Oklo Inc.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
Oklo does not yet have a moat; it has a licensing head start, a funded balance sheet and a pipeline of intentions. The Aurora sodium-cooled fast reactor is under NRC combined-licence review after the Principal Design Criteria topical report was approved on an accelerated schedule, the Groves test reactor reached criticality under the DOE pilot programme, and about $3.0B of liquidity funds the Aurora-INL build toward a 2028 start. None of that is an operating reactor, a licence, or a binding power contract, and every barrier Oklo is climbing is one TerraPower, Kairos, X-energy and NuScale are climbing alongside it.
Oklo's case rests on being first through a licensing process that peers are running at the same time, funded by equity rather than customers:
- Licensing Progress Is Real but Shared: Oklo's first combined licence application was denied by the NRC in 2022 for lack of information. The rebuilt effort has gone better: a Phase 1 readiness assessment found no significant gaps, the Principal Design Criteria topical report for the Aurora powerhouse was approved in under half the usual review time, and a customised combined licence application for Aurora-INL went to the NRC in the second quarter of 2026. That is a head start, not a lock-in. Kairos holds construction permits, TerraPower is licensing Natrium in parallel, and NuScale already has an approved design, so the NRC barrier Oklo is clearing is one every serious peer will clear too.
- Operating Evidence Begins at Groves: The Groves isotope test reactor in Texas reached first criticality under the DOE Reactor Pilot Program less than a year after groundbreaking, the first reactor in the programme to do so on private land. It is a low-power machine built to generate operating experience and support isotope production, not a power plant. Aurora-INL, the first 75 MWe powerhouse, has DOE approval of its preliminary documented safety analysis and site excavation under way, with start-up still targeted for 2028. Until Aurora runs, Oklo's build-own-operate model has no cost, availability or uptime record for a customer to underwrite.
- A Pipeline of Intentions, Funded by Shareholders: Oklo cites roughly 14 GW of customer interest, but most of it is non-binding: the 12 GW Switch master power agreement and a string of letters of intent. The one binding commitment is Meta's prepayment for fuel and early work on a campus of up to 1.2 GW in Pike County, Ohio, with first power targeted as early as 2030, and Meta has signed nuclear deals with other developers as well. Meanwhile shares outstanding rose from 160.5M at year-end 2025 to 185.1M by June 30, 2026 as at-the-market sales built the $3.0B cash pile, so shareholders rather than customers are paying for the build.
Moat Verdict
Oklo's pillars are physical and regulatory, so AI cannot disrupt them. AI data-centre power demand is also the reason its customer pipeline exists, which makes the company a net beneficiary of AI. The problem is that none of those pillars has formed yet: the licensing lead is shared with peers, the contract book is mostly non-binding, and there is no operating fleet, so the moat score measures an early head start, not a durable advantage.
45.0 resilient · 65.0 vulnerable · 80/20 = 49.0 · = 49
Oklo sells power and isotopes, not software; there is no user-trained interface or workflow, so this moat category does not apply.
Not software business logic: the Aurora reactor design and fuel-cycle engineering are physical hardware and licensing know-how, which the framework scores under regulatory lock-in and scale economics rather than here.
Oklo controls no unique public data source; this moat category does not apply to a reactor developer.
Advanced-reactor designers, NRC licensing specialists and fast-reactor operators are a thin talent pool, and Oklo has assembled a team that took an application from a 2022 denial to an accelerated design-criteria approval; but TerraPower, Kairos, X-energy and the utilities hire from the same pool, so it is a hiring advantage, not an exclusive one.
Oklo plans to sell power, fuel recycling and isotopes together under a build-own-operate model, but no customer buys that bundle yet; there is nothing to rate until the first powerhouse runs.
The sodium fast reactor physics Oklo builds on comes from the public EBR-II operating record that TerraPower also draws on, and Oklo has no operating power reactor generating a dataset of its own, so this category does not apply yet.
Oklo is further along than most advanced-reactor developers: the NRC approved its Aurora Principal Design Criteria topical report in under half the usual review time and has the Aurora-INL combined licence application under review, and Groves reached criticality under DOE authorisation. But no licence has been granted, the first application was denied in 2022, and Kairos, TerraPower, X-energy and NuScale are working through the same NRC and DOE pathways, so the barrier is held by peers too.
Electricity and isotopes are sold bilaterally; Oklo's product does not become more valuable as more customers buy it, so this moat category does not apply.
For a power seller the contract book is the moat, and Oklo's is mostly intentions. The 12 GW Switch agreement and most of the ~14 GW pipeline are non-binding; Meta's prepayment for up to 1.2 GW in Ohio is the only binding commitment, and Meta, Google and Amazon have each signed nuclear deals with other developers, so buyers are multi-sourcing rather than locking in.
Oklo is not the authoritative record for any information function; this moat category does not apply to a reactor developer.
Oklo has no operating fleet: its only reactor is a low-power test unit, and the first 75 MWe Aurora is still under construction. The factory-built, small-unit cost case is a plan, while incumbent nuclear operators such as Constellation spread fuel, outage and staffing costs across a fleet of about 22 GW.
