Summary
China’s Cyberspace Administration (CAC) in 2024 issued directives compelling major Chinese technology companies, including ByteDance and Alibaba, to halt testing and cancel orders of advanced Nvidia semiconductor products, notably the RTX Pro 6000D-FT GPUs. This regulatory move reflects escalating geopolitical tensions between China and the United States amid ongoing U.S. export controls designed to restrict China’s access to cutting-edge artificial intelligence (AI) hardware. The directive underscores China’s intent to assert greater control over foreign technology imports, citing national security concerns such as alleged vulnerabilities and potential backdoors in Nvidia’s chips, claims that Nvidia has firmly denied.
The affected Nvidia products, including the Blackwell-based RTX 6000D and the AI-focused H20 chip, are central to China’s AI development ambitions but have become focal points in the intensifying technological rivalry. Despite Nvidia’s efforts to adapt its offerings for the Chinese market and maintain substantial sales—accounting for about 13% of its revenue—the regulatory clampdown has disrupted supply chains and challenged the company’s market presence in China. Concurrently, Chinese firms are accelerating the development of indigenous semiconductor technologies, with ByteDance investing over $20 billion in AI infrastructure and collaborating with partners like Broadcom and TSMC to produce proprietary AI processors, while Alibaba commits more than $53 billion yuan toward chip innovation.
This dynamic has triggered strategic adjustments by Chinese companies, who navigate a complex environment of partial reliance on foreign GPUs—often accessed through foreign cloud services—and aggressive investment in domestic alternatives to mitigate the impact of U.S. sanctions and Chinese regulatory scrutiny. The broader context involves a heightened global semiconductor competition where advanced chips are viewed as critical assets for national security and technological leadership, intensifying policy responses from both countries that shape the future of AI hardware markets worldwide.
Background
In recent years, escalating geopolitical tensions between the United States and China have significantly impacted the semiconductor industry, particularly in the realm of advanced artificial intelligence (AI) hardware. The U.S. government has imposed stringent export controls aimed at restricting China’s access to cutting-edge semiconductor technology, including high-performance graphics processing units (GPUs) critical for AI development. These measures have targeted some of China’s most prominent technology firms, including ByteDance and Alibaba, effectively curtailing their ability to procure and test advanced Nvidia products such as the RTX Pro 6000D-FT chips.
ByteDance, the parent company of TikTok, has reportedly planned a substantial investment exceeding $20 billion in AI infrastructure, with around $7 billion earmarked for accessing advanced Nvidia GPUs via cloud services, data centers, and even submarine cables. However, direct purchases of Nvidia GPUs by ByteDance have been prohibited under U.S. sanctions, prompting the company to explore alternative means such as renting Nvidia GPUs located in foreign cloud data centers, thereby circumventing some restrictions. To reduce reliance on foreign technology, ByteDance is also collaborating with Broadcom to develop proprietary AI processors, expected to be manufactured by Taiwan Semiconductor Manufacturing Company (TSMC) using advanced 5-nanometer process technologies, with mass production anticipated by 2026.
Similarly, Alibaba is intensifying efforts to develop homegrown AI chips amid the constraints imposed by U.S. export controls. The company is reportedly working on new AI-focused semiconductor designs as part of a broader strategy to enhance its AI capabilities and infrastructure investment, which is projected to reach at least 380 billion Chinese yuan ($53.1 billion) over three years. Alibaba and other Chinese firms, such as Baidu and Huawei, have also pursued open-source AI models to bypass U.S. restrictions, decentralize development, and engage global talent for refining their technologies. This trend reflects China’s broader ambition to remain competitive in the global AI race despite technological embargoes.
Nvidia itself remains a central figure in this geopolitical contest. The company has agreed to pay a percentage of its revenues from chip sales in China to the U.S. government to resume some transactions, such as the sale of its H20 chips, which, while not its most powerful, are highly sought after by Chinese AI firms. Nvidia has also maintained substantial inventory levels and continues to engage in sample deliveries to Chinese clients, signaling ongoing demand despite regulatory hurdles.
Chinese Government Directives on Nvidia Products
In 2024, China’s Cybersecurity Authority, specifically the Cyberspace Administration of China (CAC), issued directives to major Chinese technology companies, including ByteDance and Alibaba, to halt testing and cancel orders of Nvidia’s RTX Pro 6000D-FT chips. This move was part of a broader regulatory effort targeting advanced semiconductor technologies amid growing geopolitical tensions and concerns over national security risks.
