Nvidia grasping to hold onto China’s AI chip market – Asia Times
Nvidia CEO Jensen Huang was in Beijing on April 17 two weeks after the Trump administration imposed a sales ban on his company to China.
Huang, whose visit merited a reception by Chinese Vice Premier He Lifeng at the Great Hall of the People, also met with Ren Hongbin, chairman of the China Council for the Promotion of International Trade (CCPIT ), and Liang Wenfeng, founder and CEO of DeepSeek.
He added that” We welcome more US businesses, including Nvidia, to deepen their presence in the Foreign market and utilize their talents these to gain an edge in international contest.” He is also a part of the Politburo. In response, Huang remarked,” We look forward to deepening our presence in China and supporting the development of the local software ecosystem.”
In a split conference arranged by CCPIT, Huang told Ren,” We hope to continue to collaborate with China”. He was quoted by Chinese media as saying,” Nvidia will continue to make every effort to improve its product lineup in accordance with regulatory requirements and will steadfastly offer the Chinese marketplace,” according to him.
Huang officially discussed how Nvidia may give DeepSeek with AI chips that would satisfy both the company’s needs and regulation requirements during a meeting with Liang.
Nvidia issued a statement saying,” We regularly meet with government leaders to discuss our company’s products and technology”, but these were not regular meetings. The business is now at the core of US-China commerce and technology disputes that are becoming more and more thorny.
Nvidia had a busy year. In the evening of April 15, Nvidia revealed that imports of its H20 AI computers and similar devices to China and other countries of concern then require a permit from the US government, an attempt that “addresses the danger that the included products may be used in, or diverted to, a computer in China”.
Nvidia’s promote price dropped 6.3 % to US$ 105.10 in after-hours investing. By Thursday’s close ( Friday was a holiday ), it was down$ 104.42, bringing its year-to-date decline to$ 26.7 %.
With no permit likely to be granted, Nvidia stated that its results for the current fiscal quarter ending on April 27 “are expected to contain up to around$ 5.5 billion of costs associated with H20 materials for inventory, purchase commitments, and associated reserves”.
AMD’s MI308 AI accelerators are subject to the same new restrictions, which dropped 7.1 % in after-hours trading on Tuesday and finished the week down 27.5 % from the beginning of the year. ADM anticipates publishing special charges in the neighborhood of$ 800 million.
Intel’s Gaudi 3 processor is also affected. In the month to Thursday, Intel’s share price dropped 27 % due to this and other factors.
AMD’s MI308 was created with less performance in mind to meet previous US government export restrictions, similar to those on Nvidia’s H20. So were redesigned versions of Intel’s Gaudi 2 and Gaudi 3 AI accelerators.
The US Commerce Department’s Bureau of Industry and Security ( BIS ) has placed a cap on the performance of AI processors that can be exported to China for the third time since October 2022, and Nvidia, AMD, and Intel have created new, less effective chips to lower the cap after Nvidia, AMD, and Intel have created new, less efficient versions of their chips. President Trump is implementing the same policy as President Joe Biden in this regard.
For Nvidia, the restriction timeline runs as such:
- October 2022: The Biden administration halted exports of Nvidia’s most advanced AI processors, the A100 and H100 GPUs.
- November 2022: Nvidia introduced the bogeyed-down A800, which was made compliant with BIS requirements for sale in China.
- March 2023: Nvidia launched the , H800, a low-performance version of the H100.
- November 2023: BIS halted exports of A800 and H800 GPUs.  ,
- March 2024: Nvidia launched , the H20, which met the new, tighter BIS requirements.
- April 2025: Trump administration; exports; of; H20; GPUs; and a blockade.
First, the reasons for this chain of events are that, despite the limitations, Chinese demand remained strong despite the performance of the available chips, and that, second, Chinese AI capabilities continued to advance despite the restrictions.
In addition to revealing the behind-the-curve incompetence of US government analysis and response, it demonstrates that, in the case of semiconductors, anything the Chinese want to buy, the US will refuse to sell, all the while complaining about its trade deficit with China.
Another wave of McCarthyite paranoia was sparked by the astonishing success of the Chinese AI model DeepSeek, which was trained using Nvidia H800 chips. According to The New York Times in January, DeepSeek” created a less expensive, competitive chatbot with fewer high-end computer chips than US behemoths like Google and OpenAI, showing the limitations of chip export control.”
