Oklo Inc
Build-own-operate independent power producer: develops Aurora powerhouses, sells electricity + heat via 20+yr PPAs for recurring revenue (analogous to a solar IPP); adjacent radioisotope sales (Atomic Alchemy) and a planned fuel-recycling business. No revenue yet — first reactor targeted ~2027-2028.
The thesis on this name
State of Data-Center Power
The highest-optionality SMR datacenter play — a build-own-operate microreactor model (sell power, not reactors) with a ~14-18GW pipeline anchored by Switch, Equinix and Meta, May-2026 NRC design-criteria approval, an upsized 75MWe design, and targeted first operation late-2027/2028 at Aurora-INL (fact). Levelized cost guided $40-90/MWh and ~$7,000/kW capex — credible vs NuScale's ~$20,000/kW (fact). This is explicitly an OPTION, not a near-term driver: real power is years out and execution/licensing risk is high. Size it as frontier optionality on the 2028-30 datacenter-nuclear narrative, with falsifiable milestones — never as a near-dated toll.
State of Data-Center Power
Pre-revenue SMR call option on datacenter baseload — ~14GW pipeline (Switch/Equinix/Meta) + May-2026 NRC design approval, but revenue is years out.
State of Data-Center Power
Pre-revenue 14GW pipeline + NRC approval; venture-style convex sleeve, capped tiny for 35% bear probability.
State of Nuclear Energy
The SMR leader call-option: the only advanced-fission name with both the balance sheet (~$2.5B cash) and the licensing momentum to plausibly reach commercial operation this decade, riding the AI-data-center power bid. The build-own-operate model means it sells electricity, not reactors — bigger TAM if it works. But it is pre-revenue with an ~$9-12B mcap; almost all the value is terminal/narrative.
State of Nuclear Energy
A venture-style call option on advanced-fission commercialization: best-funded pure-play (~$2.5B cash, 25yr+ runway), fastest licensing momentum, Aurora-INL physically under construction — but pre-revenue at ~$10B.
State of Nuclear Energy
The best-capitalized advanced-fission pure-play: ~$2.5B cash post the Jan-2026 $1.18B raise, 25yr+ runway at guided $80-100M burn, DOE Nuclear Safety Design Agreement approved, NRC PDC topical accepted in 15 days, Aurora-INL construction physically progressing under the Reactor Pilot Program. Sized as a venture call: real licensing momentum + an NVIDIA/AI-data-center demand narrative, but zero commercial kWh and an ~$9-12B mcap on no revenue. Starter only; accumulate on the inevitable 30-40% drawdowns.
Earnings, margins, COGS & capex
Oklo is pre-revenue: it books no product sales and funds a widening cash burn (Q1 FY26 net loss $33.1M, operating loss ~$51.2M) entirely from a ~$2.54B war chest raised via equity. The model is capex-heavy upfront (build-own-operate reactors), with recurring PPA cash flow only beginning once the first Aurora reaches commercial operation, targeted 2027-2028. Near-term 'value' is regulatory de-risking (May-2026 NRC design-criteria approval, July-2026 criticality target) and pipeline conversion, not earnings. The thesis is entirely a function of execution against the licensing and construction timeline.
Revenue trend
Margins
flat
down (opex rising on headcount + SBC)
down (loss widening YoY)
down
COGS structure
No COGS today — pre-operations. Future cost structure once reactors run: HALEU fuel (metallic U-Zr ~19% enrichment; first cores from recovered EBR-II material via Centrus/DOE), O&M and staffing of each powerhouse, and depreciation of the build-own-operate plant base. As an IPP, gross margin should be high per-MWh once a plant is commissioned (fuel is a small share of nuclear LCOE), but heavy upfront capex makes the model depreciation- and financing-cost-driven. (estimate — no operating history)
Capex
FY26 guidance: $350–450M cash used in investing (fact, Q1 FY26) funding the first Aurora-INL build, the Aurora fuel-fabrication facility at INL, the Aurora-Ohio (Meta 1.2GW) and Aurora-Eielson sites, plus Atomic Alchemy/Groves isotope reactor. Build-own-operate means Oklo carries the plant capex itself rather than selling reactors — capital intensity is structurally high and scales with each fleet addition.
Latest earnings
EPS −$0.19, in line with consensus (fact). For a pre-revenue developer the print is judged on milestones/cash, not EPS; shares fell ~6% after-hours on the loss/dilution.
