AI at war: US strikes Iran with Anthropic's Claude hours after Trump ban

Just hours after President Donald Trump ordered federal agencies to stop using artificial intelligence tools developed by Anthropic, US forces in the Middle East relied on the company’s Claude system during a major air operation targeting Iran, according to a report by The Wall Street Journal.

What happened

According to the Journal, US Central Command (Centcom), which oversees American military operations in the Middle East, has been using Claude in active operational environments.

AI System supported a range of battlefield tasks, including:

• Intelligence assessments
• Target identification
• Simulated battle planning and scenario modelling

The airstrikes took place only hours after Trump directed federal agencies to halt the use of Anthropic’s tools. The order followed defense secretary Pete Hegseth’s decision to designate the company a “supply-chain risk.”

Why it matters

The situation exposed a striking contradiction.

While the White House publicly labelled Anthropic a security concern, the Pentagon was still relying on its technology in a live military operation.

Claude’s use in intelligence fusion, predictive targeting and scenario simulation reflects the growing role of automation in modern military planning and the difficulty of quickly replacing such systems once deployed.

Trump has allowed a six-month phaseout period. Experts cited by the Journal said replacing AI systems embedded in classified military networks could take months or longer.

Between the lines

Tensions between the administration and Anthropic have been building for months.

During contract negotiations, Anthropic refused to grant the Pentagon blanket permission to use its systems for all lawful purposes. CEO Dario Amodei said Thursday:

“We cannot in good conscience accede to their request.”

The company drew red lines around potential uses such as domestic mass surveillance and autonomous lethal weapons.

Pentagon officials insisted that the military would operate within the law and argued that private contractors should not set national security policy.

Trump escalated the dispute Friday, writing:

“The United States of America will never allow a radical-left, woke company to dictate how our great military fights and wins wars.”

Defense officials also warned of “civil and criminal consequences” if Anthropic failed to cooperate during the transition.

A pattern in covert ops

Claude’s reported role in the Iran operation is not the first time the system has been used in high-profile missions.

According to the Journal, AI-assisted intelligence analysis using Claude also supported the operation that led to the capture of Venezuelan leader Nicolas Maduro.

The precedent helps explain why officials expect the removal of Claude from government systems to take months.

The system is reportedly integrated through partnerships with data analytics contractors such as Palantir Technologies, making any transition technically complex.

The pivot to OpenAI

As Anthropic’s relationship with the administration deteriorated, rival firms moved quickly to fill the gap.

The Journal and The New York Times reported that OpenAI reached an agreement with the Defense Department allowing its models to operate in classified environments — access Anthropic had previously secured.

OpenAI chief executive Sam Altman said:

“We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements.”

He also said:

“In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome.”

Under the agreement, OpenAI will allow use of its models for all lawful purposes while building in safeguards against domestic surveillance and autonomous weapons applications.

Elon Musk’s xAI has also received approval for classified use.

The supply-chain shock

Hegseth’s decision to designate Anthropic a “supply-chain risk” is a label typically applied to foreign adversaries. Applying it to a US technology company marks an unusually aggressive step.

Hegseth wrote:

“America’s warfighters will never be held hostage by the ideological whims of Big Tech.”

Anthropic pushed back, saying:

“We believe this designation would both be legally unsound and set a dangerous precedent for any American company that negotiates with the government.”

Some analysts warned the move could discourage private-sector cooperation with the Pentagon.

Dean Ball, a former Trump AI adviser, told the Journal:

“This is a dark day in the history of American AI. The message sent to the business community and to countries around the world could not be worse.”

Jack Shanahan, who previously led military AI programmes, said:

“This is about Anthropic not being one of the favored companies and they’re going to pay the price for not bowing down and not signing on the dotted line.”

The big picture

The Iran operation highlights several broader shifts.

AI is already operational

AI tools are no longer experimental. They are being used directly in combat planning and operational decision-making.

Vendor politics shape battlefield capability

Which companies the Pentagon works with — Anthropic, OpenAI, xAI or others — is increasingly shaping the architecture of military systems.

Guardrails vs sovereignty

Anthropic sought contractual limits on how its technology could be used.

The Pentagon insisted that operational authority must remain entirely with the government.

The dispute ultimately became a matter of national policy.

What’s next

The Pentagon is expected to begin phasing out Claude over the next six months, while OpenAI and xAI expand their presence in classified environments.

Anthropic is preparing legal challenges to the supply-chain designation, and contractors connected to Claude are assessing their exposure.

For now, Centcom operations continue.

The most striking detail may be that even as Washington declared Anthropic persona non grata, its AI was still being used to simulate battle scenarios in an active theater.

AI-driven warfare is no longer theoretical. It is operational, deeply integrated and politically contested and the Iran strikes showed that even a presidential order cannot immediately unplug it.