Blackstone and Google are planning to launch a new U.S.-based cloud infrastructure company focused on delivering AI compute services built around Google’s Tensor Processing Units (TPUs). The venture combines Blackstone’s large-scale infrastructure financing and data center development capabilities with Google’s AI hardware, software, and cloud expertise. Blackstone plans an initial $5 billion equity investment to support the first 500 MW of capacity, targeted to come online in 2027, with additional expansion planned over time.
The new company will offer compute-as-a-service access to Google TPUs outside of the standard Google Cloud platform, creating an additional channel for enterprises, AI labs, financial firms, and hyperscale customers seeking accelerated AI infrastructure. Google said it will provide the TPU hardware stack along with associated software and operational support. The move reflects growing industry demand for alternative AI compute providers as hyperscalers and infrastructure investors race to secure power, land, networking, and specialized silicon for next-generation AI workloads.
Blackstone also named longtime Google infrastructure executive Benjamin Treynor Sloss as CEO of the new venture. Sloss spent more than two decades helping build and operate Google’s global infrastructure platform. The announcement further expands Blackstone’s role in AI infrastructure markets, where private capital firms increasingly back hyperscale data centers, power infrastructure, and AI cloud platforms to meet surging demand for training and inference capacity.
• Blackstone committed an initial $5 billion in equity capital
• First deployment targets 500 MW of AI infrastructure capacity in 2027
• Platform will provide TPU-based compute-as-a-service offerings
• Google will supply TPU hardware, software, and operational services
• Benjamin Treynor Sloss, formerly of Google infrastructure operations, will serve as CEO
• TPU access expands beyond direct consumption through Google Cloud
• Target customers include AI labs, enterprises, financial firms, and HPC operators
“Together, we’re accelerating AI transformation and providing more options for organizations to access accelerated compute capability,” said Thomas Kurian, CEO of Google Cloud.
🌐 Analysis: The partnership highlights a broader shift toward dedicated AI infrastructure platforms financed by private capital and optimized around proprietary accelerator architectures. While NVIDIA GPUs continue to dominate AI training deployments, hyperscalers increasingly push custom silicon strategies to improve performance-per-watt, reduce supply constraints, and vertically integrate AI services. Google’s TPUs already power internal Gemini workloads and selected cloud customers, but this venture extends TPU availability into a more independent infrastructure consumption model resembling emerging AI utility platforms.
🌐 The scale is also notable from a power and infrastructure perspective. A 500 MW AI deployment rivals some of the largest hyperscale campus announcements in the market and underscores how AI infrastructure buildouts increasingly resemble utility-scale industrial projects. Blackstone has already invested heavily in digital infrastructure and data centers, while competitors including KKR, Brookfield, BlackRock, and DigitalBridge continue expanding exposure to AI-driven power, networking, and compute infrastructure assets.