All In Discussion - Davos - Maria Bartiromo, David Sacks, Michael Kratsios - Recap and Notes - January 22 2026

January 22 2026

The State of the AI Race & US Leadership

• Current Standing: The U.S. is currently leading the global AI race. David Sacks notes that American companies continue to innovate with better models, chips, and data centers.
• Strategic Pillars: Michael Kratsios outlines three pillars of the U.S. strategy:

1. Innovation: Maintaining a lead over competitors.

2. Infrastructure: Building the necessary physical support (data centers/energy).

3. Export: Sharing American technology with the world.

• Policy Shift: The speakers highlight a shift from the previous administration. President Trump’s approach promotes "permissionless innovation," contrasting with the Biden administration's executive order and "diffusion rule," which added hundreds of pages of regulations that treated AI as a highly regulated industry,,.

Infrastructure, Energy, and Economics

• Demand vs. Bubble: Sacks dismisses fears that AI infrastructure is a bubble comparable to the "dark fiber" crash of the late 90s. Unlike unused fiber, there are no "dark GPUs"; every chip installed is being used to generate tokens due to massive demand,.
• Energy Consumption: The race has evolved from just an AI race to a "power race". China is expanding its grid rapidly (doubling in 10 years), while the U.S. grid has only grown 2-3%.
• Power Generation Solution: To avoid raising electricity rates for consumers, the administration encourages AI companies to act as power companies by building their own power generation "behind the meter",.
• Economic Impact:

◦ Infrastructure buildout added approximately 2% to GDP growth last year.

◦ Allowing data centers to generate their own power can lower residential rates by selling excess power back to the grid and amortizing fixed costs over a larger supply.

◦ Microsoft recently pledged that its data centers would not cause residential rates to increase.

Regulation: Federal vs. State

• The "Patchwork" Problem: There are over 1,200 AI bills moving through state legislatures, creating a fragmented regulatory environment (a "patchwork").
• Impact on Startups: This fragmentation hurts early-stage companies most, as they cannot navigate 50 different rulebooks, unlike large incumbents.
• Federal Strategy: The goal is to establish a "lightweight federal standard" that preempts conflicting state laws to create a single national rulebook,.
• State Authority: States would retain authority over specific areas like child safety and local permitting.

Use Cases and Future Trends

• Coding & Knowledge Work:

◦ Coding assistants have seen major breakthroughs (e.g., Anthropic’s Claude Code).

◦ 2026 Prediction: A productivity boom for knowledge workers is expected. AI will move beyond coding to generating spreadsheets, PowerPoints, and websites, effectively becoming a personal digital assistant capable of executing tasks across files and emails,,.

• Scientific Discovery (Genesis Mission):

◦ The administration is pushing "AI for Science" to double R&D output over the next decade.

◦ Key areas include fusion energy (accelerating simulations), material science (space exploration and nuclear energy), and healthcare (drug discovery).

• Automotive: Self-driving technology (Waymo, Tesla) has hit a new inflection point regarding quality.

Global Competition (China & Europe)

• The China Threat:

◦ US Advantages: The U.S. is ahead in models (~6 months), chips (~2 years), and equipment (~5 years).

◦ China’s Advantages: China leads in energy grid expansion and "AI Optimism" (83% of Chinese citizens view AI as beneficial vs. 39% of Americans),.

◦ China's Strategy: China is reportedly blocking Nvidia chips to protect its domestic champion, Huawei, hoping to dominate the local market before exporting globally.

• Winning the Race: Victory is defined by market share. If the world runs on American chips and models, the U.S. wins; if it runs on Huawei and Deepseek, the U.S. loses.
• Export Strategy: The "American AI Export Program" aims to provide turnkey AI solutions to countries that cannot build frontier models but need inference capabilities. This prevents the "Huawei effect" where inferior technology wins through subsidies and ubiquity,,.
• Europe: The speakers criticize Europe’s "precautionary principle" and over-regulation (e.g., the EU AI Act passed before ChatGPT existed). They argue European regulators act like "main characters," whereas the U.S. views entrepreneurs as the main characters,,.

Risks, Culture, and Philosophy

• AI Optimism vs. Pessimism: Low AI optimism in the West is blamed on negative media portrayals (e.g., The Terminator) and tech leaders focusing on "doom and gloom" scenarios, which fuels regulatory frenzy.
• Political Bias ("Woke AI"):

◦ Sacks warns of "Orwellian" scenarios where government uses AI for surveillance or censorship.

◦ The administration rescinded the previous executive order that included DEI requirements for AI models (citing the Google Gemini historical inaccuracy controversy).

◦ New policy: The federal government will not procure politically biased AI.

• Jobs and Abundance: Regarding Elon Musk’s view that AI will make jobs obsolete, Sacks agrees directionally that AI will lead to massive abundance and higher living standards, though a moneyless society is not imminent. He likens the potential future to the abundance seen in Star Trek.