AI Model Commoditization Is Here, Says Investor Who Found Facebook

The venture capitalist who first spotted Facebook as a six-person Harvard startup now sees a familiar pattern emerging in artificial intelligence—and believes the companies building AI models will capture far less value than those using AI to solve specific problems.

Chi-Hua Chien, co-founder of consumer-focused Goodwater Capital, points to Google's recent price cut for its AI subscription product as evidence that commoditization has already begun. The search giant dropped its monthly fee from $7.99 to $4.99 while doubling storage, signaling that price competition for AI services is now underway.

Historical patterns repeat

Chien's thesis draws on decades of technology cycles. During the PC era, web boom, and mobile revolution, infrastructure companies consistently captured less value than application companies. In the web era, infrastructure providers generated $400 billion in market cap while applications created $3.1 trillion—representing 88% of new value. The mobile era showed similar dynamics: $700 billion for infrastructure versus $3.7 trillion for applications like Netflix, Spotify, Meta, Uber, and Airbnb.

Remarkably, infrastructure market caps peaked in 2000 and haven't surpassed that level in nominal dollars 26 years later, according to Chien.

The personalization advantage

Goodwater's portfolio reflects this application-first strategy. The firm has invested in entertainment companies generating hundreds of millions in annual recurring revenue where customers don't even recognize they're using AI products. Chien also highlighted Midi Health, a women's health company using AI to expand access to hormone replacement therapy by overcoming provider shortages.

"What LLMs do is basically two things: they make it possible for you to process large amounts of context and make sense of it all, and they allow you to do personalization down to the individual, cost effectively," Chien explained in an interview with TechCrunch.

The gap between frontier AI models and what runs locally on smartphones is shrinking rapidly. Two years ago, the lag was 18 to 24 months. Now it's six months, and Chien expects it to compress to three months within a year.

Physical experiences as counterbalance

Chien is also betting on in-person experiences as a reaction to infinite digital content. Goodwater has invested in Bump, from the founders of Snap-acquired Zenly, which helps people interact in physical spaces using digital information. The firm also backs Fever, a European live events company that started with niche experiences like candlelight concerts before expanding mainstream.

"What do people crave in a world where there's an infinite supply of digital content? They crave the thing that is most constrained, which is real human contact, real-world experiences," he said.

Why it matters

Chien's perspective carries weight given his track record identifying transformative consumer platforms early. His argument that AI infrastructure will commoditize while applications capture value challenges the current venture capital focus on foundation models and suggests a strategic shift toward companies using AI to solve specific vertical problems. For business leaders evaluating AI investments, this framework suggests prioritizing domain-specific applications over general-purpose model capabilities.

These details were first reported by TechCrunch.