Further reports say NVIDIA is still considering investing in OpenAI, now comes a further report saying OpenAI is not very satisfied with the performance of NVIDIA chips, and is looking at other alternative chips.
It is said that OpenAI is not satisfied with the speed offered by NVIDIA chips in providing answers to users, nor in use in conversations and various other artificial intelligence tasks.
This move is also seen to continue to cloud the relationship between the two parties. NVIDIA chips are still widely used for the purpose of training the latest models, but now many artificial intelligence model developers are looking for alternatives to chips involving the inference process.
For your information, inference process support is done when a service responds to user queries. It also requires high memory to allow it to pull data and perform mathematical calculation operations. Google is seen to have a fairly good chip for the purpose of this inference process.
Today, Google and Anthropic are using Google's TPU chip for this inference purpose. Previous rumors also said that Meta may also use a Google-made chip.
At this time, it is unknown whether this will impact NVIDIA's future investment in OpenAI or vice versa.
