OKLO — OKLO INC.

Oklo has broken ground on the first privately-owned commercial fast reactor in American history, signed a binding 1.2-gigawatt power agreement with Meta, and accumulated $2.6 billion in cash against zero debt — all without generating a single dollar of revenue. At $48.65 per share and an enterprise value of approximately $10.8 billion, the market is pricing the option on a successful nuclear future at a level that demands NRC licensing and on-schedule construction to justify. Interesting but requires specific catalysts to be actionable.

---

The electricity crisis facing American data centers is no longer a planning assumption — it is an operating constraint. Artificial intelligence model training and inference have driven data center power consumption to approximately 41 gigawatts in 2025, up from 33 gigawatts in 2024, and projections for 2030 range from 80 to 120 gigawatts depending on the pace of model deployment. The gap between projected demand and committed supply is not small: Deloitte estimates a 17 gigawatt shortfall by 2030 even accounting for all planned capacity additions. The utility grid, built around baseload coal and gas with moderate load growth, was not designed to absorb this kind of concentrated demand from facilities that require power continuously at high density. The technology companies building these facilities have responded by signing nuclear power agreements at a pace that would have seemed implausible three years ago.

This environment has been extraordinarily favorable for any credible nuclear development story. Amazon signed agreements with Kairos Power and Energy Northwest. Google partnered with Kairos. Microsoft contracted capacity from Constellation. Meta announced 6.6 gigawatts of nuclear deals in a single announcement in January 2026. The underlying logic is consistent across all of them: data centers need always-on, carbon-free power in locations that can be sited close to the facility, and nuclear is the only technology that satisfies both conditions. The question is not whether there is demand for what nuclear developers are promising. There is. The question is whether the companies promising it can build it, on schedule, at the required cost, with regulatory approval — a sequence of conditions that the history of advanced reactor development suggests cannot be taken for granted.

The small modular reactor market is simultaneously a $7.5 billion industry today and a speculative architecture for a possible future. More than 150 companies globally are developing SMR designs across six different cooling methods and multiple fuel types. The number of designs is not a sign of health; it is a sign of a market still in the pre-consolidation phase, before the physics of manufacturing economics and regulatory certification have narrowed the field. The winning designs will be those that achieve factory-scale production first, exploit the Wright's Law cost reductions that come with volume, and accumulate the operational data that convinces regulators and customers to standardize. The losers will be the companies that spent years in development and ran out of capital before the consolidation point arrived. Among the current roster, only NuScale has a full NRC design certification for a commercial SMR — approved in January 2023 for the 50 MWe design and May 2025 for the 77 MWe design — though NuScale's first commercial deployment remains years away. The others, including Oklo, are still in the licensing pipeline.

Oklo's model differs from every other SMR developer in one important structural respect: it does not intend to sell reactors. Instead, it builds, owns, and operates its Aurora powerhouses and sells electricity under long-term power purchase agreements — a power-as-a-service approach that converts a capital equipment sale into a multi-decade recurring revenue stream. The Aurora is a sodium-cooled fast reactor fueled by high-assay low-enriched uranium metallic fuel, scaled to 75 megawatts of electrical output. Its design heritage traces directly to the Experimental Breeder Reactor II (EBR-II), which operated at Idaho National Laboratory from 1964 to 1994 and accumulated thirty years of operational data that informs Aurora's passive safety design. The reactor is self-stabilizing and self-controlling through negative reactivity feedback — in the event of a loss of cooling or other abnormality, the physics of the fuel and coolant automatically reduce the nuclear reaction without human intervention or external power.

The fuel recycling capability is Oklo's most differentiated technical claim. Fast reactors can fission a wider range of uranium isotopes than conventional light-water reactors, and Oklo's pyroprocessing capability — electrochemical fuel recycling that extracts remaining energy from spent nuclear fuel — could theoretically reduce long-term fuel costs substantially and close the nuclear fuel cycle. Oklo completed an end-to-end demonstration of this process at laboratory scale, and in 2025 announced the Tennessee Advanced Fuel Center, a proposed $1.68 billion privately-funded fuel recycling facility. Whether pyroprocessing can be scaled from laboratory demonstration to commercial facility is a meaningful open question; the technology has a sixty-year scientific history but no commercial precedent in the United States.

