Blackstone and Google Form Joint Venture to Provide Data Center and TPU Computing Power at Scale
Blackstone and Google have established a joint venture to provide large-scale data center capacity and Google's latest TPU computing power. Customers will be able to access the latest TPU through this joint venture, not limited to Google Cloud.
Blackstone has initially committed $5 billion in equity, aiming to launch the first batch of 500MW capacity by 2027.
AI companies and infrastructure investors in the market are accelerating their deployment of dedicated computing power. Blackstone is quickly entering the AI data center market through this joint venture, benefiting both Google and Blackstone while putting short-term pressure on single cloud service providers. Funding is rapidly concentrating on AI-specific data centers and TPU infrastructure.
Source: Public Information
ABAB AI Insight
Blackstone has previously heavily invested in data center assets, and this collaboration with Google continues its transition from traditional infrastructure to AI power and computing. Google, on the other hand, aims to expand its ecosystem influence by opening up TPU supply and accelerating capacity construction with Blackstone's capital, both targeting the long-term shortage of computing power for AI training and inference.
In terms of capital strategy, Blackstone will directly use the $5 billion equity for data center construction and power support, while Google provides TPU hardware and operational technology, motivated by shifting AI computing power from a closed model in Google Cloud to broader institutional access, forming a closed loop from equity investment to long-term stable computing power leasing income.
Similar to Microsoft and OpenAI's Stargate project and other deep ties between hyperscalers and private capital, Blackstone and Google are currently placing AI infrastructure in a leading position, transitioning from cloud giants' monopoly to a combined expansion of capital and technology, driving the industry from a single cloud provider to a diversified dedicated computing power platform.
Structural judgment: Essentially, this is a concentration of capital. The extreme demand for high-performance computing power and electricity in AI allows a few joint entities with substantial capital and technical resources to efficiently lock in large-scale capacity, as the threshold for 500MW-level projects is extremely high, forcing AI training resources to concentrate heavily on platforms like Blackstone-Google that have long-term capital and supply chain advantages.
ABAB News · Cognitive Law
When capital meets technology, computing power equals scale.
$5 billion first, 500MW to follow.
The moment TPU opens, AI infrastructure enters a new era.