Microsoft Raises 2026 AI CapEx to $190 Billion as Infrastructure Costs Surge

Microsoft has revised its 2026 capital expenditure forecast upward to $190 billion, marking one of the largest infrastructure spending commitments by any single company in technology history. The increase includes $25 billion attributed directly to surging memory and storage component prices, which have climbed sharply as demand for AI training and inference infrastructure outstrips supply chain capacity.

The numbers are staggering in context. Microsoft has spent $97 billion across the last four quarters and generated $37 billion in AI annual recurring revenue. The gap between spend and return — a $60 billion delta — is starting to attract Wall Street scrutiny, even as the company maintains that long-term AI infrastructure investments will pay off.

For the broader industry, Microsoft's spending is a leading indicator of where AI infrastructure costs are heading. The component price surge affects every builder: memory and storage represent 30-40% of the cost of an AI training cluster. If Microsoft is absorbing a $25 billion price increase, the impact on smaller operators and cloud customers downstream will be proportionally significant.

Meanwhile, Amazon disclosed that its custom silicon business — Graviton, Trainium, and Nitro combined — now exceeds a $20 billion annual run rate, growing 100% year-over-year. That makes AWS the third-largest datacenter chip business globally behind NVIDIA and AMD. The strategic implication: cloud providers are vertically integrating to control their own silicon supply chain, reducing dependence on NVIDIA and potentially stabilizing costs long-term.