OKLO

Good-pond/bad-fish risk: the AI-driven nuclear baseload demand is a genuinely large and durable market, but Oklo must execute a first-of-a-kind advanced fast reactor on schedule, secure a HALEU fuel supply chain that barely exists, and convert non-binding pipeline into contracted revenue — all with a share count growing ~2.5x over three years and zero revenue earned to date.

Oklo designs compact fast-spectrum nuclear fission reactors — the Aurora powerhouse — and plans to own, operate, and fuel them, selling electricity to customers under long-term power purchase agreements. The Aurora uses High-Assay Low-Enriched Uranium (HALEU) metal fuel, runs at up to 15 MWe electrical output per unit, and is designed for co-location at data centers, industrial facilities, and government sites. Oklo does not sell reactors; it sells electrons, making each commissioned unit a long-duration recurring revenue asset once operational. The company has zero commercial revenue in any reported period through FY2025.

Oklo's near-term execution plan has five simultaneous tracks:

1. Aurora-INL demonstration unit: The first physical Aurora powerhouse is under construction at Idaho National Laboratory under a DOE Reactor Pilot Program (RPP) authorization — a pathway that bypasses the standard NRC licensing process for this single demonstration unit. Groundbreaking was September 22, 2025. First operations are targeted for late 2027 / early 2028. Q1 2026 capex spend was $32.8M on this project.

2. NRC commercial licensing: Every Aurora unit beyond the INL demonstration requires NRC licensing. Oklo has completed an NRC readiness assessment and received NRC Principal Design Criteria topical report approval on an accelerated 15-day review (May 6, 2026). The Part 52 Combined License Application (COLA) for the first commercial unit had been targeted for 2025 submission but has slipped to H2 2026. A new NRC Part 57 microreactor licensing framework (proposed rule published May 1, 2026) could create a faster commercial licensing template, but the final rule is 12-18 months away.

3. Customer pipeline: Oklo has announced agreements totalling a reported 14 GW of potential demand — principally a 1.2 GW campus agreement with Meta (described as binding in principle, contingent on PJM interconnection) and a 12 GW non-binding master power framework with Switch. No executed take-or-pay PPA with specified pricing and a delivery date has been publicly disclosed.

4. Fuel supply: Aurora requires HALEU. The only US commercial producer (Centrus Energy) operates at demonstration scale (~20 kg/yr). Oklo is building its own Aurora Fuel Fabrication Facility at INL under three DOE fuel-line pilot project selections. The DOE opened separate talks with Oklo in May 2026 on potential plutonium fuel use (a mid-2030s pathway with nonproliferation complexity).

5. Capital: The company completed a $1.18B ATM offering in January 2026, raising cash to $788.4M at FY2025 year-end. A second $1.0B ATM was filed May 13, 2026. Management guided $80-100M operating cash use plus $350-450M capex in 2026 — total cash consumption of $430-550M in a single year.

Income statement (FY, USD millions)

*FY2022 net income positive due to non-operating items (warrant fair-value or similar); operating loss was still -$1.8M.

Operating losses have doubled or more each year since FY2022. Q1 2026 operating loss was $51.2M — a pace consistent with $180-200M+ for FY2026.

Cash flow (FY, USD millions)

Balance sheet highlights

Cash increased from $97.1M (FY2024) to $788.4M (FY2025) entirely from the $1.18B ATM offering completed January 2026 — the increase reflects capital raises, not operations.

Runway: At guided 2026 total cash use of $430-550M (operating + capex), the $788.4M cash balance provides approximately 17-22 months of runway before the next required capital raise — potentially exhausted before Aurora-INL reaches commercial operation. This is not going-concern territory, but it means dilution is structurally built into the model.

Share dilution

Shares have grown ~2.5x in under three years. A second $1.0B ATM filed May 2026 confirms this trend continues.

Current: ~$12.0B market cap, zero revenue, operating loss of $139.3M (FY2025). No traditional valuation multiple is applicable. The market is pricing option value on the demand thesis plus regulatory progress.

Comps: NuScale Power (SMR) — the only other publicly traded SMR developer — trades at ~$800M market cap with an NRC-certified design and no construction start. Oklo is priced at 15x NuScale's market cap with fewer hard milestones reached and a more novel (unproven) reactor design. This does not mean Oklo is wrong — Oklo's fast-reactor design and DOE-authorized construction site are genuine differentiators — but the relative valuation illustrates the magnitude of premium being paid for the narrative.

Analyst targets: Consensus $101 (25 analysts, MarketBeat). Goldman Sachs: $65. UBS: $60. Wolfe Research: Peer Perform. The dispersion — with the more bearish targets below the current price of $73.47 — reflects genuine disagreement on execution probability, not just timing.

Short interest: 16.45% of float (Barchart, late April 2026) — elevated, and the highest among US utilities above $2B market cap. This creates potential short-squeeze amplification on positive catalysts, and persistent downside pressure on negative ones.

