TerraPower
Reactor technology developer-owner: designs and licenses the Natrium reactor + storage system, co-funds the first-of-a-kind plant with DOE cost-share and hyperscaler/utility offtake-plus-development funding; long-term model is selling standardized plants and fuel/services to utilities and data-center customers. Small adjacent line in medical isotopes (Actinium-225, TerraPower Isotopes).
The Jun 18, 2025 $650M round (NVIDIA's NVentures joining Bill Gates and HD Hyundai) did not disclose a valuation, so it cannot be plotted; the $3.8B mark predates the Mar 2026 NRC construction permit, Apr 2026 construction start, and Jan 2026 Meta agreement. Hiive secondary marks exist but are not officially confirmed. Public comp Oklo trades near $8.5B (Jul 2026).
Earnings, margins, COGS & capex
TerraPower is a private, pre-commercial-revenue developer. It has raised ~$1.5B in disclosed equity since 2022 ($830M in 2022 — $750M in August plus $80M in November — at a $3.8B valuation, co-led by SK and Bill Gates; $650M in Jun 2025 with NVentures joining, terms undisclosed) plus up to ~$2B in DOE Advanced Reactor Demonstration Program cost-share for Kemmerer Unit 1. No income statement is public. Cash flows are project-development outflows until Kemmerer Unit 1 reaches commercial operation (company target 2030); the Jan 2026 Meta agreement adds customer funding to support Natrium deployment, with initial units for Meta as early as 2032. A small medical-isotopes business (Ac-225) generates undisclosed revenue.
Revenue trend
Margins
n/a (pre-revenue)
flat-negative until first COD (target 2030)
COGS structure
Not applicable yet; future plant cost structure dominated by nuclear-grade EPC (Bechtel), reactor enclosure and heavy components (HD Hyundai Heavy Industries selected May 2026 as preferred manufacturer for Reactor Enclosure System components), sodium/molten-salt systems, and HALEU fuel (up to 19.75% enriched) — the single most supply-constrained input.
Capex
Kemmerer Unit 1 is the dominant spend: FOAK plant roughly $4B all-in, cost-shared ~50% by DOE ARDP (up to ~$2B). Nuclear construction officially began Apr 23, 2026 after the Mar 4, 2026 NRC permit (~1,600 construction workers expected at peak). Sodium test facility and non-nuclear energy-island work preceded the permit.
Latest earnings
n/a
Company-stated milestones: Kemmerer Unit 1 completion targeted 2030; Meta initial units as early as 2032
- NRC construction permit (Kemmerer Unit 1)
- Granted Mar 4, 2026 — first-ever for a commercial non-light-water US power reactor
- Nuclear construction start
- Apr 23, 2026, Kemmerer, WY
- Contracted/optioned pipeline
- Meta up to 8 plants (up to 2.8 GW baseload); PacifiCorp 2 units in 2023 IRP + up to 5 studied by 2035
- Last disclosed valuation
- $3.8B (2022 round: $750M Aug + $80M Nov); Jun 2025 $650M raise terms not disclosed
- DOE cost-share
- Up to ~$2B ARDP for the demonstration plant (~50% of FOAK cost)
Growth drivers
- AI/data-center electricity demand — Meta agreement (Jan 9, 2026) for up to 8 Natrium plants / up to 2.8 GW baseload (up to 4 GW with storage boost), initial units as early as 2032 and rights to energy from up to six further units (2.1 GW) targeted by 2035
- Coal-to-nuclear repowering with PacifiCorp — 2023 IRP selects two additional Natrium units by 2033 (sites indicated in Utah); joint study covers up to 5 more by 2035
- US policy tailwind — 2025 executive orders on nuclear, DOE $2.7B enrichment awards (Jan 2026) to build domestic HALEU/LEU supply
- First-mover regulatory validation — first NRC construction permit ever for a commercial non-light-water power reactor (Mar 2026); final safety evaluation issued Dec 1, 2025, roughly 8 months ahead of schedule
- International deployment via HD Hyundai / Hyundai E&C manufacturing and commercialization agreements (May 2026 framework + joint NOAK roadmap)
- Load-following differentiation — molten-salt storage boosting 345 MWe to 500 MW for 5+ hours makes Natrium complementary to renewables-heavy grids
- Adjacent isotopes business (Ac-225 for targeted alpha cancer therapy) with near-term commercial revenue
Bull & bear
TerraPower is the de-risked pure-play on US advanced nuclear: it has the only commercial advanced-reactor construction permit in the country, steel going into the ground since April 2026, roughly half its FOAK bill paid by DOE, a Meta framework worth up to 8 plants, and a utility owner (Berkshire's PacifiCorp) already planning two more units in its resource plan. If Kemmerer reaches COD anywhere near the 2030 target, TerraPower owns the reference design for coal-site repowering and AI baseload, and its last-disclosed $3.8B mark (2022) predates every major de-risking milestone.
