AI accelerates market past $100B quarterly cloud spend
The cloud computing market is growing faster than in previous years, setting new quarterly revenue highs and showing that generative AI (GenAI) workloads are driving most of the recent growth for major cloud providers. Enterprise spending on cloud infrastructure services (IaaS and PaaS) surpassed the $100 billion mark for the first time in Q3 2025, reaching $106.9 billion.
Q1–Q3 2025 cloud market data indicate a competitive market, where AWS remains the largest share, while Azure and Google Cloud are expanding at faster rates, primarily driven by investments in artificial intelligence AI and ongoing platform integration work.
Market share dynamics in 2025#
The “Big Three” collectively control approximately 62% of the global market. However, the momentum has clearly shifted toward players who can capitalize on the demand for high-performance AI compute.
Cloud Service Provider | Year-over-Year Growth (approx.) | |||
AWS | 32% | 30% | 29% | 17–20% |
Microsoft Azure | 23% | 20% | 20% | 27-33% |
Google Cloud (GCP) | 10% | 13% | 13% | 32–35% |
Alibaba Cloud | 4% | 4% | 4% | Stable |
Oracle Cloud (OCI) | 3% | 3% | 3% | Stable |
Amazon Web Services (AWS): AWS retains undisputed revenue leadership with a $132 billion annual run rate. Critically, its revenue growth re-accelerated to 20% year-over-year (Y-o-Y) in Q3 2025. This demonstrates that its aggressive capital expenditure on specialized AI infrastructure (like Trainium2 chips and liquid cooling) is translating into renewed sales momentum, pushing back against competitive pressure.
Microsoft Azure: Azure’s AI strategy is driving immense success in the enterprise. Its cloud services revenue grew robustly by 33% Y-o-Y (Q3 FY25), with AI services contributing a substantial 16 percentage points of that growth (48% of Azure’s total cloud growth). This strong performance validates its deep partnership with OpenAI and its unique ability to integrate AI directly into business workflows.
Google Cloud: GCP continues to be the fastest-growing major provider (34% Y-o-Y revenue growth to $15.2 billion in Q3). Most impressively, its operating income grew by 85% to $3.6 billion. This signals that its decade-long investment in proprietary AI (TPUs and Gemini models) is finally delivering high-margin, sustainable returns, supported by a growing contract backlog of $155 billion.
What’s driving the growth in 2025 and beyond#
The market re-acceleration in 2025 is a direct result of unprecedented
1. Amazon Web Services (AWS)#
AWS is executing a massive, all-in pivot on AI infrastructure to defend its market leadership.
Current Growth Drivers (2025) | Future Strategy (2026 and Beyond) |
Shift in compute resources: Amazon is directing the majority of its capital expenditure toward building AI-ready data centers and expanding GPU capacity. This massive scale-up is already translating into increased demand capture, helping AWS re-accelerate to 20% Y-o-Y growth in Q3 2025. | Proprietary silicon dominance: Aggressively deploying its custom chips, including the Trainium2 AI training chip and the Graviton4 general-purpose CPU. The goal is to achieve higher performance and better margins by controlling the entire hardware stack. |
Advanced cooling systems: To support AI’s intense power demands, AWS is standardizing liquid and immersion cooling across its newest facilities. This increases compute density per megawatt and ensures efficient monetization of generative AI workloads. | Doubling capacity by 2030: AWS has committed to investing $12.7 billion in cloud infrastructure in India by 2030, aiming to reduce the carbon footprint of IT workloads for Indian customers by up to 96% through the use of renewable energy. |
OpenAI partnership: AWS secured a massive partnership with OpenAI in Q4 2025. OpenAI will use AWS’s infrastructure as a “backbone” for its AI ambitions, providing a major revenue boost and critical validation of AWS’s capacity strategy. | AI utility focus: Shifting its identity from a mere cloud vendor to an AI utility. The goal is to secure a durable competitive advantage by controlling one of the most constrained resources in the AI ecosystem: specialized compute capacity, particularly high-performance AI accelerators. |
S3 enhancements: AWS launched S3 Tables to deliver the first cloud object store with built-in Apache Iceberg support, offering up to 3 times faster query performance and 10 times higher transaction throughput compared to self-managed Iceberg tables on S3. | Unified vector data pipeline: The S3 Vectors (still in preview) provide the first cloud object storage with native support for storing, indexing, and querying vector embeddings. This reduces RAG costs by up to 90% compared to traditional vector databases for less-frequently accessed data. The S3 Vectors will become the durable, cost-optimized “cold tier” for all large-scale vector datasets and AI agent memory, eliminating vector database overhead for most batch and retrieval workloads. |
2. Microsoft Azure#
Azure’s growth is inseparable from its strategy of integrating AI into every layer of the enterprise stack, making it the default choice for Microsoft’s vast customer base.
