TECH NEWS – The Chinese company is facing serious complications due to the AI chip export restrictions imposed by the U.S. President.

Since taking office, Donald Trump has made it a central priority to ensure the United States maintains its leadership in the global AI chip race. This led the administration to swiftly implement export restrictions on several of Nvidia‘s top-selling chips in China, including the H20 AI accelerator. Deprived of vital computing resources, DeepSeek is reportedly encountering major hurdles with its upcoming model, R2, and CEO Liang Wenfeng is dissatisfied with the current performance of the large language model (LLM).

Despite China obtaining AI chips through various backdoor methods, the most recent round of export controls has effectively choked the supply of top-tier accelerators to the mainland. That’s a major setback for DeepSeek, which previously shook the industry with its R1 model, triggering a dip in Nvidia’s market cap worth billions, fueled by rumors of DeepSeek’s limited funding. As a result, anticipation for the R2 model was massive.

Right now, DeepSeek has not announced a specific launch date for R2 in China. Reports suggest that much of the domestic inventory of Nvidia H20 GPUs is currently being utilized by cloud clients running R1, leaving the company with a shortage of accelerators that makes rolling out R2 even more difficult. The U.S. restrictions are clearly succeeding in slowing China’s AI growth.

There have been past claims linking DeepSeek to the Chinese military, alleging it accesses high-end chips via Asian shell companies — evidence that the company is doing everything it can to acquire Nvidia‘s top AI hardware. But those efforts appear to be falling short.

Meanwhile, Nvidia has yet to roll out a domestic-market solution, and while alternatives like Huawei exist, DeepSeek cannot rely solely on them for widespread adoption.

Source: WCCFTech, The Information