The White House has scrapped a landmark AI executive order that would have required tech companies to submit their most powerful AI models to federal agencies for safety review before public release. The order was pulled hours before it was due to be signed, following last-minute calls to President Trump from Elon Musk, Mark Zuckerberg, and former White House AI advisor David Sacks.

The decision marks a sharp turn in US AI policy and sends a clear signal: for the foreseeable future, American businesses deploying AI systems will not face federal pre-release safety requirements.

What the Order Would Have Done

The draft executive order had two key provisions. First, it would have given federal agencies up to 90 days of access to the most powerful AI models before their public launch — enough time to assess security risks, test for vulnerabilities, and flag concerns. Second, it framed participation as voluntary but signaled that government procurement would favor companies that cooperated.

Tech companies pushed back hard. Industry lobbyists argued the 90-day window was commercially unworkable and proposed cutting it to 14 days. Behind the scenes, Musk, Zuckerberg, and Sacks reportedly made direct calls to the President, arguing the review process would slow AI development and hand competitive ground to China.

The order was pulled without explanation. Musk, via social media, denied being the one who killed it — though he was among those who made calls before the decision.

The Bigger Picture

The collapse of this executive order is the latest sign that Washington cannot agree on even basic guardrails for AI development. With no federal framework emerging, the United States now sits well behind Europe and parts of Asia on AI governance.

The EU AI Act is fully applicable by August 2026, with binding requirements on high-risk AI systems, transparency obligations, and prohibited practices already in effect. Colorado’s comprehensive AI legislation takes effect June 30, 2026. California has its own frameworks moving through the system.

What this means in practice: businesses operating across jurisdictions will face a patchwork of requirements rather than a single federal standard. Companies with EU exposure still need to comply with the AI Act. Companies in Colorado and California will face state-level rules. The gap between US and European approaches is widening.

What This Means for Business

For enterprise teams deploying AI agents, automation tools, or AI-driven products, the scrapping of this order has a few concrete implications.

Short term, there is less friction. Without a pre-release federal review requirement, AI vendors can ship faster. Tools like AI agents, voice automation systems, and custom AI applications will not face a mandatory federal clearance process. That is good news for companies that want to move quickly.

Medium term, the risk is fragmentation. Without a coherent federal framework, businesses with any international exposure will need to navigate multiple regulatory regimes simultaneously. EU customers and partners require AI Act compliance. State laws add another layer. Legal and compliance teams will carry a growing burden that was supposed to be simplified by a national standard.

The deeper issue is trust. Safety reviews are not just a government concern. Enterprise buyers evaluating AI vendors increasingly ask about internal red-teaming, third-party audits, and model governance. The absence of federal validation shifts that burden onto procurement teams. Expect AI vendors to invest more in self-certified safety documentation to fill the credibility gap left by the lack of government oversight.

For businesses considering AI adoption, the message is practical: the regulatory environment in the US remains permissive, but do not assume that permissive means permanent. The EU AI Act and state-level laws are real, and the compliance story will only get more complex before it gets simpler.

The Enterprise DNA Perspective

At Enterprise DNA, we help businesses implement AI responsibly — not because regulators require it, but because it produces better outcomes. AI agents that are well-governed, transparent, and properly tested perform better than ones that are rushed out the door.

The absence of federal oversight does not change the fundamentals of good AI deployment. Whether you are building an AI agent workforce, implementing voice AI for customer interactions, or using AI to automate back-office workflows, the core principles remain the same: know what your systems are doing, test before you deploy, and keep humans in the loop for high-stakes decisions.

If you want the playbook other teams are using with Claude and Codex right now, grab the free Working With Claude field guide. Download it here.