Sells electricity and isotopes at contracted prices; buyers pay for firm power, not a brand premium.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
Oklo does not yet have a moat; it has a licensing head start, a funded balance sheet and a pipeline of intentions. The Aurora sodium-cooled fast reactor is under NRC combined-licence review after the Principal Design Criteria topical report was approved on an accelerated schedule, the Groves test reactor reached criticality under the DOE pilot programme, and about $3.0B of liquidity funds the Aurora-INL build toward a 2028 start. None of that is an operating reactor, a licence, or a binding power contract, and every barrier Oklo is climbing is one TerraPower, Kairos, X-energy and NuScale are climbing alongside it.
Growth Score
There is effectively no revenue to grow yet. Q2 2026 brought the first reported revenue, $1.2M of mostly acquired services, while the net loss nearly doubled to $48.5M and full-year 2026 guidance for operating cash burn rose to $120–150M, with capital spending of $400–500M. Near-term revenue comes from isotopes, most likely from the NRC-licensed Idaho radiochemistry laboratory from early 2027, and power revenue waits on Aurora-INL's 2028 start. Any growth rate fitted to a base this small saturates the curve, so the growth score is carried by the rising burn and the severe risk discount rather than by the CAGR.
Valuation Score
At ~$38 the stock is down about 80% from its 52-week high near $194 and sits between a $20 bear case and a $45 base case. Roughly $3.0B of the ~$7B market cap is cash and securities, so the market is paying around $4B for a reactor that has not been licensed and a pipeline that is mostly non-binding. That is cheaper than it was, not cheap: the discount to base is a discount to a target that already assumes the NRC review and the Aurora-INL build go broadly to plan.
A Head Start, Not a Moat
Oklo's case rests on being first through a licensing process that peers are running at the same time, funded by equity rather than customers:
- Licensing Progress Is Real but Shared: Oklo's first combined licence application was denied by the NRC in 2022 for lack of information. The rebuilt effort has gone better: a Phase 1 readiness assessment found no significant gaps, the Principal Design Criteria topical report for the Aurora powerhouse was approved in under half the usual review time, and a customised combined licence application for Aurora-INL went to the NRC in the second quarter of 2026. That is a head start, not a lock-in. Kairos holds construction permits, TerraPower is licensing Natrium in parallel, and NuScale already has an approved design, so the NRC barrier Oklo is clearing is one every serious peer will clear too.
- Operating Evidence Begins at Groves: The Groves isotope test reactor in Texas reached first criticality under the DOE Reactor Pilot Program less than a year after groundbreaking, the first reactor in the programme to do so on private land. It is a low-power machine built to generate operating experience and support isotope production, not a power plant. Aurora-INL, the first 75 MWe powerhouse, has DOE approval of its preliminary documented safety analysis and site excavation under way, with start-up still targeted for 2028. Until Aurora runs, Oklo's build-own-operate model has no cost, availability or uptime record for a customer to underwrite.
- A Pipeline of Intentions, Funded by Shareholders: Oklo cites roughly 14 GW of customer interest, but most of it is non-binding: the 12 GW Switch master power agreement and a string of letters of intent. The one binding commitment is Meta's prepayment for fuel and early work on a campus of up to 1.2 GW in Pike County, Ohio, with first power targeted as early as 2030, and Meta has signed nuclear deals with other developers as well. Meanwhile shares outstanding rose from 160.5M at year-end 2025 to 185.1M by June 30, 2026 as at-the-market sales built the $3.0B cash pile, so shareholders rather than customers are paying for the build.
Moat Verdict
Oklo's pillars are physical and regulatory, so AI cannot disrupt them. AI data-centre power demand is also the reason its customer pipeline exists, which makes the company a net beneficiary of AI. The problem is that none of those pillars has formed yet: the licensing lead is shared with peers, the contract book is mostly non-binding, and there is no operating fleet, so the moat score measures an early head start, not a durable advantage.
45.0 resilient · 65.0 vulnerable · 80/20 = 49.0 · = 49
Oklo sells power and isotopes, not software; there is no user-trained interface or workflow, so this moat category does not apply.
Not software business logic: the Aurora reactor design and fuel-cycle engineering are physical hardware and licensing know-how, which the framework scores under regulatory lock-in and scale economics rather than here.
Oklo controls no unique public data source; this moat category does not apply to a reactor developer.
Advanced-reactor designers, NRC licensing specialists and fast-reactor operators are a thin talent pool, and Oklo has assembled a team that took an application from a 2022 denial to an accelerated design-criteria approval; but TerraPower, Kairos, X-energy and the utilities hire from the same pool, so it is a hiring advantage, not an exclusive one.
Oklo plans to sell power, fuel recycling and isotopes together under a build-own-operate model, but no customer buys that bundle yet; there is nothing to rate until the first powerhouse runs.
