a16z’s Recommendations for the National AI Action Plan

This week, a16z shared our recommendations with the White House Office of Science and Technology Policy (OSTP) for how the United States can implement a competitiveness agenda that will enable it to continue to lead the world in AI development. Little Tech has an important role to play in strengthening America’s ability to compete in AI in the future, just as it has been a driving force of American technological innovation historically. We shared our comments as part of the OSTP’s request for information as it gathers policy ideas to inform and develop a new AI Action Plan.

AI has the promise to transform our world and lives for the better. AI-powered services could help doctors create new medicines and treatments to improve health outcomes, enhance how students learn in school and at home, improve transportation infrastructure, and the list goes on. As we’ve seen throughout history, from Edison and Ford to Hughes and Lockheed, startups are the vanguard of American technological leadership. This AI innovation cycle is no different. Consider, for instance, that more than 5,000 new AI startups in the United States were funded between 2013 and 2023 alone. And all of the recent, cutting edge products in AI have come from startups, not Big Tech incumbents.

a16z stands firmly on the side of Little Tech. It’s imperative that US policymakers create a regulatory framework that allows Little Tech to compete with larger companies with more resources and deeper pockets. If not, we run the risk of stagnation and surrendering America’s leadership position to other countries, particularly China. To that end, we proposed three policy pillars to guide the OSTP’s development of an AI policy agenda rooted in American competitiveness:

1. Adopt a policy framework that prioritizes American competitiveness, including establishing the federal government’s leadership role in regulating a national AI model market.

American startups have always been a principal driver of American competitiveness and we believe, will play a critical role in America’s future ability to compete in AI. The US AI Action Plan should support the ability of Little Tech to compete on a level playing field with better-resourced incumbent platforms.

AI models are crucial to America’s national security and economic competitiveness, its geopolitical objectives, and the overall welfare of our country. Because the AI development market is inherently a national one with potential significant impacts in commerce, national security, and foreign relations, the federal government–not individual states–must lead in promoting and regulating a national AI market. This approach will provide certainty for innovators looking to build products that serve people throughout the country. Of course, states also have an important role to play in AI governance by policing harmful conduct within their borders.

As the recent emergence of DeepSeek R1 showed, failing to prioritize competitiveness may slow American AI development and allow other nations to catch up. Policy decisions we make now will determine whether the most important AI technologies of the future are built in the United States, or by a foreign adversary like China. To outcompete other countries and win the AI race, the United States must recognize that entrepreneurship is the cornerstone of AI leadership and ensure Little Tech has a fair shot to build, compete, and thrive.

2. Regulate harms, not AI model development.

For decades, the United States’ approach to regulating technology has been based on how that technology is used–not how it’s made. For instance, there’s no law dictating how to build a computer. But the law does prohibit people from using that computer to hurt someone else.

This is the same approach lawmakers should use in AI: enforce existing laws to prohibit harms–while identifying any gaps that may exist–and punish bad actors who violate the law, rather than forcing developers to navigate onerous regulatory requirements based on speculative fear. There are no exceptions in the law for AI. And if someone is accused of violating the law, using AI is not a defense.

Regulating model development by imposing burdensome compliance requirements will make it harder for Little Tech to compete with larger platforms. Some startups might have small legal teams, but some have no lawyer on staff at all. For Little Tech, navigating complex legal frameworks or reporting requirements isn’t just hard; it’s a competitive threat.

Importantly, regulating model development does not directly protect consumers, the primary objective many lawmakers cite in weighing how best to regulate AI. To protect consumers from AI misuse, policymakers instead should focus on enforcing a host of laws already on the books related to consumer protection, civil rights, antitrust, fraud, and deceptive trade practices, and filling in gaps as necessary. In addition, we urge the administration to clarify that existing copyright law protects the ability of developers to train models. Getting this right is critical to ensuring American competitiveness in AI.

3. Invest in AI infrastructure and talent

To make it easier for startups and researchers to build and study AI models, the United States should establish a National AI Competitiveness Institute to provide them with access to the infrastructure resources they need, including compute, data, and evaluation tools. This is analogous to the approach taken in earlier eras of computing, for example, when the federal government established a National Center for Supercomputing Applications at University of Illinois Urbana-Champaign, which resulted in the development of the first web browser. At the same time, the United States should make investments to strengthen talent pipelines through workforce development initiatives. This could include AI literacy and training programs or public-private initiatives to support new jobs created for the AI economy such as AI labelers, something the Chinese government has done which provides Chinese AI companies a competitive advantage today. It’s also vital that the United States continue to support open-source models which promote innovation and competition by reducing barriers to entry and providing transparency.

It’s encouraging to see US policymakers take seriously the enormous potential of AI to drive advancements across society, industry, and government. We have no doubt that they, like us, want to advance our national security and economic interests by enabling Little Tech to compete. We believe these recommendations are the best path forward for a National AI Action Plan to ensure the United States retains its place as the world’s AI leader for generations.

For more, read our full submission to the OSTP.