Stock Market

Blackstone invests $5bn in AI infrastructure powered by Google tensor processing units

Trade
Image: Pixabay

20 May 2026

Blackstone, the world’s largest private owner of data centres, is investing $5 billion (€4.3 billion) together with Google in a new AI infrastructure company. The as-yet-unnamed venture will use Google’s tensor processing units (TPUs), specialised chips designed for AI computations, to deliver substantial computing power. Blackstone aims to bring 500 MW of computing capacity online by 2027 and to further scale up operations over time.

The new company will be led by Benjamin Treynor Sloss, former chief programs officer at Google. Although details of the ownership structure have not yet been disclosed, sources suggest that Blackstone will hold a majority stake. The venture has already identified potential locations for data centres, some of which are currently under construction.

The investment underscores Blackstone’s aggressive strategy in the AI landscape, building on its recent partnership with Anthropic. The deal is also a clear illustration of the growing competition between Google and Nvidia for dominance in AI hardware. Google’s TPUs are being positioned as a direct alternative to Nvidia’s widely used graphics processing units (GPUs).

 

advertisement



 

Specialised hardware for AI

Although Google still uses Nvidia GPUs within its cloud infrastructure, it has been actively developing its own TPUs to reduce its dependence on Nvidia. Similar efforts are under way at other tech giants such as Amazon Web Services, which is likewise striving to develop its own semiconductors. Google’s first TPU was produced in 2015, highlighting the company’s early commitment to bespoke hardware solutions.

Unlike GPUs, which are designed for general-purpose computing applications, TPUs are optimised for specific AI applications, such as agentic AI models. Well-known customers using Google’s TPUs include Anthropic and Citadel Securities, alongside Google’s own Gemini AI model. Nvidia’s GPUs, originally developed for rendering graphics in computers and games consoles, excel at parallel processing by breaking complex problems down into smaller tasks.

The sharp increase in demand for GPUs following the launch of OpenAI’s ChatGPT helped turn Nvidia into the most valuable company in the world. Earlier this month, however, Google briefly overtook Nvidia in terms of market value.

Analysts believe that Alphabet, Google’s parent company, is well positioned in AI thanks to its robust in-house AI development capabilities, extensive distribution network and profitable cloud operations.

Business AM

Read More:


Back to Top ↑