OpenAI is planning to purchase artificial-intelligence chips developed by Google, according to a report by The Information cited by Reuters on October 28. The reported move would give the ChatGPT maker another source of computing capacity as demand for AI training and deployment continues to grow.
The report did not disclose the size, timing or financial terms of a potential purchase. Neither company had publicly confirmed the reported plan at the time of publication, so the transaction should not be treated as a completed deal.
Google’s Tensor Processing Units, or TPUs, are designed for machine-learning workloads and compete with Nvidia’s widely used GPUs. OpenAI has relied heavily on Nvidia hardware, while the reported interest in Google’s chips points to a broader effort across the industry to secure enough computing power and reduce dependence on a single supplier.
The two companies have already expanded their commercial relationship in a separate area. OpenAI announced on October 21 that its models and Codex workflows would become available through Google Cloud’s Model Garden. That partnership is confirmed; the reported chip purchase remains unconfirmed.
