Winning the AI Race
What can we learn from the Senate hearings?
On May 8, 2025, a group of tech executives including Sam Altman (OpenAI), Lisa Su (Advanced Micro Devices - AMD), Michael Intrator (CoreWeave), and Brad Smith (Microsoft) testified at the Senate Committee on Commerce, Science, and Transportation on “Winning the AI Race: Strengthening U.S. Capabilities in Computing and Innovation.” You can get a full transcript here.
I asked ChatGPT to summarize the main remarks and organize the testimonials along ten common themes. I hope this makes for an easier read of more than three and a half hours of prepared remarks and Q&A.
Enjoy!
Chairman Sen. Ted Cruz (R-TX) - Opening and Closing Remarks
Senator Cruz emphasized the need for the U.S. to outpace China in AI through innovation, not regulation. He criticized European-style regulations and President Biden’s executive actions as harmful to U.S. AI competitiveness. Cruz proposed a regulatory sandbox modeled after the early internet era to remove barriers and promote growth. He stressed the importance of domestic infrastructure, export markets, and AI adoption, framing the AI race as critical to economic and national security.
In his closing remarks, Cruz proposed a “10-year learning period” during which states would refrain from passing comprehensive AI regulations. He suggested this as a form of federal preemption to avoid a patchwork of state laws and give AI developers a consistent national framework. Cruz argued that this approach would allow the federal government and industry time to better understand the technology and craft appropriate policy.
Committee Ranking Member Sen. Maria Cantwell (D-WA) – Comments
Senator Cantwell advocated for an open, democratic AI ecosystem led by the U.S. She highlighted the importance of investing in computing power, data, and public-private partnerships, especially through the NSF and the CHIPS and Science Act. Cantwell also emphasized export controls as national security tools—not trade tactics—and stressed urgency in establishing global standards, infrastructure, and workforce pipelines, including energy and electricians for data centers.
Sam Altman (OpenAI)
1. Regulation and Fragmentation
Altman warned that requiring AI developers to pre-clear systems before deployment would be 'disastrous,' expressing strong concern about excessive regulation slowing innovation. While open to some form of sensible federal standards, he pushed back on the idea of the National Institute of Standards and Technology (NIST) imposing rigid frameworks. Altman advocated for clear federal rules to avoid a patchwork of state-level policies, aligning with Senator Cruz’s call for a 'light-touch' regulation.
2. Standards
Altman stated that while NIST involvement 'could be helpful,' he remained skeptical of mandated testing regimes or rigid rulebooks. He prefers flexible, use-case-driven guidelines to ensure innovation continues without regulatory gridlock.
3. Environmental and Climate Considerations
Altman highlighted OpenAI’s collaborations with national laboratories, noting that AI tools are helping accelerate energy innovation and climate science. He positioned AI as part of the solution to energy and environmental challenges.
4. U.S. Leadership vs. China
He strongly advocated for a U.S.-led AI model grounded in democratic values like openness and transparency, contrasting it with authoritarian approaches like China’s. Altman called for American AI to prevail globally, framing it as a strategic imperative.
5. Export Controls
Altman emphasized that policy certainty is crucial for global diffusion of U.S. AI. He warned against overreaching controls that could hamper U.S. companies and lead allies to adopt less secure alternatives.
6. Innovation and Research Institutions
Altman praised national labs and universities, saying OpenAI tools are already used to advance scientific research. He supported further federal investment in research and emphasized enabling domestic manufacturing of chips and compute systems.
7. Energy and Data Center Infrastructure
He referenced the 'Project Stargate' initiative, a $500 billion AI infrastructure investment in Abilene, Texas, to illustrate the massive energy and data center needs of frontier model training. Altman called for faster permitting and utility reforms.
8. Workforce and Immigration
Altman urged reforms to make it easier for top-tier global talent to work in the U.S. He also noted that over one-third of U.S. college-aged users rely on OpenAI tools for learning, indicating AI’s potential in workforce development.
9. Risks and Responsible Development
OpenAI leads the industry in red-teaming and transparency, Altman claimed. He emphasized AGI’s dual-use risks and the necessity of safeguards, even while maintaining optimism about its transformative potential.
10. Capital and Investment
Altman cited OpenAI’s multibillion-dollar infrastructure plans and stressed the importance of regulatory clarity to attract continued domestic and international investment.
Lisa Su (AMD)
1. Regulation and Fragmentation
Su did not explicitly critique regulation but emphasized the need for policies that enable rapid innovation and scale-up of AI compute infrastructure. She highlighted the importance of public-private cooperation to ensure innovation remains fast-paced.
2. Standards
Su strongly endorsed open ecosystems, where hardware and software from different vendors interoperate. She framed openness as essential to fostering competition, accelerating innovation, and ensuring secure, transparent systems.
3. Environmental and Climate Considerations
She stressed that expanding AI infrastructure must go hand-in-hand with investments in clean, reliable energy. Su called attention to the rising energy footprint of AI data centers and urged planning to keep energy sources affordable and sustainable.
4. U.S. Leadership vs. China
Su framed AI leadership as a global competition with no guaranteed winner. She underscored that the U.S. must maintain its lead by accelerating innovation, manufacturing capacity, and talent pipelines.
