The US government is arguing with itself over who gets to police the most powerful AI models on the planet. And the outcome could reshape how companies like OpenAI and Anthropic bring products to market.
The Trump administration is weighing an executive order that would significantly expand intelligence agencies’ role in evaluating frontier AI systems, a move driven by escalating cybersecurity concerns around advanced models. The catch: key parts of the administration can’t agree on who should actually be in charge.
The turf war
At the center of the dispute is a straightforward question with complicated answers: should the Commerce Department or the intelligence community take the lead on assessing whether cutting-edge AI models pose national security risks?
The Commerce Department has already started pre-deployment testing of AI models from major tech companies, including Google and Microsoft. That work is part of a broader AI governance push that’s been building throughout 2026.
But spy agencies want a bigger seat at the table, arguing that their expertise in threat assessment makes them better suited to evaluate models that could be weaponized for cyberattacks, disinformation campaigns, or other national security threats. In English: the people who track foreign adversaries think they should also be tracking the tools those adversaries might use.
The executive order could land as soon as May 13, 2026. If signed, it would formalize intelligence agencies’ expanded role and potentially require companies like Anthropic and OpenAI to submit their models for pre-release reviews. That’s a meaningful escalation from the current regime, where commercial AI development has operated with relatively light federal oversight.
Think of it like the difference between a restaurant getting a health inspection after opening versus needing government approval before it can serve its first meal. The stakes, and the delays, are fundamentally different.
Why crypto should be paying attention
On the surface, a fight over AI oversight might seem disconnected from digital assets. It’s not.
The same administration pushing for tighter AI controls has been quietly building an expansive surveillance apparatus for crypto markets. On April 28, 2026, the Department of Homeland Security announced a crypto intelligence network connecting 70 agencies with major platforms including Ripple and Coinbase. That network reportedly covers 85% of centralized crypto transaction volumes, enabling real-time tracking and, in some cases, the potential seizure of funds.
The pattern here is hard to miss. National security concerns are becoming the primary lens through which the administration views emerging technology, whether that’s large language models or blockchain protocols. And when national security gets invoked, deregulation tends to take a back seat.
This represents a notable evolution from the administration’s earlier posture. On January 23, 2025, Trump signed an executive order creating a digital assets working group and explicitly banning federal promotion of central bank digital currencies. That order was broadly interpreted as pro-industry, a signal that the government would foster innovation rather than constrain it.
Eighteen months later, the emphasis has shifted considerably. The digital assets working group still exists, but it’s now operating alongside a DHS surveillance network that can monitor the vast majority of centralized trading activity in real time. The vibe has changed.
What this means for investors
For market participants, the implications cut in two directions simultaneously.
On one hand, increased regulatory clarity around AI could benefit tokens and projects associated with secure AI infrastructure. If the government is going to mandate security reviews of AI models, the tools and platforms that facilitate compliance become more valuable. Projects building at the intersection of AI and blockchain, particularly those focused on verifiable computation or secure model deployment, could see increased interest as institutional capital looks for regulatory-friendly entry points.
On the other hand, the surveillance expansion poses real challenges for privacy-focused projects and decentralized platforms. A crypto intelligence network covering 85% of centralized transaction volumes creates strong incentives for users to move toward decentralized alternatives, but it also puts a target on those alternatives. When DHS can track most of what happens on Coinbase and Ripple, the next logical step is extending that capability to decentralized exchanges and privacy chains.
The broader risk is that national security priorities begin to crowd out the innovation-friendly rhetoric that characterized the administration’s early approach to both AI and crypto. Regulatory frameworks designed primarily around threat prevention tend to produce compliance burdens that favor large incumbents over smaller, more experimental projects. Google and Microsoft can absorb the cost of pre-deployment AI reviews. A three-person startup building an open-source model probably can’t.
The same dynamic applies in crypto. Coinbase and Ripple have the legal teams and compliance infrastructure to work within a DHS monitoring framework. A decentralized protocol with no corporate entity and no compliance department faces a fundamentally different calculus.
Investors should watch whether the executive order, if signed, includes any carve-outs for open-source AI development or sets thresholds below which models are exempt from review. Those details will determine whether this is a targeted intervention aimed at the most powerful systems or a broad expansion of federal oversight that touches everything from frontier models to the AI tools embedded in DeFi protocols. The distance between those two outcomes is enormous, and the administration hasn’t made clear which direction it’s heading.
