Hyperscaler Capex Surge: The $690 Billion AI Infrastructure Sprint
In early 2026, the core narrative supporting the artificial intelligence boom—aggressive capital expenditure by the world's largest technology companies—remains not only intact but is accelerating. The five largest U.S. cloud and AI infrastructure providers (Amazon, Alphabet, Microsoft, Meta, and Oracle) have collectively guided to record-breaking capital expenditures of between $660 billion and $690 billion for calendar year 2026. This represents a near-doubling of the aggregate $380 billion spent in 2025.
The primary constraint on this massive buildout is no longer demand, but physical infrastructure: power availability, site permitting, and hardware supply. For instance, Microsoft has disclosed a massive $80 billion backlog of Azure orders that it cannot fulfill due to power constraints, rather than a lack of market demand. Hyperscalers consistently report that their capacity is being fully monetized as soon as it is brought online. This unprecedented wave of capital spending directly feeds the record-breaking revenues of merchant chipmakers like Nvidia (see Nvidia's Record Q1 FY2027: Parabolic Demand Driven by Agentic AI) and custom ASIC designers like Broadcom (see Broadcom's Q2 FY2026: Reaffirming a $100 Billion AI Runway for FY2027).
Key Quotes
"The five largest US cloud and AI infrastructure providers – Microsoft, Alphabet, Amazon, Meta, and Oracle – have collectively committed to spending between $660 billion and $690 billion on capital expenditure in 2026, nearly doubling 2025 levels." — Nick Patience, The Futurum Group
"Microsoft is tracking toward $120 billion or more in fiscal 2026, having already spent $37.5 billion in its most recent quarter alone. The company disclosed an $80 billion backlog of Azure orders that cannot be fulfilled due to power constraints, suggesting demand is outpacing even its aggressive build-out pace." — Nick Patience, The Futurum Group
Interpretation
This massive capital allocation indicates that the "AI capex bubble" fears of late 2024 and 2025 have been overridden by a competitive sprint to secure compute capacity. Hyperscalers are treating GPU and data center capacity as a strategic, managed risk rather than a procurement commodity. The fact that demand remains supply-constrained (particularly by power grid capacity) suggests that the revenue runway for hardware providers like Nvidia remains highly robust through 2026 and into 2027.