Descriptions:
Peter Diamandis sits down with Alvin Wang Graylin — senior fellow at the Asia Society Policy Institute, Stanford HAI digital fellow, and a 35-year veteran of both US and Chinese tech ecosystems — for a wide-ranging examination of the US-China AI race. Graylin argues the competition is better understood through a game-theoretic “stag hunt” lens (coordination under mutual risk) rather than a classic prisoner’s dilemma, with cooperation as the only path to safety as ASI timelines shorten.
Graylin shares observations from the 2026 World AI Conference in Shanghai, where President Xi delivered the opening keynote, and from direct meetings with Chinese AI lab leadership and senior regulators. A key insight: Chinese institutions are not behaviorally acting as though ASI is imminent, in contrast to some Western lab cultures. He also offers a blunt retrospective on US chip export controls — arguing they inadvertently rescued struggling Chinese GPU companies like Biren and Cambricon by forcing domestic data centers to buy their products, while training compute simply migrated to servers in Hong Kong and Europe overnight.
The episode also previews the September 24 US-China AI safety dialogues in Washington, DC, which Graylin is helping to organize. He closes with analysis of what he calls a $1.7 trillion AI sector valuation bubble, warning that a private credit correction financing US data center buildout could create unpredictable geopolitical pressure — including scenarios involving Taiwan.
📺 Source: Peter H. Diamandis · Published August 18, 2026
🏷️ Format: Podcast