Fair value range: $45-$95 (scenario-weighted DCF framework). The low end reflects a scenario where the COLA slips beyond 2027, first commercial revenue is pushed past 2032, and dilution continues at current pace — discounting long-dated, uncertain cash flows at a high rate produces a valuation well below current levels. The high end reflects near-perfect execution: COLA filed and accepted by end-2026, Aurora-INL criticality 2027-2028, one binding hyperscaler PPA converting with disclosed pricing, bringing the company into the infrastructure capital repricing regime (8-12x long-duration EBITDA) by 2029-2030. The current price of $73.47 sits near the midpoint of this range — not egregiously overvalued on the bull case, but pricing in substantial execution success years before it occurs.

Institutional: Major institutional and ETF ownership is consistent with a high-profile pre-revenue story in an AI-adjacent sector — specific percentages not in the XBRL facts file.

Insider selling: CEO Jacob DeWitte has sold Oklo stock under a 10b5-1 plan (adopted March 31, 2025) with the following confirmed tranches: January 5 (~$9M), January 9 (~$10.2M), February 2 (~$1.5M), April 1 (~$10.1M), May 1 (~$14M). Total confirmed sales in the 5-month window: ~$44M+. The plan structure reduces (but does not eliminate) the informational inference — the plan predates the current stock level. However, monthly sales at $10-14M per tranche at a pre-revenue company with accelerating losses is an unusual cadence and a persistent headline risk. Sources: investing.com insider trading reports; StockTitan Form 4 filings.

Board: Sam Altman (OpenAI CEO) as board chair is a genuine deal-flow and attention asset in the AI-energy procurement moment — he is in the decision-making network of the hyperscalers Oklo is targeting.

• Construction is real. Aurora-INL groundbreaking September 22, 2025 at Idaho National Laboratory. Q1 2026 capex was $32.8M. DOE approved the Aurora Fuel Fabrication Facility safety analysis in December 2025. This is not a promise — it is a worksite with spending.
• Regulatory velocity is genuine. NRC PDC topical report approved in 15 days (May 6, 2026) — an order of magnitude faster than typical NRC timelines. NRC readiness assessment completed with no significant gaps. Part 57 proposed rule creates a potential future repeatable licensing template.
• The only publicly traded pure-play advanced fast reactor. Any capital allocation to advanced nuclear in public equities flows disproportionately here. TerraPower (most direct tech peer: sodium fast reactor) is private. NuScale's light-water design is a different technology class.
• Pipeline from names that matter. Meta (1.2 GW, Ohio) and Switch (12 GW framework) are not random counterparties — they are the largest AI infrastructure builders globally. Even partially converting this pipeline creates a long-duration revenue base.
• Build-own-operate captures recurring value. Each commissioned Aurora unit under a 20-year PPA is a long-duration cash flow stream. At scale, this is an infrastructure business that would be repriced at lower discount rates than a technology developer — significant valuation upside if execution occurs.
• Cash runway covers the critical 2026-2027 window. $788.4M cash with the next ATM filed provides sufficient runway to reach the Aurora-INL demonstration milestones.
• Short squeeze amplifier. 16.45% short interest means positive catalysts can produce outsized price reactions.

Red Flag 1 — Zero revenue, accelerating losses (HIGH) Revenue: $0 in every reported period (FY2021-FY2025). Operating losses: -$1.8M (FY2022) → -$18.6M (FY2023) → -$52.8M (FY2024) → -$139.3M (FY2025). Q1 2026 alone: -$51.2M, pacing toward $180-200M+ for FY2026. These are not capital investment losses — they are R&D and SG&A ahead of a revenue base that will not exist for years.

Red Flag 2 — Persistent, aggressive dilution (HIGH) Shares outstanding grew ~2.5x in under three years (69.2M end-2023 to 173.9M March 2026). A second $1.0B ATM filed May 13, 2026 confirms structural dilution continues. Every dollar of runway is purchased with shareholder dilution. Guided 2026 cash use of $430-550M (combined operating + capex) will likely require another raise before the demonstration unit reaches first operations.

Red Flag 3 — Cash runway conditional on no slippage (HIGH) At $788.4M cash and $430-550M guided 2026 total cash use, runway is approximately 17-22 months from end-FY2025 — potentially exhausted before Aurora-INL commercial operation. Any construction delay, cost overrun, or NRC COLA slip compresses this and triggers a dilutive raise into potentially weaker market conditions. The company acknowledged in Q1 2026 earnings it will require additional financing.

Red Flag 4 — Commercial NRC license not filed; prior application was denied (HIGH) The INL unit operates under a DOE RPP authorization — not NRC licensed, covering only this one demonstration unit. Every commercial Aurora requires either NRC Part 52 COLA (multi-year process) or the new Part 57 framework (still in rulemaking). The COLA for the first commercial unit has not been filed as of the data date. Oklo's prior NRC application was denied in January 2022 for insufficient safety and emergency planning information. NRC timelines for novel designs routinely run 5-7 years.

Red Flag 5 — HALEU fuel supply has no commercial solution (HIGH) Centrus Energy (the only US HALEU producer) operates at ~20 kg/yr demonstration scale. Oklo's own Aurora Fuel Fabrication Facility and DOE fuel-line pilots are in progress but nowhere near commercial scale. The Q1 2026 earnings call acknowledged fuel availability for mid-term commercial opportunities requires government collaboration. The plutonium fuel alternative (DOE talks opened May 2026) is a mid-2030s pathway with nonproliferation and cost-recovery complexity. Without fuel at commercial scale, a built reactor fleet cannot operate.

