President Trump said the United States “might” take government stakes in OpenAI and Anthropic along the lines of the Intel deal, while ruling out nationalizing top frontier AI labs, according to a White House interview TIME published on Oct. 1 from a Sept. 28 conversation.
Asked whether he would nationalize leading AI labs, Trump answered “No,” per TIME’s transcript. Asked why Washington would not take shares in OpenAI and Anthropic the way it did with Intel, he said: “I might. Maybe I could do that. I have many deals like that—I do that.” The exchange announced no negotiations, ownership percentage, dollar amount, board or governance package, funding source, or closing date for any AI-lab stake.
The model Trump pointed to is the 2025 U.S.–Intel arrangement. Per Intel’s Aug. 22, 2025 announcement, the United States agreed to invest about $8.9 billion for roughly 433.3 million shares — about 9.9% — funded by unpaid CHIPS Act grants and Secure Enclave awards, as a largely passive holding with no board seat and a warrant that could add about 5% more if Intel loses majority foundry control. In the TIME interview, Trump said he had sought “10%” and claimed the deal made the United States “like $60 or $70 billion”; that paper-gain figure is his estimate, with no calculation or valuation date supplied in the transcript. CHIPS conversion is the Intel funding story — not an announced path for OpenAI or Anthropic stakes.
July 2026 reporting, including the Financial Times as carried by Reuters and others, said OpenAI had discussed offering Washington roughly a 5% stake in early talks. Reuters, citing sources, reported then that the administration and Anthropic had not discussed a government stake in Anthropic. Those accounts are background, not a signed October deal, and neither company had an on-record reply to Trump’s TIME “might” line located for this piece.
Trump’s same answer also cited requiring a “cut” for the United States on certain Nvidia chip sales — export-linked revenue share in his telling, not the same instrument as equity. This piece is about ownership and industrial-policy upside talk, not the White House’s already-covered voluntary AI self-police track or OpenAI safety-staff exits.
What the TIME exchange settles is tone, not a term sheet: nationalization off the table in his words; Intel-style minority stakes left as a presidential “might,” with every concrete term still to be disclosed if a real proposal appears.