Cloud infrastructure spending accelerates as AI demand fuels hyperscaler expansions in 2025

Global cloud infrastructure expenditure soared to $95.3 billion in Q2 2025, driven by surging AI workloads, enterprise migration, and expanding collaborations among leading providers like AWS, Microsoft Azure, and Google Cloud amid persistent market momentum.

Global spending on cloud infrastructure services surged to $95.3 billion in the second quarter of 2025, reflecting a 22% increase year on year, driven predominantly by soaring demand for artificial intelligence (AI) workloads along with continuing legacy system migrations and cloud-native enterprise expansion. This marks the fourth consecutive quarter where cloud spending growth has exceeded 20%, highlighting sustained market momentum amid rapid AI adoption and evolving enterprise cloud strategies.

Industry research from Canalys, part of Omdia, underscores that the expanding AI consumption, renewed traditional workload shifts, and scaling by digital-first firms are primary growth engines. Hyperscale cloud providers—Amazon Web Services (AWS), Microsoft Azure, and Google Cloud—remain dominant, collectively capturing 65% of the global market. Customer expenditure on these platforms climbed 27% year on year, with Microsoft Azure and Google Cloud both posting quarterly growth above 30%, while AWS grew by 17%, consistent with prior quarters. Despite AWS’s slower rate, it registered the largest absolute increase in revenue.

Investment plans from these hyperscalers illustrate their commitment to meeting AI-driven demand. Google increased its capital expenditure target for 2025 from $75 billion to $85 billion, while AWS projects over $100 billion in spending this year, and Microsoft anticipates around $80 billion within its current fiscal cycle. Alphabet’s cloud division, buoyed by AI developments such as its Gemini AI model and in-house chip advancements, posted a 32% revenue increase recently, further cementing its cloud growth trajectory.

According to Canalys Senior Analyst Yi Zhang, customer AI service needs are maturing from a focus on availability and ease of use to prioritising flexibility and tailored model choices. Enterprises increasingly seek multi-model approaches, switching between AI models to balance performance, cost, and application fit optimally. Reflecting this, major cloud vendors are broadening their AI model portfolios; AWS Bedrock, Azure AI Foundry, and Google Vertex AI now offer extensive proprietary and third-party models catering to diverse industry requirements.

The competitive landscape is marked by ‘coopetition’—a blend of competition and collaboration. Rachel Brindley, Senior Director at Canalys, notes that while vendors race to advance AI models and product features, they also collaborate on compute capacity and model distribution. For instance, AWS Bedrock aggregates models such as Anthropic’s Claude and OpenAI’s GPT, while OpenAI utilizes Google Cloud’s infrastructure to enhance computing capacity, leveraging complementary strengths to satisfy accelerating AI demand. This synergy aims to offset constraints like power and semiconductor shortages, which have limited the expansion of computing capacity.

Looking at individual vendors, AWS holds a 32% market share with steady growth and a robust $195 billion backlog as of June, indicating strong sustained demand. AWS recently launched innovations including Amazon Bedrock AgentCore and a new AI marketplace category featuring over 800 products designed to simplify AI agent deployment. Microsoft Azure, commanding a 22% share, achieved 39% yearly growth, driven by traditional migrations, cloud-native expansions, and AI adoption. Azure AI Foundry recently expanded its model suite to include offerings from OpenAI, DeepSeek, Meta, and others, with more additions forthcoming. Microsoft’s global data centre footprint also continues to grow, now surpassing 400 facilities in over 70 regions.

Google Cloud, with an 11% market share, posted 34% growth, propelled by a surge in high-value contracts exceeding $250 million and a doubling of billion-dollar deals compared to the previous year. Its order backlog surged to $108.2 billion by mid-year. Google’s capital expenditure hike to $85 billion for 2025 aligns with the launch of AI models like Gemini 2.5 Flash and Pro, including the cost-efficient Flash-Lite variant. The Gemini platform reports over 450 million monthly active users, with a 50% quarter-on-quarter increase in daily requests. Additionally, reports indicate Google Cloud may join OpenAI’s supply network to support growing computational demands, highlighting deepening industry collaborations.

Geographically, AI demand is also rising robustly in Mainland China, where cloud infrastructure spending hit $11.6 billion in Q1 2025, a 16% increase year on year. Cloud providers there are intensifying investments in AI infrastructure and model development, employing strategies such as turnkey AI model access, partner ecosystem expansions, and AI agent development platforms to meet local enterprise needs.

Despite these optimistic trends, the cloud sector faces challenges including supply chain constraints for semiconductors and power shortages, potentially limiting infrastructure expansion. Nevertheless, the rapid evolution of AI workloads and enterprise digital transformation continue to drive unprecedented investment and innovation across the global cloud ecosystem.

📌 Reference Map:

• Paragraph 1 – ^[1], ^[5]
• Paragraph 2 – ^[1], ^[2], ^[3], ^[5]
• Paragraph 3 – ^[1], ^[6], ^[7]
• Paragraph 4 – ^[1]
• Paragraph 5 – ^[1]
• Paragraph 6 – ^[1], ^[5]
• Paragraph 7 – ^[1], ^[6], ^[7]
• Paragraph 8 – ^[1], ^[4]
• Paragraph 9 – ^[1]

Source: Noah Wire Services