AI-related capital expenditure by major hyperscalers continued accelerating, raising important questions around valuation, returns and capital intensity.
The largest technology platforms, the hyperscalers that dominate cloud computing and AI services, have undergone a quiet but important transformation. For most of the past decade, these businesses were celebrated for combining strong revenue growth with modest capital requirements and abundant cash flows returned to shareholders. Now, the ongoing build-out of AI capability has fundamentally changed what they do with the cash they generate.
The scale of investment now underway is striking. AI-related capital expenditure across the major hyperscalers is expected to run into the high hundreds of billions of US dollars in 2026 alone. In some cases, individual companies are directing close to their entire operating cash flow back into infrastructure. The asset-light platforms of a few years ago are today morphing into utilities: deploying enormous capital upfront, against returns that are real but still some years away.
