a16z's Recommendations for the National AI Action Plan

Jai Ramaswamy, Collin McCune, and Matt Perault

Posted March 14, 2025

This week, a16z shared its recommendations with the White House Office of Science and Technology Policy (OSTP) for how the United States can implement a competitiveness agenda to maintain its global leadership in AI development. The firm emphasized the crucial role of "Little Tech" (startups) in strengthening America's AI capabilities, drawing parallels to historical technological innovations driven by startups.

As part of the OSTP's request for information to inform a new AI Action Plan, a16z highlighted AI's potential to revolutionize various sectors, including healthcare (drug discovery), education, and transportation. The firm noted that over 5,000 AI startups in the U.S. received funding between 2013 and 2023, and that recent AI advancements have largely originated from these startups rather than established Big Tech companies.

a16z advocates for a regulatory framework that supports Little Tech's ability to compete with larger, more resourced companies. Failure to do so risks stagnation and ceding leadership to other nations, particularly China.

To guide the OSTP's AI policy development, a16z proposed three key policy pillars:

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

• Support for Little Tech: The U.S. AI Action Plan should foster an environment where startups can compete on a level playing field with larger incumbents.
• Federal Leadership: Given AI's national security, economic, and foreign relations implications, the federal government, not individual states, should lead in promoting and regulating a national AI market to provide regulatory certainty for innovators.
• State Role: States should continue to play a role in AI governance by addressing harmful conduct within their borders.
• Competitiveness Imperative: Policy decisions made now are critical to ensuring the U.S. leads in future AI technologies. Prioritizing competitiveness and entrepreneurship is essential to outcompeting other nations and winning the AI race.
• Avoiding Stagnation: The emergence of foreign AI models, like DeepSeek R1, underscores the need to prioritize competitiveness to prevent slowing U.S. AI development.

2. Regulate harms, not AI model development.

• Focus on Use, Not Creation: The U.S. should continue its historical approach of regulating technology based on its use, not its development. For example, there are no laws dictating how to build a computer, but laws exist to prevent its misuse.
• Enforce Existing Laws: Policymakers should enforce existing laws related to consumer protection, civil rights, antitrust, fraud, and deceptive trade practices to address AI-related harms. Any necessary gaps in these laws should be identified and filled.
• Punish Bad Actors: The focus should be on punishing bad actors who violate the law, rather than imposing burdensome regulatory requirements on developers based on speculative fears.
• No AI Exception: Using AI should not be a defense for violating existing laws.
• Competitive Threat: Regulating model development with onerous compliance requirements poses a competitive threat to Little Tech, which may lack extensive legal resources.
• Consumer Protection: To protect consumers, focus should remain on enforcing existing laws and addressing gaps, rather than regulating model development.
• Copyright Protection: The administration should clarify that existing copyright law protects developers' ability to train models, which is critical for U.S. AI competitiveness.

3. Invest in AI infrastructure and talent

• National AI Competitiveness Institute: Establish an institute to provide startups and researchers with access to essential infrastructure, including compute, data, and evaluation tools. This mirrors past government initiatives, such as the National Center for Supercomputing Applications, which spurred innovation like the first web browser.
• Talent Pipeline Strengthening: Invest in workforce development initiatives, including AI literacy and training programs, and public-private partnerships to support new AI-related jobs (e.g., AI labelers).
• Chinese Competitive Advantage: Acknowledge that the Chinese government's investment in AI talent and jobs provides its AI companies with a competitive advantage.
• Support Open-Source Models: Continue supporting open-source models to foster innovation, competition, and transparency by reducing barriers to entry.

Conclusion:

a16z believes these recommendations offer the best path forward for a National AI Action Plan to ensure the United States remains the global AI leader for generations to come. The firm emphasizes that enabling Little Tech to compete is crucial for advancing national security and economic interests.

For the full submission, refer to a16z's full submission to the OSTP.

Contributors:

• Jai Ramaswamy: Chief Legal Officer at Andreessen Horowitz, overseeing legal, compliance, and government affairs.
• Collin McCune: Head of Government Affairs at Andreessen Horowitz, managing the firm's political and policy strategy.
• Matt Perault: Head of Artificial Intelligence Policy at Andreessen Horowitz, guiding the firm's AI policy strategy and assisting portfolio companies.

More From These Contributors:

• We’re Leaving Delaware, And We Think You Should Consider Leaving Too
• NSF RFI: Fostering American AI Leadership Through Research and Innovation
• To Help AI Startups Compete, Use Revenue-based Thresholds
• AI Model Facts: Transparency that Works for Little Tech
• SEC RFI: Creating a Safe Harbor and Crowdfunding Regime for Collectible Tokens (NFTs)

Disclaimer: The views expressed are those of the individual personnel and not necessarily those of a16z. Information from third-party sources is believed to be reliable but not independently verified. This content is for informational purposes only and not legal, business, investment, or tax advice. Past performance is not indicative of future results.