DeepSeek is having trouble accessing NVIDIA's cutting-edge AI chips, as a new report says that US export controls are delaying China's AI ambitions.
U.S. Export Controls Have Limited the Flow of AI Chips to Mainland China, Which Has Stalled the Release of DeepSeek's R2
Since President Trump took office, a key focus has been on ensuring that the US stays ahead in the AI chip race, and that motive led the administration to promptly impose export controls on some of NVIDIA's best-selling chips in China, the H20 AI accelerator. By essentially being stripped away from computing power, DeepSeek is now claimed to have been facing troubles with its next R2 AI model, since a report by The Information claims that CEO Liang Wenfeng is not happy with the performance of the LLM, and Chinese CSPs cannot readily adopt the R2 model, due to lack of NVIDIA chips.
Despite China's role in getting access to AI chips through various "workarounds", there's no doubt that the recent round of export controls has restricted the flow of high-end accelerators to the mainland, which has bothered DeepSeek. The company made huge strides with its previous model, the DeepSeek R1, which erased billions in market cap of NVIDIA after it was rumored that the company utilized confined financial resources, so there was a lot of hype towards DeepSeek's next model, presumably the R2.
The report claims that DeepSeek currently doesn't have a definite time for when the R2 AI model could be pushed out in the Chinese markets, since domestic CSPs are finding it difficult to accumulate enough computing resources to power the upcoming model. It is claimed that a significant stockpile of NVIDIA's H20 AI GPUs in the domestic market is currently occupied with cloud customers running DeepSeek R1, and with the lack of accelerators, deploying R2 becomes even more difficult. It's safe to say that US export controls have done their job limiting China's AI expansion.
In a previous report, it was claimed that DeepSeek is involved with the Chinese military, and has access to AI chips through "shell companies" in Asia, which shows that the Chinese AI firm is trying all means to get its hands on NVIDIA's high-end AI chips, but is apparently failing to do so. NVIDIA has yet to introduce a solution for domestic markets, and with alternatives like Huawei, DeepSeek cannot rely solely on them for mass deployment.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
