Oklo Inc.

Investment Thesis

Oklo is developing advanced nuclear microreactors — the Aurora powerhouse, a sodium-cooled fast fission design delivering 15-75 MWe — purpose-built to solve the AI data center power crisis. The thesis is elegantly simple: hyperscalers need 24/7 carbon-free baseload power at massive scale, renewables cannot provide it, natural gas is politically and environmentally unacceptable at the required scale, and only advanced nuclear can close the gap. Oklo is the most advanced pure-play developer of commercial fast fission microreactors in the United States, with a test reactor (Groves) already built in ~229 days, a first commercial power reactor under construction at Idaho National Laboratory (INL), and $2.5B in cash to fund development through the critical proof-of-performance window.

The market currently prices Oklo at $7.95B market cap with zero revenue — a pre-revenue deep tech bet that has already de-risked substantially since initial coverage. Key milestones that have been crossed: Groves test reactor built (concrete poured, DOE Final DSA approved), INL Aurora site excavation complete (PDSA approved June 2026), NRC Pre-Application Topical Report (PDC) approved May 2026 on accelerated timeline, Meta signed a 1.2GW prepayment deal for Pike County Ohio, Switch signed a 12GW Master Power Agreement, and Sam Altman retains a ~6% stake (post-dilution) providing strategic alignment with AI infrastructure buildout. Each milestone reduces technical and regulatory risk. The remaining binary risk is first criticality at Groves — imminent as of this writing (July 2026) — and first commercial power at INL targeted for late 2027.

Physical AI / Value-Chain Relevance

Oklo maps to Layer 2 of the Physical AI taxonomy — Grid, Power & Thermal Infrastructure. In the physical AI value chain, compute is useless without power. The hyperscaler buildout of AI data centers is running into physical grid constraints: 2-5 year interconnection queues, transmission bottlenecks, and the simple fact that the US grid was not designed for 500MW+ data center clusters. Oklo's Aurora microreactor is designed to bypass these constraints entirely, providing colocated 24/7 baseload power at the data center site without grid interconnection dependencies.

The sodium-cooled fast fission design offers two key advantages over conventional light-water SMRs: higher operating temperature (enabling more efficient power conversion and process heat applications) and intrinsic safety features (low-pressure sodium coolant eliminates the need for large containment structures and emergency coolant injection systems). The 15-75 MWe scale is designed to match data center module sizing — one Aurora per data center hall, scalable in 15 MWe increments.

Oklo's thesis is not about winning a technology beauty contest — it is about having the fastest path to licensed, commercial nuclear power for AI loads. The Groves test reactor (229 days from groundbreaking to completion) demonstrated that Oklo can build faster than any nuclear project in modern US history. If this speed can be replicated for commercial units, Oklo becomes the default power solution for hyperscalers facing multi-year grid interconnection delays.

Catalysts

Binary catalyst — Groves criticality (any day now): Groves test reactor received its Final DSA (Documented Safety Analysis) approval from the DOE in July 2026. First criticality is the single most important catalyst in Oklo's near-term timeline — it provides the first physical proof that the Aurora design works as modeled. Success here removes the biggest overhang on the stock (technology risk) and opens the door to full NRC Construction Permit Application.

Near-term — NRC licensing progress: The PDC topical report approval in May 2026 accelerated the regulatory pathway. Next milestone is submission of a full Construction Permit Application (COL application for NRC). Approval would mark the first NRC-licensed fast reactor in US history.

Medium-term — INL first power (late 2027): The first commercial Aurora at INL is under active construction with Kiewit as EPC partner. Excavation is complete, PDSA approved June 2026. First power by late 2027 would validate the commercial deployment model.

Commercial catalysts: Additional hyperscaler power purchase agreements beyond Meta (1.2GW) and Switch (12GW). Any deal with Microsoft, Amazon, or Google for dedicated nuclear capacity would validate that Oklo's approach is scaling.

Positioning / What the Market May Be Missing

Oklo's market positioning is poorly understood by the broader market. Most investors categorize it as a "SMR stock" alongside NuScale, GE Hitachi, or X-energy. But Oklo is fundamentally different: it is not selling small light-water reactors. The fast fission design gives it fuel flexibility (can use HALEU or recycled fuel) and a fuel utilization efficiency that is dramatically higher than conventional designs — Oklo's fuel recycling pathway (Tennessee Advanced Fuel Center, $1.68B roadmap) could reduce long-term fuel costs by extracting more energy per unit of uranium.

What the market may be missing is that Oklo's $2.5B cash position at a $132M annual burn rate provides ~19 years of runway. This is not a pre-revenue biotech racing against a cash-out date. Oklo has the financial staying power to weather NRC licensing delays, regulatory appeals, and construction cost overruns that would bankrupt less capitalized developers.

The market also underestimates the power bottleneck thesis. AI inference compute demand is growing at 10x per year according to NVIDIA. Every 100MW of AI compute needs ~150MW of power after cooling and overhead. At current hyperscaler expansion plans, total AI power demand in the US could exceed 50GW by 2030. The grid is not ready. Natural gas is carbon-exposed. Nuclear is the only 24/7 carbon-free solution at scale. Oklo is the most advanced pure-play developer targeting exactly this use case.

Risks and What Invalidates the Thesis

1. Groves reactor criticality failure. If Groves fails to achieve criticality, or has a safety incident during startup testing, the technical case for fast fission microreactors is substantially weakened. This is the single largest binary risk in the thesis.

2. INL cost overrun or timeline slip. If the first commercial Aurora at INL exceeds budget by >50% or slips past 2029, the economic case for distributed nuclear at data center scale is undermined. Kiewit partnership mitigates this but does not eliminate it.

3. NRC licensing denial. If the full Construction Permit Application is denied or indefinitely stalled, Oklo becomes a science project rather than a commercial venture. The May 2026 PDC approval on accelerated timeline is encouraging but not a guarantee of eventual NRC approval.

4. HALEU fuel supply failure. The domestic HALEU (High-Assay Low-Enriched Uranium) supply chain does not exist at commercial scale yet. Centrus is Oklo's primary supplier under a Letter of Intent with deliveries expected from 2029 — but if HALEU fuel fabrication fails to scale, Oklo's fuel supply is constrained.

5. Customer attrition. Meta's 1.2GW prepayment and Switch's 12GW MPA are non-binding framework agreements. If Meta or Switch walk away, the commercial thesis loses its anchor customers and the market will question real demand.

What to Watch Next

Groves test reactor criticality is the single most important binary event on Oklo's calendar — monitor for DOE announcements or company press releases. Then track INL construction progress: foundation pouring, reactor vessel installation, and the NRC COL application submission. On the commercial side, watch for customer announcements beyond Meta/Switch — any new hyperscaler PPA is strong positive signal.

Also watch the HALEU supply chain. Centrus production milestones and DOE funding for HALEU enrichment infrastructure are leading indicators of fuel availability when Oklo's first commercial reactors need to be loaded (~2028-2029). Fuel recycling progress at the Tennessee Advanced Fuel Center is a longer-term signal but important for the cost-advantage narrative.

Finally, monitor the broader regulatory and political environment. The DOE Reactor Pilot Program accelerated pathway was a Biden-era initiative — any Trump administration pivot away from advanced nuclear would create headwinds. Conversely, bipartisan nuclear support in the ADVANCE Act and the IRA's nuclear tax credits (ITC/PTC) provide a policy floor regardless of election outcomes.