Global cloud market surpasses $107 billion in Q3: AWS share drops to 29%, Microsoft and Google hold their ground.

Introduction

In the third quarter of 2025, the global cloud infrastructure services market delivered a stunning report card: total revenue reached $107 billion, up 28% year-over-year, and up $7.6 billion quarter-over-quarter, marking the largest single-quarter increase ever. Data released by Synergy Research Group shows that AWS, Microsoft Azure, and Google Cloud collectively hold a 62% market share, but the competitive landscape is quietly shifting—AWS's share fell further from 31% in the same period last year to 29%, while Microsoft and Google remained stable at 20% and 13%, respectively. This is not just a change in numbers; it also reflects the deep structural evolution of the cloud market driven by AI computing power. For enterprise IT decision-makers, understanding the drivers behind these changes is more important than remembering the market shares themselves.

Background: AI Ignites the Growth Engine of the Cloud Market

After experiencing a growth slowdown in 2023, the global cloud computing market re-entered a high-growth trajectory in 2024-2025. John Dinsdale, chief analyst at Synergy Research Group, said that the year-over-year growth rate in Q3 2025 has risen for the eighth consecutive quarter, and is more than 10 percentage points higher than in the same period of 2023. For a market already exceeding $100 billion in size, this acceleration is unusual.

The core force driving this growth is undoubtedly artificial intelligence. In his analysis, Dinsdale noted that AI is "deeply imprinted" on these record-breaking numbers. Generative AI has not only directly driven demand for specialized services such as GPUaaS (GPU as a Service)—whose revenue growth rate has exceeded 200% annually—but has also indirectly boosted consumption of traditional cloud services. From model training to inference deployment, from data analysis to agent applications, AI workloads are becoming the largest incremental source of cloud resource consumption.

At the same time, cloud vendors' capital expenditures on AI infrastructure continue to rise. AWS generated $33 billion in revenue this quarter, up 20% year-over-year; Microsoft Intelligent Cloud reported quarterly revenue of $30.9 billion, up 28%; and Google Cloud generated $15.2 billion in revenue, up 34%. Despite strong growth, it is worth noting that AWS's revenue growth rate was lower than the market average of 28%, directly leading to a further contraction in its market share.

Technical Analysis: GPUaaS and "AI-Native" Cloud Architecture

The explosive growth of the cloud market essentially stems from a shift in computing paradigm. Traditional cloud services are dominated by general-purpose computing (CPU), while the AI era poses massive demand for high-performance computing (GPU/TPU). The emergence of GPUaaS allows enterprises to rent computing power on demand without making a one-time huge investment in purchasing GPU clusters. This model greatly lowers the barrier to AI experimentation and production, and is especially favored by SMEs and startups.From an architectural perspective, GPUaaS is giving rise to an “AI-native” cloud service model. For example, Kubernetes clusters are beginning to natively support GPU scheduling, serverless platforms have added AI inference functions, and object storage has optimized throughput performance for training datasets. The direct impact of these underlying changes on enterprises is that AI infrastructure that once required a specialized MLOps team to manage can now be invoked like an ordinary API.

However, technological convenience also brings new challenges. First, although GPU compute costs are pay-as-you-go, sustained long-term use is no cheaper than building your own cluster. Enterprises need to establish cost models to evaluate combinations of peak load and reserved instances. Second, AI model training involves a large amount of data movement; if the cloud provider’s data transfer and storage services are not efficient enough, training efficiency will decline. Finally, security and compliance pressures multiply—AI models may involve sensitive data, and enterprises must ensure that cloud platforms meet data privacy and sovereignty requirements.

Enterprise Impact Analysis: Cost, Deployment, and Strategic Choices

For CIOs and CTOs, the Q3 data provides new evidence for a multi-cloud strategy.

In terms of cost, the growth of AI services means that the structure of enterprise cloud bills is changing. Traditionally, compute and storage were the main expenditure items; now, GPU instances and AI inference calls may quickly account for the bulk of budgets. CloudTechDaily recommends that enterprises place AI workloads in a separate cost center and use FinOps practices for fine-grained governance. For example, for non-real-time inference tasks, Spot instances can be used to reduce costs; for long-term training tasks, committed use contracts can be used to obtain discounts.