Red Flag 6 — Pipeline is largely non-binding (HIGH) The widely-cited 14 GW customer pipeline consists primarily of non-binding LOIs and master frameworks. Switch 12 GW: non-binding master power framework, no committed capital, no site-specific PPAs. Meta 1.2 GW: binding in principle, but contingent on PJM interconnection (months to over 1 year per Q1 2026 earnings). Zero executed take-or-pay PPAs with specified pricing and delivery dates have been announced. NuScale's cancelled UAMPS Carbon Free Power Project is the directly relevant precedent for how quickly LOIs can evaporate.

Red Flag 7 — CEO insider selling, high cadence (MEDIUM) CEO Jacob DeWitte sold ~$44M+ under a 10b5-1 plan in a 5-month window (Jan-May 2026), with monthly tranches of $10-14M. The 10b5-1 structure reduces the informational read — the plan predates the current price level — but monthly monetization at this scale at a pre-revenue company is unusual and a persistent headline risk.

Red Flag 8 — Valuation at 15x the only public peer (MEDIUM-HIGH) At ~$12B market cap vs NuScale's ~$800M, Oklo trades at roughly 15x the only other publicly traded SMR developer, with NuScale holding an NRC-certified design and Oklo having none. Goldman Sachs ($65) and UBS ($60) targets are both below the current price. Short interest of 16.45% reflects a significant, well-capitalized skeptical camp.

• The promotional narrative packages optionality as contracted pipeline. The Switch 12 GW agreement was described by Oklo as "one of the largest corporate power agreements in history" — but it is a non-binding master framework with no committed capital. The 14 GW pipeline figure is widely cited without distinguishing binding from non-binding. The gap between a framework and a contract is measured in years and hundreds of millions of committed dollars.
• Oklo missed its own 2025 COLA filing target. The company stated intent to submit its NRC COLA in 2025. It has not been filed as of the research date, now targeting H2 2026. This is a one-year slip on a company's own stated regulatory milestone — worth noting against the regulatory momentum narrative.
• The DOE Reactor Pilot Program criticality deadline was widely misread in retail media. A DOE RPP milestone was reported as applying to Oklo's commercial Aurora units in 2026 — the commercial units target 2027-2028. This misreading inflated near-term sentiment on a milestone that was never Oklo's commercial catalyst.
• FY2022 net income was +$3.9M despite a -$1.8M operating loss, due to non-operating items (likely warrant fair-value adjustments at SPAC formation). This positive NI figure should not be interpreted as operating profitability.
• Sam Altman's role is board chair, not CEO. Media coverage frequently implies a closer operational relationship than a board chair position represents. DeWitte is CEO and the day-to-day operator.
• TerraPower's Wyoming sodium fast reactor project — the most direct technology comparable — targets 2030 commercial operations at 345 MWe. If TerraPower (private, Bill Gates-backed) reaches criticality first, it will provide a real-world technology proof point for the sector but also demonstrate a faster execution path than Oklo's 15 MWe microreactor format.

Oklo occupies a genuine and narrow structural niche: the only publicly traded advanced fast reactor developer with physical construction underway, a DOE-authorized demonstration site, and named hyperscaler customers in a market where AI-driven electricity demand is one of the most durable secular themes of the decade. The regulatory progress in 2025-2026 — groundbreaking, NRC PDC approval in 15 days, Part 57 proposed framework — is real and not promotional.

The problem is the price. At ~$12B market cap and zero revenue, with $139.3M in operating losses in FY2025 accelerating, $82.2M in operating cash burn, shares outstanding growing 2.5x in three years, a HALEU fuel supply chain that barely exists, commercial NRC licensing not yet filed, and a customer pipeline composed almost entirely of non-binding frameworks, the current valuation prices near-perfect execution on a 5-year timeline across five simultaneous dependencies. It leaves almost no margin for the kind of slippage that is routine in first-of-kind advanced nuclear construction.

The honest upside: if Aurora-INL achieves criticality in 2027-2028, the commercial COLA is filed and accepted, and one hyperscaler LOI converts to a binding PPA with disclosed pricing, the stock re-rates toward the consensus $101 target. The honest downside: if any single dependency slips 12-18 months — as NuScale's UAMPS project illustrates is entirely possible — the stock compresses toward the Goldman/UBS range of $60-65 or below.

For a long-horizon holder who genuinely believes in the nuclear-AI baseload thesis and can tolerate ongoing dilution and a 5+ year pre-revenue period, the structural moats are real. For anyone with a shorter time horizon or lower distress tolerance, the risk-reward at current levels is asymmetric in the wrong direction: the bear case catalysts are closer and more probable than the bull case catalysts.

This is a binary-outcome, long-duration bet on first-of-kind execution. The thesis is coherent. The price demands perfection.

Research, not investment advice. Figures sourced from SEC filings and public data; verify before acting.