Anthropic Denies Discussing Government Stake
Claire Weston
The Trump administration and Anthropic have never discussed a government stake in the company, Reuters reports — contradicting earlier speculation and exposing a clear split among AI giants on the 'equity-for-policy' question.
What actually happened?
Reuters, citing people familiar with the matter, reports that the U.S. government and Anthropic never held any talks on a government equity stake, directly contradicting earlier market rumors.
This means → Anthropic's relationship with Washington remains at the regulatory level, with no equity negotiation on the table.
As of publication, neither Anthropic nor the U.S. Commerce Department has commented.
What is OpenAI proposing?
The Financial Times previously reported that OpenAI is discussing giving the U.S. government a 5% stake, aiming to clear political hurdles by offering Washington a financial upside.
Kevin Hassett, director of the National Economic Council, told CNBC: "We are talking to all the big AI companies about how they think they can share upside with the American people."
In plain terms = OpenAI's play is: hand the government a slice, get a smoother path — but Hassett disclosed no specific terms.
Why are the two companies diverging?
Anthropic denies any talks; OpenAI actively explores a deal. The two have taken visibly different stances on equity-for-policy.
This reflects a deeper reality: the AI industry has no unified playbook for dealing with government oversight — each company is betting on a different political route.
This means → For investors, "government stakes in AI companies" is not an inevitable industry trend — it is a variable still in play.
Export controls just eased — what does that signal?
This week the U.S. Commerce Department lifted export restrictions on Anthropic's Claude Fable 5 and Mythos 5, ending a high-stakes standoff over export controls.
The government's core concern in tightening reviews of new model releases is preventing advanced AI models from being misused by military-intelligence agencies in China, Russia, and elsewhere.
Yet new-model review submissions remain voluntary — Washington wants oversight but lacks a mandatory mechanism.
In plain terms = The export-control thaw shows Anthropic and the government are not on hostile terms, but that does not mean they are heading toward an equity arrangement.
Content is for reference only, not financial advice.