Google Cloud explodes with 103% growth as TPU systems fuel $30.7 billion in quarterly sales

Google Cloud Achieves Record 103% Revenue Growth

Google Cloud delivered unprecedented financial results in Q3, with revenue more than doubling to $30.71 billion, marking a staggering 103% year-over-year increase from $15.16 billion1

1

2

2

. This represents the highest growth rate in Google Cloud's history, cementing its position as a formidable player in the AI infrastructure market. Operating income surged 209% to $11.12 billion, up from $3.6 billion in Q3 20252

2

. The division now boasts 18 product lines each generating more than $1 billion in annual revenue, according to CEO Thomas Kurian1

1

.

Tensor Processing Units Drive Extraordinary Demand

TPU systems emerged as a meaningful contributor to Google Cloud sales, with Kurian describing "extraordinary demand" from AI labs, capital markets firms, financial institutions, and enterprises2

2

. Nearly 100 customers currently use TPUs, with over 100 significant customers starting on Google Cloud Platform1

1

2

2

. The latest TPU 8t and 8i models deliver up to 2.7x better training performance and 80% better inference price performance compared to previous generations1

1

. Kurian emphasized that TPU systems represent far more than chip sales, combining specialized hardware, compute, storage, networking, system software, compilers, and optical networking into integrated solutions for AI training and inference2

2

.

Strategic Partnerships Expand TPU Distribution

Google Cloud is aggressively broadening TPU availability beyond its own cloud platform. The company recently partnered with Oracle and established an agreement with Crux AI, a joint venture with Blackstone, to deliver TPU systems to customers1

1

2

2

. This strategy allows customers to deploy TPU systems in their own data centers and through third-party facilities, significantly expanding the addressable market. Notable TPU customers include Anthropic, Apple for foundation model pre-training, frontier model companies like Mirendil and Physical Intelligence, capital markets firms Citadel Securities and Jump Trading, universities like Caltech, and public sector providers including Cassava and Cirrascale1

1

.

Software Optimization Accelerates Adoption

Google Cloud's deliberate approach to TPU deployment emphasizes software optimization alongside hardware innovation. The recent introduction of TorchTPU, developed in collaboration with Meta, enables developers to train, serve, and deploy PyTorch models on TPUs using familiar tools1

1

. This optimization makes portability seamless and opens a much larger market by supporting the widely-used PyTorch framework. The strong demand for AI systems reflects Google's integrated approach to building the right silicon, optimizing software for better performance and portability, and broadening distribution channels1

1

.

Gemini Enterprise Gains Traction Across Enterprises

Gemini Enterprise demonstrated significant momentum, with over 2,600 customers using at least 100 billion tokens and nearly 75 customers consuming at least 10 trillion tokens over the past 12 months1

1

. Nearly 90% of the Fortune 100 now use Gemini Enterprise, and customers who adopted the platform in Q3 2025 more than quadrupled their spend year-over-year1

1

. Google Cloud has sold more than 100 million seats for knowledge work across all AI-related services. Gemini Enterprise now includes deeply integrated agents across Google's entire cloud portfolio, enabling companies to research, answer questions, write and run code, and automate workflows with strong security controls and access to frontier models1

1

.

End-to-End AI Infrastructure Sets Google Apart

Ben Kessler, CEO of 66degrees, a Google Cloud channel partner, emphasized that no other hyperscaler offers the comprehensive end-to-end AI infrastructure that Google provides. "Google is the one hyperscaler that brings everything together," creating chips, data centers, the agentic data stack, AI models, agent development kits, monitoring capabilities, and Gemini Enterprise2

2

. Kurian attributed Q3 momentum to this integrated AI portfolio of chips, models, data, security, and agent platforms all designed to work together2

2

. While TPU systems garnered significant attention, Kurian noted that Google Cloud Platform accounted for the vast majority of revenue, with very strong growth even excluding TPU system sales2

2

.