DeepSeekV4 will be released in the next few weeks or will run completely on Huawei chips

April 4 news, According to the latest US technology media TheInformation report, China AI start-up DeepSeek will release a new generation flagship model V4, Will run completely on Huawei self-developed chips, Is regarded as China to promote semiconductor autonomy process important milestone. The V4 is expected to be released in the next few weeks and will run entirely on Huawei chips. DeepSeek worked with Huawei and chip designer Cambrian for several months to rewrite some of the underlying code of the model to achieve compatibility with domestic chips. It is worth noting that DeepSeek did not open V4 advance testing rights to Nvidia this time, but only invited domestic chip companies to participate in early optimization, breaking industry practice. China technology giants such as Alibaba, ByteDance, and Tencent have purchased Huawei's latest Ascend950PR chips in advance, with a total order volume of hundreds of thousands. They plan to deploy V4 through cloud services and integrate it into their own AI applications. The surge in demand pushed the chip's price up by about 20%. The V4 adopts a Hybrid Expert Architecture (MoE), with a total number of approximately 1 trillion parameters, and about 37 billion parameters are activated per inference. It can support multimodal input of text, images, and code while maintaining low latency. Last year, DeepSeek released low-cost models V3 and R1, which triggered a sharp drop in global technology stocks, causing the market to question whether U.S. AI companies need to spend billions of dollars in purchasing computing power. As a result, V4 has attracted international attention.