AI Bubble Warning: Oracle $ORCL, Data Center Debt Risk for U.S. Investors

A critical perspective on the artificial intelligence sector suggests that the current investment boom is out of alignment with underlying economic realities. The core issue, according to a prominent AI critic, lies in the fundamentally broken cost structure of Large Language Models (LLMs). These models consume computational resources – measured in 'tokens' – at rates that defy traditional software subscription models, where costs are predictable and don't escalate based on usage or outcome.

While consumers and businesses typically pay fixed monthly fees for software, LLMs operate on a metered basis, with costs accruing even when the output is unhelpful. Many AI companies initially subsidized these true operational costs to attract users, offering subscriptions that allowed token consumption far exceeding the actual fee. Reports indicate substantial losses for major AI players, with one analyst finding billions in deficits despite significant revenues. This practice of selling services below their operational cost creates an illusion of profitability.

The unsustainability of this model became evident when enterprise customers transitioned to token-based billing. A notable example involved a large transportation company exceeding its annual AI budget within a single quarter, prompting its Chief Operating Officer to question the value proposition of AI services given their escalating, unpredictable costs. This challenge is widespread across AI startups, many of which remain unprofitable because users are unwilling to pay the full, high cost of token usage.

Further exacerbating the issue is the immense capital expenditure required for AI data centers. Building these facilities takes years and billions of dollars, funded largely by private credit. However, the primary customers for these data centers are the same unprofitable AI companies. This reliance on a small, financially struggling client base, financed by private credit and ultimately pension funds, raises concerns about potential systemic contagion if the AI market experiences a downturn. Investors are cautioned to assess the equity implications, especially for IT infrastructure providers and those with exposure through non-public credit markets.