OpenAI to lease Google’s AI chips

 

 

 

OpenAI has reportedly started using Google’s artificial intelligence chips to power ChatGPT and other products. Reuters reports that the move is part of OpenAI’s efforts to diversify chip suppliers and reduce its reliance on Nvidia, which currently dominates the AI hardware market.

OpenAI is currently one of the largest customers of Nvidia’s graphics processing units (GPUs). The graphics processing chips are used for training models and computing. However, as the demand for computing power increases, OpenAI is now looking for alternatives. As part of this, OpenAI has started using Google’s Tensor Processing Units (TPUs), a significant shift not only in its hardware strategy but also in its reliance on cloud services, the Reuters report claims.

Earlier this month, Reuters reported that OpenAI was planning to use Google Cloud to meet its growing computational needs. The collaboration between the two leading AI rivals surprised the tech world. The new report comes on the heels of that. Until now, Google has limited its use of its in-house tensor processing units (TPUs) to its own use. The deal with OpenAI comes as it expands its external availability. Google's move to lease the TPUs marks the first time OpenAI has used non-Nvidia chips.