The Regulatory Moat: Anthropic's Push for Tougher State AI Laws

### Analysis: The Rules of the Game

**Opinion:** In the high-stakes race for AI dominance, the most effective weapon isn't always a better model—it's a more restrictive rulebook. Anthropic's current trajectory suggests a classic moat-building strategy. By aggressively lobbying for state-level regulations that mandate expensive third-party audits and strict transparency, Anthropic is effectively attempting to codify its own operational standards into law. For a company now valued at nearly $1 trillion, these requirements are manageable; for a leaner, more aggressive startup, they represent a prohibitive cost of entry. By framing these hurdles as 'safety,' Anthropic can potentially stifle the agility of smaller rivals under the guise of civic responsibility.

***

### The Roundup

**The Push for 'Tougher' Standards** As Wired first reported, Anthropic is currently urging U.S. states to adopt more stringent AI regulations, arguing that previous efforts are already obsolete. Cesar Fernandez, Anthropic’s head of US state and local government relations, told Wired that while the transparency-focused laws passed in New York and California in 2025 were a starting point, self-reporting is no longer a sufficient safety measure for the most powerful systems.

Anthropic has expanded its support beyond simple transparency. Wired reports the company has endorsed an Illinois measure requiring third-party safety audits for AI labs. Furthermore, the company has backed a Massachusetts policy that would mandate these audits and grant the state's attorney general the power to seek injunctive relief against non-compliant firms.

**The 'Regulatory Capture' Debate** This pro-regulation stance has sparked a fierce debate over the company's motives. Wired notes that David Sacks, a technology adviser to President Donald Trump and former White House AI czar, has explicitly accused Anthropic of employing a "sophisticated regulatory capture strategy." Sacks claims the company is using "fear-mongering" to pass cumbersome laws that trap smaller startups in red tape, thereby securing Anthropic's leadership position.

In response, Cesar Fernandez told Wired that the company only supports bills targeting "large AI model developers." While definitions vary by bill, Fernandez notes these typically apply to firms with over $500 million in annual revenue and those that have spent hundreds of millions on development. He argued it is "hard to imagine a startup" meeting those specific thresholds.

**The Competitive Landscape** Wired points out that while many startups may be exempt, several well-funded competitors could potentially hit those thresholds. The outlet specifically names Mistral, Thinking Machines Lab, and Safe Superintelligence as entities that have raised billions of dollars, positioning them as potential rivals to Anthropic and OpenAI.

**The Federal Friction** There are limits, however, to what Anthropic thinks states should do in the absence of federal regulation. Wired reports that in a policy document released last month, Anthropic suggested governments should have the power to block the deployment of unsafe AI models. However, the company was reportedly not a supporter when the Trump administration recently told Anthropic to suspend access to its two most powerful models for foreign nationals.