Ben Thompson on Big Tech, China, and the AI Boom Running Out of Money - [Invest Like the Best, EP.487]
Summary
Patrick O'Shaughnessy discusses AI geopolitics, arguing that US dominance is dangerous and that current US-China tensions reflect underappreciated supply chain dependencies. He analyzes the capital intensity of AI, comparing it to railroad booms and noting how tech giants are shifting from high-margin software to capital-heavy infrastructure. The conversation covers the risks of compute shortages, the oligopolistic nature of memory and chip manufacturing, and the challenge of monetizing AI through advertising versus subscriptions.
Topics discussed
Sponsors: Ramp and Felix by Rogo
The problematic nature of a US-only AI victory
Global AI equilibrium and the China chip gap
Capital intensity, equity issuance, and the railroad analogy
Verifiability in AI and the potential of Neuralink
Sponsors: Vanta and Ridgeline
Discovery vs. distribution and enterprise AI pricing models
The Dropbox playbook: Consumer virality to enterprise sales
Compute supply cycles and commodity market dynamics
Memory makers, TSMC's monopoly, and risk transfer
Hyperscalers building custom chips and AWS's evolution
Apple's device strategy vs. ambient AI and frontier winners
Historical parallels: IBM, Oracle, and Meta's social moat
Meta's advertising value and the impact of ATT
NVIDIA's commodity risk and sovereign cloud deals
Listen ad-free on Castria