Gartner has crunched the numbers on AI datacenter investment and found a widening gap between spending and revenue. The firm estimates AI providers will spend close to $1 trillion on infrastructure in 2026 and another $6 trillion between 2027 and 2030. To make that capital pay off at a 12% return on invested capital, operators would need to generate $8.5 trillion in revenue over the period. Gartner's forecast puts actual enterprise and consumer AI revenue at no more than $3.6 trillion.
The shortfall raises the risk of large asset write-downs, especially as newer GPUs devalue older hardware. Gartner expects the pain to fall hardest on neoclouds, the GPU-as-a-service providers that specialize in AI compute, predicting that 80% of them will be acquired or shut down by 2030. Hyperscalers are seen as better positioned because their businesses are more diversified beyond AI services.
The report adds to growing unease in financial markets over AI capex. Bain & Company has estimated the industry needs $6 trillion in annual revenue by 2031 to sustain infrastructure buildout, while Jefferies has flagged pricing pressure from open-source and Chinese models. The IMF and Bank for International Settlements have also warned about the potential for an AI bubble to spill into the broader economy.