Without Federal Preemption, California Democrats Will Hand China The AI Race On A Silver Platter

The United States is engaged in the most consequential technological competition in its history, a race with China to develop and deploy artificial intelligence at scale, and it is on the verge of losing that race not to Chinese ingenuity, not to Chinese investment, and not to Chinese espionage, but to the California state legislature. More than 400 AI-related bills are currently moving through statehouses across the country, with California Democrats leading the charge toward the most expansive and burdensome regulatory frameworks on the table. The arithmetic of corporate compliance means that the most restrictive jurisdiction in any fragmented regulatory landscape effectively sets national policy by default, because no company can afford to build different products for 50 different legal regimes. That dynamic threatens to make Sacramento the de facto AI regulator for 330 million Americans, while Beijing looks on with undisguised satisfaction. The Trump Administration’s National Policy Framework for Artificial Intelligence, developed by AI and Crypto Czar David Sacks and released this week, is the intervention that can stop that from happening, but only if Congress acts.

To understand why the preemption problem is so acute, consider what California did with automobile emissions. Beginning in the 1970s, California leveraged a unique Clean Air Act waiver to set vehicle emission standards more stringent than federal ones. Because automakers could not profitably manufacture one fleet for California and another for the other 49 states, they built to California’s specifications everywhere. A state containing roughly 12% of the US population became the emissions regulator for 100% of US consumers, without any democratic mandate from the other 49 states. California Democrats ran the table on American automotive policy for decades, and the rest of the country had no meaningful recourse. Now replace “emissions standards” with “AI regulation” and the same dynamic is unfolding in real time, with far higher stakes. Colorado passed a sweeping AI liability bill in 2024 imposing extensive obligations on developers of “high-risk” AI systems, so aggressive that the governor sent it back for revision. California is actively advancing proposals that would extend liability chains to model developers themselves for downstream third-party misuse. The regulatory Cambrian explosion is already underway, with California Democrats at the leading edge of it, and Beijing is watching.

The mechanics are worth making explicit. A company developing an AI model does not build a different model for Colorado and a different one for California and a different one for Texas. The rational response to a fragmented regulatory landscape is to comply with the most demanding jurisdiction everywhere, because that single compliance regime eliminates the need to track 50 different legal frameworks simultaneously. The result is that the most aggressive state legislature sets the de facto national standard, without any mandate from the other 49 states. Right now, that legislature is California’s, controlled by Democrats who have shown no reluctance to impose maximally burdensome requirements on the technology industry even as that industry represents America’s primary strategic asset in the competition with China. This is precisely the GDPR scenario the Sacks framework moves to quarantine. When the EU enacted its sweeping data regulation in 2018, most major companies adopted compliance globally rather than geofencing it, because uniform compliance was cheaper than differentiation. China’s AI developers faced no such constraint. Baidu, Huawei, and the constellation of firms operating under Chinese Communist Party direction were free to develop, iterate, and scale without the friction that democratic regulatory systems impose. Without congressional action on the Sacks framework, a fragmented American state AI regulatory landscape will reproduce that dynamic domestically, handing Chinese developers a structural competitive advantage that no amount of American ingenuity can fully overcome.

The Sacks framework is surgically designed to address this problem without trampling on legitimate state interests. It explicitly preserves state authority over traditional police powers, including laws protecting children, preventing fraud, and protecting consumers. States retain zoning authority over AI infrastructure placement and full authority over their own government use of AI. What states would be prohibited from doing is regulating AI development itself, on the grounds that it is an inherently interstate phenomenon with direct foreign policy and national security implications. States would also be prohibited from penalizing AI developers for a third party’s unlawful conduct involving their models, a principle with an obvious analogy in existing law: we do not hold Ford liable when a drunk driver kills someone with a Ford. The line is drawn precisely at the point where state regulation would compromise America’s ability to win the AI race, and not one inch further.

The child protection provisions illustrate how carefully that line is drawn. The framework explicitly carves out state child protection laws from preemption, including prohibitions on AI-generated child sexual abuse material. It builds on the Take It Down Act, championed by First Lady Melania Trump, which extended federal protections against non-consensual intimate imagery to AI-generated deepfakes. Congress is called upon to require age-assurance mechanisms for platforms likely accessed by minors and to mandate features reducing exploitation and self-harm risks. California Democrats have routinely invoked child safety as justification for sweeping AI regulation. The Sacks framework answers that concern directly and comprehensively, then draws the line clearly against using child safety as a pretext for the broader regulatory agenda that would cripple American competitiveness.

The energy provisions carry particular relevance for the China competition. China is building AI data center infrastructure at a pace Western observers have consistently underestimated, partly because Chinese developers face no permitting delays of the kind that routinely add years to US infrastructure projects. The framework’s answer is elegant: streamline permitting so AI developers can generate their own power on-site and behind the meter, reducing grid demand rather than increasing it. Ratepayers are protected not by taxing the data center but by requiring it to feed itself. The infrastructure gets built, rather than waiting years for environmental reviews that Chinese competitors are not subject to. California has historically been among the most permitting-hostile jurisdictions in the country, causing multiple major technology infrastructure projects to route around the state entirely. Without the Sacks framework becoming law, California retains the practical ability to veto the infrastructure buildout that American AI dominance requires.

The anti-censorship provisions address a threat that is simultaneously domestic and geopolitical. The framework calls on Congress to prohibit the federal government from coercing AI providers into altering content based on partisan or ideological agendas, and to create a private right of action for Americans who suffer from such government-directed censorship. One of the central claims of Chinese Communist Party information strategy is that American democracy is hypocritical, that US technology companies censor political expression on behalf of the government while publicly claiming to champion free speech. That claim draws credibility from documented instances of federal agencies pressuring social media platforms to suppress disfavored content. A large language model trained or prompted to favor particular political perspectives does not merely amplify one viewpoint, it presents that viewpoint as the output of a neutral, authoritative intelligence. Codifying First Amendment protections against AI censorship does not merely protect American conservatives. It deprives China of one of its most effective rhetorical weapons in the global competition for democratic legitimacy.

All of these provisions together constitute a coherent national policy that no individual state can produce acting alone, and that the most activist states are actively undermining right now. This is not primarily a federalism debate. It is a national security debate. A patchwork of 50 state AI regimes will not merely burden American companies with compliance costs. It will create a regulatory environment so complex and unpredictable that the decisive advantages in the global AI race migrate to jurisdictions, most notably China, that face no such friction. Beijing does not need to outinvent Silicon Valley. It simply needs Sacramento to outregulate it. California Democrats, armed with the best intentions and no apparent awareness of the geopolitical consequences of their ambitions, are actively pursuing that outcome right now, in the current legislative session, with more bills in the pipeline. David Sacks and the Trump Administration have done their part. They produced a framework that is at once protective of American rights and ruthlessly focused on American victory. The question now is whether Congress will act on it before California and its Democratic allies finish the job that Beijing cannot do for itself.

If you enjoy my work, please subscribe: https://x.com/amuse.

Sponsored by the John Milton Freedom Foundation, a nonprofit dedicated to helping independent journalists overcome formidable challenges in today’s media landscape and bring crucial stories to you.

READ NEXT: Iconic Movie Star Dies At 86, Leaves Lasting Hollywood Legacy