The thesis
Most nuclear startups want to sell reactors. Oklo wants to sell electricity. That single choice defines the company: rather than shipping hardware to a utility and walking away, Oklo intends to build, own, and operate its Aurora powerhouses and bill customers for power under long-term contracts, the way an independent power producer does. It is a harder, more capital-intensive path — the company carries construction risk and cash burn that a pure vendor avoids — but it is also a bet that firm, dispatchable clean power is about to become one of the most valuable commodities in the economy, and that whoever owns the generating asset captures the upside.
The tailwind that turned that bet from contrarian to consensus is artificial intelligence. Hyperscale data centers need round-the-clock, carbon-free electricity at a scale and reliability that wind and solar alone cannot deliver, and they are willing to sign decade-long agreements to secure it. Oklo's non-binding order book has swelled toward roughly 14 gigawatts of prospective demand, anchored by data-center developers — a pipeline that, on paper, dwarfs the company's actual output today, which is zero. Oklo remains pre-revenue with no operating plant, so the entire story hinges on converting a certified concept and a stack of letters of intent into financed, licensed, generating steel.
From an MIT lab to a Sam Altman SPAC
Oklo was founded in 2013 by Jacob DeWitte and Caroline Cochran, a husband-and-wife pair of MIT-trained nuclear engineers — DeWitte, now CEO, was a PhD student at the time; Cochran, now COO, held a master's in nuclear engineering. They named the company after Oklo, Gabon, the only place on Earth known to have hosted a natural, self-sustaining fission reaction, roughly two billion years ago. The founding idea was to revive the metal-fueled fast reactor — a technology proven decades earlier at Idaho's Experimental Breeder Reactor II but never commercialized — and package it as a small, factory-style product.
The company's most consequential early backer was Sam Altman, who served as board chair from 2015 and co-led AltC Acquisition Corp., the special-purpose acquisition company that took Oklo public on the NYSE in 2024 under the ticker OKLO. Altman stepped down as chair in April 2025, a move framed as clearing the way for potential dealings between Oklo and OpenAI without conflicts of interest. The road to the public markets was not smooth: in 2022 the NRC denied Oklo's first combined-license application 'without prejudice,' a setback that forced the company to rethink its regulatory strategy — a detour that, in hindsight, pushed it toward the DOE-led licensing path it is riding today.
The technology: a small fast reactor with a long fuel appetite
The Aurora powerhouse is a compact, liquid-metal-cooled fast fission reactor, modeled on the sodium-cooled EBR-II and fueled with metallic high-assay low-enriched uranium (HALEU). Each unit is designed to produce between 15 and 75 megawatts of electricity — small enough to site next to an industrial load or data center campus, and configurable in clusters to reach hundreds of megawatts. Because it runs on a fast neutron spectrum, Aurora can burn fresh HALEU, recycled material, or down-blended fuel, and can extract far more energy from a given quantity of fuel than a conventional water reactor.
That fuel flexibility is the technology's strategic core, not just an engineering footnote. A fast reactor that can consume recycled and surplus material lets Oklo attack the industry's chronic HALEU supply bottleneck from two directions at once — buying enriched fuel and manufacturing its own from reprocessed streams. Oklo has begun locking in the conventional side with a June 2026 letter of intent from Centrus Energy to supply enough domestic HALEU to power up to five Aurora units, with deliveries starting in 2029, while separately pursuing recycling and surplus-plutonium pathways for the longer term.
The business model and the moat
Oklo's revenue is meant to come from selling power and heat, not reactors. The commercial pipeline reflects that: a non-binding 12 GW master power agreement with data-center developer Switch running out to 2044, a prepayment agreement with Meta tied to a planned 1.2 GW campus in Pike County, Ohio, and additional letters of intent with Equinix, Prometheus Hyperscale, and Diamondback Energy. If even a fraction converts to binding contracts, it would represent one of the largest advanced-nuclear order books in the industry — but every one of those agreements is today conditional and revenue-free.
