Oil crisis? What oil crisis? IT spending de-coupled from wider war shock
Gartner sees accelerating growth in IT spending, powered by cloud and AI infrastructure investment
A day after the International Energy Agency (IEA) said the US/Israel/Iran war was creating the worst energy crisis ever faced by the world, Gartner increased its growth forecasts for global IT spending by nearly three percentage points.
The spike in the cost of oil and gas caused by the on-going conflict has no direct relationship to IT spending, argues John-David Lovelock, Distinguished VP Analyst at Gartner. There is possible indirect impact on both business and consumer confidence, but the numbers show little evidence of that yet, he told The Register.
Gartner forecasts worldwide IT spending will grow to 13.5 percent in 2026 to reach $6.31 trillion. In February, it forecast 10.8 percent growth to reach $6.15 trillion.
“The price of oil has very little to do with it,” Lovelock said. “The closest direct effect we can get is in some countries, the price of electricity will come up. There's some countries in Asia-Pacific where it's an issue. To say that a business or a technology provider is going to change their plans based on the price of electricity just isn't [true]. The only other effect that we could be looking for is business confidence change, and we don't see that occurring either.”
However, outside of cloud infrastructure and AI spending, IT investment is relatively normal. Spending on consumer IT is growing at only 4.1 percent and has been hit by the increase in memory costs.
Even enterprise IT spending growth was in the range of 7 percent. “We only get the 13.5 percent when we include that being spent by the hyperscalers. Without the datacenter spend, without the AI optimized servers, overall spend is growing 7.2 percent this year,” he said.
“All the extra spending is the tech providers, the hyperscalers, still building out these datacenters, still building new foundational technology together for us to use for AI and we're still building the AI models,” he said.
Enterprise IT users are likely to pay for that investment in the long run. They are set to come out of 2026’s “trough of disillusionment” in terms of AI projects.
- Iran claims US used backdoors to knock out networking equipment during war
- AI is reshaping Britain's datacenter map away from London
- Investors are going nuclear to keep UK's AI datacenters fed
- Renewables reached nearly 50% of global electricity capacity last year
"We're probably out of it by the beginning of 2027; expectations are at their lowest point right now. CIOs are looking for [AI] software from their incumbent [software providers]. We are looking at smaller projects that help bring immediate improvement. From the AI agent point of view, this is the year where we see a lot more in the way of project failures. We're coming down towards the trough for agents. That's part of the cycle. The next cycle, which is getting to the plateau [of productivity] and finding out what we can do with these things and becoming reliable, predictable and stable. That's still a three-year cycle to come; it'll probably take us till 2030," he said.
However, the forecast depends on the assumption that the war is a "relatively short-lived conflict."
Only if the war continued throughout the year, and the world lost confidence in the economy, would some CIO budgets stall. However, even if that happens, "we're unlikely to see any pause in the technology providers' race [for AI] because they're on a four-year cycle here," Lovelock said. ®
More about
More about
Narrower topics
- AIOps
- AWS
- Azure
- Cloud Migration
- Cloud native
- Content delivery network
- DeepSeek
- Digital Ocean
- Disaster recovery
- EC2
- Edge Computing
- FinOps
- Fusion Power
- Gemini
- Google AI
- Google Cloud Platform
- GPT-3
- GPT-4
- G Suite
- Hybrid Cloud
- IaaS
- iCloud
- Kubernetes
- Large Language Model
- Machine Learning
- MCubed
- Multicloud
- Neural Networks
- NLP
- Nuclear Power
- Open Compute Project
- OpenStack
- Paas
- Private Cloud
- Public Cloud
- PUE
- Retrieval Augmented Generation
- Serverless
- Software defined data center
- Star Wars
- Systems Approach
- Tensor Processing Unit
- TOPS
- Virtualization
- vSphere
Broader topics
More about
More about
More about
Narrower topics
- AIOps
- AWS
- Azure
- Cloud Migration
- Cloud native
- Content delivery network
- DeepSeek
- Digital Ocean
- Disaster recovery
- EC2
- Edge Computing
- FinOps
- Fusion Power
- Gemini
- Google AI
- Google Cloud Platform
- GPT-3
- GPT-4
- G Suite
- Hybrid Cloud
- IaaS
- iCloud
- Kubernetes
- Large Language Model
- Machine Learning
- MCubed
- Multicloud
- Neural Networks
- NLP
- Nuclear Power
- Open Compute Project
- OpenStack
- Paas
- Private Cloud
- Public Cloud
- PUE
- Retrieval Augmented Generation
- Serverless
- Software defined data center
- Star Wars
- Systems Approach
- Tensor Processing Unit
- TOPS
- Virtualization
- vSphere


