Chamath Palihapitiya argues that the AI model business lacks lasting advantage due to rapid commoditization, suggesting that large valuations for companies like Anthropic and OpenAI are a 'mathematical mistake.' He believes the real value lies in the application layer above and infrastructure layer below the models, benefiting companies like Alphabet and Nvidia. This perspective challenges the high valuations of frontier AI model developers and highlights the potential for a shift in value capture within the AI ecosystem.
Chamath Palihapitiya's comments suggest a significant re-evaluation of the AI market's value chain. He posits that the rapid commoditization of AI models, exemplified by Moonshot AI's Kimi K3, means that the 'model layer' itself will not sustain high valuations. This directly impacts companies like Anthropic and OpenAI, which are currently seeking public market funding based on their frontier models. In the short term, this could create skepticism around their IPOs. Long-term, it implies that the real beneficiaries will be companies providing the underlying infrastructure (like Nvidia) or building applications on top of these increasingly commoditized models (like Alphabet, which sells cloud services). The key risk for traders is overvaluing model-centric AI companies, while the opportunity lies in identifying infrastructure and application layer plays.