Gartner tech spending 2026 forecast rises as AI costs hit customers
Gartner raised its 2026 tech spending forecast to $6.37 trillion as AI infrastructure costs flow into software, hardware and cloud prices.
By Dominic Okoye · Staff Writer
· 3 min read
Gartner tech spending 2026 estimates now point to $6.37 trillion in global outlay, up 14.2% year on year, as AI infrastructure investment pushes vendor spending and customer prices higher. John-David Lovelock, a Distinguished VP Analyst at Gartner, told The Register that technology companies’ own technology spending is already about $1 trillion and is expected to rise 34.7% in 2026.
The revised number is higher than Gartner’s prior estimates this year: $6.31 trillion in April and $6.15 trillion in February. Gartner attributes the uplift to faster overall growth, but the details are uneven. Some of the increase reflects more units and more cloud demand. Some reflects vendors passing through higher input costs.
What is driving Gartner's tech spending 2026 forecast?
The main driver is the AI infrastructure buildout by technology companies, especially spending to equip datacenters for expected demand from AI products and services. Gartner’s figures do not include the datacenter buildings themselves or cooling systems, which means the full capital burden around AI capacity is broader than the spending total captures.
Lovelock described the AI infrastructure program to The Register as larger than major historical infrastructure projects, including US highways and European rail. He said the industry is moving from spending on information technology toward spending on what he called “intelligence technology.” That framing is Gartner’s view, but the dollars behind it are already showing up in customer budgets.
Infrastructure as a service is the clearest beneficiary. Gartner said IaaS, one segment of cloud computing, is on track to grow 29.3% this year to $287 billion. The market grew 25.3% in 2025, according to Gartner, so the category is accelerating as providers prepare capacity for AI workloads.
Other categories are moving more slowly. Devices, including both consumer hardware and business laptops, are forecast to grow 9.8%. Gartner said a meaningful part of that increase comes from higher prices as memory and chips become more expensive. Services are expected to grow 5.3%, while telecoms are forecast at 4.4%.
How are enterprise customers paying for AI infrastructure?
Enterprise customers are seeing higher software and hardware prices as vendors add AI features and cover the cost of new infrastructure. Lovelock told The Register that CIOs are “extremely concerned” about price increases from vendors and are resisting where they can.
Gartner’s view is that buyers have had the most success pushing back in IT services. Lovelock said that when service providers add AI to their offerings, customers are rewarding them with lower price points rather than accepting an automatic premium. That suggests AI packaging alone is not enough to raise prices in every part of the market.
Software is a more difficult category for buyers. Enterprise software companies are adding AI capabilities to existing products and partnering with model providers such as OpenAI and Anthropic. The source data does not disclose how much of vendors’ AI-related price increases is tied to model costs, infrastructure costs, margin protection or new revenue.
There are open questions about whether customers will keep absorbing these increases. Lovelock pointed to the possibility that some AI spending may be defensive, rather than a clean new revenue stream. He cited Google’s addition of Gemini to search as an example where AI could be viewed as protecting an existing market position from AI challengers rather than creating incremental revenue.
Customers and developers are also responding to model pricing. Several model builders have moved from capped subscriptions to usage-based billing, according to the report. Lower-cost models from China are entering the market, and developers are using open source models where appropriate to reduce reliance on proprietary foundation models.
The unresolved question is whether the tech industry can keep funding the AI infrastructure buildout if customers resist the price increases needed to pay for it. Lovelock told The Register that this remains “the big open question” and said it has not been investigated or answered well.
This story draws on original reporting from The Register.
