OpenAI recently started renting Google’s artificial intelligence chips for Power ChatgPT and its other products, a source close to Reuters told Friday.
ChatgPT manufacturer is one of the largest buyers of NVIDIA’s Graphic Processing Units (GPUS), using AI chips to train models and also for inference computing, a process in which an AI model uses its trained knowledge to make predictions or decisions based on new information.
OpenAi planned to add the Google Cloud service to meet its growing computing capacity needs, Reuters reported exclusively earlier this month, marking a surprising collaboration between two important AI competitors.
For Google, the agreement occurs when expanding the external availability of its internal tensor processing units (TPUs), which were historically reserved for internal use. This helped Google win customers, including Big Tech Player Apple, as well as startups such as Anthropic and Safe Superintelligence, two competitors of Chatgpt Maker released by former OpenAi leaders.
The change to rent the Google Signs TPUs the first time OpenAI has used significantly non-novidia chips and shows Sam Altman’s change of company to rely on Microsoft’s date centers. It could potentially increase TPUs as a cheaper alternative to Nvidia GPUs, according to the information, which reported the development previously.
OpenAi expects TPUS, which will rent on Google Cloud, will help reduce the cost of inference, according to the report.
However, Google, an OpenAi competitor in the AI race, is not renting its most powerful tpus to its rival, reporting, citing a Google Cloud employee.
Google refused to comment while OpenAi did not respond immediately to Reuters when contacted.
The addition of OpenAi by Google to its customer list shows how the technology giant has capitalized on its internal AI technology, from hardware to software, to accelerate the growth of its cloud businesses.
© Thomson Reuters 2025