The defensible advantage Oklo is building is regulatory and industrial rather than purely technical. By running its first plants through the DOE's Reactor Pilot Program at Idaho National Laboratory — a path that authorizes construction on federal land through a safety-basis review rather than the traditional NRC combined-license gauntlet — Oklo is compressing a timeline that has historically stretched a decade or more. Pairing that head start with a build-own-operate model means the company's moat is less any single patent than the accumulating stack of approvals, sites, fuel contracts, and manufacturing capacity that a newly funded rival cannot simply buy off the shelf.
Vertical integration: fuel, isotopes, and in-house manufacturing
Rather than depend on a fragile advanced-nuclear supply chain, Oklo is trying to own much of it. In June 2026 it acquired ARMEC, an Oak Ridge, Tennessee precision-manufacturing and engineering firm with roughly 40 fabricators, machinists, and technicians, bringing in-house welding, machining, and fabrication capability for reactor and fuel programs. It is also developing an advanced fuel center in Tennessee — backed by up to $1.68 billion of planned investment — to recycle used fuel and fabricate the metallic fuel Aurora needs, and in mid-2026 signed its first third-party offtake MOU with TRISO maker Standard Nuclear and was selected for the DOE's Surplus Plutonium Utilization Program alongside European fast-reactor developer newcleo.
The second leg of diversification is radioisotopes. Through its 2025 acquisition of Atomic Alchemy, Oklo is building a business to supply medical, industrial, research, and defense isotopes — a market with real revenue today, unlike power generation, which is still years out. Atomic Alchemy secured an NRC materials license in March 2026 enabling initial sales, and its Groves Isotope Test Reactor in Lockhart, Texas cleared its final DOE safety analysis on July 1, 2026, leaving only a readiness review before Oklo targets first criticality — potentially making it one of the first privately funded advanced reactors to operate on private land with commercially sourced fuel.
Financials: an enormous war chest against zero revenue
Oklo's financial profile is unusual: a company with essentially no revenue and a multi-billion-dollar valuation. In the first quarter of 2026 it booked $0 in revenue and a net loss of roughly $33.1 million, consistent with a developer still years from commercial power sales. What sets it apart from most pre-revenue peers is the balance sheet — about $2.54 billion in cash and securities at the end of Q1 2026, built largely through public equity issuance rather than debt, giving it a long runway to fund construction, acquisitions, and licensing without immediate financing pressure.
That war chest is a direct product of investor enthusiasm for the AI-nuclear thesis, which has carried Oklo's market capitalization to roughly $9 billion despite the lack of an operating plant — a valuation that prices in successful execution years in advance and leaves little margin for delay. The founders themselves became billionaires on paper as the stock climbed. The bull-bear debate reduces to a single tension: an unusually well-capitalized company with a vast prospective market and a faster regulatory path, set against the hard reality that it has not yet sold a kilowatt-hour and must convert its pipeline before the enthusiasm — or the cash — thins.
What to watch
The central near-term catalyst is the Aurora powerhouse at Idaho National Laboratory. The DOE approved its Preliminary Documented Safety Analysis on June 11, 2026 — the third of four major Reactor Pilot Program steps — leaving only the final Documented Safety Analysis before the DOE can authorize operation. Watch whether Oklo secures that final approval and completes construction on schedule toward a targeted first deployment in late 2027 to 2028; slippage there would ripple through the entire valuation.
Two other threads will signal whether the model is real. First, Groves: a successful first criticality of the Texas isotope test reactor would give Oklo an operating asset and a revenue-generating isotope business well ahead of its power plants. Second, contract conversion — the moment any of the Switch, Meta, Equinix, or Prometheus agreements hardens from a non-binding letter into a financed, binding power purchase agreement, the story shifts from promise to backlog. Alongside those, watch the fuel-supply and manufacturing build-out (Centrus HALEU deliveries, the Tennessee recycling center, and EPC planning with partners such as Kiewit) that must all mature for Oklo to deliver the power it has already promised on paper.
Sources
- Oklo — Q1 2026 Form 10-Q (financials, Aurora output)
- Oklo — 2025 Form 10-K (history, business model, customers)
- Utility Dive — Oklo 12 GW Switch supply agreement
- Forbes — MIT founders, Sam Altman, and Oklo's AI-nuclear rise
- Oklo — DOE approves Aurora-INL Preliminary Documented Safety Analysis
- BusinessWire — DOE approves Groves Isotope Test Reactor safety analysis