The Google Cloud logo appears on a high-resolution digital monolith at the entrance to the Google Pavilion during Mobile World Congress on March 5, 2026 in Barcelona, Spain. In the 2026 edition, Google Cloud showcases the evolution of the “IQ era” by announcing new Agentic AI solutions for communications, powered by the Gemini 3.1 Pro model and Vertex AI Agent Builder. This exhibit focuses on Google Distributed Cloud deployments running on TPU v6 (Tensor Processing Units) to enable AI at the network edge. Key demos include a “network digital twin” powered by Spanner Graph and BigQuery. This enables operators such as MasOrange and Deutsche Telekom to use autonomous agents for real-time root cause analysis and predictive network self-healing. The stand highlights Google’s unified data foundation, which leverages graph neural networks to move from traditional monitoring to proactive intent-based network management. (Photo by Joan Cros/NurPhoto, Getty Images)
Null Photo | Null Photo | Getty Images
black stonethe world’s largest private owner of data centers, will invest $5 billion in capital in a new artificial intelligence infrastructure company. googlethe New York-based asset management company announced on Monday.
Google will supply tensor processing units (chips dedicated to processing artificial intelligence calculations) to the U.S.-based new company, which will bring an initial 500 megawatts of computing power online by 2027, with “plans to expand significantly over time,” Blackstone said in a statement.
“This new company has great potential as it helps us meet unprecedented computing demands,” John Gray, Blackstone’s president and chief operating officer, said in a statement.
The unnamed company will be led by Benjamin Trainor Sloss, who most recently served as Google’s chief program officer. A Google spokesperson declined to comment on whether Google would maintain a direct leadership role in this area.
The Wall Street Journal, which first reported on the venture ahead of Blackstone’s public statement, said the private equity giant would own a majority stake, citing sources familiar with the matter.
Blackstone did not disclose the venture’s ownership structure in a statement and did not respond to CNBC’s request for comment by the time of publication.
The magazine also reported that the joint venture has already identified potential locations for data centers, some of which are under construction.
Blackstone, which has more than $1.3 trillion in assets under management, is actively investing across the AI ecosystem and launched a venture similar to Anthropic earlier this month.
Shares of Alphabet and Blackstone rose about 1% in premarket trading Tuesday.
The deal highlights the growing competition between Google and Nvidia in AI hardware. Google’s TPUs have long been seen as the company’s answer to industry leaders. Nvidiagraphics processing unit.
While Google continues to use Nvidia’s GPUs across its cloud architecture, the company is looking to reduce its dependence on Jensen Huang’s Nvidia by developing its own TPUs.
Other tech giants: Amazon Web services have similarly sought to develop their own semiconductor chips. Google was an early proponent of building hardware in-house, producing its first TPU in 2015.
Unlike GPUs, which are broadly suitable for general-purpose applications, Google touts its TPUs as specialized for more efficient processing of a narrower range of applications, such as agent AI applications. Google runs Gemini AI models on TPU, and its TPU customers include Anthropic and Citadel Securities.
First developed in 1999 to render graphics in computers and game consoles, Nvidia’s GPUs work by breaking down complex computing problems into smaller tasks and solving them simultaneously.
Demand for GPUs skyrocketed after the launch of OpenAI’s ChatGPT in 2022, propelling Nvidia to the position of the world’s most valuable company in 2024.
Earlier this month, Google briefly overtook Nvidia in market capitalization. Analysts told CNBC that Google’s parent company Alphabet remains well-positioned in the AI space thanks to its in-house AI development, extensive sales network and profitable cloud division.
