Cloud market grows 25 per cent as enterprises move beyond AI pilots
Global spending on cloud infrastructure services reached a record $102.6 billion in the third quarter of 2025, according to new research from Omdia, as enterprises shifted their focus from early AI experimentation toward large-scale production deployment.
This performance represents a 25 per cent year-on-year growth and marks the fifth consecutive quarter where the sector has maintained a growth rate above 20 per cent, highlighting the underlying strength of the technology landscape.
The shift in demand has prompted major hyperscalers to move away from competing solely on incremental model performance, focusing instead on platform-level capabilities that support multi-model strategies and the reliable operation of AI agents in real-world environments.
AWS, Microsoft Azure, and Google Cloud maintained their market rankings during the quarter, collectively accounting for 66 per cent of all global spending and delivering a combined 29 per cent year-on-year growth.
Amazon Web Services (AWS) led the market with a 32 per cent share, reporting its strongest performance since 2022 as revenue growth reaccelerated to 20 per cent amidst easing compute supply constraints and high demand from its Anthropic partnership.
Microsoft Azure followed as the second-largest provider with a 22 per cent market share and a robust 40 per cent revenue jump, bolstered by its renewed partnership with OpenAI and the expansion of its Azure AI Foundry ecosystem.
Google Cloud maintained its third-place position, delivering a 36 per cent growth rate and increasing its market share to 11 per cent, driven largely by its enterprise-grade AI offerings and a backlog that reached $157.7 billion by the end of September.
Order backlogs across all three major providers continued to rise, reinforcing a healthy demand environment as companies move toward the scaled deployment of enterprise-grade applications.
“Collaboration across the ecosystem remains critical,” said Rachel Brindley, Senior Director at Omdia, regarding the current state of the market.
Multi-model support is increasingly viewed as a production requirement rather than a feature, as enterprises seek resilience, cost control, and deployment flexibility across generative AI workloads, the director added.
The complexity of real-world deployment has also led hyperscalers to step up investment in agent build-and-run capabilities, providing standardized building blocks that many enterprises currently lack.
“Many enterprises still lack standardized building blocks that can support business continuity, customer experience, and compliance at the same time, which is slowing the real-world deployment of AI agents,” said Yi Zhang, Senior Analyst at Omdia.
This is where hyperscalers are increasingly stepping in, using platform-led approaches to make it easier for enterprises to build and run agents in production environments, the analyst noted.
Recent launches such as AWS AgentCore and Microsoft’s Agent Framework reflect this direction, aimed at helping firms operate AI agents more efficiently.
AWS specifically strengthened its stack during its re:Invent 2025 event, introducing the Nova 2 model family and frontier agents like Kiro, which acts as a virtual developer for software teams.
Microsoft has seen similar success with its agent framework being utilized by customers like KPMG to enhance audit processes, while Google Cloud launched Gemini Enterprise in October 2025 to integrate its model family with no-code development tools.
Expansion into new regions also continued during the quarter, with AWS launching a new zone in New Zealand and Microsoft announcing infrastructure plans for Malaysia and India.
As the market enters 2026, the focus remains on transforming satisfied customers into vocal advocates through trusted digital stewardship and the delivery of hyper-personalised services.