Gold bars to be exempt from tariffs, White House clarifies
Investing.com -- OpenAI has started using Google’s (NASDAQ:GOOGL) artificial intelligence chips to help power ChatGPT and related services, marking its first significant shift away from exclusive reliance on Nvidia (NASDAQ:NVDA) hardware, according to a report by The Information. The move is part of a broader strategy by the AI company to reduce its dependence on Microsoft(NASDAQ:MSFT)-managed infrastructure.
Through Google Cloud, OpenAI is renting Google’s tensor processing units (TPUs) with the aim of cutting the costs associated with inference computing, the execution of models after training is completed. The decision could offer Google’s TPUs a higher profile as a cost-effective alternative to Nvidia’s widely used graphics processing units (GPUs), which dominate the AI sector.
Previously, OpenAI sourced Nvidia chips primarily via partnerships with Microsoft and Oracle (NYSE:ORCL) to train and deploy its models. While Google is providing some TPU capacity, it is reportedly not offering its most powerful versions to OpenAI, according to sources cited by The Information.
That limitation suggests Google’s most advanced TPUs remain reserved for internal use, including work on its own large language models under the Gemini project. For OpenAI, access to earlier versions of the TPUs still represents a step toward infrastructure diversification amid growing industry demand.
It’s still unclear whether OpenAI will use Google chips for model training or limit them to inference workloads. As competition increases and resource constraints deepen, a hybrid-use infrastructure could provide new flexibility for scaling.
The arrangement highlights the evolving dynamics of the AI hardware landscape, where companies like Google are leveraging years of investment in both software and custom silicon. For OpenAI, the addition of Google as a chip supplier broadens the ecosystem around its technology stack and addresses growing concerns over availability and cost of compute resources.