The RTX Pro 6000D, a Blackwell-based GPU specifically tailored for the Chinese market, was designed to circumvent export restrictions on Nvidia’s earlier H20 AI chips. Despite initial interest from several major Chinese firms, which had planned to order tens of thousands of these GPUs and had begun testing and verification processes, the directive abruptly stopped these activities. The CAC’s order marked a stronger stance compared to earlier guidance that had only limited the use of the H20 chips.
Chinese authorities summoned companies such as Tencent, ByteDance, and Baidu to explain their purchase intentions and assess potential information security risks related to these advanced chips. The concerns included allegations that the chips might contain security vulnerabilities or backdoors that could allow remote shutdowns or user location tracking. Nvidia has consistently denied these claims, with CEO Jensen Huang stating that no such backdoors exist in their products.
This regulatory pressure reflects China’s broader strategy to control the inflow of cutting-edge semiconductor technology while promoting domestic alternatives. Domestic chipmakers, including Huawei and Baidu, have been asked to report on their products’ competitiveness relative to Nvidia’s GPUs. The directive has had notable economic repercussions, including stock price drops for affected companies and disrupted supply chains.
The context of these measures is the intensifying US-China technological rivalry, with export controls imposed by the US on Nvidia’s H20 chips over fears of military applications by China. The Chinese government’s response illustrates its commitment to increasing scrutiny of foreign technology imports and tightening antitrust and security regulations in advanced tech sectors.
Specific Nvidia Products Affected
Several Nvidia products have come under scrutiny and regulatory pressure in China, leading to directives for major Chinese technology companies such as ByteDance and Alibaba to halt testing and procurement activities. Among the most prominently affected is the Nvidia RTX Pro 6000D-FT, a high-performance GPU that belongs to Nvidia’s professional visualization lineup. Chinese tech firms have been explicitly instructed to cease any ongoing engagements with this model as part of a broader clampdown on American semiconductor imports.
The RTX 6000D, Nvidia’s first Blackwell-based GPU designed specifically for the Chinese market, is another critical product impacted by these developments. It features advanced specifications such as GDDR7 memory, a 4nm manufacturing process by TSMC, and a memory bandwidth of approximately 1,100GB/s. Despite Nvidia’s initial plans to sell around two million units to offset losses caused by restrictions on the H20 AI chip, the RTX 6000D has not met its sales or strategic objectives due to the unfavorable regulatory environment and market conditions in China.
The Nvidia H20 AI chip, which is the most powerful AI processor Nvidia is currently permitted to sell in China, also faces significant challenges. Although less powerful than Nvidia’s H100 or Blackwell series, the H20 still outperforms many domestic Chinese alternatives. Nonetheless, Chinese authorities have raised concerns over the chip’s security, specifically regarding potential backdoor access, leading to intensified scrutiny and effective bans on its sales and deployment within government or security-related projects. This has caused many Chinese companies to reduce or suspend orders, citing political and regulatory sensitivities despite the absence of an outright ban.
These actions are part of a wider effort by Chinese regulators to curb reliance on foreign semiconductor technology and encourage domestic alternatives. This has involved not only halting sales but also convening domestic chipmakers such as Huawei and Baidu to assess and compare their products against Nvidia’s offerings, further intensifying competitive pressures on Nvidia’s market presence in China.
Company Responses and Strategic Adjustments
ByteDance, Alibaba, and other major Chinese technology firms have responded to regulatory pressures and geopolitical tensions involving Nvidia products with a combination of public statements, strategic shifts, and intensified efforts toward developing homegrown semiconductor technologies.
ByteDance, the owner of TikTok, has denied reports of halting Nvidia GPU usage despite U.S. export restrictions that prohibit the company from purchasing Nvidia GPUs directly or using American cloud services for this purpose. Instead, ByteDance continues to utilize Nvidia GPUs located in cloud data centers outside China and plans to significantly expand this usage, reportedly aiming to spend up to $7 billion on access to such GPUs. Concurrently, ByteDance is collaborating with U.S.-based Broadcom and Taiwan Semiconductor Manufacturing Company (TSMC) to develop its own advanced AI processors, including separate chips for training and inference workloads. These processors are projected to be manufactured using TSMC’s N4/N5 process technologies, with mass production expected to begin around 2026. This development aligns with ByteDance’s broader strategy to reduce reliance on Nvidia and American semiconductor technology while investing over $20 billion in AI infrastructure.
