Analysts Project Trillions in AI Spending Amid Revenue Gap

Financial analysts project massive capital expenditure for artificial intelligence infrastructure through the end of the decade, though views differ on the immediate economic viability. JPMorgan estimates cumulative AI data center spending will range between $5.5 trillion and $10 trillion by 2030, noting that accelerating revenues among AI companies have improved the cycle's viability over the last six months. The bank expects AI cloud and model providers to reach a combined revenue run-rate of $1.6 trillion by the end of 2026.

Conversely, Goldman Sachs estimates cumulative spending will reach $7.6 trillion between 2026 and 2031, warning of a significant gap between infrastructure costs and generated revenue. The firm suggests the industry must scale annual recurring revenue beyond $1 trillion by 2030 to avoid massive hardware write-downs. This environment places Nvidia Corporation at higher risk as hyperscalers develop internal chips to reduce reliance on external GPUs.

Software providers are attempting to bridge this gap through scaled deployment. Microsoft has integrated AI into existing products, with AI operations exceeding a $37 billion annual run rate in the March quarter and Microsoft 365 Copilot surpassing 30 million paid seats by June. Additionally, OpenAI and Anthropic generated combined annualized revenue of over $105 billion by August 2026. Analysts suggest the industry may enter a digestion phase that shifts value from hardware providers to software monetization.