Cloud Giants Set Nearly $595B Capex Pace As AI Capacity Lags
Amazon, Alphabet and Microsoft now point to roughly $595 billion in 2026 capital-spending measures as AI demand keeps cloud capacity tight, though The Register notes the figures use different accounting bases.
AI demand is turning cloud growth into a capital-spending contest.
The Register reported that Amazon, Alphabet and Microsoft now point to 2026 spending measures that add up to roughly $595 billion, even though the companies use different accounting bases and say capacity still trails customer demand.
The pressure is visible in both revenue and infrastructure guidance.
AWS took in $42.2 billion in the quarter ended June 30, 2026, up 36.7 percent year on year; Google Cloud reached $24.8 billion; and Microsoft commercial cloud revenue reached $59.3 billion.
Amazon Raises Its AI Infrastructure Plan
Amazon gave the clearest upward revision.
Andy Jassy told analysts that the company now expects about $220 billion in cash capital expenditure in 2026, up from an earlier plan of about $200 billion, with most of the spending tied to AI and AWS.
The higher figure is not presented as enough capacity.
Jassy told analysts that memory costs pushed the forecast higher and that Amazon still would not have enough capacity to meet all demand in 2026, with the same dynamic likely in 2027.
That constraint changes how enterprise buyers read the cloud cycle.
Faster AWS growth points to strong demand, but the company is also warning that memory supply and data-centre build-out remain bottlenecks rather than back-office cost lines.
Google And Microsoft Add Different Spending Signals
Alphabet also lifted its full-year 2026 capital-spending guidance.
Chief Financial Officer Anat Ashkenazi put the new range at $195 billion to $205 billion, compared with the previous $180 billion to $190 billion range, and tied the increase to faster capacity delivery for demand growth.
Ashkenazi framed the cloud unit as a supply-constrained business after three years of capacity additions.
Core GCP, AI solutions and AI infrastructure were among the drivers of Google Cloud's $24.8 billion quarterly revenue.
Microsoft's number is harder to compare directly.
The company put expected calendar 2026 capex at about $175 billion, after a lease-accounting shift moved more future data-centre leases into operating-lease treatment rather than finance leases.
Amy Hood described Microsoft's calendar-year 2026 investment expectations as otherwise unchanged outside the lease treatment.
Capital spending in the next quarter, Microsoft's Q1 FY27, should exceed $50 billion including the reclassification impact.
Supply Bottlenecks Move Into Enterprise Planning
The Register cautioned that the three forecasts are not like-for-like: Amazon cites cash capex across several businesses, Alphabet gives whole-company guidance and Microsoft's figure reflects lease accounting as well as direct expenditure.
Even with that caveat, the combined total shows how much infrastructure the largest cloud providers expect AI demand to require.
Hardware scarcity is part of the same story.
Memory, GPUs and hard disks remain in short supply, and Jassy argued that suppliers are prioritising the largest customers, pushing some companies with on-premises infrastructure toward cloud providers.
Cloud customers are therefore buying into a market that is expanding and constrained at the same time.
Provider differentiation now depends on who can secure enough memory, accelerators, power and leased data-centre capacity to turn AI demand into usable cloud supply.
