Overview
This deep dive compares three nuclear fission companies as investments in the AI-datacenter power buildout: Valar Atomics, Aalo Atomics, and Oklo. All three sell the same top-level story — small, factory-built reactors delivering firm, behind-the-meter power to AI datacenters — but they differ radically in technology, regulatory strategy, commercial traction, team, and price.
The analysis assumes the valuations specified in the prompt: $2–3B for Valar and Aalo (Valar’s actual April 2026 Series B priced it at $2.0B; Aalo’s last disclosed mark was ~$450M in August 2025, with an undisclosed May 2026 Series B-II, so $2–3B for Aalo is an assumed, not observed, price) and Oklo at its public market valuation of ~$9.0B (NYSE: OKLO, ~$51.80/share as of 2026-07-06).
This is research, not investment advice. All figures are as of early July 2026 and sourced in footnotes on each page.
Pages
- Market Context — the AI power crunch, hyperscaler nuclear deals, and the 2025–26 regulatory revolution (DOE Reactor Pilot Program, NRC reform) that makes these companies investable at all
- Oklo — the public benchmark: Aurora sodium fast reactor, 14 GW customer pipeline, $2.5B cash, first power targeted late 2027
- Valar Atomics — the velocity bet: TRISO-fueled HTGRs, “gigasites,” second startup reactor to go critical, Nvidia demo, $2B Series B
- Aalo Atomics — the capital-efficiency bet: MARVEL-derived sodium microreactors, LEU fuel supply chain, criticality on July 4, 2026, on ~$136M raised
- Head-to-Head & Verdict — dimension-by-dimension comparison, valuation math at the assumed marks, scenario analysis, and portfolio conclusions
Key Findings
- The July 4, 2026 criticality race reset the field. Under the DOE Reactor Pilot Program, four startup test reactors achieved criticality by the Executive Order 14301 deadline: Antares (Mark-0, June 4), Valar (Ward 250, June 2026), Deployable Energy (Unity, early July), and Aalo (Aalo-X, July 4). Valar and Aalo are two of only four venture-backed companies ever to take a self-built reactor critical — a milestone Oklo itself has not yet reached.1
- But criticality ≠ commercial power. These are ~100 kWt-class zero/low-power test units under DOE authorization, not NRC-licensed commercial plants. The bridge from DOE test reactor to selling commercial electrons is an NRC license, and an April 2026 NRC proposed rule to fast-track licensing of DOE/DOD-authorized designs is the single most important regulatory catalyst for both private names.2
- Oklo is the de-risked-but-crowded play. ~$9.0B market cap, ~$2.5B cash, zero revenue, a ~14 GW non-binding customer pipeline (Switch 12 GW, Meta 1.2 GW Ohio, Equinix 500 MW with a $25M prepayment), first Aurora (75 MWe) under construction at INL targeting power in late 2027–early 2028. The stock is down ~72% from its late-2025 high of $193.84, which materially improves the entry versus six months ago.3
- At equal $2–3B pricing, Aalo screens as the better risk-adjusted private bet. It reached criticality on ~$136M of disclosed capital (vs. Valar’s ~$600M), uses commercially available 5% LEU fuel (Urenco contract — the first US advanced-reactor commercial enrichment contract) rather than supply-constrained HALEU or TRISO, has the strongest technical founder (CTO Yasir Arafat, who led INL’s MARVEL program), a named utility PPA negotiation (Idaho Falls Power, 7 reactors), and a 2027 reactor-plus-datacenter demonstration at INL.
- Valar is the highest-variance position. Unmatched narrative velocity (first startup nuclear electricity, Nvidia demo on July 1, 2026, defense-tech investor syndicate, Philippines regulatory arbitrage), the most capital, and the most aggressive regulatory posture (co-plaintiff suing the NRC) — but also the thinnest nuclear engineering bench, credible third-party criticism of its claims, and a commercial unit (~5 MWe HTGR) whose per-unit economics depend entirely on unproven mass manufacturing.
- Fuel is the hidden differentiator. Oklo needs HALEU (first core from recovered EBR-II material; Centrus deliveries only from 2029). Valar needs TRISO compacts (tiny US fabrication base). Aalo runs on 5% LEU UZrH — the only one of the three that can buy fuel from today’s commercial supply chain.
Snapshot
| Oklo | Valar Atomics | Aalo Atomics | |
|---|---|---|---|
| Valuation (this doc) | ~$9.0B public | $2–3B (actual: $2.0B, Apr 2026) | $2–3B (assumed; last disclosed ~$450M) |
| Founded | 2013 | 2023 | 2023 |
| Reactor | Aurora — 75 MWe sodium fast reactor, heat pipes, metallic HALEU | Ward — helium-cooled HTGR, TRISO fuel, ~5 MWe class | Aalo-1 — 10 MWe sodium-cooled, UZrH (TRIGA-type) LEU fuel |
| Datacenter product | Powerhouse + PPA (build-own-operate) | “Gigasite” clusters of hundreds of reactors | Aalo Pod: 5 × Aalo-1 = 50 MWe per pod |
| Criticality achieved | Not yet (first power late 2027–28) | ✅ Ward 250, June 2026 (100 kWt) | ✅ Aalo-X, July 4, 2026 |
| Regulatory path | NRC COLA in review + DOE route at INL | DOE pilot + NRC lawsuit + Philippines | DOE pilot + full NRC COLA planned 2026 |
| Fuel | HALEU (constrained) | TRISO (constrained) | 5% LEU (commercial) |
| Anchor customers | Meta 1.2 GW, Switch 12 GW, Equinix 500 MW (all non-binding) | Nvidia collaboration (exploratory 30 MW) | Idaho Falls Power PPA negotiation (~75 MW); INL datacenter 2027 |
| Cash / capital | ~$2.54B cash | ~$600M raised (incl. $110M debt) | >$136M disclosed + undisclosed B-II |
Footnotes
-
The deadline arrives: Checking in on the Reactor Pilot Program — ANS Nuclear Newswire; DOE — U.S. Department of Energy Reactor Pilot Program ↩
-
NRC proposed rule for licensing reactors authorized by DOE, DOD — ANS Nuclear Newswire ↩
-
Oklo Inc. (OKLO) Stock Price & Overview — stockanalysis.com; Oklo Stock Is Down 72% From Its 52-Week High — TIKR ↩