Lates News

According to multiple sources, Ideal Auto is exploring the development of its own cloud inference chip, which will adopt the same data flow architecture as the smart driving chip; the project is still in its early stages. AI data centers typically use GPUs to complete model training and inference, but some data centers have deployed cloud inference chips. An industry insider stated that cloud inference chips can take on some of the inference tasks that originally run on GPUs, such as data processing, testing, simulation of smart driving models, and request processing for large language models. The part involving parameter updates in model training still requires training chips. (Late Auto)