Alibaba has similarly accelerated efforts to cultivate domestic semiconductor capabilities amid growing geopolitical risks. The company is reportedly working on new AI-focused chips, building on its prior releases such as the Hanguang 800 inference chip introduced in 2019. Alibaba plans to invest heavily in AI infrastructure over the next several years, with announced commitments exceeding $53 billion to enhance its technological independence. These initiatives reflect a broader trend among Chinese technology firms to prioritize homegrown semiconductor innovation in response to tightening U.S. export controls and increasing political pressures.
In official statements, Nvidia has emphasized its compliance with all applicable laws and expressed willingness to cooperate with government agencies as export controls evolve. Nvidia’s CEO, Jensen Huang, acknowledged the complex geopolitical environment, stating the company’s presence in any market depends on governmental permissions and underscored his understanding of the larger diplomatic context influencing U.S.-China relations. Nvidia has also attempted to adapt its product offerings for the Chinese market, exemplified by the introduction of the RTX 6000D GPU, a Blackwell-based professional visualization product designed to comply with export regulations. However, reports indicate that this product has not achieved the expected market reception in China, partly due to Chinese firms accessing more powerful computing solutions through other means and their own developing chips.
Chinese authorities have reportedly urged domestic companies, including Baidu, ByteDance, and Tencent, to avoid using Nvidia’s H20 AI chips, especially for government-related applications. Nonetheless, these companies have not been formally ordered to cease all purchases of such products. Chinese firms maintain that they do not rely on American chips for critical government operations, mirroring the U.S. government’s approach to Chinese technology. This stance reflects ongoing efforts by Chinese companies to balance regulatory compliance with strategic autonomy amid escalating U.S.-China technological rivalry.
Together, these responses illustrate a multifaceted strategic adjustment by Chinese technology companies, combining continued, albeit limited, use of foreign technology with aggressive investments in indigenous semiconductor development to navigate and mitigate the impact of export controls and geopolitical challenges.
Broader Industry and Market Implications
The recent developments involving China’s directive urging companies like ByteDance and Alibaba to halt testing and orders of Nvidia products occur within a complex landscape shaped by geopolitical tensions and evolving market dynamics. The U.S. has intensified export restrictions on advanced AI chips to China, significantly impacting the availability of cutting-edge semiconductor technology. This has compelled Chinese firms to increase their reliance on domestically developed AI chips, even as leading companies such as Alibaba and Baidu continue to incorporate Nvidia hardware for their most sophisticated AI models.
The restriction on Nvidia products not only affects supply chains but also influences strategic partnerships. ByteDance, for example, has sought to mitigate procurement risks by strengthening ties with U.S. chipmakers like Broadcom, aiming to secure customized 5 nanometre application-specific integrated chips (ASICs) manufactured by Taiwan’s TSMC. This move reflects attempts to navigate export control constraints while maintaining access to advanced technologies compliant with U.S. regulations.
Nvidia’s business in China, which accounted for approximately 13% of its total sales last year, now faces heightened uncertainty amid these regulatory pressures and reciprocal trade measures. Despite CEO Jensen Huang’s repeated visits to China and efforts to foster goodwill, regulatory developments such as China’s State Administration for Market Regulation (SAMR) rulings suggest that diplomatic overtures alone may not suffice to stabilize Nvidia’s position in the market.
From a broader perspective, the technological competition between the U.S. and China has intensified the strategic importance of semiconductor technologies. Chinese government agencies and industry groups have highlighted vulnerabilities in foreign processors, including Intel, AMD, and Nvidia, further motivating Beijing to accelerate domestic semiconductor development and reduce reliance on foreign suppliers. Chinese tech firms acknowledge that Nvidia’s chips outperform domestic alternatives, underscoring the technological gap they face despite increasing investment in home-grown AI hardware.
This ongoing struggle reflects a larger narrative where advanced semiconductors are not only commercial commodities but also critical assets in national security and technological leadership. Both countries’ policies appear geared toward leveraging regulatory frameworks and diplomatic negotiations to gain competitive advantages in the rapidly growing AI sector, with China simultaneously seeking to diminish dependence on U.S. technology while challenging foreign dominance in AI markets.
