DeepSeek V4 will be released in the coming weeks and may run entirely on Huawei chips.

April 4 news: According to the latest report by U.S. technology media The Information, DeepSeek, an AI startup in China, will soon release its new-generation flagship model V4, which will run entirely on Huawei’s in-house developed chips. It is seen as an important milestone for China’s push toward semiconductor self-reliance.
V4 is expected to be released within the next few weeks and will run entirely on Huawei chips. For this purpose, DeepSeek has been working with Huawei and chip designer Cambricon Technology for several months, rewriting parts of the model’s underlying code to achieve compatibility with domestic chips.
Notably, DeepSeek did not open up early testing access to V4 to Nvidia this time. Instead, it only invited domestic chip companies to participate in early optimization, breaking with industry convention.
Chinese tech giants such as Alibaba, ByteDance, Tencent, and others have already placed advance orders for Huawei’s latest Ascend950PR chips, with total order volume reaching tens of thousands of units. They plan to deploy V4 through cloud services and integrate it into their own AI applications. Demand surges have driven the chip’s price up by about 20%.
V4 uses a mixture-of-experts (MoE) architecture, with a total parameter count of about 1 trillion. For each inference, it activates about 37 billion parameters, supporting multimodal inputs of text, images, and code while maintaining low latency.
Last year, DeepSeek released low-cost models V3 and R1, which triggered a global selloff in technology stocks and led the market to question whether U.S. AI companies need to spend tens of billions of dollars to buy computing power. V4 has therefore drawn significant international attention.