Pricing basis
Per the prompt: Valar and Aalo at $2–3B each; Oklo at its public ~$9.0B. One asymmetry must be stated up front because it drives the whole conclusion: Valar’s $2B is an observed, market-clearing price (April 2026 Series B). Aalo’s $2–3B is an assumption sitting 4–7× above its last disclosed round (~$450M, Aug 2025; May 2026 B-II undisclosed). Oklo’s $9B is a liquid daily price, down 72% from its high. Same number, three very different epistemic states.
Dimension-by-dimension
1. Technology risk
| Oklo | Valar | Aalo | |
|---|---|---|---|
| Physics pedigree | EBR-II sodium fast reactor, 30 yrs of heritage | HTGR/TRISO — proven fuel form (AVR, HTR-PM) | TRIGA UZrH fuel + sodium pool — most-demonstrated combo |
| What’s actually been built | Nothing critical yet; 75 MWe FOAK under construction | 100 kWt test unit, critical, powering loads | Critical test reactor at INL, 10 months build |
| Gap to commercial unit | One scale step (in construction) | ~750× thermal scale-up + helium power conversion | ~10 MWe unit next door (2027), then 5-pack pod |
| Core uncertainty | FOAK construction/cost | Mass-manufacturing economics of 5 MWe units | UZrH at power-reactor duty cycle; pod integration |
Ranking: Aalo ≥ Oklo > Valar. Oklo’s design maturity is highest, but Aalo’s fuel/coolant conservatism plus a live reactor gives it the least physics risk per dollar. Valar’s distance from test article to product is the largest in the group.
2. Fuel supply chain — the quiet decider
- Aalo: 5% LEU (Urenco contract, GNF fabrication). Orderable today from the commercial industry. Only company of the three with zero exotic-fuel dependency.
- Oklo: HALEU. First cores from recovered EBR-II material (finite, government-gifted); Centrus deliveries start 2029; long-term answer is Oklo’s own recycling facility (huge if true, capital-intensive if not).
- Valar: TRISO. US fabrication capacity (TRISO-X, BWXT) is small and largely committed; a thousands-of-reactors gigasite implies building a fuel plant nobody has scoped publicly.
Ranking: Aalo ≫ Oklo > Valar. In every historical nuclear ramp, fuel was on the critical path. This is Aalo’s most underpriced advantage.
3. Regulatory position
- Oklo: only company with a commercial COLA in active NRC review, an approved PDC topical report (on an accelerated clock), and DOE authorization for unit 1. Two independent paths to selling power.
- Aalo: DOE authorization achieved (Aalo-X critical), full COLA planned for 2026 — the Oklo playbook, three years behind but moving faster per month.
- Valar: DOE authorization achieved, but the commercial path is a bet on litigation (Texas v. NRC) and/or the April 2026 proposed rule bridging DOE-authorized designs into NRC licenses, plus a Philippines hedge. Highest political beta in the group — biggest winner if the deregulatory wave continues, most stranded if it reverses.
Ranking: Oklo > Aalo > Valar.
4. Commercial traction
- Oklo: ~14 GW pipeline. Non-binding, but with real names (Meta 1.2 GW Ohio with prepayment mechanics, Switch 12 GW, Equinix 500 MW + $25M cash) and a 2027–28 first-revenue date.
- Aalo: Idaho Falls Power PPA negotiation (75 MW, ~2030), Project Ascension datacenter pairing (2027), NRG/Tishman/Hitachi as strategic channels. Real but small counterparties.
- Valar: Nvidia collaboration (30 MW exploratory), zero disclosed offtake. The demo was historic; the order book is empty.
Ranking: Oklo ≫ Aalo > Valar.
5. Team
- Oklo: deepest bench, 13 years of institutional learning including a COLA denial survived; public-company infrastructure.
- Aalo: best founder-market fit in the private pair — Arafat is arguably the most credentialed microreactor engineer of his generation (MARVEL, eVinci); Loszak supplies product/velocity culture.
- Valar: strongest operators-from-adjacent-hard-tech (Relativity, Ultra Safe Nuclear alumni) and the best capital-raising founder, but the thinnest nuclear engineering leadership and a documented pattern of overclaiming.
Ranking: Oklo ≈ Aalo > Valar (different strengths: institutions vs. founder).
6. Capital position
- Oklo: $2.54B cash, ATM access, ~$100M/yr operating burn — funded through first fleet deployments.
- Valar: ~$600M raised (incl. $110M debt) — funded through the next 2–3 technical gates.
- Aalo: ~$136M disclosed + undisclosed B-II — needs a large round within ~12–18 months; financing risk is its principal risk.
Ranking: Oklo ≫ Valar > Aalo.
Valuation math at the given marks
Oklo at $9.0B
EV ≈ $6.5B net of cash. An operating GW at $100/MWh, 90% CF, 40–50% EBITDA margin is worth ~$4–6B EV at infrastructure multiples. The stock therefore prices ~1.2–1.6 GW of successful NOAK deployment — roughly “Meta Ohio works, on time, on economics.” Everything else (Switch’s 12 GW, Equinix, isotopes, fuel recycling) is optionality. After the 72% drawdown this is a defensible, not cheap, price; at the $193 peak it priced ~4–5 GW of flawless execution. Consensus targets ($83–101) imply the street underwrites roughly one additional GW of pipeline conversion.
