[co-author: Annabelle Darby]
On July 23, 2025, the White House released “Winning the Race: America’s AI Action Plan” (Action Plan), which describes itself as a strategic roadmap intended to foster US leadership in AI and “articulate policy recommendations that [the] Administration can deliver for the American people to achieve the President’s vision of global AI dominance.”
Originally foreshadowed in one of the Trump Administration’s first Executive Orders, “Removing Barriers to American Leadership in Artificial Intelligence,” the Action Plan has undergone a comprehensive process prior to its release, including consultation with industry and requests for public comment.
The Action Plan is framed as an executive branch initiative and outlines actions that are considered realistically achievable over the remainder of the current administration’s term. While the Action Plan does not introduce new statutory obligations, it signals a substantial policy shift within the US that is expected to shape the regulatory environment, federal procurement, and international AI diplomacy for the foreseeable future.
Modeled on familiar US industrial policy playbooks, the Action Plan is structured around three pillars:
- “Accelerate AI Innovation”
- “Build American AI Infrastructure”
- “Lead in International AI Diplomacy and Security”
We outline several key provisions of the Action Plan below and consider next steps for organizations seeking to align existing AI governance programs with the new federal direction.
Pillar I: “Accelerate AI Innovation”
The most immediately actionable component of the Action Plan, Pillar I is centered on rapidly advancing AI innovation by removing regulatory barriers, promoting open-source and open-weight models, and accelerating adoption across both the private and public sectors.
The federal government aims to act as catalyst and customer, fostering a permissive environment for AI development while prioritizing values such as free speech, security, and workforce empowerment. It aims to do so by removing layers of “bureaucratic red tape” found in existing AI policy. This includes leveraging federal AI fundings against states that have enacted “burdensome AI regulations that waste these funds” by redirecting it to more “innovation-friendly” states.
Beyond encouraging deregulation, the Action Plan outlines several other strategies to accelerate innovation and encourage adoption of AI. The Action Plan emphasizes the US’s support of next-generation manufacturing, citing its importance in robotics, drones, self-driving cars, and other advanced technologies that could bolster the US’s position as a world-leading jurisdiction for AI. The Plan notes that funding and investment of these technologies will be prioritized, with the goal of ushering in a “new industrial renaissance.”
The Action Plan aims to further promote AI innovation by redirecting federal resources toward the establishment of a comprehensive scientific network for AI development and advancement in the US. Revisions to existing risk management frameworks provided by the National Institute of Standards and Technology (NIST) and the launching of a National AI Research Resource pilot are expected to follow, with the goal of accelerating maturation of compute markets and increasing access to key resources (e.g., data, models, and software).
Pillar I encourages increased adoption of AI within the federal government, with a specific focus on the unique needs of the Department of Defense (DOD). It also outlines an AI evaluations ecosystem, recommending guidelines to be published through NIST and the development of AI testbeds and regulatory sandboxes around the country.
Pillar II: “Build American AI Infrastructure”
Pillar II emphasizes that US leadership in AI depends on robust physical infrastructure – computing power, energy, secure facilities, and a skilled workforce. It underscores the importance of developing the energy capabilities necessary to keep pace with AI innovation.
Specifically, the Action Plan outlines steps to streamline permitting for data centers and semiconductor manufacturing, modernize the power grid to support energy-intensive AI applications, and revitalize domestic chip production.
Efforts also include deploying high-security environments for sensitive government data and establishing technical standards to protect against nation-state threats. There are recommendations to advance the use of AI to bolster cybersecurity efforts, particularly concerning critical infrastructure protection.
To support this infrastructure, the Action Plan launches a national skills initiative to train workers for critical roles and strengthens cybersecurity and incident response capabilities. In doing so, Pillar II aims to ensure the US establishes the foundational assets and talent needed to sustain and scale an AI-driven economy.
Pillar III: “Lead in International AI Diplomacy and Security”
Pillar III outlines several actions that it claims will “drive adoption of American AI systems, computing hardware, and standards throughout the world,” while also striking a balance between vigorous export promotion and tough export controls.
Most notably, the Plan establishes a new program at the Department of Commerce aimed at gathering proposals from industry consortia for full-stack export packages. Once approved by the Department, various government agencies will work to facilitate deals abroad.
The Plan also sets forth robust export controls and promotion of American AI standards. Notably, it recommends “plug[ging] loopholes” in existing semiconductor controls, particularly in relation to component subsystems. More broadly, the Plan focuses on strengthening enforcement of existing export controls, working to ensure that the export controls regimes of foreign partners are aligned with those of the US, and countering Chinese influence in international governance bodies in relation to AI.
Finally, this pillar sets forth a plan – to be led by the Department of Commerce’s Center for AI Standards and Innovation – to evaluate the national security risks in frontier models, with an emphasis on chemical, biological, radiological, nuclear, and cyber threats that emanate from malign foreign influence. In relation to biosecurity, all institutions receiving federal funding for scientific research will also be required to use nucleic acid synthesis tools and synthesis providers that have robust nucleic acid sequence screening and customer verification procedures.
Companion Executive Orders
The Action Plan should be read alongside three complementary Executive Orders that the Trump Administration has positioned as the keystones of its broader AI-policy framework. Individually, each EO advances a distinct strategic objective; taken together, they provide important context for both the scope and the likely pace of future regulatory and commercial activity of AI in the US.
- “Declaring a National Energy Emergency” (January 20, 2025) addresses essential AI infrastructure, including Department of Energy (DOE) data-center initiatives, energy, and permitting hurdles the industry currently faces.
- “Establishing the US Investment Accelerator” (March 31, 2025) seeks to invigorate foreign adoption of US AI technologies and compete against foreign technologies.
- Modeled on prior social media bias orders, “Removing Barriers to American Leadership in AI” (January 23, 2025) mandates that AI procured by the federal government be “neutral and not biased,” potentially reshaping technical evaluation criteria for vendors of foundation models.
Key takeaways
- A pro-innovation regulatory posture: Unlike the prior administration’s approach, the Action Plan’s touchstone is acceleration, and regulatory obligations will be recalibrated where they are perceived to slow or impede deployment.
- Conditional federal funding as a pre-emption tool: By tying funds to “innovation-friendly” state environments, the administration could indirectly harmonize an otherwise fragmented state landscape without invoking formal pre-emption.
- Infrastructure bottlenecks in the crosshairs: The combination of NEPA streamlining, DOE-hosted data centers, and grid-modernization initiatives signals recognition that compute and power availability are strategic assets on par with intellectual property.
- Heightened scrutiny of model “neutrality”: Vendors seeking federal contracts for large language models likely should anticipate detailed questionnaires or third-party audits concerning content filtration, reinforcement learning procedures, and dataset curation.
- An expanding international footprint: The Plan’s export focus frames the US AI stack as a turnkey solution for allied countries, creating new market opportunities while heightening compliance obligations under US export control regimes.
Planning ahead
As part of its wider Action Plan rollout strategy, the White House is expected to follow the Plan’s publication with a promotional campaign, featuring Director Michael Kratsios, Deputy Director Lynne Parker, Senior Policy Advisor Dean Ball, and others, to amplify the Plan’s goals nationwide.
While the Action Plan stops short of statutory reform, it signals the Trump Administration’s intention to prioritize speed, scale, and strategic advantage in AI.
As many of the Plan’s commitments will be operationalized through agency guidance and procurement policies, often on compressed timelines, it is unclear how they will translate into companies’ compliance requirements. Organizations are encouraged to consider proactively establishing internal AI governance that is framed against industry best practice in anticipation of further developments to the US AI regulatory landscape.
[View source.]