FY2026: cash used in operations $80–100M; cash used in investing $350–450M. Reaffirmed ~14GW pipeline; targeting Aurora-INL + Groves criticality by ~July-4-2026 and first commercial operations 2027-2028. (fact, Q1 FY26)
- Cash + securities
- $2.54B (31-Mar-26)
- Total liabilities
- $64.9M (near debt-free)
- Customer pipeline
- ~14GW (non-binding)
- Net loss / EPS
- $33.1M / −$0.19
Growth drivers
- ~14GW non-binding pipeline conversion to binding PPAs (Switch up to 12GW MPA through 2044; Equinix 500MW with $25M prepay; Meta 1.2GW Ohio campus targeting ~2030)
- NRC licensing progression — May-2026 Principal Design Criteria approval → Aurora-INL COLA acceptance/approval → fleet-replication standard design
- First criticality (DOE pilot, July-2026 target) and first commercial Aurora operations (2027-2028) — the revenue ignition point
- Vertical integration into HALEU fuel fabrication + fuel recycling (lower long-run fuel cost, supply security)
- Atomic Alchemy radioisotope sales — a near-term, smaller revenue stream independent of the power timeline (NRC materials license granted Mar-2026)
- AI-datacenter baseload demand secular tailwind (hyperscaler power scarcity, behind-the-meter nuclear interest)
Bull & bear
Oklo is the best-positioned pure-play to monetize AI-datacenter baseload demand: a recurring-revenue IPP model, ~14GW of marquee pipeline, a near-debt-free $2.5B balance sheet, and accelerating NRC/DOE regulatory de-risking that could ignite first revenue by 2027-2028.
- Regulatory inflection: May-2026 NRC Principal Design Criteria approval (accelerated) + July-2026 criticality target materially de-risk the licensing path that killed the 2022 application
- Demand is real and blue-chip: Switch (up to 12GW MPA to 2044), Equinix (500MW, $25M prepay), Meta (1.2GW Ohio, PJM interconnection filed) — AI power scarcity favors firm baseload
- Build-own-operate yields decades of high-margin PPA cash flow per plant once commissioned, plus optionality from fuel recycling and Atomic Alchemy radioisotopes
- ~$2.54B cash, only ~$65M liabilities — funds the first reactors without immediate financing distress and several years of runway
- Standardized fleet + federal support (DOE loans/pilot, NRC modernization) could compress per-unit cost/time and scale faster than skeptics expect
Oklo is a richly-valued (~$9.4B) pre-revenue developer whose first-of-a-kind reactor has never operated; the thesis rests on a non-binding pipeline and a multi-year, dilution-funded build where any schedule slip, cost overrun, or licensing setback resets the equity.
- No revenue, widening losses (Q1 FY26 net loss $33.1M), and first commercial cash flow not until 2027-2028 at the earliest — the timeline is the whole story and it has historically slipped in nuclear
- ~14GW 'pipeline' is non-binding LOIs/MPAs, not contracted revenue; hyperscalers retain full optionality on gas, grid, geothermal, or rival reactors
- Funded by serial dilution (~$1.18B ATM in one quarter) — shareholders keep paying for a capex-heavy build with no product yet
- First-of-a-kind execution risk is severe: no Aurora has run, the 2022 NRC application was denied, and FOAK nuclear is the canonical cost/schedule-overrun category
- Valuation prices flawless delivery; a rate shock or risk-off tape raises dilution cost and could starve a multi-year build — a textbook 'story stock' on a long fuse
What it is worth
Option-value / scenario DCF on risk-adjusted future PPA cash flows (no current earnings or revenue to anchor multiples); cross-checked vs SMR peers (NuScale ~$3.6B) on pipeline-GW and balance-sheet terms.
$15-30/sh
schedule slip / cost overrun / licensing setback or a risk-off capital-markets shock resets the de-rated, dilution-funded story
$45-75/sh
current ~$55 range; market prices steady milestone progress with execution/dilution discount intact
$120-190+/sh
flawless licensing + first reactors on time (2027-2028), pipeline goes binding, fleet replication scales; revisits prior 52wk high ($193.84)
Worth is the probability-weighted NPV of converting ~14GW of pipeline into operating, owned reactors generating multi-decade PPA cash flow — gated by NRC approval, FOAK execution, and dilution; not supportable on any near-term financial metric.
