Three hyperscalers, three capacity strategies
Amazon Web Services, Google Cloud, and Microsoft Azure have each committed markedly different levels of capital for the next 12 months in their latest filings and earnings call.
While Amazon CEO Andy Jassy spoke of spending $200 billion in 2026 on AI, chips, and possibly low-orbit satellites as data centers, Google CFO Anat Ashkenazi said during an earnings call that the company will commit around $180 billion to replacing aging servers and building new data centers.
Microsoft, which operates on a July–June fiscal year, has yet to formally disclose its total capital expenditure plans through June 30, 2026. The company reported capex of $34.9 billion in its first fiscal quarter, and $37.5 billion in second. Its CFO, Amy Hood, has indicated that capital spending is expected to moderate in the coming quarters, leading industry analysts to revise their estimates for Microsoft’s full-year capex to around $100 billion.
Each of these figures reflect distinct priorities of hyperscalers, especially how they are preparing their cloud platforms for the next phase of AI-driven demand.
AWS is using capital expenditure to lock down the physical constraints that will shape future cloud capacity, including power, silicon, land, and water, Gogia said, adding that this signals a move beyond incremental expansion toward utility-scale infrastructure, signaling a strategy aimed at institutionalizing AI demand rather than simply responding to it.
Microsoft and Google, however, are taking more targeted approaches.
