Oklo (OKLO): what it does, how it plans to make money, and the licensing risk

The standard $OKLO story is a syllogism: AI needs power, nuclear is power, therefore buy the nuclear names. The syllogism is true and useless. It explains why the whole Nuclear / SMR bubble re-rated, not why you would own Oklo specifically over a utility that already sells electrons today. The half the hype skips is the half that decides Oklo's outcome: it has no power revenue, its Aurora reactor has never run commercially, and the thing gating the entire thesis is a regulatory authorization, not a demand curve.

Oklo Inc. trades on NYSE at $35.87 (close of 2026-10-02, QA data), down 72% over one year and about 4% above its 52-week low of $34.38. On the 186.0M shares the last 10-Q cover reported (as of 2026-08-04), that is a market cap of roughly $6.7B, against $3.0B of cash and marketable securities on the June 30 balance sheet. That is still a large number for a company that has not yet sold a kilowatt-hour. What you are pricing is optionality on a specific commercialization arc, so the arc is what matters.

What does Oklo do?

Oklo is building small nuclear reactors aimed at the one customer that suddenly cannot get enough firm power: the data center. Its Aurora design is a sodium-cooled fast reactor sized at 15-75 MWe per unit (the 10-Q for Q2 2026 says sizes of 100 MWe and higher are possible), small enough to site in clusters (multi-reactor campuses) rather than as one giant plant. The pitch is firm, 24/7, carbon-free power delivered near the load, on a timeline the grid cannot promise.

The technical bet is that a small, factory-repeatable reactor beats the bespoke-megaproject economics that have made large nuclear so slow and so over-budget. That is the same bet the rest of the SMR field is making; Oklo's specific wager is the fast-reactor design plus the business model below.

It is not yet an operating power company. It is a design, a site at Idaho National Laboratory, a DOE safety basis, a small isotope test reactor that reached criticality in August 2026, and a pipeline of customer interest. That gap between "designed" and "operating" is the whole investable question.

How Oklo plans to make money

Here is the detail that separates Oklo from the "SMR vendor" framing, and it matters: Oklo does not plan to sell reactors. It plans to build, own and operate them and sell the electricity under long-term contracts.

The visible commercial pathway is data-center power. Per the Q2 2026 10-Q, the named pipeline is: a 12 GW Master Power Agreement with Switch (December 2024); a prepayment agreement with $META (January 5, 2026) toward a 1.2 GW power campus in Pike County, Ohio, with Meta's funding earmarked to secure fuel for the first phase; non-binding letters of intent with $EQIX (Equinix), Diamondback Energy and Prometheus Hyperscale; and a Notice of Intent to Award for at least 5 MWe at Eielson Air Force Base in Alaska. The company's own plan of operations still lists "converting such preliminary agreements into PPAs" as work to do, so the binding, priced power contract is not there yet.

The catch is that none of this is power revenue yet. The $1.2M of Q2 2026 revenue came from two small engineering firms Oklo acquired in June 2026 (ARMEC and Creative Engineers), not from electricity. The model is elegant precisely because it is deferred: no power sold until a plant is authorized, built and running, and the company's stated target for the first powerhouse is 2028, which the 10-Q itself calls "ambitious". Everything before that is spending: management guides 2026 operating cash burn at $120M to $150M and capital expenditure at $400M to $500M.

Where Oklo sits in the nuclear cluster

Oklo is a pure-play member of QA's Nuclear / SMR bubble (Altman-backed, primary-bubble classification) and the broader Nuclear theme. It is one of the cleanest ways to express the AI-power-shortage thesis, which is also why it trades as a high-beta version of it.