SWOT
Strengths
- Fortress balance sheet — ~$2.54B cash/securities, near debt-free — multiple years of runway at guided ~$80-100M op burn (FY26)
- Regulatory momentum — May-2026 NRC Principal Design Criteria approval on an accelerated track (<half normal timeline); DOE pilot-program selection
- Differentiated build-own-operate IPP model → recurring multi-decade PPA cash flows vs one-time reactor sales
- Blue-chip non-binding demand — ~14GW pipeline (Switch 12GW, Equinix 500MW, Meta 1.2GW) anchored to AI-datacenter baseload need
- Vertical integration — in-house HALEU fuel fab + fuel recycling + Atomic Alchemy radioisotopes diversify beyond power
Weaknesses
- Zero product revenue and a widening loss — entirely pre-commercial; first cash flows years out
- Heavy, recurring equity dilution (e.g. ~$1.18B ATM in Q1 FY26) funds the build-own-operate capex
- Execution risk on a first-of-a-kind reactor — no Aurora has ever operated; 2022 NRC application was previously denied
- Pipeline is non-binding LOIs/MPAs, not contracted revenue — convertibility unproven
- Valuation (~$9.4B) discounts flawless multi-year delivery; no near-term financial anchor
Opportunities
- AI/datacenter power scarcity could pull forward binding behind-the-meter nuclear PPAs at premium prices
- Fleet standardization (one approved design replicated) compresses cost/time per incremental reactor
- Radioisotope market (cited ~$55.7B by 2026) gives an earlier, smaller revenue path via Atomic Alchemy
- Federal tailwinds: DOE loans/pilot program, NRC licensing modernization, pro-nuclear policy
- Fuel recycling (closing the fuel cycle) is a long-run cost + national-security moat if it works at scale
Threats
- Schedule slip / cost overrun on first reactor — the classic nuclear-project risk — would reset the thesis
- Competing SMR designs further along on licensing (NuScale has NRC design approval)
- Customer optionality: hyperscalers may favor gas, geothermal, grid PPAs, or rival reactors if Oklo slips
- Capital-markets dependence: a risk-off / rate shock raises dilution cost and could starve the build
- HALEU/fuel supply-chain constraints (enrichment capacity, recovered-material logistics) gating deployment
Moats, dependencies & bottlenecks
Moats
Medium-high once a design is approved May-2026 NRC Principal Design Criteria approval + DOE pilot selection are real barriers, but design approval ≠ a built/operating plant, and rivals (NuScale) hold NRC design approval too. Moat widens sharply only after fleet COLA approval + first operation.
if contracts go binding 20-44yr PPAs (Switch to 2044) create durable switching costs — but today they are non-binding; the moat is prospective, not yet realized.
if it works at scale In-house fuel fabrication and closing the fuel cycle could lower long-run cost and lock supply in a constrained HALEU market — unproven commercially.
Marquee anchors (Switch, Equinix, Meta) + NVIDIA/LANL collaboration give brand and demand visibility, but relationships are early and non-exclusive.
$2.54B near-debt-free balance sheet lets it self-fund first reactors — an advantage over thinner-capitalized SMR peers, but contingent on capital-markets access.
Dependencies
Revenue cannot begin without COLA approval to build/operate; the 2022 application was denied, so this is the gating dependency.
DOE/EBR-II recovered material, fuel-fab facility) First cores depend on scarce ~19%-enriched HALEU and the not-yet-operating Aurora fuel-fab facility; enrichment/logistics constraints can gate deployment.
Equinix, Meta) The entire revenue thesis rests on non-binding pipeline becoming contracted offtake; hyperscalers retain optionality.
Build-own-operate capex is funded by dilution ($1.18B ATM in Q1 FY26); a risk-off shock raises cost of capital and could starve the build.
No Aurora has been built or operated; FOAK nuclear is the canonical cost/schedule-overrun risk.