Current Growth Drivers (2025) | Future Strategy (2026 and Beyond) |
OpenAI integration and Copilot: The deep integration of OpenAI models into the Azure stack via Azure OpenAI Service is the No. 1 growth catalyst. This partnership enables enterprises to leverage cutting-edge GPT models, complemented by Azure’s built-in enterprise governance and security features. | Improvements in Maia and Cobalt: Azure is rolling out improvements in specialized AI (Maia) and general-purpose (Cobalt) chips. The Maia chip is designed to power Microsoft’s internal AI workloads, such as Microsoft Copilot and Azure OpenAI Service, reducing the company’s reliance on external vendors like NVIDIA. The Azure Cobalt 200 is a 132-core ARM-based CPU designed to offload compression and cryptography tasks from general-purpose cores, thereby improving overall system performance. |
Hybrid cloud leadership (Azure Arc): Azure’s robust hybrid cloud platform, Azure Arc, continues to attract large enterprises with existing on-premises infrastructure. The Azure Storage with Arc Anywhere expands Azure Arc-enabled data services to manage Block, File, and Blob storage on any Kubernetes cluster (including those on AWS/GCP), pushing data sovereignty into multi-cloud environments. | Arc-native data fabric: With Azure Arc data fabric, Azure aims to abstract the storage location entirely, allowing developers to query or access data without knowing whether it resides on-premises, at the edge, or in an Azure region. This also allows companies to keep massive, proprietary knowledge bases on-premises or at the edge, drastically cutting the high cost of transferring petabytes of data into the cloud for indexing and retrieval. |
Massive investment: Microsoft is maintaining an aggressive infrastructure spending pace, successfully scaling profit alongside infrastructure—a testament to its ability to monetize these investments quickly. | Sustainability focus: Targeting a carbon-negative operational goal by 2030. This will influence future data center design and energy procurement strategies to appeal to increasingly ESG-focused enterprises. |
3. Google Cloud (GCP)#
GCP is leveraging its heritage in AI/ML and data analytics, making it the preferred cloud for companies with highly intensive or cutting-edge AI and data workloads.
