An event last week drew attention from many tech media outlets, but the full layers of the story may not have been completely explained.
The Trump administration designated Anthropic as a supply chain risk — effectively kicking the company out of the U.S. military contracting system.
The reason? Anthropic refused to sign a contract that required them to remove safety guardrails, for uses that included domestic mass surveillance and fully autonomous weapons systems.
What Anthropic Said
Anthropic's position was straightforward: We have safety use terms for our AI systems, and these two scenarios are not within permitted use, so we will not sign this contract.
A federal court then intervened, temporarily blocking this supply chain risk designation — the judge found issues with the policy decision.
But a temporary block is not a resolution; the legal battle continues.
Newsom's Countermove
Against this backdrop, California Governor Newsom signed Executive Order N-5-26 on March 30.
On the surface, the order is about standards for California's government procurement of AI products, but in reality, Newsom is using procurement power as a policy tool to push back against the federal government's AI policy direction.
Specific provisions:
- AI product vendors must certify to California that their systems will not be used for illegal content distribution, discriminatory bias, or infringement of civil rights
- Requires the California Department of Technology to issue guidelines for AI-generated content watermarking — a first in the U.S.
- Requires state agencies to provide employees with vetted generative AI tools
California is the largest market for AI products and services in the U.S. Newsom said California's procurement standards can effectively become industry-wide benchmarks — because doing government business requires compliance.
The Broader Context of This Conflict
This is not an isolated incident.
In December 2025, Trump signed an executive order directing the U.S. Attorney General to establish a task force to challenge state AI-related laws, aiming to create a minimally restrictive national uniform standard.
On March 20, 2026, the White House released the National AI Policy Framework, further pushing for congressional legislation to preempt state AI regulations deemed to hinder national competitiveness.
Newsom's response: Fine, if the federal government won't regulate, I'll use procurement power to do it.
Legal experts predict this could ultimately reach the Supreme Court, potentially becoming one of the most consequential legal cases in U.S. tech history.
What This Means for the AI Industry
Short term: Anthropic is currently excluded from U.S. military AI procurement, but the situation remains fluid due to the court's temporary ruling.
Medium term: The divergence between federal and California AI regulatory paths is growing. Companies must navigate not one set of rules, but two — and they are moving in completely opposite directions.
Long term: If the Supreme Court ultimately rules that the federal government can preempt state AI laws, California's regulatory experiment will be severely constrained. If state authority is preserved, California's standards could indeed become the de facto national standard.
All of this is unfolding within 2026.
Sources: As Trump rolls back protections, Governor Newsom signs first-of-its-kind executive order to strengthen AI protections (California Governor's Office); California, not Trump, will decide how risky its AI startups are, Newsom says (CalMatters, via CocoLoop); Federal Preemption: President Trump Signs Landmark AI Executive Order to Dismantle State Regulations (FinancialContent)