Executive Order N-5-26 requires AI vendors seeking California state contracts to demonstrate safeguards against algorithmic bias, civil rights violations, and illegal content — and gives the state authority to override federal blacklists of AI companies.
Governor Gavin Newsom signed Executive Order N-5-26 on Monday, March 30, establishing new requirements for artificial intelligence companies seeking to do business with California’s state government. The order arrived against a charged backdrop: a federal judge in San Francisco had just days earlier temporarily blocked the Pentagon’s blacklisting of Bay Area AI company Anthropic, and President Trump had publicly warned states against regulating AI. Newsom moved in the opposite direction.
Signed at a ceremony in San Francisco, the order uses California’s procurement power — rather than waiting for legislation — to set AI governance standards with a 120-day clock for implementation.
What the Order Requires
The order directs the Department of General Services and the California Department of Technology to develop new vendor certifications within 120 days. AI vendors selling to the California state government must prove they have safeguards against algorithmic bias, civil rights violations, and illegal content, or risk being barred from state contracts.
The categories of misuse that vendors must address through written attestations include the distribution of illegal content — including child sexual abuse material — AI models that embed discriminatory bias, violations of civil rights and free speech protections, and threats to user privacy and data security. Unlike broad legislative declarations about responsible AI, the order converts those concerns into a form, a contract, and a decision gate. Vendors that cannot document their safeguards will not be eligible for state business.
The order will also enable the state to separate its procurement authorization process from the federal government’s if needed, and direct the state to leverage AI to improve government service delivery, increase transparency, and strengthen accountability.
That separation provision is significant. It means California can maintain or expand its use of AI vendors even if the federal government has placed them on a supply-chain risk list — a direct response to Washington’s recent actions against San Francisco-based Anthropic.
Where the California Chief Information Security Officer concludes a federal ban is improper, the Department of General Services and the California Department of Technology must jointly issue guidance ensuring that departments and agencies can continue to easily procure from that company.
Watermarking AI-Generated Content — A First in the Nation
One provision of the order has drawn particular attention from policy analysts and communications professionals: the Governor directs the California Department of Technology to create recommendations and best practices for watermarking AI-generated images or manipulated video consistent with state law — described as the first of its kind nationwide.
The watermarking guidance is intended to address misinformation risks posed by synthetic media, and builds on existing California law covering AI-generated content in political advertising and other high-stakes contexts. When finalized within the 120-day window, the guidance will apply to state agencies and create a model that other jurisdictions — and private sector companies — are expected to reference.
The Anthropic Backdrop
The executive order did not emerge in a vacuum. In late February, Defense Secretary Pete Hegseth labeled San Francisco-based Anthropic a supply-chain risk to national security after the company refused to allow its Claude AI model to be used for autonomous weapons or domestic mass surveillance. The designation — a label previously used only against companies linked to foreign adversaries — barred government contractors from doing business with Anthropic, and President Trump subsequently ordered all federal agencies to sever ties with the company.
Anthropic argued that the government’s actions had already cost the company $180 million in collapsed deals, with over 100 enterprise customers expressing confusion and doubt following the supply chain risk designation.
On March 27, U.S. District Judge Rita Lin, serving on the federal bench in the Northern District of California, issued a 43-page ruling temporarily blocking the Pentagon’s designation. Lin wrote that “nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government.” She added that the measures appeared designed to punish Anthropic and constituted “classic First Amendment retaliation.”
Newsom signed his executive order three days later. While the governor made no specific mention of Anthropic in the order’s text, the CISO override mechanism — allowing California to independently assess and continue procuring from a company the federal government has flagged — reads as a direct architectural response to the federal-state friction the case exposed.
Newsom said California’s commitment is to stop bad actors who want to exploit people’s data, undermine security, and violate civil rights. “California leads in AI, and we’re going to use every tool we have to ensure companies protect people’s rights, not exploit them or put them in harm’s way,” he said in a statement. “While others in Washington are designing policy and creating contracts in the shadow of misuse, we’re focused on doing this the right way.”
California’s AI Footprint and What This Signals
The market weight behind this executive order is substantial. California is home to 33 of the world’s top 50 privately held AI companies and accounts for 25% of U.S. AI patents, according to Newsom’s office. When California sets procurement standards, it reaches not just a state bureaucracy but one of the deepest concentrations of AI talent, infrastructure, and capital anywhere in the world.
The order builds on prior legislative action. In September 2025, Newsom signed the Transparency in Frontier AI Act, which required large frontier AI developers to publish safety frameworks and report critical safety incidents, with fines of up to $1 million per violation. More than 20 California AI statutes covering employment, healthcare, education, and pricing algorithms took effect in January 2026. Executive Order N-5-26 extends that framework into the state’s buying power, adding procurement as a third enforcement vector alongside legislation and regulatory oversight.
The order also calls for expanding the Engaged California program. Launched as a tool to help the state communicate with communities affected by the 2025 Los Angeles wildfires, the platform is now being deployed statewide to solicit input from all Californians on how the state should respond to AI’s impact on the workforce. The Engaged California rollout is expected in the coming months and represents one of the first large-scale government efforts to use a digital democracy platform to shape AI workforce policy.
The Federal-State Fault Line Deepens
The executive order makes explicit what California’s posture has been since at least early 2025: the state does not intend to wait for federal consensus on AI governance, and it is prepared to diverge from Washington’s approach when it judges that approach to be harmful or inadequate.
President Trump revoked former President Biden’s 2023 executive order on the safe and trustworthy development of AI in January 2025, eliminating mandatory red-teaming for high-risk models, structured oversight for AI in critical infrastructure, and safety reporting requirements for frontier developers. California has steadily moved in the opposite direction in the period since.
The 120-day clock on vendor certifications means the new procurement architecture will take formal shape before summer. For AI companies with California state contracts — or those seeking them — the practical implications are immediate: document your safeguards, prepare for attestation, and expect California to conduct its own assessment independent of whatever Washington says about your company.
Newsom pointed to California’s position as the world’s fourth-largest economy in framing the stakes. “California’s always been the birthplace of innovation. But we also understand the flip side: In the wrong hands, innovation can be misused in ways that put people at risk.”
That framing — California as both the home of AI and the regulator of its risks — is not new. But the executive order gives it institutional weight it did not previously carry. The state is no longer simply hosting the AI industry. It is now conditioning the terms on which that industry does business with the government of the state that built it.