The sodium fast reactor physics Oklo builds on comes from the public EBR-II operating record that TerraPower also draws on, and Oklo has no operating power reactor generating a dataset of its own, so this category does not apply yet.
Oklo is further along than most advanced-reactor developers: the NRC approved its Aurora Principal Design Criteria topical report in under half the usual review time and has the Aurora-INL combined licence application under review, and Groves reached criticality under DOE authorisation. But no licence has been granted, the first application was denied in 2022, and Kairos, TerraPower, X-energy and NuScale are working through the same NRC and DOE pathways, so the barrier is held by peers too.
Electricity and isotopes are sold bilaterally; Oklo's product does not become more valuable as more customers buy it, so this moat category does not apply.
For a power seller the contract book is the moat, and Oklo's is mostly intentions. The 12 GW Switch agreement and most of the ~14 GW pipeline are non-binding; Meta's prepayment for up to 1.2 GW in Ohio is the only binding commitment, and Meta, Google and Amazon have each signed nuclear deals with other developers, so buyers are multi-sourcing rather than locking in.
Oklo is not the authoritative record for any information function; this moat category does not apply to a reactor developer.
Oklo has no operating fleet: its only reactor is a low-power test unit, and the first 75 MWe Aurora is still under construction. The factory-built, small-unit cost case is a plan, while incumbent nuclear operators such as Constellation spread fuel, outage and staffing costs across a fleet of about 22 GW.
Sells electricity and isotopes at contracted prices; buyers pay for firm power, not a brand premium.
Growth Analysis
Growth Drivers
Key Risk
Everything waits on decisions that have not been made. The NRC has not yet ruled on the Aurora-INL combined licence, and it denied Oklo's first application in 2022. The first-of-a-kind reactor could slip past its 2028 start. The ~14 GW pipeline has to convert into binding power contracts while hyperscalers sign with rival developers. HALEU beyond the first core depends on DOE allocations and enrichment capacity still being built. Each delay also means more dilutive equity at a lower share price.
Score Derivation
90.0 base + 1.3 trajectory − 4 margin − 15 risk = 72
Base 90 (step-change from a ~$1M quarterly base; the curve saturates above 30% and cannot resolve a pre-commercial developer) + 1.3 trajectory (isotope revenue starting is the one accelerating revenue line; power revenue is stable at zero until Aurora-INL starts and the binding customer book is stable at a single Meta prepayment) − 4 compressing margin (observed: Q2 net loss $48.5M vs $24.7M, full-year operating burn guidance raised from $80–100M to $120–150M) − 15 severe risk (combined licence not yet granted after a 2022 denial, first-of-a-kind reactor, non-binding pipeline, fuel supply unsecured beyond the first core) = 72
Price Scenarios (12–24 Months)
Valuation Analysis
Earnings and revenue multiples do not apply; there is no power revenue until Aurora-INL starts. The useful anchors are enterprise value net of the ~$3.0B cash pile, the pace of cash burn against guidance, and share count, since at-the-market sales added about 25M shares in the first half of 2026. Watch the NRC review milestones on the Aurora-INL application and the conversion of any part of the Switch agreement into a binding contract. $45.
Where We Are vs Targets
Loading live price…
The NRC review drags or raises design questions, Aurora-INL slips past 2028, and the market values Oklo closer to its cash as dilution keeps adding shares at lower prices.
- NRC requests for additional information on the Aurora-INL application push the licensing decision well beyond Oklo's schedule, and Aurora start-up moves from 2028 into 2029–2030
- Operating cash burn runs above the raised $120–150M guidance and capex above $500M, so Oklo keeps selling stock through its at-the-market programme and the share count climbs past 200M
- Hyperscalers sign binding offtakes with rival developers such as TerraPower, Kairos and X-energy while Switch and the other letters of intent stay non-binding, leaving Meta as the only paying customer
The NRC review of the Aurora-INL application progresses on schedule, construction continues toward a 2028 start, and isotope revenue begins in 2027, earning back part of the 2026 de-rating without any power revenue yet.
- The NRC completes the first step of its two-step Aurora review and issues a licensing schedule consistent with a 2028 start, with no major design issues raised
- Isotope revenue begins from the Idaho radiochemistry laboratory in early 2027 and Groves operating data supports the planned isotope production facilities
- At least part of the Switch agreement or another letter of intent converts into a binding power purchase agreement, and cash burn stays within guidance without another large equity raise
Oklo wins the first NRC combined licence for an advanced fission reactor, converts gigawatts of its pipeline into binding contracts, and the market prices it as the leading Western advanced reactor developer again.
- The NRC issues the Aurora-INL combined licence ahead of schedule on its accelerated pathway, making Oklo the first advanced fission developer licensed to build and operate a power reactor
- Switch and at least one more hyperscaler sign binding multi-gigawatt power agreements with prepayments, following the Meta structure, and the second and third Aurora sites move into licensing
- DOE finalises access to surplus plutonium or additional HALEU, which removes the fuel constraint on the next units and supports Oklo's fuel-recycling plans