Key Takeaways
- • Google and Blackstone will establish a $25 billion AI cloud joint venture, with Blackstone committing $5 billion in equity capital and the partnership targeting total computing investment of $25 billion including leverage.
- • The venture represents Google's strategic pivot to commercialize its proprietary Tensor Processing Unit (TPU) AI chips beyond internal use, directly competing in the Nvidia GPU-dominated AI infrastructure market.
- • Benjamin Trainor Sloss, a veteran of Google's site reliability engineering division, has been named CEO of the new unnamed U.S.-based company, which will operate large-scale AI data centers.
- • The joint venture plans to secure 500MW of AI data center power capacity by 2027—equivalent to the electricity consumption of a medium-sized city—with plans for gradual expansion to meet enterprise demand.
- • This partnership differs from Google's recent TPU supply agreements with Anthropic and Meta by constructing integrated large-scale AI data center and cloud infrastructure rather than supplying chips alone.
Summary
Google and Blackstone, the world's largest private equity firm, are establishing a $25 billion artificial intelligence cloud joint venture to build competing infrastructure in a market currently dominated by Nvidia-based providers. According to the Wall Street Journal, Blackstone will invest $5 billion in equity capital initially, with the total investment reaching approximately $25 billion when including leveraged financing. The partnership represents a significant shift in Google's strategy to externalize its proprietary AI chip technology and compete directly in the booming AI infrastructure sector.
Benjamin Trainor Sloss, who has led Google's site reliability engineering efforts, will serve as CEO of the new company, which will be incorporated in the United States. The venture targets deployment of 500 megawatts of AI data center power by 2027—a capacity equivalent to powering a medium-sized metropolitan area with hundreds of thousands of residents. The companies plan to scale capacity incrementally thereafter to accommodate anticipated large-scale demand from enterprise AI operations.
The core strategic objective centers on commercializing Google's Tensor Processing Unit (TPU), a proprietary AI-only semiconductor previously deployed primarily for internal Google services including search, YouTube, and the Gemini AI platform. While Google has recently signed TPU supply contracts with Anthropic (developer of Claude) and Meta (Facebook parent), the Blackstone partnership operates at a different scale—constructing end-to-end cloud infrastructure and data centers rather than supplying chips in isolation. This model positions Google to compete more directly with CoreWeave and other emerging AI infrastructure companies that have captured significant market share from hyperscalers seeking alternatives to Nvidia's GPU ecosystem.
The venture addresses a critical market dynamic: explosive demand for AI computing capacity has outpaced available infrastructure, creating a favorable environment for alternative chip providers like Google to penetrate the market. By pairing Google's chip technology with Blackstone's capital and infrastructure expertise, the partnership creates a vertically integrated offering capable of attracting enterprise customers seeking TPU-based alternatives to GPU-centric solutions. The $25 billion investment scale reflects the magnitude of capital requirements to compete meaningfully in this infrastructure-intensive sector.
Benjamin Trainor Sloss, who has led Google's site reliability engineering efforts, will serve as CEO of the new company, which will be incorporated in the United States. The venture targets deployment of 500 megawatts of AI data center power by 2027—a capacity equivalent to powering a medium-sized metropolitan area with hundreds of thousands of residents. The companies plan to scale capacity incrementally thereafter to accommodate anticipated large-scale demand from enterprise AI operations.
The core strategic objective centers on commercializing Google's Tensor Processing Unit (TPU), a proprietary AI-only semiconductor previously deployed primarily for internal Google services including search, YouTube, and the Gemini AI platform. While Google has recently signed TPU supply contracts with Anthropic (developer of Claude) and Meta (Facebook parent), the Blackstone partnership operates at a different scale—constructing end-to-end cloud infrastructure and data centers rather than supplying chips in isolation. This model positions Google to compete more directly with CoreWeave and other emerging AI infrastructure companies that have captured significant market share from hyperscalers seeking alternatives to Nvidia's GPU ecosystem.
The venture addresses a critical market dynamic: explosive demand for AI computing capacity has outpaced available infrastructure, creating a favorable environment for alternative chip providers like Google to penetrate the market. By pairing Google's chip technology with Blackstone's capital and infrastructure expertise, the partnership creates a vertically integrated offering capable of attracting enterprise customers seeking TPU-based alternatives to GPU-centric solutions. The $25 billion investment scale reflects the magnitude of capital requirements to compete meaningfully in this infrastructure-intensive sector.
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.