Global spending on cloud infrastructure services reached USD90.9 billion in Q1 2025, marking a 21% year-on-year increase, according to Canalys estimates.
Enterprises have recognised that deploying AI applications requires a renewed emphasis on cloud migration. Large-scale investment in both cloud and AI infrastructure remains a defining theme of the market in 2025.
Meanwhile, to accelerate the enterprise adoption of AI at scale, leading cloud providers are intensifying efforts to optimise infrastructure, most notably through the development of proprietary chips aimed at lowering the cost of AI usage and improving inference efficiency.
In Q1 2025, the ranking of the top three cloud providers (AWS, Microsoft Azure, and Google Cloud) remained unchanged from the previous quarter, with a combined market share accounting for 65% of global cloud spending. Collectively, the three hyperscalers recorded a 24% year-on-year increase in cloud-related spending.
Growth momentum diverged among the top players. Microsoft Azure and Google Cloud both maintained growth rates of over 30% (although Google Cloud’s growth slowed slightly from the previous quarter), while AWS grew by 17%, a deceleration from 19% growth in Q4 2024.
This deceleration was largely driven by supply-side constraints, which limited the ability to meet rapidly rising AI-related demand. In response, cloud hyperscalers have continued to invest aggressively in AI infrastructure to expand capacity and position themselves for long-term growth.
Overall, the global cloud services market sustained steady growth in Q1 2025 as enterprises sharpened their focus on two strategic priorities: accelerating cloud migration, either by shifting additional workloads or reviving stalled on-premises transitions, and exploring the adoption of generative AI. The rise of generative AI, which relies heavily on cloud infrastructure, has, in turn, reinforced enterprise cloud strategies and hastened migration timelines.
“As AI transitions from research to large-scale deployment, enterprises are increasingly focused on the cost-efficiency of inference, comparing models, cloud platforms, and hardware architectures such as GPUs versus custom accelerators,” said Canalys Senior Director Rachel Brindley. “Unlike training, which is a one-time investment, inference represents a recurring operational cost, making it a critical constraint on the path to AI commercialisation.”
“Many AI services today follow usage-based pricing models—typically charging by token or API call—which makes cost forecasting increasingly difficult as usage scales,” added Canalys Analyst Yi Zhang. “When inference costs are volatile or excessively high, enterprises are forced to restrict usage, reduce model complexity, or limit deployment to high-value scenarios. As a result, the broader potential of AI remains underutilised.”
To address these challenges, leading cloud providers are deepening their investments in AI-optimised infrastructure. Hyperscalers including AWS, Azure, and Google Cloud have introduced proprietary chips such as Trainium and TPU, and purpose-built instance families, all aimed at improving inference efficiency and reducing total cost of AI.
Amazon Web Services (AWS) maintained its position as the market leader in Q1 2025, capturing 32% of global market share and recording a 17% year-over-year increase in revenue. Its AI business continues to grow at a triple-digit annual rate, though it remains in the early stages of development.
In March, AWS introduced a price-cutting strategy to promote the adoption of its Trainium AI chips over more costly NVIDIA-based solutions, highlighting Trainium 2’s 30–40% price-performance advantage. The company also accelerated the expansion of its Bedrock service, adding Anthropic’s Claude 3.7 Sonnet and Meta’s Llama 4 models, and became the first cloud provider to fully manage DeepSeek R1 and Mistral’s Mixtral Large.
Further underscoring its long-term commitment to global infrastructure, AWS announced a capital investment of over US$4 billion in May 2025 to establish a new cloud region in Chile by the end of 2026.
Microsoft Azure remained the second-largest cloud provider in Q1 2025, holding a 23% market share and delivering strong year-over-year growth of 33%. Microsoft reported a 16-point growth rate lift to Azure from AI, marking the largest single-quarter uplift since Q2 2024.
In April, Azure announced the availability of the GPT-4.1 model series on both Azure AI Foundry and GitHub, further broadening developer access to advanced AI capabilities across its ecosystem. Azure AI Foundry, Microsoft’s platform for building and managing AI applications and agents, is now used by developers at more than 70,000 enterprises. The platform processed over 100 trillion tokens this quarter, a fivefold increase year-over-year.
Microsoft has also focused on lowering the cost of AI adoption, reporting a nearly 30% improvement in its AI performance at constant power consumption and a reduction of over 50% in cost per token. As part of its ongoing global infrastructure expansion, it opened new data centres in 10 countries across four continents during Q1.
Google Cloud, the world’s third-largest cloud provider, maintained a 10% market share in Q1 2025 and delivered strong year-over-year growth of 31%. As of March 31, its revenue backlog reached USD92.4 billion, marking a slight decline from the previous quarter.
This decrease was primarily attributed to supply constraints, particularly in compute capacity, that limited Google Cloud’s ability to fully meet customer demand. In March, Google introduced the Gemini 2.5 model series, with Gemini 2.5 Pro receiving widespread acclaim for its leading benchmark performance and top ranking on Chatbot Arena. With enhanced reasoning and coding capabilities, the model opens new possibilities for both developers and enterprise users.
Since the beginning of the year, active usage of Google AI Studio and the Gemini API has surged by over 200%, reflecting strong developer adoption and growing demand for generative AI solutions. Google also launched a new cloud region in Sweden (it’s 42nd globally) and committed USD7 billion to expand its Iowa data centre, further supporting its growing AI and cloud workloads.
Canalys defines cloud infrastructure services as the sum of bare-metal-as-a-service, infrastructure-as-a-service, platform-as-a-service container-as-a-service and serverless that are hosted by third-party providers and made available to users via the internet.
