Why the Next AI Winners Will Be Applications, Not Infrastructure

Chi-Hua Chien, co-founder of Goodwater Capital and an early investor in Facebook, argues that the current AI gold rush is following a well-trodden historical path. While much of the industry is currently obsessed with the 'model layer'—the infrastructure providers building foundational AI—Chien contends that this sector is rapidly commoditizing. Drawing parallels to the PC, web, and mobile eras, he suggests that the vast majority of long-term economic value will be captured by application companies rather than those selling the underlying AI technology.

Historical data supports Chien’s skepticism toward infrastructure-heavy investments. In both the web and mobile cycles, application-focused companies generated significantly more market capitalization than their infrastructure counterparts. Chien believes that as the gap between advanced AI models and on-device capabilities shrinks to just a few months, the competitive advantage of owning a proprietary model will evaporate. Consequently, the true winners of the AI era will be businesses that leverage AI to solve specific human needs, rather than those simply providing the tools to build them.

Beyond his market outlook, Chien offers a critical perspective on the current state of venture capital. He observes that the industry is becoming increasingly 'meme-ified' and polarized, driven by the vertical integration of massive firms that no longer feel the need to maintain traditional syndicate decorum. He also addresses the prevalence of 'fast follow' funding rounds, noting that aggressive valuation marketing is often a symptom of excess demand rather than fundamental business health. For founders and investors, Chien’s insights serve as a reminder that market frothiness often masks the underlying reality: sustainable value is built through consumer utility, not just technological novelty.