Nvidia has reportedly halted production of its H200 artificial-intelligence (AI) chips intended for China amid rising political tensions between the United States and China and Beijing\u2019s policy of supporting domestic chips.<\/p>\n
The company has reallocated manufacturing capacity at Taiwan Semiconductor Manufacturing Company (TSMC) away from producing H200 chips to its next\u2011generation Vera Rubin hardware, the Financial Times reported<\/a>, citing two people familiar with the situation.\u00a0<\/p>\n The report said Nvidia decided it could not remain in regulatory limbo between the US and China and had to focus on products with clearer market prospects, particularly as demand for its most advanced chips remains strong.<\/p>\n US President Donald Trump said<\/a> on December 8 last year that\u00a0Washington would allow exports of the H200 chips to China for civil use. Media reports said Chinese internet giants Alibaba, Tencent and ByteDance were seeking to purchase 400,000 units.<\/p>\n However, after Trump formally approved limited H200 exports to China on January 13, Chinese customs authorities informed<\/a> Nvidia that the chips would not be allowed to enter the country. In late January, Beijing fine-tuned<\/a> its line by saying that Chinese companies can purchase the H200 but should consider local chips first.\u00a0As of now, no H200 chips have been \u200csold<\/a> yet to Chinese customers, according to US officials.\u00a0<\/p>\n Some Chinese commentators said<\/a> Beijing\u2019s decision to curb H200 imports was due to the United States\u2019 interception of 1.8 million barrels of Venezuelan oil destined for China on December 20 last year.<\/p>\n \u201cA tanker that had just left Venezuela was intercepted on the high seas, with armed personnel boarding and redirecting it,\u201d Xia Yuanqi, a Shanghai\u2011based columnist, writes<\/a> in an article. \u201cThe ship carried a Panamanian flag, the operator was based in Hong Kong and the cargo owner pointed to a Chinese petrochemical company. The interception was less about the oil and more about demonstrating US law enforcement\u2019s power.\u201d<\/p>\n \u201cThe seized heavy crude may be worth more than a billion dollars at market prices, but Nvidia\u2019s H200 chips cost 10 times more,\u201d he says. \u201cNvidia wanted Chinese firms to buy its H200 chips. It definitely felt the pressure.\u201d<\/p>\n \u201cBy seizing Venezuela\u2019s oil, the US wants to show its muscle to the whole world,\u201d a Guizhou-based writer says<\/a>. \u201cBut China no longer accepts this. Over the past two years, China\u2019s chip technology has advanced rapidly. Besides, we also hold\u00a0the rare earth card.\u201d<\/p>\n \u201cThe US wants to use the H200 chips to squeeze the final profits from China, but the Chinese market has changed. The global supply of AI computing power has also been diversified,\u201d he says.<\/p>\n Ascend 950PT<\/p>\n Chinese media and commentators had previously welcome<\/a>d the prospect of Nvidia\u2019s H200 chips entering the Chinese market. They said the H200 would be used for training AI models such as DeepSeek, while domestic chips would handle AI inference, a\u00a0stage when trained models respond to user queries or perform real\u2011time tasks.<\/p>\n Jin Canrong, a professor and vice dean of the School of International Relations at Renmin University of China, said<\/a> on December 19 that Trump had given China a good opportunity to upgrade its AI sector.<\/p>\n After the seizure the following day of the Venezuelan tanker,\u00a0Jin changed<\/a> his views. He said on December 25 that China might ultimately not need the H200 chips, as domestic chipmakers were preparing substitutes that could soon enter the market.\u00a0<\/p>\n On the same day, the Fudan Development Institute published<\/a> a research report stating that Trump\u2019s approval of H200 exports was primarily a strategy to help Nvidia preserve its dominance in the AI chip market, rather than a sign that Washington was easing its technological blockade against China.\u00a0<\/p>\n Some Chinese commentators said banning Nvidia\u2019s hardware could provide more market space for Huawei Technologies\u2019 Ascend AI chips.<\/p>\n \u201cChina\u2019s ban on Nvidia\u2019s H200 chips seems to be a single issue, but it reflects the broader technology competition,\u201d says<\/a>\u00a0a Guangdong-based columnist writing under the pen name \u201cHY Skywalk.\u201d\u00a0\u201cWashington\u2019s approach is to offer China advanced technology and maintain its chipmakers\u2019 market share. Beijing\u2019s response is to endure short\u2011term pain while accelerating development of core technologies so the country controls its own technological future.\u201d<\/p>\n He points out that Huawei\u2019s Ascend 910B, Cambricon\u2019s Siyuan 590 and Biren Technology\u2019s BR100 are now being used for AI training and inference. He says Huawei\u2019s Ascend roadmap includes the Ascend 950PR planned for 2026, followed by the Ascend 960 in 2027 and the Ascend 970 in 2028.<\/p>\n High\u2011bandwidth memory<\/p>\n Nvidia\u2019s Hopper AI chips, including the H100 and H200, are more powerful than most Chinese alternatives largely because they rely on high\u2011bandwidth memory (HBM), a type of dynamic random access memory (DRAM) produced by South Korea\u2019s SK Hynix. In a personal computer, insufficient DRAM can cause slowdowns or system freezes.<\/p>\n Hopper chips use SK Hynix\u2019s HBM3 and the enhanced HBM3e, while newer Blackwell processors use HBM3e. The next\u2011generation Vera Rubin architecture is expected to adopt HBM4. The higher the bandwidth of these memory chips, the faster the data can move between AI processors during training.<\/p>\n SK Hynix began<\/a> mass production of HBM3 in June 2022 for Nvidia\u2019s H100 GPU. It was only in December 2024 that\u00a0the Biden administration blocked South Korea from exporting HBM chips and related manufacturing equipment to China.<\/p>\n Huawei said<\/a> last September that it developed its own HBM chips, called HiBL 1.0 and HiZQ 2.0. It said HiBL 1.0 is expected to be used in the Ascend 950PR chip, scheduled for release in the first quarter of 2026. <\/p>\n It remains unclear how Huawei and its supplier, ChangXin Memory Technologies (CXMT), could advance HBM technology so quickly.<\/p>\n South Korean police have repeatedly arrested former employees of chipmakers for stealing semiconductor secrets and passing them to China.\u00a0<\/p>\n In April 2024, a Chinese national who had worked at SK Hynix for about a decade was arrested<\/a> at a South Korean airport after allegedly printing 3,000 pages of the company\u2019s confidential documents for Huawei.<\/p>\n Last May, a former SK Hynix employee, surnamed Kim, was accused<\/a> of copying more than 11,000 documents at the company\u2019s China office and later attempting to apply for a job at Huawei. Last December, South Korean prosecutors indicted<\/a> 10 people suspected of leaking Samsung\u2019s HBM technology to CXMT.\u00a0<\/p>\n On January 19 this year, South Korean police said<\/a> they arrested 378 suspects linked to the leaking of HBM\u2011related technologies and detained six of them in 2025. They said more than half of last year\u2019s overseas technology leak cases were linked to China.<\/p>\n Read: Beijing to approve Nvidia H200 imports, flagging overreliance<\/a><\/p>\n Follow Jeff Pao on Twitter at\u00a0@jeffpao3<\/a><\/p>\n \t