OpenAI said it has no current plans to deploy Google’s in-house Tensor Processing Units at commercial scale, clarifying that the chips are only being used in early testing. A spokesperson told Reuters on 30 June that introducing new hardware broadly would require extensive architectural and software changes. The ChatGPT maker remains reliant on Nvidia graphics processors and AMD accelerators to meet growing demand and continues to rent capacity from cloud specialist CoreWeave. OpenAI is simultaneously designing its own silicon and expects to reach the “tape-out” design-completion milestone later this year. Reuters had reported earlier this month that OpenAI had signed up for Google Cloud services, fuelling speculation that the company might shift more workloads to its rival’s TPUs. Google declined to comment on the latest clarification.