The $1 Trillion De-escalation:
- The High-Stakes Thaw: A months-long, severe dispute between the Trump administration and AI powerhouse Anthropic is showing signs of easing as the company aggressively prepares for a blockbuster public debut that could value it at $1 trillion.
- The Core Clash: The relationship ruptured earlier this year after Anthropic refused to allow the U.S. military to use its AI models for domestic surveillance and fully autonomous weapons. The government retaliated by slapping Anthropic with a national security blacklist and a unprecedented “supply-chain risk” designation by the Pentagon in March—the first time a U.S. company has ever received this label. The blacklisting legally bars tens of thousands of defense contractors from using Anthropic’s software.
Signs of a Symbolic & Political Turnaround:
- White House Access: Tensions began to cool after Anthropic CEO Dario Amodei visited the White House in mid-April. In a major symbolic pivot, Amodei was invited to the White House for the signing of a sweeping executive order on artificial intelligence, which President Trump officially signed this Tuesday.
- Formulating Federal Policy: Anthropic employees met with U.S. Treasury Secretary Scott Bessent this spring. These high-level discussions directly helped Trump’s team draft the June 2, 2026 executive order, which mandates that leading AI developers hand over their most advanced models for government cybersecurity testing.
The “Mythos” Threat & Ongoing Legal War:
- Supercharging Cyberattacks: Anthropic has been holding briefings with National Cyber Director Sean Cairncross regarding Mythos, its most advanced AI system. The company explicitly warned the government that Mythos has the terrifying capability to supercharge AI-enabled cyberattacks against critical U.S. infrastructure (including banks and hospitals).
- The Legal Standoff: Despite the political thaw, the underlying structural damage remains. Anthropic is still vigorously suing the Department of Defense in court over the supply-chain risk label, with both sides filing conflicting court briefs as recently as this Thursday.
- The Pentagon Cold Shoulder: Highlighting the remaining friction, Anthropic executives were noticeably excluded from a crucial April 27 Army-run AI cyberattack simulation that included direct rivals like Google and OpenAI.