Assessing Oklo's competitive moat requires honesty about what exists and what does not. The EBR-II heritage is real, the physics of the design are validated by decades of operational data, and the power-as-a-service model is structurally attractive once commercial operations begin. What does not yet exist is a deployed commercial Aurora reactor, an NRC combined operating license for the Aurora design, or a proven supply chain for the HALEU fuel the design requires. The NRC rejected Oklo's first combined license application in January 2022 — a formal rejection, rare in nuclear regulatory history — citing insufficient information on maximum credible accident scenarios and safety system classifications. The company restarted its pre-application process and has since made genuine progress: the NRC accepted Oklo's Principal Design Criteria topical report under an accelerated fifteen-day review timeline, and the DOE approved multiple facility design agreements in 2025. But the COLA for Aurora-INL has not yet been submitted. The licensing pathway for a sodium-cooled fast reactor is the first of its kind the NRC has encountered at the commercial scale; the review will take time it has never before taken for this reactor type. The moat, if it materializes, will be a decade's head start in operational data and regulatory precedent for fast reactors. That moat does not yet exist.

The financial profile of Oklo is that of a company spending aggressively to build the infrastructure for a business that has not yet begun. Revenue in 2024 was zero. Revenue in 2025 was zero. The company expects its first commercial revenue — from medical isotopes produced at the Groves isotope test reactor in Texas — in late 2026, representing a small fraction of the eventual power sales business. Operating cash burn in 2025 was $82.2 million, within management's guidance range of $65 to $80 million. The 2026 guidance is $80 to $100 million for operating expenses and $350 to $450 million for capital expenditures — a total cash use of $430 to $550 million, representing a six-fold acceleration from 2025. To fund this, Oklo completed a $540 million ATM equity program in 2025 and in January 2026 completed a second program raising $1.182 billion in net proceeds. The balance sheet as of early 2026 holds approximately $2.6 billion in cash and marketable securities against zero debt — a funded position, but one that will be consumed at significant pace as Aurora-INL construction advances. GAAP net loss in 2025 was $139.3 million, of which $41.8 million was non-cash stock-based compensation; the cash operating loss of approximately $82 million is the relevant figure for evaluating cash consumption.

The management picture at Oklo is complicated. Jacob DeWitte and Caroline Cochran co-founded the company, serve as CEO and COO respectively, and each hold beneficial ownership of approximately 10% — a meaningful alignment of founder interest with shareholder outcomes. Sam Altman, who served as chairman and whose involvement was widely credited with accelerating the company's commercial partnerships, stepped down in April 2025 to avoid conflicts of interest as OpenAI explored potential power supply relationships with Oklo. The leadership team has demonstrated genuine execution capability: they navigated a regulatory rejection, rebuilt the NRC relationship, secured major customers including Meta, broke ground on Aurora-INL, and raised over $1.7 billion in equity capital across 2025 and early 2026. Against this record, the insider selling stands out. In a ninety-day window from December 2025 through February 2026, insiders sold approximately $152 million in stock — including $69 million by Caroline Cochran (COO) and $5.5 million by the CFO — concurrent with the ATM equity program that was simultaneously diluting public shareholders. The sales were conducted under Rule 10b5-1 trading plans, which provide pre-planned selling schedules, but the scale and timing relative to the ATM raises a legitimate question about how the founders themselves assess value at the prices at which public investors are buying.

The growth runway argument for Oklo rests on a pipeline that is large in headline terms and substantially non-binding in legal terms. The table below shows the trajectory that matters: customer demand building alongside a cash consumption profile that is about to accelerate sharply.

