The Trump administration says its conflict with Anthropic has come to an end. U.S. Commerce Secretary Howard Lutnick told Axios on September 1, 2026, that the administration now trusts the company behind Claude and believes it has done what officials asked.
“We trust Anthropic,” Lutnick said. He added: “They’ve done what we asked. They’re back on the right side. So the answer is: Yes.” The comments close months of open conflict between the U.S. government and Anthropic over national security and AI safeguards.
A sharp turn after months of conflict
The change follows a series of actions that placed Anthropic at the center of a dispute with the federal government. On February 27, 2026, President Trump posted a directive ordering every federal agency to cease all use of Anthropic’s technology, with a six-month phase-out.
Secretary of War Pete Hegseth also announced that he was directing the Department of War to designate Anthropic a supply-chain risk to national security. He ordered that no contractor, supplier, or partner doing business with the U.S. military could conduct any commercial activity with the company.
Lutnick’s September 1 statement shows the administration has moved from that ban toward cooperation. His remarks did not describe a new agreement, but they made the administration’s position clear: Anthropic has met the requests that led to the dispute.
That shift was visible at the G20 Innovation Ministerial in Chapel Hill, North Carolina. The U.S. Department of Commerce and the White House Office of Science and Technology Policy co-hosted the gathering on September 1 and 2, 2026, bringing commerce and technology ministers from more than 20 countries together.
Anthropic presents data centers as a public benefit
At the ministerial on September 2, Lutnick introduced Anthropic co-founder Tom Brown to the assembled G20 ministers. Brown praised President Trump’s recent social media post supporting data center development, saying, “I really love Trump’s post from earlier this week… where he was pointing out that the data centers are just an enormous source of prosperity.”
Brown said data centers produce jobs and tax revenue while bringing more power onto the grid. His remarks connected Anthropic’s business plans to the administration’s broader support for building data center infrastructure.
Brown also highlighted Anthropic’s support for the administration’s ratepayer pledge. The company outlined that commitment in a February 11, 2026 policy post, promising to cover electricity price increases faced by consumers from its data centers.
The pledge includes several parts. Anthropic committed to paying for all grid upgrades needed to interconnect its data centers, procuring net-new power generation, and investing in curtailment systems. Together, those commitments address the costs and power demands tied to the company’s data center operations.
The U.S. pushes a lighter touch on AI rules
The G20 meeting also carried a broader message about how governments should regulate AI. The U.S. pressed G20 members to take a hands-off approach and avoid creating rules for the technology.
Those goals align with the thinking of the world’s biggest AI companies, nearly all of which are American. U.S. tech adviser Michael Kratsios advocated for countries to adopt the “Carolina Principles,” which set out a limited approach to new AI regulation.
Countries that signed on agreed to avoid writing entirely new regulations for AI. Instead, they would focus on rules for “novel” situations involving the technology.
That approach places the emphasis on specific new uses and problems rather than building a broad set of rules for AI as a whole. It also gives the U.S. position at the G20 a clear connection to the interests of American AI companies, including Anthropic.
Anthropic’s return to the administration’s good side now sits alongside that regulatory push. The company went from a federal phase-out order and a national security designation to a public introduction before G20 ministers within months.
Lutnick’s words captured the change in simple terms: “They’ve done what we asked.”
Based on
