Foundational AI models will commoditize into a utility layer where companies buy "intelligence on the fly." The real, sustainable profit will be captured by application companies that leverage various models to solve specific business problems, as most enterprises lack the expertise to use raw models effectively.

Drawing a parallel to Web3's 'Fat Protocols' thesis, today's large AI models are capturing the majority of value in the tech stack. As these 'fat' models become more capable, the applications built on top become 'thinner,' serving primarily as simple wrappers or marketing channels rather than creating defensible value.

Creating frontier AI models is incredibly expensive, yet their value depreciates rapidly as they are quickly copied or replicated by lower-cost open-source alternatives. This forces model providers to evolve into more defensible application companies to survive.

Similar to how blockchain protocols like Bitcoin and Ethereum accrued more value than the apps built on them, AI foundation models are getting 'fatter.' They are absorbing more capabilities, allowing users to perform complex tasks in a single step within the base model, reducing the need for specialized application-layer companies.

Mobile networks built expensive global infrastructure with massive usage but captured little value as profits moved "up the stack" to apps. Foundation models, despite huge CapEx, face a similar risk of becoming a commoditized infrastructure layer with low pricing power.

Comparing AI to 1995-era internet bandwidth, the hosts argue that selling raw 'intelligence' is a low-margin, commodity business. The significant financial upside will be captured not by the infrastructure providers, but by the creators who build novel applications and experiences using that intelligence as a building block.

If AI makes intelligence cheap and universally available, its economic value may collapse. This theory suggests that selling raw AI models could become a low-margin, utility-like business. Profitability will depend on building moats through specialized applications or regulatory capture, not on selling base intelligence.

Much like 'big data' evolved from a competitive advantage into a widely available commodity, AI models will likely follow the same path. So many sources will offer powerful models that they will cease to be a unique differentiator or a durable moat for businesses.

Contrary to the 'winner-takes-all' narrative, the rapid pace of innovation in AI is leading to a different outcome. As rival labs quickly match or exceed each other's model capabilities, the underlying Large Language Models (LLMs) risk becoming commodities, making it difficult for any single player to justify stratospheric valuations long-term.