Blackstone and Google Form Joint Venture for New TPU Cloud

Blackstone, a leading alternative asset manager, and Google have entered into a significant joint venture to create a new U.S.-based company. This new entity will focus on delivering compute-as-a-service, specifically leveraging Google Cloud's Tensor Processing Units (TPUs) to meet the escalating global demand for AI-optimized computing capabilities. The partnership provides organizations with an expanded set of options for accessing Google's advanced TPU technology, which has been integral to AI development and deployment for over a decade. Under the terms of the agreement, Blackstone is making an initial equity capital commitment of $5 billion from funds it manages. The ambitious goal is to bring the first 500 megawatts (MW) of data center capacity online by 2027, with substantial plans for future expansion. Google will be a key enabler, supplying the essential hardware, including its custom-built TPUs, along with comprehensive software and services. This support is designed to accelerate the new company's growth and ensure it can effectively address the burgeoning need for accelerated computing, drawing upon Google's extensive technical and domain expertise in the field of artificial intelligence. Thomas Kurian, CEO of Google Cloud, emphasized that this collaboration with Blackstone is a direct response to the growing demand for TPUs, which are engineered for optimal efficiency and performance in the AI era. He stated that the joint effort will accelerate AI transformation and offer more avenues for organizations to access high-performance compute capabilities. Benjamin Treynor Sloss, a long-standing Google executive with over two decades of experience in building and operating Google's global infrastructure, has been appointed as the CEO of this newly formed company. This venture highlights Google's strategic intent to further cultivate the AI economy by making its foundational TPU technology more widely accessible. TPUs are crucial for the training and inference of sophisticated AI models and currently power workloads for numerous leading AI labs, capital market firms, and companies running complex high-performance computing applications, including Google's own Gemini models and other AI-driven products that serve billions of users globally.