Oklo

Phase A — Understand the business

Lens 1 · Company Overview

Oklo is a pre-revenue advanced-fission developer commercializing the "Aurora" powerhouse — a metal-fueled, liquid-metal-cooled fast reactor producing 15–75 MWe (roadmap to 100 MWe+), built on the legacy of the EBR-II fast reactor the US government ran for 30 years. Founded 2013 (as Oklo Technologies); went public May 9, 2024 via SPAC merger with AltC Acquisition Corp. (Sam Altman's vehicle).

The differentiator is the business model, not just the reactor: Oklo intends to be designer + builder + owner + operator, selling power (electricity and heat) under long-dated PPAs rather than selling or licensing reactor designs to utilities — the inverse of the traditional nuclear model (Westinghouse/GE selling 600–1,000 MWe LWR designs) and even of most SMR peers (NuScale licenses its design). Management's pitch: recurring PPA revenue, project-financeable assets, and "operating cash flow positive from the first year of operation due to anticipated favorable unit economics" — a claim to hold them to, since no unit exists.

Three business lines, one reportable segment:

1. Powerhouses (Aurora) — the core. First unit: Aurora-INL, a 75 MWe reactor at Idaho National Laboratory. Broke ground Sept 2025; targeted commercial ops late 2027 to early 2028. Kiewit Nuclear Solutions is lead constructor; Siemens Energy supplies the steam-cycle equipment.
2. Fuel recycling + fabrication — vertical integration to secure fuel. Plans a fuel-recycling facility in Oak Ridge, TN (the "Advanced Fuel Center," roadmap of up to $1.68B investment) and a pilot fuel-fabrication facility at INL.
3. Radioisotopes — via Atomic Alchemy (acquired Feb 2025 for $28.4M, mostly stock). The "Groves" isotope test reactor hit substantial completion in 229 days; first criticality targeted July 4, 2026, and this is the source of Oklo's first-ever revenue — commercial isotope contracts "expected later in 2026".

Customers / pipeline (all non-binding). The headline logos are letters of intent and master agreements, NOT binding PPAs: Switch 12 GW Master Power Agreement (Dec 2024); Meta 1.2 GW prepayment agreement for a Pike County, Ohio campus (Jan 2026); LOIs with Equinix, Diamondback Energy, Prometheus Hyperscale; tentatively selected for Eielson Air Force Base (Alaska, ~5 MW cogen); exploring recycling + power sales with TVA. Contract-structure red flag: as of the 10-Q, Oklo has "not... entered into any binding power purchase agreement with any customer". The one hard commercial term on the books is a $25.0M non-refundable ROFR prepayment from an unnamed data-center counterparty (Feb 2024 LOI), carried as a "right of first refusal liability" — it buys the counterparty a 36-month right of first refusal + most-favored-nation pricing, not committed volume.

Lens 2 · Supply Chain

The commercial-layer files (bottlenecks.md, supply-chain.md) are missing for the datacenters topic, so this maps from the filings + web — but Oklo's chain is unusually concrete for a pre-revenue name because the constraint (fuel) is the whole story.

Upstream fuel (the chokepoint): Oklo's fast reactors need HALEU (high-assay low-enriched uranium, 5–19.75% U-235) or plutonium-based fuel until recycled fuel is available. HALEU is not available at commercial scale in the West — this is the single most binding constraint in the entire advanced-nuclear sector. Oklo's multi-pronged fuel strategy:

• Centrus Energy (LEU) — LOI signed June 18, 2026: Centrus to supply enough domestic HALEU for up to five Aurora powerhouses for multiple years, from its American Centrifuge Plant in Pike County, Ohio (co-located with the Meta campus); deliveries begin 2029. Centrus is the only US HALEU producer — a genuine single-source dependency for near-term fuel.
• DOE HALEU award — 5 metric tons of HALEU recovered from irradiated EBR-II fuel at INL, for the first Aurora.
• Surplus plutonium — DOE dilute-and-dispose material; DeWitte framed the available surplus as "equivalent to 160–200 tons of HALEU," a strategic bridge fuel (subject to DOE authorization).
• Recycling — the long game (used fuel → new fuel), commercial facility targeted early 2030s.

