Why the biggest AI winners won't be selling AI

Chien, a co-founder of Goodwater Capital, suggests that the market is already entering an era of intense price competition, evidenced by companies like Google slashing subscription costs. Drawing parallels to the PC and mobile eras, he notes that infrastructure companies rarely capture the lion's share of market value. In the mobile cycle, for instance, infrastructure generated roughly $700 billion in market cap, whereas application companies—such as Uber, Meta, and Spotify—created $3.7 trillion.

For Chien, the next wave of success depends on using artificial intelligence to solve supply-constrained problems rather than simply selling the models themselves. He points to firms like Midi Health, which uses AI to scale specialized hormone replacement therapy, as the model for future growth. By integrating AI into specific sectors like healthcare or entertainment, companies can deliver deeply personalized experiences that customers value for their utility rather than their technical sophistication. As the gap between cloud-based frontier models and on-device processing shrinks to just three months, the focus is shifting from what the technology can do to how it can be applied to human-centric challenges.