The AI Gold Rush's Hidden Cost: Cloud Giants Claim a Major Share
A new analysis from Barclays sheds light on a critical financial dynamic in the artificial intelligence sector: a substantial portion of AI companies' revenue doesn't stay with them.
Nearly 40 Cents on the Dollar Goes to Compute
For every $100 in revenue generated by AI model companies, approximately $35 to $40 flows out as payment for "inference" computing power to the three leading cloud providers: Amazon Web Services, Microsoft Azure, and Google Cloud Platform. Inference refers to the computational resources consumed when a trained AI model processes live user requests to generate outputs.
This positions cloud infrastructure as one of the single largest cost centers for AI firms, an expense that scales directly with user adoption of their applications.
Cloud Providers Reap High Margins as AI Labs See Surge
The business is highly lucrative for the cloud giants. From that $35-$40 stream, they capture an estimated $10 to $20 in operating profit, translating to operating margins between 35% and 45%. This underscores their powerful position in the AI compute supply chain.
Simultaneously, the margins for AI labs offering paid inference services have skyrocketed. Their profitability has jumped from low double-digits to a range of 50% to 65% or higher, with adjusted gross margins surging by 30 to 50 percentage points year-over-year.
The Peak May Not Last: Intensifying Competition Ahead
Barclays analysts note that the reported margin estimates might even be conservative, with actual figures potentially higher. However, this peak profitability is unlikely to be permanent.
As competition among frontier AI models intensifies and the global supply of computing power continues to expand, market dynamics are expected to shift. Analysts project that margins—for both cloud providers and AI labs—will gradually moderate from their current highs, leading to a new equilibrium in industry profit sharing.
The report highlights a fundamental truth of the current AI boom: the providers of the foundational "picks and shovels"—cloud infrastructure—may be securing a more stable and profitable position than many of the "miners" themselves.