Midstream (construction): Kiewit (EPC / lead constructor), Siemens Energy (steam turbine-generator), plus steel and other first-of-a-kind components. Filings repeatedly flag reliance on "a limited number of suppliers" for "highly specialized... first-of-a-kind or sole use" components.

Downstream (buyers): hyperscalers and data centers (Meta, Switch, Equinix, Prometheus), oil & gas (Diamondback), the US military (Eielson AFB), and utilities (TVA). The demand side is credible and hungry — AI-datacenter power demand is the tailwind — but conversion from LOI → binding PPA is unproven. Chokepoint summary: fuel upstream (HALEU, single-source Centrus near-term), FOAK construction risk midstream, LOI-to-PPA conversion downstream. Names present — this lens passes.

Lens 3 · Competitive Advantages (moats)

For a company with no product in operation, "moat" is mostly prospective. Honest read:

• Regulatory head-start (partial, real). Oklo was the first advanced-fission company to submit a custom Combined License Application (COLA) to the NRC (March 2020) — but it was denied without prejudice in 2022, and the resubmission is still pending. More durable is the DOE authorization pathway: Aurora-INL was approved to proceed under DOE purview (via the Reactor Pilot Program), which grants authority to construct + operate while maintaining safety standards, effectively routing around the NRC bottleneck for the first unit. Oklo also employs eight former NRC staff. This is a genuine, if fragile, edge over peers who must go through full NRC licensing.
• Vertical integration (fuel). Owning recycling + fabrication could secure fuel supply and add margin — a differentiator vs. NuScale/others who don't control fuel. But it's also enormous incremental capex and execution risk ($1.68B recycling roadmap) for a company that hasn't built anything yet.
• Business-model moat (BOO + recurring revenue). If it works, owning the asset and selling power creates switching costs (20-year PPAs) and recurring cash flow the licensing model lacks. If it doesn't, it exposes Oklo directly to FOAK cost/schedule overruns that utilities normally bear.
• Founder/technical depth. MIT-nuclear founders, 13 PhDs, deep DOE relationships.
• Bargaining power: weak today. Oklo needs Centrus (fuel), Kiewit (construction), DOE (fuel + site + authorization), and the NRC (license) more than any of them need Oklo. Customer bargaining power is high too — buyers hold LOIs, not commitments, and can walk. Net: the moat is a regulatory + first-mover narrative that only becomes a real moat once a powerhouse is operating and generating referenceable cash flow.

Lens 4 · Segments

One reportable segment — the CODM (CEO) reviews consolidated net loss; the business is not disaggregated for internal reporting. There is no revenue to break out (pre-revenue). The only "segment" texture is expense composition:

FY2025 total opex was $139.3M vs $52.8M in FY2024 — opex is accelerating sharply as Oklo staffs up (205 FTEs at YE2025) and moves from design into deployment. The trend is decelerating value per dollar in the sense that spend is rising fast with zero offsetting revenue; the bull reading is that this is the necessary ramp into construction. First revenue (radioisotopes) is guided "later in 2026" and will be immaterial relative to opex.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, ended 2026-03-31)

There is no revenue and no consensus revenue line to beat — this is a cash/milestone story. The print:

