OpenAI inks deal with Google Cloud

OpenAI is turning to Google for help with its growing computing needs, three sources have told Reuters. The move is unexpected given the rivalry between the two in artificial intelligence, but it reflects the massive demand for computing power to train and run AI models.

One of the three sources said the deal was finalised in May after months of discussion. OpenAI has been working to ease its reliance on Microsoft by adding other cloud providers to meet growing demand. That includes supporting large-scale efforts like its Stargate project and keeping up with commitments made in recent deals with chipmakers and infrastructure vendors.

For Google, the deal brings a new high-profile customer to its cloud platform. The company will provide extra computing resources to support OpenAI’s AI training and model operations, the sources said. This comes as ChatGPT continues to challenge Google’s core search business. Still, Google executives recently acknowledged that there may be room for more than one winner in the AI space.

OpenAI, Google, and Microsoft declined to comment.

Since launching ChatGPT in late 2022, OpenAI has faced growing pressure to scale up its compute capacity – both for training large language models and for inference, or running those models in real time. On Monday, OpenAI said its annualised revenue had reached $10 billion as of June 2025, fuelled by rising use of its tools.

Earlier this year, OpenAI teamed with SoftBank and Oracle on the $500 billion Stargate infrastructure effort and signed multi-billion-dollar agreements with CoreWeave. The company is also working on designing its own chip to cut back on outside hardware use, Reuters reported in February.

The deal with Google follows a shift earlier this year that ended OpenAI’s exclusive reliance on Microsoft’s Azure. A person familiar with the situation said OpenAI and Google had been in talks for a while, but existing agreements with Microsoft had delayed any deal. Microsoft and OpenAI are now reworking the terms of their partnership, including Microsoft’s future ownership stake.

For Google, the timing lines up with its push to expand use of its in-house chips, called tensor processing units or TPUs. TPUs were once limited to internal projects but are now being used by outside clients, including Apple and OpenAI competitors including Anthropic and Safe Superintelligence.

Adding OpenAI to its cloud customer list gives Google another way to grow its cloud business, which brought in $43 billion last year – about 12% of Alphabet’s total revenue. Google has pitched itself as a neutral cloud provider for startups looking to avoid getting locked into Amazon or Microsoft, and AI companies have become a major source of spending.

Still, Google faces its own pressures. The company is pouring money into AI and expects capital spending to reach $75 billion this year. At the same time, it’s under scrutiny from regulators and pressure from investors who want to see returns on their billions.

Internally, the deal adds another layer of complexity. Google’s AI lab DeepMind competes with OpenAI, and selling computing power to a rival can create tension over how resources are shared. CEO Sundar Pichai will have to weigh the demands of Google’s enterprise customers against those of its own AI teams.

The strain is already being felt. Google’s finance chief said in April that the company didn’t have enough cloud capacity to meet demand in the previous quarter. And while ChatGPT leads Google’s Gemini in user numbers, Pichai has played down the risk of OpenAI cutting into the company’s core business.

(Photo by Jonathan Kemper)

See also: European cloud providers play the sovereign card

Want to learn more about cybersecurity and the cloud from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London.

Explore other upcoming enterprise technology events and webinars powered by TechForge here.