In the fast-moving world of artificial intelligence, Anthropic is playing a role that feels strikingly counterintuitive for a tech giant. While much of Silicon Valley has aggressively lobbied against government oversight—fearing that red tape will strangle innovation—Anthropic is actively inviting it. Last year, the company played a pivotal role in shaping and passing new transparency laws in California and New York. Now, they are signaling that those hard-won victories are already relics of the past. As Cesar Fernandez, Anthropic’s head of U.S. state and local government relations, recently told WIRED, the rapid evolution of AI technology means that mere self-reporting is no longer enough. The company is now championing more stringent oversight, arguing that if technology is moving at breakneck speed, our policy responses must accelerate to match it.
The motivation behind this stance is rooted in Anthropic’s unique corporate DNA. Valued at nearly $1 trillion, the company is undeniably a titan of industry, yet it remains tethered to a founding mission that prioritizes the “safe transition” to a world with transformative AI over pure market dominance. This philosophy has pushed them to support some of the most rigorous proposed regulations in the country. From Illinois to Massachusetts, Anthropic is backing measures that go far beyond simple transparency, advocating for mandatory third-party audits of safety protocols and empowering state attorneys general to hold AI developers accountable. For a company that stands to lose the most from overly restrictive rules, this proactive embrace of regulation is a deliberate, albeit unconventional, strategy.
To navigate this complex political landscape, Anthropic has recruited Cesar Fernandez, a seasoned veteran of high-stakes government relations who previously cut his teeth steering policy at companies like Uber and FanDuel. Fernandez’s expertise is arriving at a critical juncture; with Congress largely deadlocked on federal AI policy, individual states have stepped into the vacuum to set their own rules. By deploying someone who understands how to win legislative battles on a state-by-state basis, Anthropic is positioning itself as the primary architect of the new AI rulebook. This gives them a significant advantage: they aren’t just reacting to laws, they are effectively shaping the environment in which their own future technology will operate.
However, not everyone in Silicon Valley sees this as a virtuous quest for safety. A growing chorus of critics, including high-profile tech figures like David Sacks, views Anthropic’s behavior through a cynical lens known as “regulatory capture.” The argument suggests that by lobbying for complex and expensive compliance requirements, Anthropic is effectively building a moat around its business. If the cost of playing in the AI field becomes high enough due to mandatory audits and strict safety mandates, smaller, hungrier startups will be forced out of the market. To these critics, Anthropic isn’t trying to save the world; they are trying to cement their status as an incumbent by rigging the game against potential future competitors.
Fernandez and the leadership at Anthropic vehemently reject this “nefarious” characterization. They point out that the legislation they support is carefully calibrated to apply only to “large AI model developers”—specifically those with hundreds of millions of dollars in development costs and significant annual revenues. From their perspective, the threshold is high enough that it won’t affect the average startup just getting off the ground. They argue that the focus is entirely on the industry giants, of which they are a member, acknowledging that companies with the power to influence society at scale should be subject to the intense scrutiny that their own massive scale demands.
The reality, however, remains nuanced and arguably somewhat gray. While Anthropic’s thresholds might shield early-stage ventures, the line between an “up-and-coming startup” and a “large developer” is blurring, especially as billions of dollars in venture capital flood the sector. Companies like Mistral and Safe Superintelligence are already operating at a scale that could eventually bring them under the umbrella of these regulations. Whether these policies are truly about protecting humanity from “catastrophic risks” or simply about controlling the supply side of the AI market remains a point of heated debate. Ultimately, Anthropic is betting that by leading the regulatory conversation, they can define the safety standards of the future—and perhaps ensure that the only players left in the game are those willing to play by their, and the government’s, sophisticated rules.
