OpenAI to press ahead with data-center expansion despite midterm sweep, CFO tells CNBC

OpenAI CFO Sarah Friar said the company will continue its U.S. data-center expansion despite heightened scrutiny after a Democratic midterm sweep; OpenAI also said it will stop using NDAs with public officials on prospective sites and share construction plans and potential effects with residents. Friar said OpenAI is raising fresh capital after business grew 70% quarter over quarter in Q3 and enterprise demand doubled. Fears of slower data-center construction wiped about $2.5 trillion off the U.S. stock market on Thursday, while Alphabet-owned Google announced a $17 billion Wyoming data-center investment.

Analysis

The key market variable is not whether AI demand exists, but whether projects can convert capital into permitted, powered capacity on schedule. OpenAI’s public-disclosure shift may reduce the political cost of opaque siting, yet it also exposes projects earlier to organized opposition; expect more front-loaded permitting and community-engagement expense, not an immediate end to capex. If local approvals become the binding constraint, value shifts toward operators with power-secured, permitted sites and equipment already under contract, while chip and networking demand becomes more back-end-loaded. Utilities and grid-equipment suppliers could benefit from sustained buildout, but local ratepayer backlash may limit pass-through.

For GOOG, Wyoming investment is evidence of continued commitment, not proof of attractive project returns; monitor capex conversion into capacity and incremental cloud/AI revenue. The claim of rapid OpenAI growth is company-reported and does not establish durable unit economics or funding terms. For ANT, IPO valuation sensitivity is a near-term read-through on private AI financing: a postponed or repriced deal could reset expectations for all capital-intensive AI ventures, even if end-user adoption remains sound.

Contrarian view: the selloff may overprice a sudden federal stop while underpricing slower, fragmented local permitting and power constraints. Days: positioning and IPO sentiment dominate. Months: local rules and financing terms matter. Over 6–18 months, execution and utilization determine whether capex supports earnings or becomes stranded capacity. The thesis weakens if permitting timelines remain stable and AI/cloud monetization accelerates; it strengthens on project delays, rising grid costs, or a materially discounted ANT offering.

Key Decisions for Investors

• Keep GOOG as a selective relative long rather than a broad AI-capex proxy; size against evidence that new capacity is translating into cloud/AI revenue, not announced spending alone.
• Do not chase ANT ahead of price discovery. Treat a delayed listing or meaningful valuation reset as a financing-sentiment signal, not by itself proof that AI demand is deteriorating.
• Watch local permitting timelines, utility interconnection queues, project cancellations, and incremental power costs over the next 1–3 months; these are better leading indicators than company statements about demand.
• Falsification: reduce the infrastructure-risk view if major projects proceed without material schedule slippage and monetization improves; increase caution if local approvals stall or ANT’s offering is delayed or repriced sharply.