On April 16, Chairman John Moolenaar (R-MI ) and Ranking Member Raja Krishnamoorthi (D-IL ) of the House Select Committee on the Chinese Communist Party released a report entitled” DeepSeek Unmasked: Exposing the CCP’s Latest Tool for Spying, Stealing, and Subverting US Export Control Restrictions”. DeepSeek poses” a serious threat to the United States ‘ national security,” according to the committee.
Chairman Moolenaar stated that:
DeepSeek isn’t just another AI app — it’s a weapon in the Chinese Communist Party’s arsenal, designed to spy on Americans, steal our technology, and subvert US law. We now know that this tool reportedly used cutting-edge Nvidia chips that shouldn’t have ended up in CCP hands and used US AI models. In order to demand answers, we’re writing a letter to Nvidia. American innovation should never be the engine of our adversaries ‘ ambitions.
However, only two days prior, on April 14, Nvidia made the announcement that it would produce up to$ 500 billion worth of supercomputers and other AI infrastructure in the US over the next four years.
In order to accomplish this, Nvidia is collaborating with Taiwanese contract manufacturers Foxconn ( Hon Hai Precision ) and Wistron to construct manufacturing facilities in Texas, Texas. The Blackwell AI processors at the heart of the AI supercomputers made there will be supplied by Taiwanese semiconductor foundry TSMC from its factories in Arizona.
Amkor and SPIL ( Silicon Precision Industries Co., Ltd. ), both located in Arizona, will take the chips together, package, and test them. As the name suggests, Amkor is Korean-American, whereas SPIL is Taiwanese.
CEO Huang said,” The engines of the world’s AI infrastructure are being built in the United States for the first time. Addition of American manufacturing improves our ability to meet the enormous and growing demand for supercomputers and AI chips, strengthens our supply chain, and increases our resilience.
That was what Trump aspired to hear. ” It’s the Trump Effect in action“, read a press release from the White House.
Although it won’t cost much and may take longer than Nvidia hopes, building AI supercomputers in the US should eventually be done with the support of so many top-notch companies.
Chinese server manufacturer H3C reported at the end of March that Nvidia H20 processors, which appear to be nearly sold out, were short. Alibaba, Tencent, ByteDance and other Nvidia customers will also be affected, indicating that the new BIS restrictions are likely to disrupt China’s AI computing industry as intended.
Additionally, Nvidia could lose up to$ 15 billion in annual sales in addition to the anticipated$ 5.5 billion in special charges, which will benefit H3C, Huawei and Cambricon, two Chinese AI chip designers, for example.
The most technologically advanced Chinese substitute to Nvidia is already Huawei’s new Ascend 910C AI processor, which is reportedly the most advanced. Cambricon, which was founded in 2016, is much smaller than Huawei but has become a darling of the Chinese stock market, rising by about five times over the past year. Huawei is not traded on the stock exchange.
DeepSeek has already been distributed throughout China to People’s Liberation Army support services, corporate services, finance and other businesses, city government, and government agencies. Rather than having been designed” to spy on Americans”, as Congressman Moolenaar claims, it aims to provide AI solutions to practical problems across Chinese society.
Meanwhile, Dylan Patel and his fellow editors at SemiAnalysis, a well-known publication, claim that Huawei’s new CloudMatrix 394 AI accelerator, which is based on the Ascend 910C,” competes directly” with Nvidia’s premium GB200 Grace Blackwell Superchip and that its “rack scale solution” is “more advanced than Nvidia’s rack scale solution”.
” Rack scale solution” refers to the complete data center system, including GPUs, servers, networking, storage, power management and cooling.
According to SemiAnalysis,” the engineering advantage is at the system level not just at the chip level, with innovation at the accelerator, networking, optics, and software layers… Huawei is a generation behind in chips, but its scale-up solution is arguably a generation ahead of Nvidia and AMD’s current products on the market.”
Although Huawei’s solutions required more electricity, SemiAnalysis came to the conclusion that” China’s power deficiencies are relevant but not a limiting factor.”
Once again, US sanctions appear to be too little, too late, and more likely to promote rather than prevent the advance of Chinese technology. Nvidia and its smaller American rivals are likely to find themselves increasingly marginalized in the booming Chinese market for AI processors unless Trump changes tack in a deal with Beijing.
Follow this author on Twitter at X: @ScottFo83517667.