• Net loss $33.1M (EPS −$0.19) vs $9.8M (−$0.07) a year ago. The "beat" reported in the press (EPS matched/beat) is against a loss estimate, not earnings — semantically pre-revenue.
• Loss from operations $51.2M (all opex) — but net interest & dividend income of $21.3M (on the $2.5B pile) offsets ~40% of the operating loss, so cash net loss is much softer than the P&L loss.
• Operating cash burn only $17.9M for the quarter — well inside FY26 guidance of $80–100M cash opex. The gap between $51.2M operating loss and $17.9M cash burn is mostly $15.6M non-cash SBC plus working-capital timing.
• Balance sheet: fortress (near-term). Cash + marketable debt securities $2,536.9M at Mar 31 (cash $1,594.1M + current MDS $614.5M + noncurrent MDS $328.3M), up from ~$1,412M at YE2025. Zero debt. Total equity $2,638.6M; accumulated deficit $273.8M.
• The raise that funded it: a $1.5B ATM program (Dec 2025), of which $1,181.9M net came in during Q1 2026 at an avg net price of $96.95/share (12.38M shares). Investing outflow was $359.0M in Q1 — but $321.2M was just parking cash into Treasuries; only $32.8M was actual capex (construction-in-progress at INL).
• Market reaction: negative despite the "beat." Stock fell ~4.4% on the print — the market is treating each equity raise as dilution and each quarter as one more of cash burn before any revenue. The stock is down ~70% from its 52-week high of $193.84 and fell ~21.8% in June 2026 even as it announced the Centrus fuel LOI and Kiewit Ohio MOU.
• Unusual vs. own history: SBC exploded 6.7x YoY and is now ~30% of opex — worth watching (see Lens 10). The Texas ground lease (Caldwell County, Mar 2026) signals a widening site footprint beyond INL/Ohio.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the shelf; drawing from the Q1 2026 call coverage. Management tone is decidedly confident and execution-framed — the recurring theme is a deliberate pivot in language from strategy to deployment: DeWitte's "Our mindset has shifted toward asset deployment." What they're leaning on:

• Milestone cadence as proof-of-progress — Groves reactor substantial completion in 229 days, July 4 2026 criticality, PDSA submitted to DOE, deep-foundation work underway at INL. When you have no revenue, hitting dated engineering milestones is the narrative, and they're hammering it.
• Fuel optionality — the "multiple fuel forms" framing (HALEU + surplus plutonium + recycling) is a direct answer to the sector's #1 bear point.
• Regulatory tailwind — heavy emphasis on the NRC's proposed Part 57 rule (fleet-based licensing, targeted 6–12 month reviews) that could convert Aurora-INL from DOE authorization to NRC licensure later in 2026.
• First revenue — carefully seeded expectation of first isotope revenue "later in 2026" to give the market a near-term "we can actually sell something" data point.

Shift over time (inferred): the through-2025 story was "pipeline and LOIs" (Switch 12 GW, Meta); the Q1 2026 story is "shovels and fuel" (groundbreaking, Kiewit, Centrus, criticality). That's a maturation of narrative — from demand-signaling to execution — which is the right direction, but it also raises the bar: the market now wants delivered milestones, not more LOIs, and is punishing the stock when it gets LOIs (the June Centrus reaction). The thing they've stopped emphasizing: near-term binding PPAs — still zero.

Lens 7 · Comps

Read-through. Oklo commands ~2.5x NuScale's market cap despite NuScale holding the only NRC-certified SMR design and Oklo not expecting an NRC-deployable design until 2027–28. The market is paying a large premium for Oklo's BOO model + fuel vertical integration + the Altman/Meta halo, not for regulatory de-risking. On an EV basis, Oklo is ~$6.5B (mkt cap $9.0B − $2.5B net cash). That is ~$490M of enterprise value per headline GW of announced (non-binding) LOI/PPA capacity (13.2 GW = Switch 12 + Meta 1.2) — i.e. the market is capitalizing letters of intent as if a meaningful fraction convert. Cash+MDS alone is $14.58/share, ~28% of the $51.74 price — so ~72% of the price is pure option value on future deployment.

Lens 8 · Stock-Price Catalysts (what moves OKLO)

Pattern from the last ~2 years:

• Sector-wide policy shocks move it hardest. White House space-nuclear mandate (Apr 2026), four pro-nuclear Executive Orders (May 2025), US Army microreactor push — all triggered sector-wide double-digit pops (OKLO +5–6% alongside NNE/SMR/LEU). OKLO trades as the high-beta expression of "nuclear-for-AI" sentiment, often more than on its own idiosyncratic news.
• Equity raises = sharp sell-offs. The Dec 2025 $1.5B ATM filing and the Q1 2026 equity offering both triggered immediate dilution-driven drops (June 2026 −21.8%; Dec 2025 −30%+). This is the defining negative catalyst — the funding mechanism is the overhang.
• Analyst target cuts bite. Goldman + Citi cuts (Citi $95→$73.50) and a Sell downgrade drove multi-percent down-days.
• Insider-sale filings (DeWitte 10b5-1 sales) coincided with pre-earnings weakness.
• Milestone/partnership news is oddly muted or fades. The Centrus fuel LOI and Kiewit MOU landed during a −22% month — the market discounted "another LOI." That's the tell: positive operational news no longer moves the stock; only cash, dilution, and licensing hard-catalysts do. The next real up-catalyst is a binding PPA or the NRC COLA acceptance/approval — not more MOUs.

---

Phase C — Judge people & books

Lens 9 · Management

• Track record. Co-founders Jacob (Jake) DeWitte (CEO, now also Chairman) and Caroline DeWitte (COO) — MIT nuclear-engineering graduate degrees, ~20 and ~15 years in nuclear respectively, DOE backgrounds. They have built a credible technical org and a formidable government-relations machine (DOE authorization, RPP/FLPP selections, site permits) — but they have never built or operated a commercial power plant, and the filings say so explicitly: "We have limited experience operating a company that builds, operates, or maintains commercial nuclear power plants... our management's decisions... may not take into account standard managerial approaches". That is an unusually candid founder-risk admission.
• Bench. CFO R. Craig Bealmear — 30+ yrs finance, ex-CFO of Renewable Energy Group and ex-CFO North America Downstream at bp; Wharton MBA. A serious public-company CFO is a plus for a capital-markets-dependent story. CTO Patrick Schweiger (40+ yrs, ex-Commonwealth Fusion). CLO William Goodwin (ex-Joby).
• Tenure & skin in the game. Founder-led since 2013; still Chairman. But insider selling is a live flag: insiders sold ~$39M net over the trailing 12 months; DeWitte sold 216k shares ($23/sh) on Mar 31 and additional open-market sales dated May 1, 2026 — under a Rule 10b5-1 plan adopted Mar 31, 2025. The 10b5-1 wrapper mitigates the optics (pre-scheduled), but consistent founder selling into a falling stock is worth noting for a name whose entire thesis is "trust us to execute a decade-long build."
• The Altman departure. Sam Altman stepped down as Chairman in April 2025 to avoid a conflict of interest ahead of a potential OpenAI–Oklo energy deal; DeWitte took the chair. Framed by the company as freeing Oklo to work with more AI/hyperscaler customers (which the Meta deal validates). Reasonable spin — but it removed the single most valuable name from the masthead of a story stock, and the shares are down heavily since.
• Capital-allocation history. Short, and mixed: raised aggressively and opportunistically (ATM at ~$97/sh — good timing, sold high); made one tuck-in acquisition (Atomic Alchemy, mostly stock — added the only near-term revenue path); parks idle cash conservatively in Treasuries (earning ~$85M/yr annualized, a real offset to burn). No buybacks (correct — they need the cash). ROE/ROIC are negative and meaningless pre-revenue.
• Archetype: classic visionary founder-CEO of a moonshot — high conviction, strong technical + political capital, unproven as an operator/capital-project deliverer. For this stage that's the right archetype to start the company and the wrong archetype to de-risk $10B+ of FOAK construction. Watch whether they bring in a heavyweight COO/CPO with megaproject-delivery experience.

Lens 10 · Forensic Red Flags

Pre-revenue, so classic revenue-recognition/receivables games don't apply — the accounting is clean and simple. What a forensic analyst actually flags here:

• Stock-based compensation is the real "earnings quality" issue. SBC was $15.6M in Q1 2026 (30% of opex), up 6.7x YoY, with $132.0M of unrecognized SBC to amortize over ~3.4 years. The gap between the $51.2M GAAP operating loss and the $17.9M cash burn is mostly SBC — so headlines citing "only $18M burn" understate the true economic cost of running the company. SBC is also a continuous dilution engine on top of the ATM.
• Dilution is structural, not incidental. Shares went from ~137.7M (Jan 2025) to 173.99M (May 2026) — +26% in ~16 months. The ATM ($1.5B authorized, fully used) plus ongoing SBC means share count marches up every quarter. This is the single most important "forensic" fact: equity issuance is the operating model's fuel line, and Goldman estimates ~$14B of lifetime capital need through the 2040s — implying years of further dilution at whatever price the market will bear.
• IPR&D / goodwill from Atomic Alchemy — $27.5M indefinite-lived intangibles (IPR&D: "Abundantia" $4.6M, "Meitner" $22.9M) + $6.6M goodwill, both valued by risk-adjusted cash flows contingent on NRC permits/licenses not yet granted. Impairment risk if those approvals stall. Not material to the whole ($34M vs $2.7B assets) but it's a soft asset dependent on regulators.
• "Other investments" $17.1M in privately held companies (SAFEs + preferred), carried at cost with no readily determinable fair value. Small, but a pre-revenue company making venture bets is a mild governance eyebrow-raise.
• Cash flow vs. earnings: no divergence concern — operating cash burn ($17.9M) is lower than the net loss because of non-cash SBC; investing outflow is mostly Treasury purchases, not spend. Clean.
• Controls: disclosure controls concluded effective; no material weaknesses; no changes to ICFR in Q1. Auditor: (not re-verified here). EGC status retained until Dec 31, 2026, so some scaled disclosure still applies.

Regulatory findings (required).

• SEC Litigation Releases / AAERs: None. Verified via EDGAR EFTS (LR + AAER) for 2021-07-06 → 2026-07-06 — total_sec_findings: 0.
• Item 3 / Legal Proceedings: Company discloses no pending or threatened legal proceedings expected to be material; "There were no contingent liabilities as of March 31, 2026".
• Non-SEC enforcement: web search ("Oklo" (FTC OR DOJ OR FDA OR CFPB OR consent decree OR settlement OR penalty) enforcement) surfaced no material enforcement actions or fines — coverage is dominated by ordinary securities-analyst commentary and shareholder-rights-firm "investigation" solicitations tied to the stock's 2025–26 decline (routine plaintiff-bar boilerplate, not filed litigation).
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-Q Item 3/Note 11 as of 2026-07-06.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection

EPS projection is not the right tool — Oklo is pre-revenue and will report GAAP losses through at least FY2028; there is no credible EPS to model into a base/bull/bear (any positive EPS is 2029+ and entirely contingent on the first powerhouse operating). Forcing an EPS number here would be fabrication. So per the pre-revenue framing, I model the cash-runway-to-catalyst and the funding gap instead — the two variables that actually determine survival and dilution.

Runway (all inputs labeled):

• Cash + MDS: $2,536.9M.
• FY2026 guided cash use: $80–100M opex + $350–450M investing = $430–550M total.
• Runway on total FY26 cash-use pace, excluding any new capital: ~4.6–5.9 years.

Funding gap (the bear's core number):

• Building the first Aurora + fuel-fab + recycling ($1.68B roadmap) + a fleet toward the 13.2 GW of headline LOIs requires capital far beyond the current pile. Goldman Sachs estimates ~$14B of capital needed through the 2040s.
• At today's ~$9B market cap, funding even a third of that via equity implies massive further dilution — or project-level debt/DOE loan guarantees Oklo has not yet secured. The base case is: many more raises, at prices set by execution.

Base / Bull / Bear on the variable that matters (deployment + funding), not EPS:

• Base: First isotope revenue (immaterial) late 2026; Aurora-INL first power late 2027–early 2028 (slips are the norm for FOAK nuclear — assume some slippage to 2028+); COLA resubmitted 2026, NRC decision 2027+; 1–2 further equity raises by 2028 to fund the capex ramp. Stock range-bound-to-down until a binding PPA or NRC hard-catalyst lands.
• Bull: Groves criticality (Jul 2026) → first isotope revenue on time; DOE authorization lets Aurora-INL deploy on schedule (2027-28); a Meta or Switch LOI converts to a binding PPA; Part 57 fleet-licensing arrives and de-risks the whole pipeline; DOE loan guarantee closes part of the funding gap → dilution fear recedes, multiple re-rates.
• Bear: FOAK construction slips to 2029+ and runs over budget; HALEU deliveries (Centrus, 2029) slip; an LOI counterparty publicly walks; a dilutive raise at a lower price resets the share count; NRC COLA stalls again. Any one triggers a re-rate toward cash value (~$14.58/sh).

Brier forecast: skipped (watchlist loop; no EPS to log, and I have not committed to a binary base case worth tracking). The catalyst to eventually track as a scoreable binary: "Oklo signs its first binding PPA (not LOI/MOU) by YE2027" — the single event that would most change the thesis.

Lens 12 · Bull vs Bear

Bull case. Oklo is the best-positioned pure-play on the one thing the AI build-out cannot manufacture fast enough — firm, clean, 24/7 power. Its BOO + PPA model is genuinely differentiated: recurring revenue, project-financeable, and aligned with hyperscaler procurement norms. The DOE-authorization pathway routes around the NRC's glacial licensing for the first unit — a real time-to-market edge. Vertical integration into fuel (Centrus HALEU LOI, surplus plutonium, recycling) attacks the sector's #1 constraint head-on. The demand roster is A-list (Meta prepayment, Switch 12 GW, Equinix, the US military). Cash is a fortress ($2.5B, zero debt) that funds the next several years and earns ~$85M/yr in interest while it waits. Policy is a durable tailwind (bipartisan ADVANCE Act, four pro-nuclear EOs, Part 57 fleet licensing). If the first powerhouse lights up on schedule and one LOI converts to a binding PPA, the story flips from "science project" to "the only advanced-fission company actually delivering power to AI" — and at that point the ~$6.5B EV looks small.

Bear case (permanent-impairment risks). (1) It's a pre-revenue FOAK nuclear build — the base rate for first-of-a-kind reactors delivering on time and on budget is terrible (Vogtle, NuScale-UAMPS cancellation). A multi-year slip + cost overrun on Aurora-INL is the modal outcome, not the tail. (2) The funding model is dilution — Goldman's ~$14B lifetime need means the equity is the ATM; every share you own gets diluted for a decade, and the market is already voting (stock −70% from peak) by refusing to reward operational milestones and punishing every raise. (3) Zero binding demand — 13.2 GW of "agreements" are LOIs/MOUs any counterparty can walk from; the company itself has never signed a binding PPA. Pre-mortem (18 months out, thesis broken): it's early 2028, Aurora-INL has slipped to 2029 on a fuel or construction snag, Oklo has raised twice more (share count 210M+) at prices below today, a hyperscaler LOI quietly lapsed, and the stock trades near cash ($15) as the market fully de-rates it to "optionality that keeps getting pushed out." Multiple assessment: at ~$490M EV per headline GW of non-binding capacity and ~72% of the price as pure option value, the stock is priced for successful deployment + meaningful LOI conversion — i.e. it is not cheap on any risk-adjusted basis; it is a levered bet on flawless execution. Contrarian view the market is refusing to see (cuts both ways): the bear market-consensus ("just another diluting pre-revenue nuclear name") may be underpricing the DOE-authorization end-run and the July-2026 first-revenue inflection — if isotopes actually start generating cash and Aurora-INL hits criticality on the DOE pathway, Oklo separates from the SMR pack faster than the −70% tape implies.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• Revenue is concentrated in a promise. There is no revenue. The first dollar (radioisotopes, late 2026) is a rounding error vs. ~$90M/yr opex. The entire equity value rests on a powerhouse that doesn't exist meeting a 2027-28 date that FOAK nuclear almost never meets.
• The moat is thinner than bulls think. NuScale — with a market cap ~60% lower — holds the only NRC-certified SMR design. Oklo's own COLA was denied in 2022 and is still not resubmitted-and-approved. "DOE authorization" is a workaround for one unit, not a fleet license; the fleet still needs the NRC. The "moat" is a first-mover narrative that a single construction slip erases.
• Most dangerous competitor bulls underestimate: not another SMR startup — it's the incumbents and gas. Hyperscalers needing firm power now will sign with operating gas + grid + existing large nuclear (Constellation/Vistra/Talen restarts) years before Oklo delivers a megawatt. Oklo's TAM assumes buyers wait until ~2028+; many won't.
• Worst capital-allocation / incentive flags: SBC at 30% of opex and 6.7x YoY; founder selling ~$39M net into a −70% stock (10b5-1 notwithstanding); the Altman conflict-of-interest exit removing the marquee name. None are fraud — but they're the pattern of a story stock monetizing its own narrative.
• What must hold for today's price: Aurora-INL on time and on budget; ≥1 LOI → binding PPA; the funding gap filled without a death-spiral raise; NRC/Part 57 cooperating. Miss any one and growth "disappoints" not by 20–30% but by a year, and a year of delay on a pre-revenue name is a re-rate to cash (~$15).
• Single scenario that permanently impairs: a public FOAK failure or multi-year cost blowout at Aurora-INL (the reference plant) — it would simultaneously break the schedule, force a dilutive raise into weakness, and shatter the "we can build-own-operate" thesis that differentiates Oklo from every design-licensor. Plausibility: moderate-to-high given the FOAK base rate.

