Nvidia Corp. (NVDA) has reinforced its partnership with OpenAI, confirming that OpenAI will not deploy Google’s in-house AI chips (TPUs) for large-scale uses. Instead, OpenAI continues to depend on Nvidia’s GPUs and AMD AI chips to meet rising computational needs. This announcement follows a recent cloud services agreement between OpenAI and Google Cloud, which allows OpenAI to leverage Google’s infrastructure for training models like ChatGPT, reducing its reliance on Microsoft Azure. Although OpenAI is testing Google TPUs, they have no plans for extensive usage, maintaining a focus on Nvidia’s technology. Meanwhile, Google is making its TPUs available for external customers, hoping to gain traction against Nvidia’s dominance in performance. This strategic partnership highlights the dynamics of the competitive AI landscape, where collaboration and technology sourcing play crucial roles in driving innovation.
Source link