Year	Pipeline (GW, cumulative)	Binding Agreements	Cash Position ($M)	Cash Burned ($M)	Key Milestone
2022	<0.1	None	~$10	$(10)	NRC rejects COLA application
2023	~0.5	None	~$15	$(19)	Pre-application process restarted
2024	~2.1	None	$282	$(38)	SPAC merger (IPO); Switch 12 GW LOI signed
2025	~14	None binding at year-end	$1,400	$(82)	Ground broken at Aurora-INL; Meta deal imminent
2026E	14+	Meta 1.2 GW (Jan 2026)	~$2,600	$(430–550)	Groves criticality (Jul 4); NRC licensing progress

The pipeline growth from near-zero to 14 gigawatts in four years reflects genuine market validation — it is not credible that Meta would sign a binding 1.2-gigawatt agreement without serious diligence on the technology. But the composition of that 14 gigawatts matters: Switch's 12-gigawatt master power agreement, signed in December 2024, is a non-binding framework under which individual binding PPAs will be negotiated as project milestones are reached. The Meta 1.2 gigawatt agreement, signed in January 2026, is the only binding commercial commitment in the pipeline. The cash burn trajectory simultaneously shows a business building toward something real and accelerating toward a capital intensity that will require either further equity raises (and dilution) or successful first-power generation before the balance sheet runs thin. At $430 to $550 million in 2026 cash use, the $2.6 billion starting position provides roughly five years of runway at the current acceleration rate — assuming no further increase in burn, which construction timelines rarely allow.

The penetration argument requires specific framing for a pre-revenue company. Data center operators will need approximately 14 to 70 gigawatts of new power capacity by 2035, depending on the projection source. Oklo's pipeline represents approximately 14 gigawatts of that total. Commercial nuclear is expected to supply roughly 10% of the incremental demand increase, implying a nuclear addressable market in the range of 6 to 18 gigawatts. On this framing, Oklo has already captured commitments that equal or exceed the total available nuclear share of data center demand growth — if, and only if, the contracts convert to operating power plants. The company has captured 0% of actual power sales, because no Aurora reactor has generated commercial power. Every gigawatt of the pipeline is contingent on technology that has not yet been commercially deployed at scale.

At $48.65 per share, Oklo's market capitalization is approximately $8.4 billion. Enterprise value, netting the $2.6 billion cash position against zero debt, is approximately $5.8 billion. Against near-zero revenue in 2026 and management's projection of perhaps $16 million in first-year power sales in 2028, any traditional earnings or revenue multiple is essentially meaningless. The investment is priced as a call option on a successful nuclear future — and the option value must be assessed against the specific conditions that determine whether it expires worthless or becomes transformative. The 52-week high was $193.84; the current price of $48.65 represents a 75% decline from that peak, itself driven in part by the capex guidance revision that clarified just how capital-intensive the next phase of development will be.

The most intelligent case against owning Oklo at any near-term price is that the critical value-creating events — NRC commercial licensing of the Aurora design and successful first-of-a-kind construction — are genuinely binary in their outcome and substantially uncertain in their timing. The NRC has never approved a sodium-cooled fast reactor for commercial operation; the review will establish precedents across a reactor type that is genuinely new to the regulatory process, and the agency's track record on novel designs suggests timelines extend further than developers project. HALEU — the enriched uranium fuel Aurora requires — has no domestic commercial supply chain; Oklo is partnered with Centrus to address this, but commercial HALEU availability is a sector-wide problem that no single company has yet solved. The answer to this bear argument is that the DOE has selected Oklo for accelerated deployment support, construction is underway with an experienced contractor (Kiewit), and the NRC is reviewing PDC materials on an accelerated timeline. The answer is real but incomplete — accelerated timelines are not the same as guaranteed timelines.

Two specific events would change the conclusion. The first is the Groves isotope reactor achieving criticality on or near its July 4, 2026 target — a near-term binary test of whether Oklo can build and operate a fast fission system on a compressed schedule. Success there does not de-risk Aurora-INL directly, but it demonstrates the team's technical and construction execution in a nuclear regulatory environment. The second is the NRC accepting and progressing on the Aurora-INL combined operating license application, establishing a regulatory pathway that converts the licensing risk from open-ended to bounded. Neither has yet occurred. Until one or both does, the investment remains an option priced for optimism on a technology that is constructing its first commercial proof point.

The technology is real. The demand is real. The construction has begun. What the investor at $48.65 is buying is the probability that a first-of-a-kind advanced nuclear reactor gets built, licensed, and operated on a schedule that commercial data centers will actually wait for — and that is a probability the next eighteen months will begin to define.