Anthropic and White House ease tensions before IPO

TL;DR:

• A months-long dispute between Anthropic and the Trump administration is showing signs of easing as the firm prepares to go public.
• The Pentagon labelled Anthropic a “supply-chain risk” in March — a first for a US company — barring contractors from using its AI, and the firm is challenging it in court.
• Anthropic is preparing a stock-market debut that could value it at around $1tn.

The relationship between Anthropic and Washington is thawing at a pivotal moment. Parts of the US government are easing a months-long standoff with the AI firm as it gears up for an IPO that could value it at roughly $1tn, according to people familiar with the matter — though the underlying Pentagon blacklist remains in force.

A defence rift still unresolved

The rupture began when Anthropic refused to let the US military use its models for domestic surveillance and fully autonomous weapons. In retaliation, the Defense Department labelled it a “supply-chain risk” in March — the first time a US company received a designation normally reserved for firms tied to adversarial nations — barring tens of thousands of contractors from using its AI. The firm is contesting the label in court, and the Pentagon is still “vigorously” defending the case. “Anytime the government signals that it’s washing its hands of a company, that’s a major problem,” said government-contracts attorney Franklin Turner.

Signs of détente have followed chief executive Dario Amodei’s April White House visit. Anthropic has discussed protecting critical infrastructure from AI-enabled cyberattacks with the national cyber director, and its staff met Treasury Secretary Scott Bessent — talks that helped shape a recent executive order asking developers to submit advanced models for cybersecurity testing. Yet the easing is partial: Anthropic was left out of an Army-run cyberattack simulation that included Google and OpenAI.

Looking forward

The saga is US-centric, but it matters for anyone tracking frontier-lab market structure — including the UK firms and investors weighing exposure to Anthropic, whose confidential IPO filing edged ahead of OpenAI’s. It is a case study in how political risk now sits alongside technical risk for AI leaders. The unresolved blacklist is the swing factor: a favourable court ruling would clear a major overhang before any listing, while a loss could dent investor confidence at exactly the wrong moment.