Lens 14 · Management Questions (ordered by information value)

1. Binding PPAs: What is the specific, dated path from your LOIs/MOUs (Switch, Meta, Equinix) to your first binding, take-or-pay PPA, and what has to be true (NRC milestone? COFD? price?) before a hyperscaler will sign one?
2. Funding gap: Goldman estimates ~$14B of capital need through the 2040s. What is your actual funding stack — equity vs. project debt vs. DOE loan guarantees — and how much further common-equity dilution should shareholders underwrite between now and Aurora-INL first power?
3. Aurora-INL schedule confidence: What is your P50 vs. P90 commercial-operations date, and what are the top three schedule risks (fuel, long-lead components, DOE authorization) that could push late-2027/early-2028 to 2029+?
4. FOAK cost: What is your current all-in capex estimate for Aurora-INL specifically, how has it moved in the last 12 months, and what's the cost curve you expect from unit #1 to unit #5?
5. DOE-to-NRC handoff: Precisely how does the DOE-authorized Aurora-INL convert to an NRC license under Part 57, and what happens to the fleet economics if Part 57 slips or arrives weaker than proposed?
6. Fuel certainty: Centrus deliveries start 2029 for up to 5 powerhouses — what fuels units #1–#2 before 2029, and how firm is the DOE HALEU/surplus-plutonium supply in quantity and timing?
7. Unit economics: You claim powerhouses can be "operating cash flow positive from year one." At what PPA price ($/MWh), capacity factor, and capex/kW does that hold, and how sensitive is it to a 25% construction overrun?
8. Radioisotopes: What is the realistic revenue and gross-margin trajectory for Atomic Alchemy in 2026–2028, and is it a real business or a milestone-signaling asset?
9. Insider selling: Why are founders selling (~$39M net LTM) at what you presumably believe are depressed prices, and should the 10b5-1 plans be paused to align with the "trust our long-term execution" message?
10. SBC: SBC is 30% of opex and diluting alongside the ATM — what's the plan to bring stock comp down as a share of spend as you scale?
11. Competitive timing: Hyperscalers can buy firm power from gas + restarted large nuclear today. Why do you win the 2028+ contract instead of losing the customer to a bridge solution now?
12. Recycling capex: The $1.68B Advanced Fuel Center is a second megaproject on top of the reactor fleet. Why pursue it now rather than after the first powerhouse proves the model, and how is it financed?
13. Balance-sheet strategy: With $2.5B mostly in Treasuries earning ~$85M/yr, what triggers deploying that cash into construction vs. holding it as a dilution buffer?
14. Governance post-Altman: How has losing Sam Altman as chairman changed your access to AI-customer relationships and capital, and who fills that convening role now?
15. Kill criteria: What specific milestone miss (a date, a cost, a failed license) would make you conclude the BOO model isn't working and change strategy?

---