New chapters in US AI policy

July 23, 2025, White House Announced AI Action Plan (Planning), an important policy document that outlines the current administration's priorities and deliverables in artificial intelligence. The plan appeared after the White House received it 10,000 public comments (RFI) in response to information requests for February 2025. Cloudflare's comment It pioneered the White House the conditions for US leadership in AI and encouraged it to support open source AI.

It's packed with lots of three pillars, 28 pages of plans.

• Pillar I: Accelerate AI innovation. It focuses on removing regulations, the possibility of adopting and developing AI, and ensuring availability of open source and openweight AI models.

• Pillar II: Build American AI infrastructure. We prioritize building high-security data centers, enhancing critical infrastructure cybersecurity, and promoting AI technology by design.

• Pillar III: International AI diplomacy and security leads. It focuses on providing access to AI to American allies and partners and strengthening AI calculation export control enforcement.

Each of these pillars outlines recommendations from various federal agencies' policy to advance the comprehensive goals of the plan. There are many things that will get your plans right. Below we cover some of the plans I think are particularly important.

The plan stands as a global competition for the US to achieve AI domination and that it is a national priority for US technology companies to become the gold standard for AI worldwide. Through the plan, President Trump commits his administration to support American workers, technology and energy to achieve that goal.

The government shares its view that it plays a useful role in shaping rules and regulations that allow private sector innovation to flourish. For CloudFlare's network to continue operating globally, the US government must shape and influence appropriate regulatory conditions. They need to balance national and economic security concerns, promote consensus industry-driven international standards, and support interoperable regulatory regimes.

In recent years, policy developments have been observed that have unnecessarily increased restrictions on U.S. technology providers, making it difficult to operate. Protectionist obligations, including data sovereignty requirements, customer data retention policies, various oversight and government access requirements, do little to improve security and innovation, and have unintended consequences. Protectionism increases the costs of businesses, limits access to world-class technology, and increases cybersecurity risks.

Implementing policies that ensure access to globally distributed edge computing networks and the freedom to choose the technology that best suits your needs will help ensure that AI is in the right conditions to thrive.

The plan supports open source and open weight AI models, promotes innovation and benefits commercial and government adoption. The plan recommends ensuring access to computing resources to enhance capabilities in the startup and academic world.

CloudFlare shares the view that open source AI models play an important role in driving innovation. As recognized in the plan, these models provide flexibility for businesses, freeing them from reliance on closed providers, allowing AI to be used with sensitive data that is not possible to export to closed models. So CloudFlare includes access to over 50 open source models. Worker AI Model Catalog.

However, accessing only open source models is not sufficient to exploit the possibilities of AI. A complete ecosystem is needed to build and deploy AI applications and tools that will mark the arrival of a new era imagined in your plans. GPU-powered inference allows CloudFlare's global network to play a key role. Having a distributed network like ours that enables AI inference with EDGE is critical for building fast and efficient AI development and next-generation AI applications.

The open ecosystem is deeply embedded in CloudFlare's DNA. Our developer platform provides powerful tools to democratize access and build and deploy applications for anyone. It provides a global network infrastructure that removes complexity and reduces barriers. This allows AI developers to innovate freely using many different AI models without relying on gatekeepers. Our commitment to making these tools easy to use reflects the plan's call to promote innovation and support AI leadership by enabling developers to globally build, deploy and scale new AI applications using open source AI models.

The plan highlights the importance of AI cybersecurity in several ways. There are two things I would like to emphasize.

First, we recommend using AI technology for cybersecurity of critical infrastructure. Using AI-assisted cyber defense tools is a forced multiplier for network defenders, and is absolutely necessary for all organizations (particularly critical infrastructure) to protect against cyber threats.

CloudFlare's network uses predictive AI and machine learning to block 247 billion cyberattacks every day. Under the theory of Defensive AICloudFlare uses information to constantly improve the effectiveness of security solutions. and ai labyrinthcreated a new tool to trap AI using AI. This is a new next-generation honeypot and cybersecurity defense tool that will use AI to disrupt crawlers and bots that ignore “no crawl” directives. Instead of blocking these bots, Ai Labyrinth directs the bots into an infinite maze of pages generated by compelling AI.

