The AI reality check: 85% of hyperscalers' cloud revenue is still non-AI

Artificial intelligence may dominate the technology industry's headlines, but it has not yet taken over the business that pays the bills.

According to Jefferies, AI-related revenue currently accounts for only around 15% of the cloud revenues of the major US hyperscalers, leaving roughly 85% coming from non-AI cloud services.

That is an important reality check at a time when investors are increasingly treating artificial intelligence as the central growth engine for Big Tech.

The underlying cloud business, meanwhile, is growing rapidly.

Microsoft, Google and Amazon generated combined cloud revenue of $126 billion in Q2 2026, up 38% year-on-year, according to a report by Jefferies’ Global Head of Equity Strategy Christopher Wood. The strength of this broader cloud business is one of the key reasons investors believe the hyperscalers can eventually generate adequate returns from their massive AI infrastructure investments.

AI is growing fast

The fact that AI represents a relatively small share of cloud revenue should not be mistaken for weak growth.

MUST READ | AI boom's hidden bill: Big Tech's spending hits $165 billion as free cash flow shrinks

Amazon CEO Andy Jassy said AWS was running at a $169 billion annualised revenue run rate in Q2, with its AI revenue run rate exceeding $25 billion. That puts AI at about 15% of AWS revenue.

Microsoft's numbers tell a similar story.

The company said its AI business had surpassed a $37 billion annual revenue run rate, up 123% year-on-year, while Azure revenue crossed $100 billion for the first time in the last fiscal year. Jefferies estimates AI accounted for about 33% of Azure revenue and around 16% of Microsoft's overall cloud revenue.

The numbers suggest that AI is growing extremely quickly, but the existing cloud ecosystem remains much larger.

That distinction matters because the hyperscalers are simultaneously embarking on an extraordinary infrastructure spending cycle.

DON'T MISS | AI may reshape Indian jobs, but physical work could offer some protection, Goldman Sachs says

Their combined capex reached $165 billion in Q2 2026, while aggregate free cash flow fell to $7 billion. Jefferies expects the latter to turn negative in Q3.

So the investment case rests on two businesses moving together: the established cloud business providing today's revenue base, and AI becoming the next major growth engine capable of eventually supporting the infrastructure spending now underway.

The bigger test

The AI industry therefore faces a deceptively simple question: how quickly can AI revenues become large enough to justify the enormous infrastructure being built today?

DO CHECKOUT | Meta's Mark Zuckerberg says AI creates jobs as data centres expands

For now, the answer remains encouraging. Jefferies estimates the S&P 500 AI basket could deliver 48% annualised EPS growth in 2026-27, more than twice the expected 23% growth for the broader index.

But the report also makes clear that the AI story is still intertwined with the much larger, older cloud business.