Trump: Anthropic No Longer a Safety Threat, but Intervention Options Remain on the Table
0xBroomberg
Trump says he no longer views Anthropic as a national-security threat, yet explicitly keeps the Defense Production Act on the table — a $900 billion-plus IPO gets a reprieve, but the regulatory sword stays drawn.
From "maybe a threat a week ago" to "very responsible" — what changed?
Trump told media he no longer considers Anthropic a national-security threat, while admitting "a week ago, maybe it was."
The turning point: a meeting with Anthropic CEO Dario Amodei at the G7 summit in France. Trump called him "very smart" and said the company has been "very responsible."
This means → the White House went from treating Anthropic almost like a foreign adversary to publicly endorsing it — in a matter of days.
What did the Commerce Department actually do?
The Commerce Department ordered Anthropic to obtain U.S. government approval before letting any foreign individuals, companies, or governments access its most advanced models — Fable 5 and Mythos 5. This is widely seen as the most significant government intervention in an AI company's operations to date.
Anthropic immediately disabled both models for all users and sent senior engineers to brief government officials on model capabilities and proliferation risks.
In plain terms = the government said: who gets to use your strongest models is our call, not yours. Anthropic chose to comply on the spot.
Who tipped off the government? Is this really about competition?
Trump revealed that the entity that flagged Anthropic to the government was Amazon — both a competitor and a partial shareholder in Anthropic.
Trump's words: "They were unhappy with what Anthropic was doing, and they were very concerned."
This reflects a commercial-competition dimension underneath the national-security framing — Big Tech rivals are using regulatory channels against each other.
The Defense Production Act is still on the table — will Trump use it?
Trump did not rule out invoking the Defense Production Act — a law that lets the president compel companies to cooperate with the government during emergencies.
His exact words: "I have a lot of tools I can use, but I'm not sure I have to use them."
This means → as long as Anthropic keeps cooperating, the coercive powers stay holstered — but the holster is open, and the government can draw at any time.
Can the IPO move forward?
Bloomberg reports that Anthropic confidentially filed IPO paperwork weeks ago. Its latest valuation tops $900 billion.
The Commerce Department restrictions had raised market fears: if the White House kept labeling Anthropic a security risk, both commercialization and the listing could take a hit.
In plain terms = Trump's softer tone gives the IPO a breathing window, but as long as the regulatory sword hangs overhead, investor uncertainty doesn't vanish.
What is the real test ahead?
Anthropic said in a statement it remains committed to working with the government to "ensure America's leadership in AI."
Washington increasingly treats the most advanced AI models as national strategic assets — the logic: keep the tech lead, control the tech spread.
This means → whether both sides can agree on AI-model safety-assessment standards is the real marker of resolution — what we have now is a temporary cool-down, not a closed case.
Content is for reference only, not financial advice.