Palantir's CEO Says OpenAI Wants to Be Nationalized. OpenAI Already Filed to Go Public.
Alex Karp says the AI labs' apocalypse talk is a pitch for a government bailout, not a safety warning. He made that case a month after OpenAI's own CFO told staff an IPO is coming in 2027 — and OpenAI had already quietly filed the paperwork Karp says it's too scared to file.
What was checked, and what could not be.
Karp's quotes are corroborated across at least four independently reported outlets (Yahoo Finance, TechRadar, Moneywise, NewsBytesApp), all citing the same September 17, 2026 CNBC interview, with matching wording on the key lines ("nationalize us, please"; "these businesses have to be nationalized..."; "you're assuming there will be an S-1"). I did not find a CNBC primary transcript I could load directly, but the consistency across multiple independent secondary reports on the same quotes gives reasonable confidence in accuracy; flagging that this piece relies on secondary reporting of a broadcast interview rather than a primary transcript.
OpenAI's June 8, 2026 confidential S-1 filing and Anthropic's June 1, 2026 filing are corroborated by multiple financial-press summaries (Yahoo Finance, CNBC, valueaddvc, ipo.club). Sarah Friar's August 19, 2026 all-hands remarks ("will be a public company in 2027") are sourced to CNBC's own August 19, 2026 report, which I take as reliable since CNBC is reporting its own network's access to the meeting.
Altman's 5% government-stake proposal is sourced to CNBC (Jul 2, 2026) and corroborated by Yahoo Finance, Time and multiple other outlets citing the same FT-originated report; I labeled it as a minority, non-controlling proposal per those reports, distinct from Karp's 50% framing.
Sanders' "American AI Sovereign Wealth Fund Act" (June 2026, 50% government voting stake) is corroborated by the same set of nationalization-roundup search results; I did not independently pull the bill text, so I'm treating its existence and headline terms as reported rather than independently verified against the legislative record.
Could not verify: the exact six-month timeline Karp claims for privately warning AI executives (his own claim, unverifiable by definition — treated as an assertion, not a fact). Could not obtain full text of the TechRadar article (page truncated on fetch) so I relied on its headline plus corroborating outlets for that specific claim, not TechRadar's full framing.
I paraphrased rather than quoted Karp's March 2026 remark at the a16z summit because the original wording uses a slur for intellectual disability; multiple outlets (Benzinga, Fortune) quote it directly, but reproducing it isn't necessary to accurately describe the argument he was making.