Artificial intelligence may dominate technology headlines, but it has yet to become the primary revenue engine for the world's largest cloud providers. According to a recent Jefferies report, AI-related services currently account for only about 15% of the cloud revenues for major US hyperscalers, leaving a substantial 85% derived from traditional, non-AI cloud offerings.
The Cloud's Enduring Core Strength
While AI captures much of the attention, the underlying cloud business continues its robust expansion. In Q2 2026, Microsoft, Google, and Amazon collectively generated an impressive $126 billion in cloud revenue, marking a 38% year-on-year increase. This strong performance in core cloud services provides a critical foundation, enabling hyperscalers to fund their massive investments in AI infrastructure.
AI's Explosive Growth Trajectory
The relatively smaller share of AI revenue should not be mistaken for weak performance; rather, it highlights the rapid growth from a nascent base. Amazon CEO Andy Jassy reported that AWS was operating at a $169 billion annualized revenue run rate in Q2 2026, with its AI revenue run rate exceeding $25 billion, representing roughly 15% of AWS's total. Microsoft presents a similar picture, with its AI business surpassing a $37 billion annual revenue run rate, a remarkable 123% increase year-on-year. Jefferies estimates AI now constitutes about 33% of Azure revenue and approximately 16% of Microsoft's overall cloud revenue.
Massive Investments for Future Growth
This dynamic—a large, established cloud business alongside rapidly expanding AI—is crucial as hyperscalers embark on an extraordinary infrastructure spending cycle. Their combined capital expenditure reached $165 billion in Q2 2026, which led to aggregate free cash flow dropping to $7 billion, with Jefferies anticipating a negative turn in Q3. The investment thesis hinges on these two businesses evolving in tandem: the current cloud revenue base supporting today's operations, and AI becoming the next major growth engine to justify the substantial infrastructure outlays currently underway.
The Critical Test Ahead
The core question for the AI industry remains: how quickly can AI revenues scale to adequately support the enormous infrastructure being built today? Early indicators are encouraging. Jefferies projects that the S&P 500 AI basket could deliver a 48% annualized Earnings Per Share (EPS) growth in 2026-27, more than double the 23% growth expected for the broader index. However, the report underscores that the AI narrative is still deeply intertwined with and dependent on the much larger, more mature cloud business.