- Regulatory moat realized, not promised: first non-light-water commercial construction permit in NRC history (Mar 4, 2026), with the final safety evaluation delivered ~8 months ahead of schedule — competitors are years behind on this exact step
- Demand is contracted-adjacent, not hypothetical: Meta up to 2.8 GW / 8 units with funding to support deployment (initial units as early as 2032, rights to six more by 2035); PacifiCorp's 2023 IRP selects 2 more units by 2033 and studies up to 5 by 2035
- FOAK risk is socialized: up to ~$2B DOE ARDP cost-share plus ~$1.5B of post-2022 private capital means unit 1 does not need project finance to complete
- Public-market comp gap: Oklo (OKLO) — pre-revenue, without a comparable-scale construction permit — carries a ~$8.5B market cap (Jul 2026), more than 2x TerraPower's stale 2022 $3.8B mark; an IPO or new priced round is a natural repricing catalyst
- Natrium's storage-boost architecture (345 to 500 MW for 5+ hours) is the only near-term US advanced design that monetizes peak pricing and firming, improving plant economics vs fixed-output SMRs
- Gates + NVIDIA + HD Hyundai + SK cap table buys time and supply-chain access that venture-funded rivals lack
TerraPower is a single-asset construction bet with no revenue, an opaque cap-table mark from 2022, and a physical dependency (HALEU) it does not control. Sodium fast reactors have a decades-long global track record of cost overruns and poor availability; if Kemmerer slips well past the 2030 target or blows through its roughly $4B budget, the Meta rights and PacifiCorp IRP selections are just options — they evaporate, and the NOAK cost story dies with them. Private investors have no liquidity and no disclosed financials while public nuclear names can be exited any day — and the public nuclear re-rating has already partially unwound (Oklo is well off its 2025 peak).
- Everything keys off one FOAK plant: a multi-year slip or major overrun at Kemmerer impairs the entire order pipeline, which is rights, options, studies, and IRP selections — not firm EPC contracts
- HALEU is the choke point: fuel load needs material HALEU before the 2030 COD target; Centrus and the Jan 2026 DOE awardees are years from TerraPower-scale output, and the 2022 Russia cutoff already cost ~2 years
- Sodium fast reactor history is unkind: Monju (Japan) and Superphenix (France) were shut down over cost and operational issues; Natrium must be the exception
- No disclosed revenue, margins, or burn; the $3.8B mark is from 2022 and the Jun 2025 round did not disclose terms — investors cannot price dilution or runway
- Competition for the same demand is intense and publicly funded: Oklo, X-energy (Amazon-backed), GE Vernova, Westinghouse — hyperscalers are deliberately multi-sourcing (Meta split its 6.6 GW across three vendors)
- Nuclear demand from AI could disappoint: gas turbines and grid renewables deliver power in 2-3 years vs 2032 for Meta's first Natrium units; the window may be filled before delivery
What it is worth
Last-priced-round mark plus public-comp cross-check (no DCF possible: no disclosed revenue, burn, or round terms since 2022)
$2-3B or below
if Kemmerer slips materially past 2030-31, HALEU deliveries miss fuel-load timing, or the public nuclear re-rating fully unwinds and private marks compress
$5-8B on a new priced round reflecting the Mar 2026 permit, construction start, and Meta agreement vs the stale 2022 $3.8B mark, benchmarked against Oklo's ~$8.5B (Jul 2026)
$10-20B implied in an IPO scenario post-permit with the Meta pipeline converting and public nuclear sentiment near its 2025 peaks (when Oklo exceeded $20B) — highly sentiment-dependent
Last disclosed valuation $3.8B (2022 round: $750M Aug + $80M Nov, co-led by SK and Gates). The Jun 2025 $650M round (NVentures, Gates, HD Hyundai) did not disclose valuation. Public comp Oklo (OKLO) — pre-revenue, without a comparable construction permit — carries a ~$8.5B market cap as of Jul 2026, down from a 2025 peak above $20B; this suggests TerraPower's 2022 mark is stale to the low side but also that the public nuclear re-rating is volatile. Accredited-investor secondaries trade on platforms like Hiive; implied secondary marks are not officially confirmed. Value is milestone-driven: it concentrates on Kemmerer COD (2030 target) and conversion of the Meta/PacifiCorp pipeline. Not financial advice.