Its closest tracked peers by 252-day correlation (QA data, refreshed for this update) are the other advanced-nuclear names: $NNE (Nano Nuclear, ~0.83), $SMR (NuScale, ~0.82) and $LEU (Centrus Energy, ~0.78, the enrichment supplier the whole field depends on, and since June 2026 a would-be HALEU supplier to Oklo under a letter of intent). The uranium miners $UUUU, $DNN and $UEC sit next, around 0.68. The one ETF QA tracks holding it is $URA (Global X Uranium), at a 5.5% weight as of 2026-10-02. They move together because they are the same macro trade: a bet that the AI power crunch pulls advanced nuclear forward by a decade. When the trade is on, they rise together; when sentiment turns, they fall together, and a pre-revenue name falls hardest.

The numbers

Pin these to the date, because a pre-revenue name's story moves faster than its fundamentals:

• Price: $35.87 at the 2026-10-02 close (QA data). Returns: -10% over one month, -25% over three months, -72% over one year. The 52-week range is $34.38 to $193.84.
• Market cap: roughly $6.7B on 186.0M shares (share count as of 2026-08-04, 10-Q cover), down from ~$8.7B in the August version of this piece. NYSE-listed, GICS Utilities / Independent Power Producers.
• Cash: $3,006.3M of cash, cash equivalents and marketable debt securities at 2026-06-30 (10-Q). Net tangible book value was $17.34 per share at that date (September 2026 prospectus supplement).
• Revenue: $1.2M in Q2 2026, all from acquired engineering services; zero from power. Q2 net loss: $48.5M; six-month net loss: $81.6M.
• Reactor: Aurora, sodium-cooled fast reactor, 15-75 MWe per unit.
• Regulatory status: DOE approved the Preliminary Documented Safety Analysis for Aurora-INL on 2026-06-11 under its Reactor Pilot Program; the NRC approved Oklo's principal design criteria topical report in April 2026.
• First powerhouse: targeted for 2028 (company target, Q2 2026 10-Q).

The market cap against the cash pile is the entire debate in one line. At roughly $6.7B, the market is paying about $3.7B above the June cash balance for the licensing, construction and contracting timeline. Bulls read that as cheap optionality on a category-defining asset; bears point out that the cash is being raised by selling stock, which is the next section.

What changed since August

The licensing path is DOE first, NRC later. The August version of this piece described an NRC combined-license application "progressing". The Q2 2026 10-Q is more precise: for the first plant at Idaho National Laboratory, Oklo is pursuing DOE authorization "as the primary pathway" to construct and operate, after Aurora-INL was selected for the DOE Reactor Pilot Program in August 2025. The DOE has approved two of the five steps (the Nuclear Safety Design Agreement early in 2026, the Preliminary Documented Safety Analysis on 2026-06-11). On the NRC side, Oklo says it is "evaluating the timing and form of future NRC applications" in light of the new Part 53 and proposed Part 57 frameworks. That matters: the DOE route can get a first plant running, but selling power commercially at scale still runs through the NRC, and no NRC license application for Aurora is on file as of this update.

A reactor actually went critical. The DOE granted startup authorization for Oklo's Groves Isotope Test Reactor in Lockhart, Texas on 2026-07-23, and it reached first criticality on 2026-08-05 (10-Q). It is a low-power test unit for isotope production, not an Aurora powerhouse, but it is the first time Oklo has taken a reactor of its own from authorization to operation, and the first reactor under the DOE pilot program to go critical on private land.

Dilution is the funding model. Oklo raised about $1.85B through at-the-market (ATM) share sales in the first half of 2026: 12.4M shares at an average net price of $96.95, then 10.7M shares at $63.51 (10-Q). The second program ($1B, opened May 2026) was fully used: an 8-K filed 2026-09-11 says Oklo sold 17,971,448 shares for about $1B through its termination on 2026-09-10, and opened a new $1B ATM the same day with a ten-bank syndicate. Shares outstanding went from 160.5M at the end of 2025 to 185.1M at June 30. The new program's prospectus illustrates the cost at the 2026-09-09 price of $42.57: selling the full $1B would take about 23.5M shares, roughly 12% more stock. The falling price is the mechanism to watch: each dollar raised now costs more shares than it did in the spring.

