Inside Anthropic CEO’s leaked memo attacking OpenAI

Hello,

This week, a private memo from Anthropic CEO Dario Amodei has spotlighted hostile relations between the leaders of the AI industry. Catch up on what you may have missed.

In a 1,600-word message to employees, Amodei criticized the Trump administration and attacked OpenAI's new Pentagon deal, accusing the company of abandoning the safety principles many AI labs once shared. Amodei later apologized publicly for the tone of the memo, as the dispute with the Pentagon continued to unfold.

The Information broke this story, revealing what Amodei told staff and exposing deeper tensions between Anthropic and OpenAI as both compete to shape the future of AI.

This isn't speculation. It's primary-source reporting—internal communications that show how decisions about AI safety, government partnerships and industry power are actually being made. Subscribe for $399 $299 and save 25% on the first year and be the first to know.

Anthropic's Memo

In his message to employees, Amodei argued the administration's hostility toward Anthropic wasn't just about policy. He wrote that the company hadn't donated to Donald Trump or offered the kind of "dictator-style praise" for the president that leaders such as OpenAI CEO Sam Altman had.

The Attack on OpenAI

Amodei also blasted OpenAI's Pentagon announcement, calling it "mendacious" and accusing the company of staging "safety theater." He argued OpenAI's safeguards against misuse were largely symbolic and said CEO Sam Altman was misleading both the public and employees about how closely the companies' safety standards aligned.

The Fallout

Anthropic's dispute with the Pentagon has already begun spilling into business relationships. Anthropic's partnership with Palantir to serve government customers could be at risk after the Pentagon labeled Anthropic a "supply chain risk."

Meanwhile, the leak of Amodei's comments has complicated attempts to reach a compromise with the Defense Department, turning a policy dispute into a public confrontation.

Watch a special episode of TITV, our daily news show, taking you inside our newsroom in real time as they cover Anthropic's Memo and what this means for the AI landscape.

Why It Matters

Our readers subscribe because they can't afford to make decisions with yesterday's information. Subscribers also get access to conversations with our journalists—including a live debrief on this story on Wednesday, where they asked questions directly.

Don't miss the reporting behind the AI industry's biggest power struggles and opportunities to join live conversations like this in the future, subscribe for $299 and save 25% on the first year.