Blog

DEEP RESEARCH · Oklo

Oklo: An Advanced Nuclear Platform Targeting the AI Power Bottleneck

A look at Aurora, HALEU, DOE pilots, cash runway, and short-seller risks

Published: 2025-12-31 · Advanced nuclear and AI infrastructure analysis · Naver Blog

Investment decisions are your responsibility. This research is not a recommendation to buy or sell.

0. Bottom line first

Oklo is not just a reactor designer. It aims to own and operate plants and sell power and heat under long-term contracts as an Energy-as-a-Service company. The AI data-center power bottleneck is a large market, but the 2027 operation target, HALEU supply, NRC licensing, and FOAK cost are major risks.

Original Oklo image 1

Official fact: The source says Oklo was founded in 2013 by MIT alumni Jacob DeWitte and Caroline Cochran, and its name comes from the 1.7-billion-year-old natural reactor phenomenon in Oklo, Gabon.

Interpretation: Oklo's investment appeal is not only the reactor design but the business model: selling long-term power and heat through PPAs without forcing customers to carry plant CAPEX.

Oklo Energy-as-a-ServiceBuild · Own · Operate
AuroraFast neutron reactor
FuelHALEU · metal fuel
CustomersData centers, industry, defense
Revenue20-40 year PPAs
A long-term power and heat sales platform, not a one-off reactor sale

1. Technology: Aurora's Design Logic

Fast Reactor

Fast neutron reactor

Maintains fission with fast neutrons, aiming for higher uranium utilization and waste-burning potential.

Sodium

Liquid metal cooling

Sodium boils around 883°C at atmospheric pressure, reducing high-pressure-system burden.

Modular

Simplification and modularity

Fewer moving parts and factory-built modules aim to reduce construction time and quality variance.

HALEU

5-20% enriched fuel

Central to the technology and also the central supply-chain risk.

Original Oklo image 2

2. Licensing and Government Tracks

The source highlights Oklo's 2020 first-of-kind COLA filing and the NRC's 2022 denial for insufficient information as an important failure lesson. In September 2025, the NRC accepted Oklo's principal design criteria topical report for accelerated review, which the source reads as an improved regulatory signal.

Official fact: The source says Oklo was selected for the DOE Reactor Pilot Program in August 2025, enabling Aurora-INL construction and operation under DOE oversight, and was selected for the U.S. Air Force Eielson AFB microreactor project.

Interpretation: DOE and defense tracks can accelerate demonstration outside the ordinary commercial-license path, but mass commercial deployment still requires safety proof and repeatable construction execution.

3. Partnerships: A Fabless Nuclear Ecosystem

AreaPartner/customerSource detail
ConstructionKiewitJuly 2025 MSA; design and construction for the INL project
Power conversionSiemens EnergyNovember 2025 binding contract for power conversion system
FuelCentrus EnergyPartnership with a U.S. HALEU license holder
Data centersEquinixLOI for up to 500 MWe
IndustryDiamondback Energy20-year Permian Basin power LOI, 50 MW
IsotopesAtomic AlchemyFebruary 2025 all-stock acquisition for USD 25 million

4. Financials and Valuation

Official fact: As of September 30, 2025, the source lists USD 410 million of cash and equivalents, USD 511.56 million of current marketable securities, and USD 261.96 million of non-current marketable securities, or about USD 1.183 billion of liquid assets.

ItemSource figureRead
Liquid assets at 2024 year-endAbout USD 227.8 millionBefore 2025 financing
Liquid assets at Sep. 30, 2025About USD 1.183 billionFollow-on and ATM funding effect
Q3 2025 net lossAbout USD 29.72 millionDeep-tech growth-company loss profile
YTD Q3 2025 net lossUSD 64.22 millionHigher R&D, G&A, and acquisition costs
Simple runwayAbout 7-10+ yearsBuffer before full construction ramp
Market capAbout USD 11-12 billion in late Dec. 2025AI infrastructure premium and overheating debate coexist
Original Oklo image 3

5. Risks: Short Report and Technical Execution

  • Kerrisdale Capital disclosed a short position in November 2024, questioning the realism of 2027 operations, HALEU cost assumptions, and management's lack of large nuclear construction experience.
  • The source accepts some criticism as valid but argues DOE RPP and Air Force selection show some government confidence in execution.
  • Liquid sodium can react violently with air or water, so design controls must be proven.
  • FOAK cost is a chronic first-unit nuclear risk. Cash may look sufficient, but long delays could accelerate burn.

6. My Conclusion

Oklo could become an energy-sector Tesla, but today it is still a pre-revenue deep-tech nuclear company. Near term, I would track NRC/DOE milestones, quarterly burn rate, HALEU supply, insider sales, and short-seller arguments. For long-term investors, the source's conclusion remains sensible: only a very limited allocation as a hedge on AI-infrastructure power bottlenecks.

Sources