Google and Blackstone are creating a joint venture to provide cloud access to Google's Tensor Processing Units (TPUs), marking a major shift in how the search giant distributes its custom AI silicon. The partnership represents an initial investment of $5 billion according to The Information, though Bloomberg and The Straits Times report the overall business could be worth $25 billion.
The venture directly targets CoreWeave and other neoclouds that have emerged as alternatives to traditional cloud providers for AI workloads. These newer entrants, many backed by Nvidia, have gained traction by providing more flexible access to scarce GPU resources that video AI companies need for training and inference.
Google's TPUs have powered the company's internal AI operations for years, but access has been limited to Google Cloud customers. This new entity upends that calculus, creating a standalone business focused solely on TPU infrastructure. The Information characterizes this as challenging "Nvidia-backed neoclouds," showing how chip architecture choices are becoming central to cloud strategy.
The timing matters. Bloomberg's analysis points to serious headwinds facing data center buildouts, including delays from local regulations and capacity constraints. These bottlenecks affect all players trying to scale AI infrastructure, possibly giving an advantage to those with existing hardware and facilities.
Get the latest model rankings, product launches, and evaluation insights delivered to your inbox.
The reported figures, $5 billion initial investment, $25 billion total value, underscore what the massive capital required for AI. For context, CoreWeave raised $7.5 billion in debt financing just last week, suggesting these dollar amounts have become table stakes for serious AI cloud competition.
What remains unclear from available reporting is the operational structure. Will this venture operate Google's existing TPU capacity, build new facilities, or both? How will pricing compare to Google Cloud's current TPU rates? These details will determine whether this represents genuine market expansion or simply financial engineering.
The partnership also arrives amid what Bloomberg describes as broader industry tensions over AI safety, regulation, and corporate governance, including ongoing legal disputes involving OpenAI. For video AI creators choosing infrastructure partners, these governance issues increasingly matter alongside raw compute availability.