reported loss overstates the funding need or merely front-loads dilution. It also shapes how much new equity each construction phase actually requires from outside investors.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Key risksWhat could break the investment case
- Valuation and margin of safetyWhat the current price assumes, and how much room for error it leaves
- SummaryOur overall judgement and what to watch next
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Oklo Inc.
Oklo Inc.
Business, moat and key questions
Open company page