5. Export Controls
While recognizing national security imperatives, Su warned that overbroad export restrictions could undermine U.S. global market leadership. She advocated for rules that are simple, consistent, and pro-innovation.
6. Innovation and Research Institutions
Su detailed AMD’s partnerships with the Department of Energy, powering the world’s two fastest supercomputers. She highlighted supercomputing’s role in advancing science and AI research.
7. Energy and Data Center Infrastructure
Su echoed others in urging expansion of AI data center capacity, including reliable energy and permitting reforms. She cited chip manufacturing and high-performance computing as critical national infrastructure.
8. Workforce and Immigration
Su called for a robust national STEM strategy and immigration policies that attract and retain global AI talent. She emphasized university partnerships and reskilling as essential to long-term competitiveness.
9. Risks and Responsible Development
Su did not dwell on existential AI risks, but implicitly endorsed responsible AI development through transparency, open standards, and public-private trust.
10. Capital and Investment
Su emphasized the capital intensity of chip manufacturing and packaging. She praised the CHIPS Act and called for additional incentives to scale domestic production and reduce dependence on foreign supply chains.
Michael Intrator (CoreWeave)
1. Regulation and Fragmentation
Intrator called for regulatory consistency and predictability, stressing that long-term infrastructure investment requires a stable policy environment. He argued that unclear or fragmented regulations across jurisdictions create barriers to scaling AI infrastructure.
2. Standards
Intrator did not focus on technical standards but emphasized the performance needs and specialized nature of AI cloud platforms. His remarks suggested that innovation in infrastructure must come before formal standards take shape.
3. Environmental and Climate Considerations
He highlighted the immense energy demands of AI infrastructure—CoreWeave operates more than 250,000 GPUs, consuming 360MW of power. He stressed the need for reforms in permitting and energy grid expansion to meet future AI demand.
4. U.S. Leadership vs. China
Intrator positioned infrastructure leadership as the foundation of global AI dominance, warning that China is heavily investing to outpace the U.S. He urged Congress to act quickly or risk falling behind in the global AI race.
5. Export Controls
Intrator supported calibrated export controls that secure national interests while enabling U.S. tech to diffuse globally. He warned that restricting access to U.S. AI tools could incentivize adoption of alternative, possibly less secure platforms.
6. Innovation and Research Institutions
Intrator highlighted CoreWeave’s partnership with Princeton University and the New Jersey AI Hub as a model public-private initiative that bridges academia, industry, and government.
7. Energy and Data Center Infrastructure
CoreWeave’s data center growth was framed as a case study in AI-scale energy needs. Intrator emphasized that AI infrastructure requires radically different design, cooling, and grid access than traditional cloud services.
8. Workforce and Immigration
Intrator advocated for stronger workforce pipelines, especially technical roles in infrastructure such as electricians, data center technicians, and networking specialists. He endorsed government support for skilling programs through community colleges and apprenticeships.
9. Risks and Responsible Development
He framed CoreWeave’s role as foundational rather than policy-setting, but emphasized the importance of secure, reliable systems as a form of risk mitigation.
10. Capital and Investment
Intrator emphasized that AI infrastructure is capital-intensive and relies on clear policy signals. He called for public-private collaboration to de-risk long-term investment in AI-specific platforms.
Brad Smith (Microsoft)
1. Regulation and Fragmentation
Smith advocated for trust-driven, innovation-focused regulation that avoids stifling diffusion. He warned that burdensome rules could delay AI’s benefits for productivity and inclusion.
2. Standards
He called for global leadership in export and interoperability standards, arguing that trust in U.S. AI tools depends on consistent rules, data protections, and ethical guidelines.
3. Environmental and Climate Considerations
Smith stressed the need to modernize the energy grid and train hundreds of thousands of electricians. He noted Microsoft’s significant U.S. investment in grid-linked infrastructure.
4. U.S. Leadership vs. China
Smith underscored the importance of becoming the first mover in both innovation and diffusion. He argued that global leadership depends not just on building the best tools, but on distributing them rapidly and earning international trust.
5. Export Controls
He emphasized that export controls should be precise and strategic, not blunt instruments. Smith advocated for rules that maintain national security while reassuring global partners.
6. Innovation and Research Institutions
Smith championed basic research funding through universities and national labs, calling them the crown jewels of U.S. innovation. He urged more public-private collaboration to accelerate breakthroughs.
7. Energy and Data Center Infrastructure
He described Microsoft’s infrastructure investment, with significant focus in the U.S., as evidence of the scale required. He linked success in AI to fast permitting and utility reform.
8. Workforce and Immigration
Smith called for a nationwide AI skilling initiative—from basic fluency in K-12 to advanced AI degrees in universities. He linked human capital directly to productivity growth.
9. Risks and Responsible Development
He framed AI as a tool to empower people, not replace them. Smith emphasized that Microsoft’s approach is centered on augmenting human capability and building trust. He underscored transparency, responsibility, and ethics as foundational principles.
10. Capital and Investment
Smith emphasized that AI leadership requires a healthy innovation ecosystem that includes both large firms and startups. He warned that regulation should avoid entrenching incumbents or excluding emerging players from market participation.