SWOT
Strengths
- Only commercial advanced (non-light-water) reactor in the US with an NRC construction permit and active nuclear construction — a multi-year regulatory lead
- Deep-pocketed, patient cap table — Bill Gates (founder/chairman, largest backer), NVIDIA's NVentures, HD Hyundai, SK Inc/SK Innovation, ArcelorMittal — plus up to ~$2B DOE cost-share roughly halving FOAK risk
- Natrium design differentiation — built-in molten-salt storage (345 MWe base, 500 MW boost) fits AI-load and renewables-heavy grids better than fixed-output SMRs
- Anchor customers on both sides of the demand story — regulated utility (PacifiCorp / Berkshire Hathaway Energy) and hyperscaler (Meta)
- Proven heavy-industrial partners — Bechtel EPC, GE Hitachi co-developed design, HD Hyundai Heavy Industries as preferred Reactor Enclosure System manufacturer
Weaknesses
- No commercial power revenue and no public financials — cash burn through at least the 2030 COD target with FOAK cost roughly $4B on unit 1
- HALEU fuel dependence — the 2022 Russian supply cutoff already slipped the original 2028 target to 2030; domestic enrichment (Centrus, General Matter, Orano) is still scaling
- One-project concentration — nearly all near-term value hangs on Kemmerer Unit 1 executing on schedule and budget
- Unproven cost curve — sodium fast reactors have a global history of cost overruns and outages (Monju, Superphenix) that Natrium must break
- Private-market illiquidity; last clean valuation mark is from late 2022
Opportunities
- US data-center power shortfall — hyperscalers signing multi-GW nuclear frameworks (Meta alone up to 6.6 GW across Vistra, Oklo, and TerraPower, announced Jan 2026)
- Coal-plant repowering — hundreds of retiring US coal sites with transmission interconnects match Natrium's siting model
- Federal support intensifying — 2025 nuclear executive orders, DOE $2.7B HALEU/LEU enrichment awards (Jan 2026: ~$900M each to General Matter, American Centrifuge/Centrus, Orano, plus $28M to GLE), ADVANCE Act licensing reform
- Export markets via HD Hyundai / Hyundai E&C manufacturing-and-commercialization roadmap and UK/Asia advanced-reactor programs
- Follow-on unit cost declines (NOAK) if Kemmerer stays near plan — the entire advanced-nuclear sector's economics thesis
- Potential IPO or new priced round repricing the stale 2022 $3.8B mark — public pre-revenue comp Oklo carries ~$8.5B (Jul 2026) with fewer hard milestones, though well off its 2025 peak
Threats
- Kemmerer schedule/cost overrun would impair the NOAK cost story and follow-on orders
- HALEU supply failing to scale by fuel-load date (inferred ~2029-30 from the 2030 COD target) — the binding physical constraint
- Competition: Oklo, X-energy, GE Vernova BWRX-300, NuScale, Kairos, Westinghouse AP1000 restarts all chasing the same hyperscaler and utility demand
- Policy reversal or NRC resourcing shifts after the 2026-28 political cycle
- Gas turbines and renewables-plus-storage undercutting nuclear on speed-to-power for data centers this decade
- State-backed Chinese fast-reactor program (CNNC's CFR-600 operating) could dominate export markets on price (context only, not an investable angle)
Moats, dependencies & bottlenecks
Moats
Regulatory head start (NRC construction permit + operating-license path) First commercial non-light-water power-reactor construction permit in US history; replicating it takes competitors years of NRC review
through FOAK completion Non-replicable for new entrants at this scale; covers roughly half of first-unit capital
sodium fast reactor + molten-salt storage integration Co-developed with GE Hitachi; storage-boost architecture is unique among near-term US designs
Berkshire/PacifiCorp, Meta, HD Hyundai) relationship-dependent Access to capital, sites, manufacturing, and offtake that pure startups cannot match
Site selection, workforce transition, and transmission reuse know-how compounds with each unit
Dependencies
General Matter, Orano, Framatome, Global Nuclear Fuel/GE Vernova GEV, ASP Isotopes ASPI, DOE allocation program) The binding constraint; 2022 Russian cutoff already delayed COD ~2 years; DOE's Jan 2026 $2.7B enrichment awards help but scale-up timing is tight ahead of the 2030 COD target
Up to ~$2B of unit-1 funding plus fuel allocation; exposure to appropriations and administration priorities
Nuclear-grade construction execution determines schedule and cost; labor ramp toward ~1,600 workers at peak
Construction permit granted Mar 2026; operating-license review is a separate future gate
Preferred manufacturer for Natrium Reactor Enclosure System components (May 2026 framework agreement); HD Hyundai group is also an investor
