The Left Hand Bans What the Right Hand Deploys

The Trump administration is fighting Anthropic in two federal courts while simultaneously telling Wall Street to use its most powerful AI model. If that sounds incoherent, it's because it is.

The ban

The timeline is worth recounting because it's only been seven weeks.

In late February, Defense Secretary Pete Hegseth gave Anthropic CEO Dario Amodei a Friday deadline: remove Claude's safety restrictions on autonomous weapons and mass surveillance, or lose a $200 million defense contract. Amodei refused. He told Hegseth the company could not "in good conscience" allow it.

Hegseth's response was unprecedented: he designated Anthropic a supply chain risk to national security - the first time an American company has received that label. Trump followed up on Truth Social, calling Anthropic a "radical left, woke company" and ordering every federal agency to "IMMEDIATELY CEASE all use of Anthropic's technology."

Anthropic sued. A California federal judge blocked the ban, writing that "nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary for expressing disagreement with the government." The Trump administration appealed to the Ninth Circuit. On April 8, a D.C. appeals court sided with the Pentagon, denying Anthropic's request to halt the blacklisting during litigation.

That's the legal state of play: Anthropic is officially designated a supply chain risk. Federal contractors are barred from doing business with the company. The government is supposed to be phasing out all Anthropic technology.

The deployment

On the same day the appeals court upheld the ban, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell summoned the CEOs of America's largest banks to Washington for an emergency briefing.

The attendees: Goldman Sachs, Citigroup, Morgan Stanley, Bank of America (CEO Brian Moynihan confirmed present), and Wells Fargo. JPMorgan CEO Jamie Dimon was summoned but unable to attend.

The message from Bessent and Powell: test Anthropic's Mythos model for vulnerability detection, fraud flagging, and compliance automation. The two most senior economic officials in the US government were actively encouraging the private sector to adopt technology from a company their own administration had designated a national security risk.

JPMorgan is already on board - it's one of the 12 founding members of Project Glasswing, Anthropic's cybersecurity coalition with $100M in API credits.

As The Next Web put it: "the left hand blacklists what the right hand is busy deploying."

The quiet workaround

It's not just banks. Federal agencies are finding ways around the ban.

The Commerce Department's Center for AI Standards and Innovation is actively testing Mythos. CISA - the agency responsible for defending federal networks - has been in discussions with Anthropic about the model's implications for critical infrastructure security.

The strategic gap enabling this: Hegseth's supply chain designation bars DoD contractors from working with Anthropic. It does not technically prevent non-defense agencies from engaging with the company, especially when the technology in question has direct relevance to their defensive mission.

Anthropic co-founder Jack Clark made this explicit at the Semafor World Economy Summit on April 14, confirming the company briefed the Trump administration directly on Mythos. His framing was careful: "We have a narrow contracting dispute. I don't want that to get in the way of the fact that we care deeply about national security." He added: "We're talking to them about Mythos, and we'll talk to them about the next models as well."

TechCrunch confirmed the briefing happened. The company is simultaneously suing the government and providing it with intelligence about its most capable model.

The valuation

While the legal fight plays out, the market has its own verdict. Bloomberg reported today that Anthropic is drawing investor offers at over $800 billion valuation. For context, that's roughly where Meta was a year ago.

The Pentagon ban hasn't dented investor confidence. If anything, Glasswing and Mythos have strengthened it - the model that the government can't officially use is the model that Wall Street CEOs are being told to adopt.

What this actually means

The incoherence is the point. Different parts of the Trump administration have different interests:

The Pentagon wants unrestricted AI for military applications. Anthropic said no. Hegseth retaliated.

Treasury and the Fed want the financial system protected from AI-enabled cyberattacks. Mythos found thousands of zero-days across every major OS and browser. That capability is more relevant to financial infrastructure than any other AI model available.

The White House wants to maintain US AI leadership globally. Banning America's second-most-valuable AI company from government work while Britain actively courts it is strategically counterproductive.

The result is a policy that simultaneously treats Anthropic as a threat and a critical national asset, depending on which office you walk into. The courts will eventually resolve the legal question. The strategic question - whether you can blacklist a company whose technology your own officials are telling banks to adopt - answers itself.

Sources:

• Fox News - Bessent, Powell Summon Wall Street CEOs for Emergency Meeting
• The Next Web - Trump Administration Banks Anthropic Mythos Pentagon Paradox
• TechCrunch - Anthropic Co-Founder Confirms Government Briefing on Mythos
• Axios - CISA Cuts, Anthropic Lawsuit Complicates Mythos Response
• Military.com - Appeals Court Rebuffs Anthropic