Current Growth Drivers (2025) | Future Strategy (2026 and Beyond) |
Gemini and Vertex AI platform: GCP’s rapid growth is driven by its comprehensive Vertex AI development platform and the launch of the Gemini 3 family of multimodal models. This attracts developers needing maximum performance and flexibility for building, deploying, and tuning AI models. | TPU and custom compute leadership: Continued rapid advancement of its Tensor Processing Units (TPUs), such as the new Ironwood TPU v7 (unveiled at Cloud Next 2025). This is key to maintaining a performance edge in large-scale AI training, a core differentiator. |
Google Kubernetes Engine (GKE) at the edge: Having invented Kubernetes, GCP offers the most mature managed Kubernetes service (GKE), which remains the standard for containerized workloads and modern application development. The edge storage ensures stateful applications running on customer-managed hardware can reliably access local storage managed by the Google Cloud Storage (GCS) data plane. | AI agent and RAG: Investing heavily in advancements for building and managing complex, multi-agent AI systems, allowing developers to create highly sophisticated, real-time AI applications using models from any vendor. Furthermore, using GCS as the unified data plane reduces the need for a separate vector database (like AWS is doing with S3 Vectors, but via a managed service wrapper), drastically lowering the maintenance and operational cost of the RAG vector component. |
Data Analytics ecosystem: The unique capabilities of BigQuery and its integration with AI and new tools like the Data Engineering Agent provide unparalleled speed and cost-efficiency for analyzing massive datasets, a foundation for all GenAI applications. | Open ecosystem and multi-cloud: Actively fostering an open ecosystem, including bringing competitors’ models like Meta’s Llama 4 and Mistral to its platform, and collaborating on interoperability with rivals (e.g., Oracle Interconnect for Google Cloud) to capture multi-cloud workloads. |
2026 forecast: The road ahead#
The cloud market is projected to maintain a strong 20%+ average annual growth rate in 2026, driven by three key themes:
1. The competition around AI infrastructure#
Hyperscalers are investing billions in capital expenditures to expand their data center capacity. In 2026, this new capacity will be deployed and is expected to generate revenue from high-demand, high-margin workloads.
Custom silicon dominance: Expect a major push for in-house chips (TPUs, Trainium, etc.) as providers focus on improving the cost-efficiency of AI inference. This shall reduce the recurring operational cost of running AI models at scale.
GPU-as-a-service acceleration: Specialized AI compute services are expected to continue experiencing triple-digit growth as businesses transition GenAI models from pilot to production.
2. Niche player specialization#
While the Big Three solidify their collective dominance, specialized vendors will thrive by carving out defensible niches.
Neoclouds rise: Providers focused solely on high-performance computing (HPC) for AI, such as CoreWeave, will continue to gain traction by offering highly optimized GPU infrastructure.
Industry clouds: Solutions focused on highly regulated sectors (e.g., financial services, healthcare) will grow, emphasizing data sovereignty and regional compliance.
3. Multi-cloud normalization#
For large enterprises, a multi-cloud strategy is a necessity for achieving strategic resilience, effective cost management, and accessing best-of-breed services. Companies will continue to strategically blend the scale of hyperscalers with the flexibility and specialization of targeted vendors.
Future challenges#
The cloud market is projected to maintain a strong average annual growth rate of 20% or more in 2026, driven by the structural changes initiated in 2025. The forecast for 2026 points beyond market competition toward material infrastructure constraints. While spending will remain robust, the ability of providers to meet demand will hinge on solving severe physical bottlenecks.
Power and supply deficit looms: Analysts predict that data center occupancy rates will peak in late 2026, as AI-driven demand outstrips supply, particularly in high-density regions. The core constraint is no longer just GPUs, but electrical grid capacity and water availability for liquid cooling, which could delay capacity rollouts and raise customer prices.
Hybrid cloud dominance: The transition to hybrid cloud will accelerate as enterprises seek operational resilience. With 70% of organizations predicted to use hybrid IT by 2027, the focus shifts to data logistics and governance across dispersed environments, utilizing AI-ready platforms to manage complexity and ensure compliance.
The rise of sovereign AI: Geopolitical and regulatory mandates will drive investment in Sovereign Cloud solutions, particularly in Europe and Asia, requiring providers to meet strict data residency and control requirements. This will be a key competitive lever, with vendors partnering with local governments and utilities to secure power and land.
The massive CapEx deployed in 2025 has created a temporary bubble of optimism. 2026 is expected to be a year of execution, as hyperscalers work to demonstrate their ability to convert planned infrastructure builds into sustained capacity while addressing significant physical and regulatory constraints, thereby shaping the operating baseline for an AI-driven cloud ecosystem.
Ready for more cloud-based learning? Be sure to check our latest and most popular Cloud Labs:
