Global spending on cloud infrastructure services reached $76.5 billion in the most recent quarter, marking a 23% increase year-over-year as enterprises accelerate migration of artificial intelligence workloads to hyperscale data centres, according to new figures from Synergy Research Group.
The surge, the sharpest quarterly rise in three years, was driven primarily by demand for GPU-accelerated compute instances capable of training and running large language models. Amazon Web Services maintained its market lead at 31%, while Microsoft Azure closed the gap to 25% on the back of its deepening OpenAI partnership.
“We are witnessing a structural shift in how enterprises think about infrastructure,” said Maria Chen, principal analyst at Cloud Economics Forum. “AI is not just another workload — it is reshaping the economics of data centre design, from power density to liquid cooling requirements.”
The build-out is putting pressure on an already strained supply chain for high-bandwidth memory and advanced networking silicon. Lead times for NVIDIA H100 and Blackwell GPUs remain at 12 to 16 weeks, prompting cloud providers to sign multi-year procurement agreements worth tens of billions of dollars.
Europe saw the fastest regional growth at 27%, fuelled by sovereign cloud initiatives and the European Union’s AI Act compliance requirements. A growing number of regulated industries, including financial services and healthcare, are opting for in-region cloud deployments to satisfy data residency mandates.
Industry observers note that the spending trajectory shows no signs of slowing. Gartner projects that worldwide IT services spending will surpass $1.5 trillion in 2026, with cloud infrastructure and platform services accounting for the largest share of new investment.
“The next phase will be about efficiency,” added Chen. “As CFOs scrutinise AI budgets more closely, cloud providers that can demonstrate measurable return on investment will separate themselves from the pack.”