Friday, November 13, 2026
Explainer

Cloud computing trends for 2026: AI demand, hybrid by default and cost control

Cloud spending is growing faster than it was before the AI boom. The trends that matter for IT budgets in 2026 all trace back to that.

A rack of servers in a server room

Gartner's forecast for 2025 put worldwide public cloud spending at $723.4 billion, growth of 21.5%, and its analysts named AI as the force accelerating the role of cloud computing. Four trends follow from that for IT leaders planning 2026 and 2027 budgets.

Worldwide public cloud spending by segment$ billion250.8299.1SaaS171.6208.6PaaS169.8211.9IaaS20242025Source: Gartner forecast, November 2024
Chart built from the figures cited in this story.

1. Infrastructure is growing fastest

Gartner's forecast broke spending down by service type:

Segment 2024 spending 2025 forecast 2025 growth
Infrastructure (IaaS) $169.8 billion $211.9 billion 24.8%
Platform (PaaS) $171.6 billion $208.6 billion 21.6%
Software (SaaS) $250.8 billion $299.1 billion 19.2%
Desktop as a service $3.5 billion $3.8 billion 11.1%
Total $595.7 billion $723.4 billion 21.5%

Infrastructure and platform services, where AI training and inference run, are growing faster than SaaS. For buyers, that means GPU capacity and AI platform services are where prices and availability will be tightest.

2. Hybrid is the default

Gartner predicts 90% of organizations will adopt a hybrid cloud approach through 2027. The practical question is no longer whether to use more than one environment but how to run them consistently: one identity system, one set of security policies and one view of costs across all of them.

3. FinOps has absorbed AI and SaaS

The FinOps Foundation's State of FinOps 2026 survey found 98% of respondents manage AI spend in some form, according to Kion's summary, and more than half plan to bring licensing and private cloud into scope within a year. Cost management is becoming a technology-wide practice rather than a cloud-only one.

4. Power is now a limit

The International Energy Agency projects data center electricity use will more than double between 2024 and 2030, to around 945 terawatt-hours, with AI as the main driver. Power and grid connections now constrain where providers can add capacity. Expect that to show up as regional capacity limits and as pressure on the price of AI compute.

What IT teams should do

  • Budget AI compute as its own line, with usage-based forecasts.
  • Standardize identity, security policy and cost reporting across every cloud you use.
  • Bring SaaS and AI spending into the same cost reviews as cloud infrastructure.
  • Ask providers about capacity in your preferred regions before committing to large AI projects.

Sources