In terms of deployment, AWS’s declining share does not mean its capabilities are waning; rather, the market now offers more choices. Microsoft Azure, through its deep partnership with OpenAI, has a first-mover advantage in enterprise-grade AI applications (such as Copilot and Azure OpenAI Service); Google Cloud continues to innovate in data analytics and machine learning platforms (such as BigQuery and Vertex AI). Enterprises should avoid a one-size-fits-all approach to a single cloud and instead choose the optimal platform based on workload characteristics. At the same time, the role of containerization and platform engineering will become even more prominent, shielding underlying cloud differences and improving portability.

Security and compliance are equally important. Cross-border AI data flows, data sovereignty laws (such as GDPR and China’s Data Security Law), and industry regulatory requirements are all prompting enterprises to be more cautious in cloud selection. Regional cloud providers (such as Alibaba Cloud and Huawei Cloud) have compliance advantages in their home markets, while global cloud providers need to offer more localized regions and certifications. Enterprises should establish a compliance framework and regularly audit cloud vendors’ security practices and data residency capabilities.From a share perspective, AWS's global cloud market share in Q3 was 29%, down from 31% in the same period last year, while it was 32% in Q3 2023 and 34% in Q3 2022. This trend clearly shows that AWS is facing increasingly fierce competition, and its growth rate is no longer keeping pace with the overall market. Microsoft and Google, however, have held their ground through their AI technology strategies—Microsoft's share has remained at 20% for several consecutive quarters, while Google has stayed stable at 13%. Neither has achieved significant growth, but neither has lost ground either, indicating that their competitiveness in the AI era has been recognized by the market.

Alibaba Cloud ranks fourth with a 4% share, growing 26% year-over-year to $4.7 billion, maintaining its dominant position in the Chinese market. Oracle ranks fifth with a 3% share, with cloud revenue growing 28%, and its database and OCI performance in AI workloads is worth watching. In addition, Salesforce, IBM, Tencent, and Huawei each account for about 2%, while Akamai, Baidu, CoreWeave, Databricks, Snowflake, OpenAI, and others also appear in the 1% tier.

This pattern of "layered competition" means that the cloud market is no longer a single-dimensional scale race. The established giants are defending their share in general-purpose infrastructure, while emerging players are building moats in areas such as AI-specific compute, data platforms, and vertical SaaS. For enterprise customers, this is actually good news—more choices mean stronger bargaining power and a richer array of innovative solutions.Fourth, sovereign clouds and regional clouds will usher in development opportunities. Against the backdrop of geopolitics and data sovereignty, enterprises increasingly need a cloud environment that can "operate independently in a specific region." This is not only a compliance requirement for multinational enterprises, but also a strategic direction for countries to promote digital autonomy. The deep cultivation of the domestic market by Chinese cloud vendors (Alibaba Cloud, Huawei, Tencent), as well as the advancement of European sovereign cloud plans, both indicate that this trend is accelerating.

CloudTechDaily Insight

The global cloud market reached $107 billion in Q3 2025 and set a historic record in quarter-over-quarter growth, which is undoubtedly a direct reflection of the AI infrastructure investment boom. But CloudTechDaily believes that what deserves more attention in this data is the structural change: AWS's continued decline in market share is not so much a loss of competitiveness as it is the market repricing the value of "cloud" in the AI era. When AI computing power becomes a core resource, the competitive dimension of cloud vendors will shift from "service availability" to "AI capability density."

For enterprise IT strategy, this quarter's data provides three key insights: First, AI workloads should be placed at the core of cloud strategy rather than at the edge, which means budgets, architecture, and teams all need to be adjusted accordingly; Second, multi-cloud strategy is no longer an option but a necessary condition for reducing risk and ensuring flexibility; Third, when considering cloud service providers, in addition to price and performance, data sovereignty, sustainability, and ecosystem integration capabilities should also be incorporated into the evaluation matrix.

The next chapter of cloud computing has begun: AI-native, multipolar, and green and sustainable. Vendors that can build differentiation across these dimensions simultaneously will become the leaders of the future. Enterprises, in turn, need to build their digital infrastructure with a more dynamic and open perspective.