Chi-Hua Chien argues the real AI opportunity lies in applications, not infrastructure. Drawing on history from PC, web, and mobile eras—where 85-88% of value went to apps—he predicts infrastructure will peak in 2025-2026 and then commoditize. Goodwater Capital is backing hyper-personalized apps in entertainment, healthcare, and real-world experiences, while noting incumbents are faster at copying than ever before.
Summarized by Podsumo
History repeats: In the web era, apps created $3.1 trillion in market cap vs. $400B for infrastructure. In mobile, apps created $3.7 trillion vs. $700B. Chien expects the same pattern for AI.
Hyper-personalization wins: Chien highlights companies like Written (Korea) creating unique entertainment experiences and MIDI Health using AI to expand access to scarce medical expertise.
Infrastructure is peaking: Chien predicts AI infrastructure value creation will peak in 2025-2026, with price competition already emerging (Google slashing AI subscription costs).
New interfaces coming: Voice and passive inputs (glasses, always-on agents) will separate input from screens—phones becoming consumption devices.
Real-world backlash: After years of digital isolation, people crave in-person experiences. Goodwater backs Bump (finding friends IRL) and Fever (live events) to capture this swing.
"In the web era, infrastructure companies created $400B; apps created $3.1 trillion. In the mobile era, infrastructure $700B; apps $3.7 trillion. So 85% of net new value went to apps—and that’s where AI’s real opportunity lies. — Chi-Hua Chien"
"In a world of infinite digital content, people crave the most constrained thing: real human contact and in-person experiences. — Chi-Hua Chien"