Legal, Economic, and Policy Considerations
The escalating tensions between the United States and China over advanced semiconductor technology have precipitated a complex web of legal, economic, and policy challenges that directly impact companies such as Nvidia, ByteDance, and Alibaba. The United States has imposed stringent export restrictions on Chinese firms, barring them from acquiring certain American technologies on grounds of national security, which includes advanced AI chips critical to China’s technological development. In retaliation, China has initiated investigations and regulatory actions targeting American suppliers, while urging local companies to reduce dependency on foreign technology and pivot towards domestic alternatives.
From a legal standpoint, Nvidia faces scrutiny amid these geopolitical pressures. Despite the U.S. allowing Nvidia to resume sales of its H20 AI chip to China, Chinese authorities have simultaneously pressured domestic tech giants to limit purchases of these products, citing security concerns and the need to bolster self-reliance. Nvidia has asserted compliance with all applicable laws and maintains cooperation with relevant government agencies as export controls evolve. However, regulatory agencies like China’s State Administration for Market Regulation (SAMR) have taken measures that appear to counterbalance recent U.S. trade blacklists targeting Chinese companies, adding layers of uncertainty for Nvidia’s operations in China.
Economically, the Chinese market represents a significant revenue source for Nvidia and similar companies, accounting for approximately 13% of Nvidia’s total sales as of the previous year. Yet, ongoing restrictions and directives—such as orders from the Cyberspace Administration of China (CAC) for companies like ByteDance and Alibaba to halt orders and testing of Nvidia’s RTX Pro 6000D AI inferencing chips—threaten this market share and the companies’ profitability. These moves reflect Beijing’s strategic effort to develop indigenous semiconductor capabilities, reduce costs, and mitigate reliance on Western suppliers. Furthermore, Chinese firms have reportedly found ways to circumvent some U.S. sanctions, as ByteDance has accessed Nvidia’s high-end processors via rental agreements with U.S. entities, highlighting legal loopholes amid tightening controls.
Policy considerations underline a broader shift in China’s technology strategy. Historically protective of its economic and technological alliances, recent political imperatives—such as regulatory crackdowns on Ant Financial Group and Alibaba’s perceived challenges to government policies—have made aggressive regulation politically necessary. Beijing’s push for domestic chip production is driven by concerns over security vulnerabilities identified in U.S.-made processors and a desire to cultivate technological independence in the face of escalating U.S. export controls. This has resulted in mandates for state-owned enterprises to source a majority of their chips domestically and intensified pressure on
Future Outlook and Developments
The evolving geopolitical tensions between China and the United States are expected to significantly shape the future landscape of semiconductor technology and AI development in China. In response to U.S. export restrictions and regulatory pressures on Nvidia products, Chinese authorities have urged domestic companies such as ByteDance and Alibaba to halt testing and orders of Nvidia GPUs, which are critical for AI and machine learning applications. This directive is part of a broader strategic move to reduce reliance on American technology and promote self-sufficiency in advanced semiconductor development.
Chinese technology firms are increasingly focusing on developing homegrown chips to circumvent these external limitations. For instance, Alibaba has been investing heavily in AI infrastructure, with plans to allocate over $53 billion yuan in the coming years to enhance its chipmaking capabilities, including the development of AI-specific inference chips like the Hanguang 800 released in 2019. Other emerging AI companies in Shanghai and Beijing similarly exemplify China’s ambitions to maintain competitiveness in the global AI race despite external constraints.
Meanwhile, Nvidia has attempted to navigate these restrictions by producing cut-down versions of its GPUs that comply with U.S. export controls. However, these models face challenges within China itself, as they may violate the country’s green energy regulations due to their high power consumption, potentially limiting Nvidia’s access to the substantial Chinese market. Nonetheless, some Chinese firms continue to access Nvidia GPUs indirectly through cloud data centers located outside China, with plans to increase such usage substantially in the near future, potentially investing up to $7 billion to secure GPU access abroad.
The current climate reflects a strategic recalibration, with Chinese industrial regions encouraging a shift away from U.S.-made chips over security and trust concerns, thus accelerating domestic semiconductor innovation and adoption. This shift underscores a growing recognition of semiconductors as critical technology with dual-use applications in both civilian and military sectors, fueling intensified scrutiny and competition worldwide. As a result, China is likely to continue investing heavily in developing advanced AI hardware and infrastructure, aiming to rival leading U.S. firms such as Nvidia and their executives like Jensen Huang.
