All-In

Inside America's AI Strategy: Infrastructure, Regulation, and Global Competition

with David Sacks and Michael Kratsios
23 Jan 2026 18 min read 1h 12m
TL;DR

The U.S. is winning the AI race across models, chips, and manufacturing equipment, but risks squandering that lead through overregulation and low public optimism (39% vs. China's 83%). The real bottleneck is energy: data centers must build their own power generation to avoid grid strain and cost increases, while a chaotic patchwork of 1,200 state bills threatens to cripple startup competition.

Key Moments
David Sacks
“There's no such thing as a dark GPU right now. Every GPU that's being put in a data center is getting used. Uh and it's being used to generate tokens and that's to power the this new generation of AI chat bots or coding assistants.”
Addressing concerns that AI data center spending mirrors the dot-com fiber bubble
▶ 2:03
Michael Kratsios
“The patchwork is actually most detrimental to early stage young companies and entrepreneurs. If you want to develop a new AI technology, if you want to build something on top of one of our great frontier models, having to figure out how to navigate 50 different rules across 50 different states creates a lot of friction and ultimately the big guys are the ones that can succeed in in that environment the best.”
Explaining why federal AI regulation trumps state-by-state approaches
▶ 4:19
David Sacks
“I mean frankly the states are going hog wild right now with regulation. There's over 1,200 bills going through state legislatores right now. I think it's very much a knee-jerk reaction.”
Criticizing the reactive wave of state-level AI regulations
▶ 5:44
Michael Kratsios
“I think in 2026, you could see that that these types of of tools again started as coding assistants, but now they become personal digital assistants. That could definitely happen this year.”
Predicting the imminent shift from task-based AI tools to autonomous personal assistants
▶ 21:40
David Sacks
“Well in China AI optimism was 83%. So 83% of the population feels that it's being more beneficial than harmful. That number in the United States is only 39%.”
Highlighting a critical psychological gap between U.S. and Chinese populations on AI adoption
▶ 25:10
About the show
›
All-In is a weekly podcast featuring venture capitalists and tech leaders discussing the most pressing issues in business, technology, and politics. In this episode, David Sacks and Michael Kratsios dive deep into America's AI strategy, covering infrastructure buildout, federal regulation of state laws, and competition with China. The conversation spans innovation policy, data center economics, and real-world AI applications transforming industries from healthcare to coding.
Takeaways
1
Data centers must own their power—economies of scale will lower rates The administration is removing regulations that forced data centers onto the public grid. When companies generate their own power and sell excess back to the grid, fixed costs amortize over larger supply pools, reducing electricity prices for all consumers. Microsoft and other hyperscalers have pledged not to increase residential rates—a model that should become standard.
2
U.S. leads AI stack but risks self-sabotage through overregulation America maintains 6-12 month leads on models, 2-year leads on chips, and 5-year leads on semiconductor equipment. However, 1,200+ state bills creating fragmented rules disproportionately harm startups—only well-resourced companies can navigate 50 regulatory frameworks. A lightweight federal standard is essential to prevent China from becoming the default AI exporter globally.
3
AI for science and knowledge work are the next productivity breakthroughs After chatbots and coding assistants, AI is shifting to domain-specific applications: fusion simulation, material science for space, medical diagnostics, and knowledge worker tools (Excel, PowerPoint generation). The Genesis mission aims to consolidate fragmented scientific data into trainable datasets—potentially doubling U.S. R&D output over a decade.