Blackstone and Google are joining forces to build a massive new artificial intelligence infrastructure company. Blackstone, the largest private owner of data centers in the world, announced on Monday that it will invest $5 billion in cash to get the joint venture off the ground.
Google will provide the hardware backbone for the new business. The tech giant will supply its custom tensor processing units, or TPUs, which it designed specifically to handle heavy artificial intelligence math. The partners plan to bring the first 500 megawatts of computing capacity online by the year 2027. They also plan to scale the operation significantly in the years that follow.
Benjamin Treynor Sloss will take the top job at the unnamed company. Sloss recently worked as the chief programs officer at Google. The Wall Street Journal reported that Blackstone will hold a majority stake in the new business. The partners have already picked out several locations for the new data centers, and construction crews are already working on some of the sites.
Jon Gray, the president and chief operating officer of Blackstone, explained the reasoning behind the massive investment. He stated that the new company has enormous potential because it helps solve the unprecedented global demand for computing power. Blackstone currently manages more than $1.3 trillion in global assets. The firm clearly loves the artificial intelligence sector, as it formed a similar joint venture with Anthropic earlier this month.
Wall Street reacted positively to the news. Shares of both Alphabet and Blackstone rose by about 1 percent during early pre-market trading on Tuesday. Investors see the partnership as a strong move for both corporations as they try to grab more control over the physical hardware that makes modern software possible.
This $5 billion deal highlights the intense and growing rivalry between Google and Nvidia. Nvidia currently dominates the market for artificial intelligence hardware with its graphics processing units. Google still buys and uses Nvidia chips inside its cloud architecture, but the search giant desperately wants to break that dependency. By building and deploying its own TPUs, Google can cut costs and control its own destiny.
Google started building its own custom silicon a long time ago. The company manufactured its very first TPU back in 2015. Google tells customers that these chips work better than standard graphics chips for specific tasks, like running advanced agentic artificial intelligence applications. Google runs its own Gemini artificial intelligence model entirely on TPUs. Big clients like Citadel Securities and Anthropic also pay to use Google’s custom chips.
Nvidia took a very different path to the top. The company developed its first graphics chips in 1999 specifically to render video games on home computers and consoles. These chips work by breaking complex computing problems into thousands of smaller pieces and solving them all at the exact same time.
That specific design turned out to be perfect for training artificial intelligence. When OpenAI launched ChatGPT in 2022, demand for Nvidia chips exploded overnight. That massive buying frenzy pushed Nvidia to become the most valuable company in the world by 2024.
However, the race for the top spot changes every day. Earlier this month, Google actually overtook Nvidia in total market value for a brief moment. Financial analysts believe Google parent company Alphabet sits in a great position to win the artificial intelligence war. The company designs its own hardware, operates a massive global distribution network, and runs a highly profitable cloud division. This new $5 billion partnership with Blackstone simply adds another powerful weapon to Google’s arsenal.