Advantages
- Near-debt-free $2.54B cash/securities — multi-year runway and self-funded first reactors (31-Mar-2026)
- Accelerated NRC design-criteria approval + DOE pilot-program selection — ahead of most non-NuScale SMR peers on real federal traction
- ~14GW of marquee, AI-datacenter-anchored demand visibility (Switch/Equinix/Meta) few competitors can match
- Recurring-revenue IPP model (own + operate) captures decades of margin per plant vs one-time reactor sales
- Optionality beyond power — Atomic Alchemy radioisotopes (NRC materials license, near-term sales) + fuel recycling
- High-profile ecosystem (NVIDIA/LANL fuel-AI collaboration, Sam Altman chairman legacy) aiding capital access and mindshare
Weaknesses
- Zero product revenue; first commercial operations not until 2027-2028 — long fuse
- Widening losses ($33.1M net loss Q1 FY26) and ongoing, large shareholder dilution
- First-of-a-kind reactor never built or operated; prior 2022 NRC application denied
- Pipeline is non-binding — no contracted, recognized revenue backlog
- Valuation (~$9.4B) leaves no margin for the schedule slips endemic to nuclear builds
- Heavy dependence on external capital + on HALEU fuel and fuel-fab readiness it doesn't yet fully control
Bottlenecks
- NRC combined-license approval and the per-site licensing/siting + environmental review cadence
- HALEU fuel availability and the still-to-be-commissioned Aurora fuel-fabrication facility
- First-of-a-kind construction — supply chain, skilled nuclear labor, and schedule for the inaugural Aurora-INL build
- Grid interconnection / site permitting (e.g. PJM interconnection queue for the Meta-Ohio 1.2GW campus)
- Capital intensity — each owned plant requires large upfront capex funded ahead of any cash flow
Top signals & trends
Top signals
The single biggest near-term de-risking event; on-time criticality validates the technical path. A slip is a major negative tell.
Moves from design-criteria approval toward an actual license to build/operate — the gate to revenue.
Turns a non-binding pipeline into contracted offtake; the proof the demand is real and bankable.
Watch runway and dilution pace; a raise above guide or off-schedule would pressure the equity.
An early, smaller revenue stream independent of the multi-year power timeline.
FOAK nuclear overruns are the canonical failure mode; first slip would reset the timeline thesis.
Trends
Core demand tailwind — firm, 24/7, carbon-free power is exactly what AI campuses need; underpins the Switch/Equinix/Meta pipeline.
Accelerated PDC approval (<half normal timeline) shows the regulatory path is genuinely speeding up for advanced reactors.
Necessary enabler but currently constrained; supply ramp is both a tailwind and a bottleneck.
Abundant investor interest funds the build but inflates valuations and invites competition (NuScale, NANO, X-energy, TerraPower, Kairos).
Favors modular, sited-at-load reactors over grid-dependent supply — structurally suits Oklo's powerhouse model.
Supports the Atomic Alchemy adjacency as an earlier, policy-favored revenue path.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
HALEU enrichment/fuel-supply partner — critical for ~19%-enriched fuel; also a JV partner.
Department of Energy / Idaho National Laboratory Host site (INL), DOE pilot-program sponsor, and source of recovered EBR-II HALEU material for first cores.
Partner on plutonium-bearing fuel validation (with NVIDIA) under the DOE pilot program.
Partner to develop US nuclear fuel-fabrication infrastructure.
Datacenter power/cooling-infrastructure collaborator for integrating reactors with AI campuses.
Energy-services partner referenced in datacenter-power buildout collaboration.
Non-binding Master Power Agreement for up to 12GW through 2044 — the anchor of the pipeline.
500MW agreement with a $25M prepayment for a 20-year PPA.
1.2GW Aurora-Ohio campus (PJM interconnection filed), targeted ~2030.
Aurora-Eielson microreactor project — government/defense baseload customer.
Healthcare / research / defense isotope buyers End markets for Atomic Alchemy radioisotopes (large cited market opportunity).
Closest US-listed SMR peer; holds NRC Standard Design Approval (77MWe module) — ahead of Oklo on certified design, lighter on datacenter pipeline. Light-water vs Oklo's fast-spectrum.
Microreactor developer (KRONOS MMR ~15MWe; CPA submitted Apr-2026) with its own HALEU fuel subsidiary; earlier-stage but overlapping microreactor niche.
Gates-backed sodium-fast reactor (345MWe) under construction in Wyoming; technically closest analog, larger scale, well-capitalized — a direct fast-reactor rival.
HTGR developer with Amazon/DOE backing and datacenter ambitions; competes for the same hyperscaler offtake.
Fluoride-salt-cooled reactor with a Google PPA and NRC construction permits in hand — a credible datacenter-nuclear competitor.
Incumbent large-reactor operators already signing datacenter PPAs (e.g. restarts, uprates) — the 'available now' alternative to waiting for SMRs.