According to media reports citing people familiar with the matter, DeepSeek V5 plans to use Huawei's Ascend 950 chips as its core training compute, with the project timeline pointing to a launch next year. At this stage, DeepSeek and the Ascend team have already completed extensive adaptation and custom development of low-level operators during the V4 phase. According to the collaboration path disclosed by the people familiar with the matter, in DeepSeek's previous-generation version, Ascend compute only handled inference tasks and had not yet entered the training process. V5 will be its first large model natively trained on Ascend 950.