The AI industry faces significant risks of a "crash" due to unsustainable capital expenses, with estimates suggesting $2 trillion annually in revenue is needed to cover infrastructure costs, a figure far from current forecasts. The industry's reliance on reciprocal investments among a few firms creates a fragile ecosystem where one failure could impact many. Funding primarily through debt, increasing public and corporate skepticism, and diseconomies of scale – where newer AI models consume more resources – further exacerbate these vulnerabilities. Moody's has already warned about the credit risks associated with high AI infrastructure spending.

A potential crash could range from a severe collapse, akin to the 2000 tech crash, wiping out substantial wealth and halting data center construction, to a milder correction. Consequences would include stranded investments for communities and utilities, leading to higher rates for consumers. Vendors supporting AI infrastructure, like Micron and Corning, would face severe repercussions, with expected revenues from fiber investments unmet.

Despite potential short-term harm, a crash could ultimately foster long-term good by forcing AI companies to prioritize efficiency and economies of scale. Just as the 2000 tech crash brought realism to the telecom market, an AI reset could lead to a more sustainable and viable AI technology, even if current developers are not the ultimate beneficiaries. The focus would shift to cost control and self-sufficiency for AI to succeed as a technology.