Second, to address the potential vulnerabilities of AI technology, the plan imposes ensuring that the US government is safe.

To ensure AI, CloudFlare has been proactive and proactive in shaping cybersecurity and risk management for AI technology. We supported and provided feedback to the National Institute of Standards' efforts to develop cybersecurity profiles for artificial intelligence. This is very important and based on us Safety with each design dedication.

We look forward to working with the administration on the proposed AI Information Sharing Analytics Center and the proposed exchange of vulnerability information.

The plan envisages the federal government to play a key role in accelerating the adoption of AI. CloudFlare is useful. As the plan points out, integrating AI can significantly enhance public services and make governments more efficient and effective. Most, if not all, federal agencies have AI chief executives, demonstrating a clear commitment to this technological change. Governments can promote their efforts by training their workforce to promote information sharing between agencies, promote best practices, and maximize the efficiency of AI.

CloudFlare can be an important partner in this journey. Our platform provides the secure, reliable and scalable infrastructure federal agencies need to deploy AI applications using full-stack AI building blocks. CloudFlare is FedRamp medium approvedand we are committed to FedRamp High. By leveraging CloudFlare's global network, federal agencies can ensure that AI initiatives are resilient, accessible and promote greater public interest.

To lead AI internationally, the plan outlines a dual strategy and presents two tensions approaches: aggressive AI exports to allies and partners, and strict restrictions on AI calculations and semiconductor exports. Meanwhile, the plan emphasizes the importance of providing a complete US AI technology stack to prevent allies from becoming rivals. It aims to solidify the global AI alliance and ensure the lasting spread of American technology.

Conversely, the plan calls for strengthening export control enforcement and cramming loopholes in order to prevent the export of sensitive technologies. The administration is attempting to use export restrictions (limitations on the goods that companies can export) to deny access to certain resources both in geostrategic competition and national security concerns. The challenge arises as excessively strict export controls can make it inadvertently difficult to export AI even to allies while aiming to deny access to attackers.

This dual approach highlights an important tightrope walking. CloudFlare, along with many other industry players, watch carefully to see how management balances these competing goals. It is consistent with our mission to provide individuals around the world with access to resources that can innovate and build applications close to their end users, in line with their mission to build a better, more connected internet. Having a globally distributed network like ours also enables US AI companies to deploy services globally. Although we appreciate the need to limit access to sensitive computing resources, excessively widespread or inaccurate controls can inadvertently suppress innovation and prevent the open exchange of ideas essential to AI development. The implementation of export controls requires meticulous balance to effectively target the enemy without hampering innovation and safely hindering the global digital ecosystem that is trying to protect.

A encouraging aspect of the planning is a clear recognition of the integral role of the private sector. The document repeatedly highlights the need for collaboration with the industry and the need for consultation with major technology companies across a range of recommended policy measures. For example, we are specifically calling for the establishment of a program within the Department of Commerce to gather proposals from the industry consortium of AI export packages. Additionally, we recommend investigating new measures “in collaboration with the industry” to enhance AI's calculation export control enforcement. This commitment to partnerships is essential to navigate the complexities of AI development and deployment. Collaboration with this industry will allow policies to be technically feasible and effective worldwide, avoiding unexpected negative impacts on the digital economy and cybersecurity.

This plan represents a critical moment for American AI leadership, and CloudFlare is ready to be a partner in shaping the future of this critical technology. We praise the focus on accelerating AI development, building robust infrastructure, and building global global diplomacy. The global nature of the Internet means that achieving these goals requires a delicate balance, especially as the business model of AI-powered webs is rapidly evolving. CloudFlare defends an approach that promotes innovation while advocating an open, secure, interoperable internet. By prioritizing consensus-driven standards and ensuring regulations do not accidentally create barriers to globally distributed AI infrastructure, it helps ensure a continuous US technology leadership and a sustainable and beneficial AI ecosystem.