Commerce Secretary Howard Lutnick announced on September 2, 2026, that AI firm Anthropic is back on the right side with the Trump administration following an earlier dispute over battlefield AI safety, military contracts, and supply-chain risk designations that a federal judge subsequently ruled unlawful.
A turbulent year of high-stakes feuding between the federal government and one of America’s leading artificial intelligence developers appears to have cooled significantly. Speaking on the sidelines of a G20 Innovation Ministerial in Chapel Hill, North Carolina, Commerce Secretary Howard Lutnick signaled a stark reversal in tone regarding AI giant Anthropic after a dramatic clash earlier this year surrounding the Pentagon’s use of proprietary language models.
Prompted by Axios on the sidelines of the G20 Innovation Ministerial on September 2, 2026, Lutnick delivered a direct assessment of the restored partnership. Lutnick said Anthropic has done what we asked
and is back on the right side,
adding we trust Anthropic
.
Military Safety Guardrails and the Federal Court Battle
The thaw follows months of severe friction rooted in technological boundaries. Defense Secretary Pete Hegseth had previously blocked Anthropic from certain military contracts because the company insisted there be guardrails in the military’s use of its AI models, including prohibiting the Pentagon from using its products on mass surveillance of Americans and autonomous weaponry. That insistence on strict battlefield safety guardrails triggered immediate retaliation from the White House.
President Donald Trump then directed all federal agencies to stop using Anthropic, calling it a radical left, woke company,
and accusing Anthropic’s Leftwing nut jobs
of making a disastrous mistake trying to strong-arm
the Department of Defense. In March, the Pentagon took the extraordinary step of labeling the company a supply chain risk. Anthropic sued over the designation in California court, accusing the Trump administration of an unlawful campaign of retaliation.
The legal pressure culminated after a federal judge ruled in favor of Anthropic, determining the Pentagon illegally punished the AI company for criticizing the government. A U.S. judge sided with Anthropic in the company’s fight with the American military over AI safety on the battlefield on August 27, ruling that the government’s actions violated Anthropic’s constitutional rights.
Tom Brown Steps Into the Diplomatic Spotlight at the G20
As executive relations thawed, leadership roles within the company shifted. Anthropic co-founder Tom Brown took center stage at the G20 ministerial event, stepping into a prominent role in dealing with Lutnick and the Trump administration. Introducing him to an international audience of economic and technology ministers, Lutnick opened the discussion by giving Brown an opportunity to introduce himself: I don't think they know you so well. Anthropic, of course, is famous, but a little story helps.
Lutnick also said on Wednesday to a room full of G20 ministers, Really excited for our conversation, so I'd like to introduce you all to Tom Brown, one of the founders of Anthropic.
Brown praised President Trump’s support for controversial data centers, telling the audience, I really love Trump's post from earlier this week . . . where he was pointing out that data centers are just an enormous source of prosperity.
Brown stressed the importance of data centers at the gathering, urging the gathered officials to grant permits to build them in their countries, noting that They produce a ton of jobs. They produce taxes. Now the way that we design them, we actually bring on more power to the grid.
His public alignment with the administration’s infrastructure priorities underscores a calculated pivot from courtroom adversaries to cooperative partners in national AI expansion.
Export Controls, Model Restorations, and Ongoing Litigation
The road to reconciliation involved intense administrative maneuvering behind the scenes. Personality differences added to tensions between the Trump administration and Anthropic earlier this year, resulting in sweeping export controls on the company’s most advanced models. Lutnick invoked export controls to restrict access to foreign nationals, citing threats to national security, prompting Anthropic to take its Claude Mythos 5 and Claude Fable 5 models offline earlier this year. Access to Fable, which is available to the public, was restored several weeks later, while Mythos access was returned to a select group of customers. Brown was reportedly instrumental in dealing with Lutnick and the Trump administration to bring the models back online.
Anthropic signed a $200 million contract with the Pentagon last year, and its Claude AI tool was the first used in the Pentagon’s classified networks. Despite the repaired executive relationship, legal entanglements have not entirely vanished. Anthropic has a second ongoing lawsuit against the Trump administration in Washington, D.C., over a separate Pentagon supply-chain designation. How that second legal challenge proceeds in the wake of the administration’s newly cordial public stance remains a crucial test for the partnership.
