IT budgets grow 13.5% in 2026. So does cloud waste

2026 is a record year for IT budgets. Gartner expects companies and governments to spend $6.31 trillion on IT, 13.5% more than in 2025. Most of the growth goes to AI: servers, data centres and software.

Growth of IT spending in 2026Total $6.31 trillion

+55.8%Data centre systems

+15.1%Software

+13.5%All IT

+8.2%Devices

More money is not the same as better IT. The same year, companies report that a growing share of their cloud spend buys nothing.

The leak in the cloud bill

In Flexera’s 2026 State of the Cloud survey, companies estimated that 29% of what they spend on cloud infrastructure and platforms is wasted. That is up from 27% a year earlier, the first rise after five years of slow improvement.

Of every 100 euros spent on cloud infrastructure

29%are wasted, up from 27% a year earlier

85%call cloud costs their top cloud challenge

The reason is not that teams stopped caring. 85% of companies call managing cloud costs their top cloud challenge, ahead of security. And 63% already have FinOps practices, the discipline of managing cloud spending. The problem is complexity: AI services are new, expensive and billed in new ways, and use of generative AI in the public cloud rose from 50% to 58% of companies in a year.

Where the money usually goes

• Servers nobody turned off. Test environments, old projects and machines sized for a peak that happened once.
• Paying full price. Steady workloads billed by the hour instead of with a one- or three-year commitment.
• Storage that only grows. Old backups, snapshots and logs nobody will ever read.
• AI without a budget. Pilots that run on the most expensive models with no limit and no owner.
• Licences for people who left. The software side of the same problem.

Five checks for this quarter

1. Give every cost an owner. Every account, project and resource should have a name next to it. Costs without an owner are the first to cut.
2. Turn off what is idle. Look for servers with very low use over the last 30 days and switch them off or make them smaller.
3. Commit where usage is steady. For workloads that run all year, commitments usually cost much less than paying by the hour.
4. Put a limit on AI. Each AI project gets a monthly budget, an alert and someone who answers for it.
5. Compare cloud with your own hardware. For stable, heavy workloads, your own servers or a private cloud can cost less. Decide with numbers, not habit.

Where to start

An infrastructure review puts the whole bill on one page: what you pay for, what you use and what you can switch off this month. For our clients the typical result is up to 30% lower IT costs with 99.9% uptime.

Sources: Gartner, April 2026 IT spending forecast; Flexera 2026 State of the Cloud Report and Flexera blog.