Valar at $2–3B
Backing out what $2.5B (midpoint) requires: at the same $4–6B/operating-GW yardstick, Valar must credibly promise ~0.5–1 GW of deployed gigasite capacity — i.e., 100–200 of its 5 MWe units operating profitably — for the current mark to be NPV-fair, before any venture-style upside. A 10× venture outcome ($25B) requires gigasites at multi-GW scale plus the synfuel/hydrogen option monetizing — outcomes bigger than Oklo has ever been worth. The price already pays for the narrative of mass manufacturing; the demo reactor, while real, supports perhaps a $500M–$1B fundamental mark on milestone-comparables (Aalo hit similar milestones at $450M+undisclosed).
Aalo at $2–3B (assumed)
Identical arithmetic applies — ~0.5–1 GW of implied successful deployment — but with three differences in Aalo’s favor: the fuel chain is commercial, the 2027 INL datacenter demo is the nearest proof event, and capital efficiency to date suggests each incremental dollar buys more progress. And one difference against: you’d be marking it up 4–7× from the last observed price yourself. If an investor can access Aalo anywhere below ~$1.5B, the risk-reward dominates Valar at $2B on nearly every dimension except balance sheet and hyperscaler narrative.
Cross-check: implied value per demonstrated watt
Deliberately crude but disciplining — market cap per demonstrated thermal watt:
| Demonstrated | Mark | $/Wt demonstrated | |
|---|---|---|---|
| Oklo | 0 (constr. underway) | $9.0B | n/a (pre-demonstration at $9B) |
| Valar | ~100 kWt | $2.0B | ~$20,000/Wt |
| Aalo | ~100 kWt-class test | $2–3B (assumed) | ~$20–30,000/Wt |
The point isn’t the absurd unit prices — it’s that all three are essentially pre-product, and the differences investors are paying for are (a) regulatory position, (b) fuel chain, (c) customer paper, and (d) team. Which is exactly the four tables above.
Scenarios (2030 horizon)
Base case — the DOE bridge works, slowly. NRC finalizes the DOE-authorized-design rule in 2027; Oklo delivers Aurora-INL in 2028 (a year late), Meta phase 1 slips to 2031–32. Aalo runs its INL datacenter in 2027–28 and converts Idaho Falls. Valar operates a several-MW gigasite pilot but hasn’t cracked volume manufacturing. Oklo 1.5–2.5× ($14–22B); Aalo ~2× if entered at $2.5B (more from lower entry); Valar flat-to-2× with high financing-round variance.
Bull case — AI load growth stays vertical and policy holds. Binding hyperscaler PPAs land on the private names too (Nvidia converts for Valar; a hyperscaler anchors Aalo pods). First mass-produced units ship. Oklo revisits/exceeds prior high (3–4×); Valar and Aalo become $10–20B companies (4–8×). Valar has the highest ceiling (synfuels TAM); Aalo the highest probability of reaching a mid-case.
Bear case — a demo incident, a policy reversal (2029 administration change), or hyperscaler capex digestion. LOIs quietly lapse; DOE pathway freezes; private rounds reprice. Oklo −50–70% (cash floor ~$14/share of net cash provides some support); Valar/Aalo down-round or worse — Aalo’s thin treasury makes it the most fragile in this branch, Valar’s $600M buys survival.
Verdict
- Core position: Oklo. It is the only investable-at-scale, liquid expression of the thesis, with the regulatory lead, the cash, and the customer paper. The post-drawdown price finally offers a rational (if still demanding) entry. Size it as the anchor; expect violent volatility around COLA milestones, first criticality at INL, and any binding PPA conversion.
- Preferred private: Aalo — with a price condition. On fundamentals (fuel chain, physics conservatism, capital efficiency, founder-CTO, nearest datacenter demo), Aalo is the better company per dollar at equal $2–3B marks, and decisively better anywhere below ~$1.5B. Its dominant risk is financing, which an investor writing the check partially solves — a rare case where the investment itself de-risks the asset.
- Optionality position: Valar. Treat it as a call option on (a) the deregulatory pathway holding and (b) mass manufacturing working — with genuine, demonstrated execution velocity and the best-connected syndicate in defense-adjacent Washington. At $2B the option is not cheap, the order book is empty, and the credibility file warrants real diligence on every claim. Appropriate at 1/3–1/2 the sizing of the other two, underwritten to power-law outcomes only.
A blended stance for a fund able to hold all three: ~55–60% Oklo / ~25–30% Aalo / ~15% Valar, with Aalo’s weight conditional on entry price and B-II terms, and a hard re-evaluation of all three on (1) the NRC’s DOE-bridge rule finalization, (2) the first binding hyperscaler PPA in the cohort, and (3) Aurora-INL first criticality.
Key monitorables (next 12 months)
- NRC final rule on licensing DOE/DOD-authorized designs — the single biggest swing factor for the private pair.
- Aurora-INL construction milestones and any slip language in Oklo’s quarterly calls; first fuel load is the re-rating event.
- Meta Ohio: site characterization start and prepayment drawdowns (converts “agreement” into cash).
- Aalo Project Ascension: turbine integration and first datacenter-powering at INL (2027 target) + disclosed terms of the B-II/next round.
- Valar → Nvidia: whether the 30 MW waterless AI factory becomes a contract with money, and any TRISO fuel-supply announcement.
- HALEU/TRISO supply news (Centrus ramp, TRISO-X capacity) — repricing events for Oklo and Valar respectively.
- Any safety or regulatory stumble at any DOE-pilot reactor — a sector-wide correlated risk that would hit the private marks hardest.