Famous Investor Andrew Chen: The AI Cycle Will Ultimately Follow the Power Law
Famous investor Andrew Chen pointed out that every platform cycle begins with a narrative of democratization and ultimately returns to a power law distribution, and AI will be no different.
He emphasized that the technological progress of society and civilization is driven by the usage of the top 10% of users, eventually benefiting the masses.
In market mechanisms, the early dividends of AI are captured by top developers and institutions, with events driving top talent and capital to concentrate on a few frontier applications. The beneficiaries are high-capability individuals and leading companies, while ordinary users face short-term pressure but will share the results in the long term.
Source: Public Information
ABAB AI Insight
Andrew Chen, as the former growth lead at Uber and a general partner at a16z, has been deeply involved in multiple cycles of social and consumer internet, having written several articles in the 2010s about the "Cold Start Problem" and platform power laws. He achieved exponential user growth at Uber through growth hacking in its early days and later shifted focus to investing in AI-native applications.
His capital strategy is to continuously allocate resources to a few high-execution founding teams and top AI toolchains, leveraging the a16z network to mobilize LP funds to support early experiments. His motivation is to bet on the historical constant of "top 10% driving innovation" rather than uniform distribution, similar to capturing power law winners in the historical shift from Facebook to TikTok in social platforms.
Similar cases include the internet's transition from Web 1.0 where everyone built websites to the oligopoly dominance of Google/Amazon, and the mobile internet's shift from an open App Store to concentrated super apps. Chen currently places AI at a critical stage of transformation from a narrative of democratization to power law control.
Essentially, this represents a transfer of pricing power: AI tools initially spread for free or at low thresholds, but the top 10% of users achieve exponential productivity leaps through compounding usage (custom agents, data flywheels, high-frequency iterations). The mechanism is that network effects and data advantages allow a few nodes to capture disproportionate value, ultimately forcing platforms to shift from broad distribution to serving high-value users, thereby achieving a delayed spillover of technological dividends from the forefront to society at large.