The article argues that whilst AI safety warnings dominate headlines, the more immediate threat may be a financial collapse of the AI industry itself. Major technology firms have accumulated massive debts—£99 billion in debt issuance this year alone—to fund rapid datacentre expansion, and the fundamental economics of their AI business models are unsustainable. This financial vulnerability could have repercussions far beyond the US, making it arguably a more pressing concern than apocalyptic warnings about AI development.

The core problem is a widening gap between declining revenue and stable costs. AI pricing has collapsed dramatically—OpenAI repeatedly cuts fees to retain customers, and token costs have more than halved since June to under $1 per million—whilst the costs of building datacentres, particularly semiconductor procurement, remain elevated. Companies are masking poor profitability through unconventional accounting measures, such as Anthropic's use of "adjusted operating income" that excludes significant costs. The situation parallels pre-2008 financial crisis dynamics, with highly leveraged firms using complex, questionable business models to justify their valuations.