The AI math isn't mathing as infrastructure spending forecast to outstrip expected revenue

The AI industry's anticipated capital spending for infrastructure is projected to far exceed the revenue it can generate from AI-driven applications, according to Gartner. The research firm's report, "Prepare for AI Market Consolidation: Capital Write-Downs Could Exceed $4 Trillion by 2032," predicts that AI datacenter operators will struggle to achieve profitable returns on their investments due to insufficient revenue.

By 2030, Gartner estimates AI providers will spend nearly $1 trillion on AI infrastructure, with an additional $6 trillion invested between 2027 and 2030. To maintain a 12% return on invested capital, the industry would need to generate $8.5 trillion in revenue, but forecasts suggest only $3.6 trillion will be available. This revenue shortfall could lead to a wave of asset write-downs, particularly for neocloud operators specializing in GPU-as-a-service AI compute.

Gartner predicts that 80% of these companies will not survive as independent entities beyond 2030. While some players may thrive and others underperform, hyperscalers are expected to better weather the shake-out due to their diversified business models. The financial markets are increasingly wary of the growing capital expenditure in AI investments, with lenders questioning the viability of these ventures.

SoftBank Group recently faced higher borrowing costs amid its investments in OpenAI, while firms like OpenAI and Anthropic face threats from open-source and Chinese AI models. Bain & Company warns that the AI industry needs to generate $6 trillion in annual revenue by 2031 to sustain infrastructure funding for anticipated demand.

Concerns about an AI bubble burst and its global economic ripple effects have been raised by the International Monetary Fund and the Bank for International Settlements.