Key Takeaways
- • Blackstone commits initial $5 billion in equity capital to joint venture with Google focused on delivering AI infrastructure powered by Google's custom Tensor Processing Units (TPUs).
- • New venture targets 500 megawatts of data center capacity online by 2027, with plans to scale significantly; total investment potential reaches approximately $25 billion.
- • Offering positions as alternative to GPU-based cloud providers like CoreWeave, providing compute-as-a-service model combining data center capacity, networking, operations, and TPU access.
- • Google supplies hardware (TPUs), software, and technical expertise; Blackstone leverages its $1.3+ trillion in assets under management and position as world's largest global data center provider.
- • Partnership addresses accelerating demand for specialized AI computing infrastructure from research labs, capital markets firms, and enterprise customers requiring high-performance computing capabilities.
Summary
Blackstone and Google announced a strategic joint venture on May 18, 2026, to establish a new U.S.-based company providing AI infrastructure centered on Google's custom Tensor Processing Units. Blackstone is committing an initial $5 billion in equity capital from its managed funds, with the venture expected to achieve total valuation of approximately $25 billion as it scales. The partnership represents a significant convergence between Blackstone's unparalleled data center portfolio and Google's specialized AI silicon capabilities.
The venture will deliver data center capacity, networking, operations, and access to Google Cloud's TPUs through a compute-as-a-service model. Google's TPUs are custom chips engineered specifically for AI workloads, optimized for both training and inference of advanced AI models. These processors have been in production for over a decade and currently power Gemini and other AI applications serving billions of users globally, in addition to supporting workloads for leading AI research laboratories and capital markets institutions.
The immediate target is to bring 500 megawatts of capacity online during 2027, with significant scaling expected thereafter. This positions the venture as a direct competitor to existing GPU-focused cloud providers like CoreWeave. By offering customers a dedicated TPU cloud option alongside Google Cloud's existing offerings, the partnership aims to increase choice and flexibility for organizations running complex AI workloads. Google will provide essential hardware, software, and technical expertise to enable rapid scaling, while Blackstone contributes its substantial asset base, operational infrastructure, and market access. For investors and institutions, this venture signals the bifurcation of AI compute infrastructure between GPU and TPU architectures, with established financial and technology powerhouses moving decisively into direct infrastructure provision.
The venture will deliver data center capacity, networking, operations, and access to Google Cloud's TPUs through a compute-as-a-service model. Google's TPUs are custom chips engineered specifically for AI workloads, optimized for both training and inference of advanced AI models. These processors have been in production for over a decade and currently power Gemini and other AI applications serving billions of users globally, in addition to supporting workloads for leading AI research laboratories and capital markets institutions.
The immediate target is to bring 500 megawatts of capacity online during 2027, with significant scaling expected thereafter. This positions the venture as a direct competitor to existing GPU-focused cloud providers like CoreWeave. By offering customers a dedicated TPU cloud option alongside Google Cloud's existing offerings, the partnership aims to increase choice and flexibility for organizations running complex AI workloads. Google will provide essential hardware, software, and technical expertise to enable rapid scaling, while Blackstone contributes its substantial asset base, operational infrastructure, and market access. For investors and institutions, this venture signals the bifurcation of AI compute infrastructure between GPU and TPU architectures, with established financial and technology powerhouses moving decisively into direct infrastructure provision.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.