According to Chi-Hua Chien, co-founder of Goodwater Capital, the future of artificial intelligence does not belong to those who sell models, but to those who integrate them into personalized experiences. With over two decades in venture capital, Chien has spotted trends that shaped entire industries. Back in 2004, as a young associate at Accel, he discovered The Facebook, demonstrating a unique ability to read human behavior at scale.
Model commoditization
Chien argues that the AI model layer is quickly becoming commoditized. Evidence includes Google's recent price cut from $7.99 to $4.99 per month for its subscription AI product, with doubled storage. In his view, companies with structural advantages in vertical integration and distribution, like Google, can compete on price and bundling for the average consumer. This mirrors previous tech cycles: in the web and mobile eras, application companies generated 88% and 84% of new value respectively, while infrastructure lagged. The real opportunities lie in applications that use AI to solve real problems.
Sponsored Protocol
Personalization as the key
Hyper-personalization is the thread separating winners from losers. Chien cites portfolio companies like Triumph, Ritten, and Flow GPT, which achieve hundreds of millions of dollars in annual recurring revenue through highly personalized entertainment experiences, without customers even perceiving AI as a core feature. Similarly, Midi Health uses AI to expand women's healthcare delivery, overcoming a shortage of specialists. In every sector where human expertise is a bottleneck, AI can cost-effectively expand access.
Sponsored Protocol
The cloud-device gap shrinks
Chien predicts that within the next year, the lag between the most advanced cloud AI models and those running locally on smartphones will shrink to three months, down from 18-24 months two years ago. This will enable truly ambient and personal experiences, though use cases are still being defined. As with the iPhone in 2007, it will take time for entrepreneurs to explore new possibilities. Large language models enable processing vast context and personalizing at the individual level with a continuous feedback loop.
Super apps and trust
Chien analyzes Facebook's failed attempts to build a super app, such as Facebook Credits and Libra. The reason is a trust gap: Americans do not trust a single app for both social and financial activities. Financial services demand high monetization but low time spent, while social media is the opposite. Bridging this psychological gap is tough. In parallel, Chien bets on the desire for in-person connection as a counterbalance to digitalization, a trend Goodwater Capital monitors closely.
Sponsored Protocol
Chi-Hua Chien, who saw Facebook coming early on, now suggests that the real AI winners will not be model sellers but companies that integrate AI into personalized, valuable applications. A message echoed by innovations seen elsewhere, such as Siri AI in iOS 27 and the 20th Anniversary iPhone, where personalization and user experience remain central.
