Anthropic is actively lobbying state legislatures for increasingly stringent AI safety regulations, pursuing a strategy that intentionally diverges from OpenAI’s efforts to establish uniform national standards. Cesar Fernandez, Anthropic’s head of U.S. state and local government relations, stated on July 15, 2026, that the company seeks to raise safety requirements for the most advanced AI systems rather than settle for consistent, baseline policies. This approach contrasts with the 'reverse federalism' strategy championed by OpenAI, which aims to bypass federal gridlock by encouraging states to mirror standardized bills nationwide.

The competition between the two firms intensified following California's 2025 law regulating advanced AI models. Anthropic was the only leading AI lab to formally endorse that legislation, while OpenAI declined to take a position before its passage. Anthropic has since used the California framework as a foundation to advocate for even tougher mandates in New York, Illinois, and Massachusetts. In late June, Anthropic endorsed a Massachusetts economic development bond bill that includes provisions requiring companies to hire independent evaluators for assessing catastrophic risks, such as potential assistance in bioweapons development, and grants the state attorney general enforcement authority.

OpenAI’s approach remains centered on creating common, enforceable safeguards that allow developers to allocate resources more efficiently, according to company spokesperson Liz Bourgeois. Despite their diverging strategies, the two organizations both supported a recent Illinois measure signed into law by Governor JB Pritzker this month, which mandates annual independent third-party audits of AI safety plans. The ongoing regulatory battle is also playing out through financial influence, with both companies associated with independent super PAC networks that have invested tens of millions of dollars into political campaigns. Anthropic additionally began making direct campaign contributions to California legislators in June, emphasizing support for candidates whose views on safety align with its internal mission.

The debate underscores the growing industry reliance on state-level regulatory clarity while Congress and the White House remain divided on federal policy. Anthropic’s push for higher safety bars stems from internal testing, such as its Claude Mythos model, which demonstrated capabilities to exploit vulnerabilities in major operating systems. These cybersecurity risks previously led the Trump administration to impose export controls on the technology until safety concerns were addressed. With the industry split, the question of whether state legislators will harmonize around a single framework or adopt disparate, escalating safety standards is expected to shape the future of U.S. AI governance.

Anthropic is actively lobbying state legislatures for increasingly stringent AI safety regulations, pursuing a strategy that intentionally diverges from OpenAI’s efforts to establish uniform national standards. Cesar Fernandez, Anthropic’s head of U.S. state and local government relations, stated on July 15, 2026, that the company seeks to raise safety requirements for the most advanced AI systems rather than settle for consistent, baseline policies. This approach contrasts with the 'reverse federalism' strategy championed by OpenAI, which aims to bypass federal gridlock by encouraging states to mirror standardized bills nationwide.

The competition between the two firms intensified following California's 2025 law regulating advanced AI models. Anthropic was the only leading AI lab to formally endorse that legislation, while OpenAI declined to take a position before its passage. Anthropic has since used the California framework as a foundation to advocate for even tougher mandates in New York, Illinois, and Massachusetts. In late June, Anthropic endorsed a Massachusetts economic development bond bill that includes provisions requiring companies to hire independent evaluators for assessing catastrophic risks, such as potential assistance in bioweapons development, and grants the state attorney general enforcement authority.

OpenAI’s approach remains centered on creating common, enforceable safeguards that allow developers to allocate resources more efficiently, according to company spokesperson Liz Bourgeois. Despite their diverging strategies, the two organizations both supported a recent Illinois measure signed into law by Governor JB Pritzker this month, which mandates annual independent third-party audits of AI safety plans. The ongoing regulatory battle is also playing out through financial influence, with both companies associated with independent super PAC networks that have invested tens of millions of dollars into political campaigns. Anthropic additionally began making direct campaign contributions to California legislators in June, emphasizing support for candidates whose views on safety align with its internal mission.

The debate underscores the growing industry reliance on state-level regulatory clarity while Congress and the White House remain divided on federal policy. Anthropic’s push for higher safety bars stems from internal testing, such as its Claude Mythos model, which demonstrated capabilities to exploit vulnerabilities in major operating systems. These cybersecurity risks previously led the Trump administration to impose export controls on the technology until safety concerns were addressed. With the industry split, the question of whether state legislators will harmonize around a single framework or adopt disparate, escalating safety standards is expected to shape the future of U.S. AI governance.