Can Tech Giants' Long-Term Nuclear Power Agreements Support an OKLO Valuation Rerating?

Analysis

The OP presents an OKLO rerating thesis based on validated nuclear power demand, faster construction progress, and cash reducing the implied valuation of the business, rather than demonstrating that the company has achieved commercial profitability. The starting point is the 20-year power purchase agreement signed by Google and Constellation on October 6, 2026, to add 890MW through upgrades to 11 existing units, with Constellation investing more than $4.3 billion and the first capacity additions expected to be delivered in 2028. The OP calculates approximately $4.83 million per MW and, together with a cited 190MW agreement by Amazon last week, argues that tech giants are willing to make long-term commitments for reliable nuclear power. However, this figure measures construction investment intensity, not the power purchase price, and cannot demonstrate that OKLO can secure equally profitable terms. (googlecloudpresscorner.com) The OP then uses the 6,831MW shortfall in PJM's capacity auction to calculate that the two agreements cover only approximately 16%, leaving a shortfall of approximately 5,751MW, and argues that new power sources must therefore be built. The shortfall figure is accurate, but it represents a reliable capacity shortfall for the 2028/2029 delivery year; nominal capacity coming online in phases cannot simply be subtracted from it, and it certainly does not constitute orders for OKLO. (insidelines.pjm.com) At the company level, the author describes Aurora as a 15–75MW sodium-cooled fast reactor that OKLO builds and owns itself to sell electricity. The planned 1.2GW project in Pike County, Ohio, is equivalent to 16 units of 75MW each. The author also cites Meta's prepayment, a letter of intent under which Centrus would supply fuel for up to five units starting in 2029, and Switch's 12GW agreement to reinforce the demand and fuel supply chain thesis. These documents exist, but their levels of maturity differ. Centrus still requires a definitive agreement, and planned capacity cannot be treated as certain revenue. (sec.gov) The OP cites the Groves test reactor in Texas achieving first criticality on August 5, 2026, less than a year after construction began, to rebut the assertion that the company has never built a reactor. This progress is substantiated, but a low-power isotope test reactor does not mean Aurora's commercial power generation economics have been validated. (oklo.com) The OP juxtaposes Idaho's DOE approval pathway, the 2027 fuel facility target, and the 2028 power plant target with Google's first capacity additions. An important omission is that the Ohio project's first phase is targeted for 2030, with the full 1.2GW targeted for 2034, and it requires NRC commercial licensing; the test reactor's construction speed cannot be extrapolated to it. (oklo.com) On valuation, the OP uses a share price of $38, a market capitalization of $7.3 billion, and estimated cash of $3.3 billion to derive approximately $17 in cash per share and a cash-adjusted business value of approximately $4 billion, or $21 per share. The author also cites annual operating cash burn of $120–150 million, capital expenditures of $400–500 million, and quarterly interest of approximately $23 million to argue that Idaho's funding risk is manageable. The author further claims that the share price has fallen approximately 80% from its October 2025 peak, short interest is approximately 20% of the float, and institutional price targets range from $41–100, but does not disclose a personal position, an independent price target, or an earnings model, and provides no images as corroboration. Catalysts include demand validation through long-term agreements, engineering approvals, and the author's expectation that November earnings will disclose ATM usage. The biggest weakness is extrapolating industry demand into OKLO's ability to deliver and treating cash that will be consumed by continued investment as valuation protection, without quantifying construction costs, electricity prices, fuel, delays, or dilution risks.

Discussion Summary

The sample of substantive discussion is insufficient, and the overall tone is skeptical. Two comments, each with a score of 5, request position disclosure, with no OP response in the provided reply tree. Some comments ask when the company will become profitable, while another asserts that it will not be profitable before 2030 without providing evidence; this cannot be treated as a verified fact. Someone suggests NLR as an alternative but does not develop the comparison. The discussion does not produce an in-depth rebuttal concerning construction costs, licensing, or valuation.

Fact Check

• supported: Google signed a 20-year agreement on October 6, 2026, supporting 890MW of capacity additions across 11 Constellation units, with investment exceeding $4.3 billion and initial delivery in 2028. — Google's official announcement confirms these principal terms. The $4.3 billion is Constellation's investment, not the total value of Google's power purchase contract. The main post also omits a separate 15-year agreement for 2,700MW of power from existing sources.
• questionable: Upgrade costs of approximately $4.8 million per MW demonstrate that tech giants are willing to pay nuclear power prices sufficient to support OKLO. — Dividing $4.3 billion by 890MW yields approximately $4.83 million/MW, so the arithmetic is correct. However, capital expenditure intensity is not equivalent to the unit price of purchased electricity, and the announcement does not disclose an electricity price that could validate OKLO's profit margins.
• questionable: PJM has a 6,831MW shortfall, and new capacity from Google and Amazon has already covered approximately 16%, leaving a shortfall exceeding 5,700MW. — PJM confirms that the 2028/2029 auction fell 6,831MW short of reliability requirements. Using the capacities of the two agreements cited in the main post yields approximately 15.8%, but nominal capacity additions, accredited reliable capacity, and actual commissioning years differ, so the shortfall cannot be considered filled on a one-for-one basis.
• supported: The 1.2GW Ohio project has a Meta prepayment arrangement and an expression of intent by Centrus to supply fuel for up to five Aurora units. — Company disclosures confirm the project capacity, prepayment mechanism, and intent to begin supplying fuel in 2029. The latter still requires a definitive contract. Ohio's first phase is targeted for 2030 and full capacity for 2034, and NRC commercial licensing is required, so this cannot be treated as supply available in 2028.
• supported: Groves achieved first criticality on August 5, 2026, so OKLO is no longer a company that has never built a reactor. — The company announcement and quarterly report confirm first criticality, achieved less than a year after construction began under DOE pilot authorization. However, it is a low-power isotope test reactor, and this does not establish that the performance, construction costs, or profitability of a commercial power reactor have been validated.
• supported: The Idaho project received preliminary safety analysis approval in June and is planned for deployment in 2028. — The quarterly report confirms the announcement of PDSA approval on June 11, 2026, and a target of deploying the first reactor in 2028. These are an intermediate regulatory step and a forward-looking plan, not final operating approval or a definite delivery date.
• supported: Cash and securities totaled approximately $3 billion at the end of June, with full-year operating cash burn of $120–150 million, capital expenditures of $400–500 million, and quarterly interest of approximately $23 million. — The quarterly report shows that cash and cash equivalents plus current and noncurrent securities totaled approximately $3.006 billion, and the spending guidance is consistent. The precise classification of the quarterly $23.209 million is net interest and dividend income, not profit from power generation.
• unclear: Approximately $320 million in additional financing was raised after the end of June, bringing current cash to approximately $3.3 billion, or approximately $17 per share, with a new $1 billion ATM also available. — The quarterly report reviewed cannot confirm subsequent financing or the latest ATM arrangement, so current cash cannot be independently verified. Simply adding financing proceeds to the earlier cash balance also ignores intervening spending, while financing increases the share count. The $17 in cash per share and $21 per share in business value are merely estimates from the main post.
• unclear: The current share price is approximately $38, down approximately 80% from its October 2025 peak; short interest is approximately 20% of the float, and analyst price targets range from $41–100. — The limited fact-check did not obtain corresponding market prices, dated short-interest statistics, or institutional reports as of October 6, 2026. These remain figures cited by the author and cannot be treated as verified real-time data.