Google takes shot at Nividia and Amazon with $5 billion Blackstone AI cloud deal, details here

Google is no longer just building AI models and cloud services quietly in the background. The company is now making one of its biggest infrastructure bets yet and it is directly aimed at weakening the grip of Nvidia and challenging the cloud dominance of Amazon. In a major move, Google has partnered with investment giant Blackstone to launch a new AI cloud company backed by an initial $5 billion investment. The venture will combine Google's custom AI chips, called Tensor Processing Units (TPUs), with Blackstone’s massive data centre infrastructure business to create a new platform for companies hungry for AI computing power.

The timing of the deal is important. The AI boom has created an enormous demand for chips and data centres, and companies are now scrambling to secure both before the next wave of AI growth begins.

"We see a generational opportunity to invest capital at scale building AI infrastructure. This new company has enormous potential as it helps to meet the unprecedented demand for compute," said Jon Gray, President and COO of Blackstone.

Google is building an AI empire beyond search

For years, Nvidia has been the undisputed king of AI chips. Most companies building AI systems relied heavily on Nvidia GPUs to train models like ChatGPT and Gemini. But things are changing fast. The next phase of AI is moving towards "inference," the process where AI models actually respond to users, generate answers, and perform tasks in real time. That market is expected to become even bigger than AI training itself.

Google clearly wants a larger share of that future.

Last month, the company introduced its latest TPU chips, including TPU 8i, its first dedicated inference-focused processor. It also launched TPU 8t, which is designed to speed up the training of massive AI models. Google says these chips can drastically reduce the time needed to develop advanced AI systems.

Now, instead of simply renting cloud servers, Google is trying to create a complete AI infrastructure ecosystem — chips, software, networking, and data centres all under one roof.

"This joint venture with Blackstone helps meet growing demand for TPUs, which are optimised specifically for efficiency and performance in the AI era. Together, we’re accelerating AI transformation and providing more options for organisations to access accelerated compute capability," said Thomas Kurian, who is the CEO of Google Cloud.

The company expects the first 500 megawatts of capacity from the project to go live in 2027. That may sound technical, but in the AI world, power capacity has become one of the biggest bottlenecks. Companies today are not just fighting for chips, they are fighting for space and electricity to run them. That is where Blackstone enters the picture.

The AI race is no longer just about chips

Blackstone is one of the world’s largest data centre investors, managing more than $1.3 trillion in assets. By teaming up with Google, the firm is betting that AI infrastructure could become one of the biggest business opportunities of the decade.

The partnership also sends a strong message to Amazon Web Services and Microsoft Azure, which currently dominate the cloud computing market. Amazon recently reported that AWS revenue jumped 28 per cent year-on-year to $37.59 billion, while Microsoft said Azure revenue surged 40 per cent. Google Cloud is growing rapidly too, with Alphabet reporting around 63 per cent growth in the segment.

To keep up with the AI race, Google parent Alphabet has already increased its planned spending for 2026 to as much as $190 billion. That money is going into AI chips, data centres, and infrastructure expansion worldwide.

The company has also been aggressively pushing its TPUs into the market through partnerships with firms like Anthropic and Meta. Until recently, Nvidia dominated nearly every part of the AI hardware business, but Google, Amazon, AMD and Intel are now all trying to build alternatives.

Even Nvidia’s dominance is beginning to face questions. Chaim Siegel, analyst at Elazar Advisors, told Reuters that the demand for AI chips remains massive, but many companies simply do not have enough data centres ready to deploy them at scale. That creates an opening for companies that can provide both infrastructure and AI hardware together.

And that is exactly what Google and Blackstone are now trying to build.