Human
SHANGHAI: In a move that may redefine Chinese artificial intelligence infrastructure, DeepSeek, a prominent domestic AI research lab, has developed its next-generation large language model to run entirely on Huawei Technologies hardware, according to recent reports.
DeepSeek's forthcoming V4 model will utilize Huawei's Ascend AI chips for its operations, The Information reported on Friday. This development represents a significant strategic shift towards utilizing purely domestic technology for advanced AI development, a goal that has gained critical importance in China amid persistent and tightening United States export controls on high-performance semiconductor components.
The successful integration of the V4 model with Huawei’s Ascend architecture is being framed by analysts as a major milestone for the Chinese tech giant's semiconductor ambitions. For DeepSeek, the collaboration ensures that the training and deployment of its leading-edge models will remain insulated from potential supply chain disruptions or future international sanctions affecting American chip designers. According to Huawei Central and others, this strategic alliance could significantly boost confidence in Huawei’s ecosystem, positioning the Ascend platform as a viable, independent alternative to industry-standard hardware, which has become difficult for many Chinese entities to acquire.
Speculation surrounding DeepSeek-V4 has intensified in recent months. While the official launch date remains unconfirmed, Chinese media reports have suggested a potential April release window for the new iteration. The lab has established a clear pattern of regular releases, indicating that preparations for the V4 launch are likely ongoing, with the V3 model having set high performance expectations.
Syndicated by The China Technology Review.
