Here’s Why Anthropic Is Urging Governments to Accelerate AI Regulation

Anthropic expressed its backing for the initial set of frontier AI safety regulations in the United States last year, achieving new transparency mandates in California and New York that faced opposition from many in Silicon Valley, who argued they would hinder the AI surge. However, Anthropic contends that these regulations might already be outdated, and the company is now advocating for even stricter rules at the state level.

“The transparency-driven safety legislation of 2025 represented a crucial beginning, but as AI systems keep advancing rapidly, the policy measures must keep pace,” said Cesar Fernandez, Anthropic’s head of US state and local government relations, in an interview with WIRED. “We believe that transparency and self-reporting are no longer adequate safety precautions for the most advanced AI systems.”

Advocating for regulation is a peculiar stance for a startup with a valuation nearing $1 trillion. Yet, Anthropic is an unconventional company. As previously noted, Anthropic’s leadership believes in building a significant enterprise focused on developing and providing access to advanced artificial intelligence to fulfill its foundational mission: “to ensure that the world safely transitions through transformative AI.”

As Anthropic has expanded, it has begun endorsing some of the strictest proposed regulations for frontier AI firms. Many of these measures aim to reduce catastrophic AI risks, including the potential for advanced models to cause financial disasters or mass casualties.

In addition to the self-reporting mandates in California and New York, Anthropic has also backed an Illinois initiative that requires AI labs to undergo safety process evaluations by third-party auditors. Most recently, the company supported a Massachusetts policy that would mandate third-party auditing for AI labs and empower the state’s attorney general to pursue injunctive relief against non-compliant firms.

I spoke with Fernandez this week to gain insight into the direction of the company’s AI policies. He joined Anthropic earlier this year, having previously served as head of US state government relations for the sports betting giant FanDuel and as a senior public policy associate at Uber. He played a significant role in helping both companies navigate policy challenges across the country. His expertise will likely be an asset for Anthropic as Congress continues to delay AI regulation and states begin to take initiative.

“Dubious” Motives

Although Anthropic’s pro-regulation stance has received acclaim from AI safety organizations, labor unions, and other allies, some leaders in Silicon Valley view its political maneuvers with skepticism.

David Sacks, the former White House AI czar and current technology advisor to President Donald Trump, has asserted that Anthropic is essentially attempting to enact cumbersome regulations that ensnare smaller AI startups in bureaucracy, thereby solidifying its own dominance in the AI sector.

“Anthropic is executing a sophisticated regulatory capture strategy based on fear-mongering,” Sacks claimed in a social media post last year. “It is primarily responsible for instigating the state regulatory frenzy that is harming the startup ecosystem.”

Fernandez categorically rejects this characterization, pointing out that Anthropic has only endorsed state AI legislation that pertains to “large AI model developers,” a term that varies in definition across different bills but generally refers to companies investing hundreds of millions in AI development and generating over $500 million in annual revenue. “It’s difficult to picture a startup meeting that criterion,” asserts Fernandez.

However, in a landscape where significant capital is now necessary to create a competitive frontier AI model, there are arguably a few promising startups that could soon meet those thresholds. Examples include Safe Superintelligence, Thinking Machines Lab, and Mistral, all of which have secured billions from investors, albeit their revenue remains modest compared to that of Anthropic or OpenAI. While these aren’t typical startups, they do pose potential competition for Anthropic.