Host utility and follow-on buyer; IRP selections are plans, not signed EPC contracts
Jan 2026 framework for up to 8 units with deployment funding; most units are rights, deliveries from ~2032
Advantages
- Only US advanced-reactor developer in commercial nuclear construction (as of 2026-07)
- ~50% of FOAK cost carried by DOE — best-capitalized risk position in the sector
- Storage-integrated design monetizes peaking and firms renewables, not just baseload
- Hyperscaler (Meta) + regulated utility (PacifiCorp) demand on the same product
- Founder-anchored patient capital (Gates) plus strategic investors (NVIDIA, HD Hyundai, SK, ArcelorMittal)
Weaknesses
- Zero commercial power revenue until the 2030 COD target at the earliest; undisclosed burn and runway
- Valuation mark stale (2022); Jun 2025 round terms undisclosed
- Single-project concentration in Kemmerer Unit 1
- Fuel cycle not vertically owned — HALEU is bought, not made
- Pipeline beyond unit 1 is rights/options/studies/IRP selections, not firm contracts
Bottlenecks
- HALEU enrichment capacity and delivery timing ahead of fuel load (inferred ~2029-30 for a 2030 COD)
- Nuclear-qualified skilled labor and supply chain (large sodium components, safety-grade fabrication)
- NRC operating-license review after construction — an unpassed gate for any commercial non-light-water plant
- FOAK cost discipline at Kemmerer (~$4B estimate) setting the NOAK price for every follow-on order
- Long-lead heavy fabrication and Reactor Enclosure System manufacturing slots (HD Hyundai)
Top signals & trends
Top signals
Nuclear construction began Apr 23, 2026 after the Mar 2026 permit; watch for schedule-slip announcements
DOE awarded $2.7B for enrichment Jan 2026; TerraPower's fuel-load timing is the test
Firm EPC contracts would validate NOAK economics; silence past 2027 would be a warning
Would reprice the stale 2022 $3.8B mark; the Oklo comp implies uplift, though the public nuclear re-rating has partially unwound from its 2025 peak
The remaining regulatory gate before startup
Technical de-risking milestones before startup
Trends
strongly positive · Hyperscalers signing multi-GW nuclear frameworks; Meta alone locked up to 6.6 GW across Vistra, Oklo, and TerraPower (Jan 2026)
Faster licensing and domestic HALEU buildout directly de-risk TerraPower's two biggest gates
Natrium's siting model reuses coal transmission and workforce; Kemmerer is the national template
positive but rate-limiting · Created the HALEU gap that delayed the project; now driving the domestic capacity TerraPower needs
positive but volatile · Oklo re-rated above $20B at its 2025 peak but sits near $8.5B as of Jul 2026 — supportive comp backdrop for eventual liquidity, well off highs
2032 first Meta deliveries leave a window rivals can fill with faster technologies
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.
Only US-licensed HALEU producer (Piketon, OH); won ~$900M DOE enrichment award Jan 2026
~$900M DOE enrichment award Jan 2026 for domestic HALEU capacity
~$900M DOE enrichment award Jan 2026; fuel-cycle services
HALEU deconversion/fuel services agreement
HALEU supply agreement via Quantum Leap Energy subsidiary
Fuel fabrication partner; Natrium co-developer via GE Hitachi
Preferred manufacturer, Natrium Reactor Enclosure System components (May 2026); group is an investor
EPC contractor for Kemmerer Unit 1
Kemmerer host utility; 2 more Natrium units in its 2023 IRP by 2033, up to 5 more studied by 2035
Jan 9, 2026 agreement: up to 8 Natrium plants / up to 2.8 GW baseload (4 GW with storage), initial units as early as 2032
Molten Chloride Fast Reactor (MCFR) development partner via MCRE at Idaho National Laboratory
Public, pre-revenue fast-reactor developer (Aurora, 15-75 MWe); Sam Altman-backed; Meta (Jan 2026) and DOD agreements; no construction permit for a Natrium-scale plant; ~$8.5B market cap Jul 2026
Only NRC design-certified light-water SMR; lost its CFPP anchor project (2023) but has RoPower (Romania) and data-center interest
HTGR (Xe-100) developer backed by Amazon (~$500M+ round, Oct 2024); Dow Seadrift industrial project; the other DOE ARDP demonstration awardee
BWRX-300 SMR under construction at OPG Darlington (Canada) and TVA Clinch River path; also TerraPower's Natrium co-developer — partner and rival
AP1000 large-reactor newbuilds/restarts and AP300 SMR; benefits from the same policy push
Fluoride-salt-cooled reactor; Google 500 MW master agreement; Hermes demo permitted and under construction at Oak Ridge
SMR-300 plus the Palisades restart — first US reactor restart, nearest-term new nuclear capacity
Context only: operating commercial-scale fast reactors outside Russia; potential export-market competitor — not an investment call