AI infrastructure leaders justify massive investments by citing a limitless appetite for intelligence, dismissing concerns about efficiency. This belief ignores that infinite demand doesn't guarantee profit; it can easily lead to margin collapse and commoditization, much like the internet's effect on media.

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.

While immense value is being *created* for end-users via applications like ChatGPT, that value is primarily *accruing* to companies with deep moats in the infrastructure layer—namely hardware providers like NVIDIA and hyperscalers. The long-term defensibility of model-makers remains an open question.

The startup landscape now operates under two different sets of rules. Non-AI companies face intense scrutiny on traditional business fundamentals like profitability. In contrast, AI companies exist in a parallel reality of 'irrational exuberance,' where compelling narratives justify sky-high valuations.

AI companies operate under the assumption that LLM prices will trend towards zero. This strategic bet means they intentionally de-prioritize heavy investment in cost optimization today, focusing instead on capturing the market and building features, confident that future, cheaper models will solve their margin problems for them.

The enduring moat in the AI stack lies in what is hardest to replicate. Since building foundation models is significantly more difficult than building applications on top of them, the model layer is inherently more defensible and will naturally capture more value over time.

The AI boom can sustain itself as long as its narrative remains compelling, regardless of the underlying reality. The incentive for investors is to commit fully to the story, as the potential upside of being right outweighs the cost of being wrong. Profitability is tied to the narrative's durability.

Products like Sora and current LLMs are not yet sustainable businesses. They function as temporary narratives, or "shims," to attract immense capital for building compute infrastructure. This high-risk game bets on a religious belief in a future breakthrough, not on the viability of current products.

While AI investment has exploded, US productivity has barely risen. Valuations are priced as if a societal transformation is complete, yet 95% of GenAI pilots fail to positively impact company P&Ls. This gap between market expectation and real-world economic benefit creates systemic risk.