Anthropic's Regulatory Push: Safety Argument or Market Strategy?

Anthropic Escalates Push for State-Level AI Regulation

Anthropic is backing a second wave of stricter state AI regulations, arguing that 2025 transparency laws are now outdated. The company, valued at nearly $1 trillion, has endorsed rules requiring third-party safety audits in Illinois and Massachusetts—including enforcement powers that would allow state attorneys general to seek injunctions against non-compliant AI labs. According to Wired AI, Anthropic’s head of US state and local government relations, Cesar Fernandez, told the publication that “transparency and self-reporting are no longer sufficient safety measures for the most powerful AI systems.”

The 2025 Baseline and Capability Acceleration

Anthropic played a significant role in securing California and New York transparency requirements in 2025, despite opposition from much of Silicon Valley, which argued the measures would constrain innovation. The company now contends that those laws, which focused on disclosure and self-reporting mechanisms, have been overtaken by the pace of model capability advancement. Fernandez framed the shift as a necessary policy response: “as the capabilities of AI systems continue to advance quickly—the policy responses need to match.” This acknowledgment implies Anthropic views the regulatory landscape as perpetually reactive rather than anticipatory.

Massachusetts Auditing and Enforcement Model

The Massachusetts policy Anthropic endorsed extends beyond Illinois’s third-party audit requirement by granting the state’s attorney general injunctive authority over non-compliant labs. According to Wired AI, this combination—mandatory audits plus direct enforcement—represents a material tightening of the regulatory posture compared to 2025 transparency bills. The move signals Anthropic’s willingness to accept enforcement mechanisms that could impose operational constraints on frontier AI development.

Regulatory Capture Allegations

Not all observers read Anthropic’s regulatory advocacy as safety-driven. According to Wired AI, David Sacks, the former White House AI czar and current technology adviser to President Trump, has characterized Anthropic’s strategy as “regulatory capture”—using compliance burdens to erect barriers that would disadvantage smaller AI startups while protecting Anthropic’s competitive position. Sacks’ framing inverts Anthropic’s stated rationale: instead of safety-first regulation, he suggests the company is engineering market consolidation through policy complexity.

Anthropic’s track record of recruiting government-relations expertise reinforces both readings. Cesar Fernandez’s previous roles at FanDuel and Uber—companies that successfully navigated or reshaped state-level regulatory frameworks—suggest policy-engineering capability rather than mere safety advocacy. According to Wired AI, Fernandez joined Anthropic earlier in 2026 specifically to lead its state and local government relations.

Why This Matters

The divergence between Anthropic’s safety rationale and skeptics’ capture theory creates a structural ambiguity: regulation that is genuinely necessary for safety may simultaneously impose asymmetric costs on smaller competitors, and regulations designed to entrench market leaders may incidentally improve safety outcomes. The empirical question—whether Massachusetts-style audits and enforcement actually reduce frontier AI risks—remains uncovered by either Anthropic’s claims or Sacks’ allegations.

For state legislators, the challenge is distinguishing signal from noise when the company pushing hardest for regulation is also the one best positioned to absorb compliance costs. If auditing and enforcement regimes become the dominant regulatory mode, their effectiveness will depend on auditor quality and independence—criteria neither Anthropic nor its critics have yet defined publicly.