Despite being a commodity business with high costs and low defensibility, AI foundation models command massive valuations. They function as a 'hope' asset where investors park capital based on narrative, similar to how gold is used in uncertain times, rather than on financial fundamentals.

The world's most profitable companies view AI as the most critical technology of the next decade. This strategic belief fuels their willingness to sustain massive investments and stick with them, even when the ultimate return on that spending is highly uncertain. This conviction provides a durable floor for the AI capital expenditure cycle.

In a rapidly evolving field like AI, waiting for mature tools is a mistake. The correct strategy is to invest now, assuming that capabilities that are almost working today will be fully functional tomorrow due to exponential, compounding progress.

The current massive investment in AI is driven by a belief that it is the most critical technology of the decade. Large companies are willing to spend billions with uncertain immediate returns simply to secure a long-term strategic position, making it a must-have expenditure that overrides normal financial discipline.

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.

The current level of spending on AI infrastructure is so astronomical that incremental improvements or narrow applications like coding assistants won't suffice for payback. The financial markets are implicitly underwriting a binary bet: either AGI is achieved, or a massive financial reckoning is inevitable.

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.

Relying on a speculative 'AI productivity miracle' to solve fundamental economic problems like the national debt is an extraordinarily high-risk strategy. Until technological advancements are reflected in actual economic data, treating them as a guaranteed solution is just 'hopium' that distracts from making necessary hard choices today.

Companies are spending unsustainable amounts on AI compute, not because the ROI is clear, but as a form of Pascal's Wager. The potential reward of leading in AGI is seen as infinite, while the cost of not participating is catastrophic, justifying massive, otherwise irrational expenditures.