Fuel gets a supply path. In June 2026 Oklo signed a letter of intent with Centrus for domestic HALEU covering multiple years of fuel for up to five Aurora powerhouses, deliveries expected from 2029 (non-binding; pricing and volume still under negotiation). In May 2026 the DOE selected Oklo for advanced negotiations under its Surplus Plutonium Utilization Program, a second possible fuel source.

The bull case

• Build-own-operate is the better model. Recurring power revenue over a plant's life beats one-time hardware sales, and avoids competing on reactor sticker price.
• Data-center demand is real and firm-power-starved. The buyers with the most urgent need for 24/7 carbon-free power are exactly the ones in the PPA pipeline.
• Design-density edge. Small, clusterable units aim at factory-repeatable economics rather than bespoke megaprojects.
• Category sponsorship. Altman backing and DOE site approval give the name credibility and a real first site, which most of the field lacks.

The bear case

• Pre-revenue and capital-intensive. Guided 2026 spend is $520M to $650M (operating cash burn plus capex), and it is funded by ATM share sales: about 30M new shares were sold across the two ATM programs completed in 2026 (12.4M plus 18.0M) before a third opened in September.
• Licensing is the historical Achilles heel. Advanced-reactor timelines have slipped repeatedly across the industry. The DOE route can get Aurora-INL running, but a commercial fleet still needs the NRC, and Oklo has not yet said when it will file.
• Unproven in commercial operation. Aurora has not yet run a commercial plant. Execution risk between design and first power is real.
• Crowded field. NuScale, X-energy, TerraPower and Kairos are all chasing the same customers; Oklo needs marquee design wins to stand out.
• High-beta basket risk. It moves with $SMR and $NNE; in a nuclear-sentiment drawdown, a pre-revenue name reprices the hardest.

How to access OKLO

$OKLO lists directly on NYSE, so any brokerage with US-market access can buy the stock. There is no clean pure-play SMR index, so broad utility or clean-energy ETFs give you almost none of this specific thesis: the microreactor bet only shows up if you hold the name (or its peers) directly. To trade the single name from a US-retail or non-US account, see /stack/ibkr for the broker QA uses for direct-exchange access.

What to watch

• The remaining DOE steps for Aurora-INL, and the first NRC filing. Three of the five DOE authorization steps are still ahead; the form and date of an NRC application for the commercial fleet is not yet announced. Either moves the thesis more than any commercial news.
• Signed PPAs vs. LOIs. The pipeline converts from "interest" to "revenue path" only when a Switch, Meta or Equinix framework becomes a binding, priced power contract. Watch for the first hard PPA, and for a definitive fuel contract with Centrus.
• The new ATM's pace. The 2026-09-11 program allows another $1B of share sales; the company reports sales at least quarterly. How fast it is used, and at what price, is the dilution read.
• Construction at INL. Kiewit is the lead constructor for the first Aurora powerhouse; the move from safety documents to a physical plant on the 2028 target is the credibility inflection.
• Q3 2026 results. The date is not yet announced by the company as of this update (the Q3 2025 date was announced on 2025-10-29). The 10-Q will show the cash balance after the summer ATM sales.
• Cluster correlation. If $OKLO decouples from the rest of the Nuclear / SMR bubble, the market is starting to price it on its own execution rather than as a basket, which would change how the whole trade behaves.

Observable reference levels on the QA chart (as of 2026-10-02): the price trades just above the 52-week low of $34.38; the editorial Fib key at $69 and the 0.786 retracement of the 52-week leg at $68.50 sit roughly 90% higher. These are reference levels, not targets.

The full teardown video is on /stocks/oklo. Subscribing to the digest on that page sends the Oklo Teardown Dossier free; the paid version is at /shop/oklo-dossier and is included in /pro, along with bubble shifts and rule-based alerts on $OKLO and its nuclear peers.

Live data on this ticker: /stocks/oklo - price, ETF holdings, bubble correlation, bot positions.

Bubble context: /bubbles/nuclear-smr - the cluster this name belongs to and how it's moving.

QuantAbundance is educational research. Nothing here is investment advice. See /disclosures.