On October 21, 2025, Dario Amodei, CEO of Anthropic, published a statement reiterating the company’s mission that artificial intelligence should advance human progress rather than pose danger. The statement outlines Anthropic’s recent government partnerships, commercial growth, regulatory positions, and responses to inaccurate claims about the company.
Key business and contract details
- Anthropic says it became the fastest-growing software company in history, with revenue run rate rising from $1 billion to $7 billion over the past nine months. Amodei stressed this growth occurred alongside what he described as thoughtful, responsible AI deployment.
- In July, the Department of War awarded Anthropic a two-year, $200 million agreement to prototype frontier AI capabilities that advance national security.
- Anthropic partnered with the General Services Administration (GSA) to offer Claude for Enterprise and Claude for Government across the federal government for $1.
- Claude is deployed on classified networks via partners such as Palantir and at Lawrence Livermore National Laboratory.
Government relations and political engagement
- Amodei stated that Anthropic publicly praised President Trump’s AI Action Plan and supported the President’s efforts to expand U.S. energy provision to help win the AI race. He said he personally attended an AI and energy summit in Pennsylvania with President Trump, where they discussed U.S. leadership in AI.
- Company executives have attended White House events: Anthropic’s Chief Product Officer joined a pledge to accelerate healthcare applications of AI, and the Head of External Affairs participated in the White House AI Education Taskforce event aimed at advancing AI fluency for teachers.
- Anthropic emphasizes bipartisan hiring practices and said its advisory council includes senior former Trump administration officials. The company asserts hiring is based on expertise, integrity, and competence rather than political affiliation.
Regulatory stance: national standard versus state-level laws
- Anthropic argues that managing AI’s societal impacts should be a matter of policy rather than politics and favors a uniform federal approach over a patchwork of state laws.
- The company opposed one proposed amendment in the One Big Beautiful Bill: a 10-year moratorium on state-level AI laws that would have blocked action without a federal alternative. That amendment was rejected in the Senate by a 99–1 vote.
- Recognizing Congress moves slowly, Anthropic supported California’s SB 53, which requires the largest AI developers to publish frontier model safety protocols. The bill exempts companies with annual gross revenue below $500 million, so it applies only to the largest firms; Anthropic supported this exemption and says it proposed an early version of it to protect startups.
Commercial restrictions and international risk
- Anthropic warned that the real risk to American AI leadership is not a single state law but enabling the People’s Republic of China (PRC) by supplying U.S. chips to PRC data centers they cannot produce themselves. The company said it agrees with leaders such as Senators Tom Cotton and Josh Hawley that such exports would aid the Chinese Communist Party’s progress in the AI frontier.
- Anthropic also stated it is the only frontier AI company that restricts selling AI services to PRC-controlled companies, foregoing significant short-term revenue to avoid enabling AI platforms and applications that could assist the Chinese Communist Party’s military and intelligence services.
Progress on model bias and neutrality
- Amodei disputed claims that Anthropic’s models are uniquely politically biased. He cited a January study from the Manhattan Institute, a conservative think tank, which found Anthropic’s then-main model (Claude Sonnet 3.5) to be less politically biased than models from most other major providers.
- He also referenced a May Stanford study on user perceptions of bias, which does not single out Anthropic; many other providers’ models were rated as more biased in that study.
- System cards for Anthropic’s latest models, Sonnet 4.5 and Haiku 4.5, indicate rapid progress toward political neutrality. At the same time, Amodei acknowledged no AI model is fully politically balanced in every reply because models learn from training data in ways not yet fully understood and developers cannot fully control all outputs.
Engagement and intent
- Anthropic said it remains committed to constructive engagement on public policy: agreeing publicly when aligned, and proposing alternatives when it disagrees. As a public benefit corporation, Anthropic frames its mission around ensuring AI benefits everyone and maintaining U.S. leadership in AI development.
- Amodei quoted Vice President JD Vance’s recent remark that AI’s effects are “probably both” good and bad and emphasized the company’s aim to maximize benefits and minimize harms. He said Anthropic is prepared to work in good faith with people of any political stripe to realize that aim.
Why this matters
The statement clarifies Anthropic’s policy positions amid public and political debate. The company’s rapid revenue growth, the large government contract, targeted commercial restrictions related to the PRC, and its public support for certain regulatory measures position Anthropic as an active participant in shaping how the United States balances AI innovation, safety, and national security.
