Nvidia CEO Defends Chip Sales to China Amid US Export Tensions and Domestic Industry Gains

Nvidia CEO Jensen Huang defends China chip sales, warning US export curbs could split AI ecosystems and boost Chinese tech independence.

Cover Image Attribute: The file photo of Nvidia Chief Executive Jensen Huang

Nvidia Chief Executive Jensen Huang has warned that restricting advanced chip sales to China could lead to a fragmented global artificial intelligence ecosystem that disadvantages the United States, even as Chinese manufacturers capture a growing share of their domestic market and US authorities pursue cases involving alleged illegal diversions of restricted technology. In recent public appearances, Huang argued that continued engagement through compliant sales helps maintain American technological leadership by encouraging developers worldwide to build on US software and hardware frameworks, rather than accelerating a shift toward non-American alternatives.
Huang made his remarks during an interview on the Dwarkesh Podcast, where he addressed concerns that access to Nvidia chips might bolster China's AI capabilities at America's expense. He emphasized that advances in artificial intelligence have stemmed largely from algorithmic and software innovations, not solely from raw hardware power. Pointing to Chinese AI models such as DeepSeek, Huang cautioned that optimization of such models for domestic hardware like Huawei's architecture could erode the software-hardware interdependence that has underpinned US dominance. "The day that DeepSeek comes out on Huawei first, that is a horrible outcome for our nation," he said. Huang added that a world in which AI models developed globally run best on non-American hardware would represent "bad news for us."
The Nvidia leader pushed back against suggestions that selling chips in China equates to ceding strategic ground, framing the issue instead as one of sustaining a broad ecosystem. He questioned policies that might force the United States to "give up the world," noting that the chip industry forms part of the broader American technology and AI leadership. Huang rejected what he described as a "loser premise" that competition in China would inevitably result in market loss, insisting that Nvidia's market share continues to grow through ongoing innovation. "You’re not talking to somebody who woke up a loser," he stated during the exchange, which observers described as tense. "That loser attitude, that loser premise makes no sense to me."
Huang drew comparisons to established computing ecosystems such as x86 and ARM, which have proven difficult to displace due to the time and resources required for migration. Unlike consumer products such as automobiles or smartphones, where switching brands can occur readily, computing platforms create sticky dependencies, he argued. "It’s our job to continue to nurture that ecosystem, to keep advancing the technology so that we can compete in the marketplace," Huang said. "Conceding a marketplace based on the premise you described, I simply can’t acknowledge that." He expressed confidence that the United States and its industry are not positioned to lose in global competition.
The debate gained additional visibility when Huang responded to comments by Anthropic CEO Dario Amodei, who in a January essay, likened unrestricted chip sales to China to "selling nuclear weapons to North Korea and then bragging that the missile casings are made by Boeing." Huang dismissed the analogy as "lunacy," clarifying that Nvidia's products are commercial chips rather than restricted materials akin to enriched uranium. He maintained that China could eventually produce comparable technology independently and that blocking sales would not prevent its progress but might instead divide the AI landscape into separate open-source and closed-source camps. "It would be extremely foolish to create two ecosystems: the open source ecosystem, and it only runs on a foreign tech stack, and a closed ecosystem that runs on the American tech stack," Huang said. "I think that would be a horrible outcome for the United States."
Amodei has contended that China lags several years behind the United States in producing frontier chips at scale and that the coming period represents a critical window for maintaining technological superiority. Huang, conversely, has estimated that compliant sales to China could contribute as much as $50 billion annually to Nvidia while preserving reliance on American technology stacks. The differing views underscore divisions within the technology sector over balancing national security priorities with commercial interests.
Against this backdrop, data from market research firm IDC indicate that Chinese GPU and AI chip producers claimed nearly 41 percent of China's AI accelerator server market in 2025, up significantly amid successive US export controls that have limited access to Nvidia's most advanced offerings. Nvidia retained the leading position with approximately 55 percent market share, shipping about 2.2 million cards out of roughly 4 million total units from Nvidia, AMD and domestic vendors. AMD held a 4 percent share, while Chinese companies collectively shipped 1.65 million cards. Huawei Technologies accounted for roughly half of the domestic shipments, followed by Alibaba's T-Head unit and others including Baidu's Kunlunxin and Cambricon. The gains reflect Beijing's push for self-reliance, including government directives favoring local procurement for AI infrastructure projects.
Nvidia has navigated these restrictions by developing adjusted products for the Chinese market while seeking regulatory approvals for broader sales. In March, Huang announced that the company had received purchase orders for its H200 processors from customers in China and was restarting manufacturing to fulfill them. "We have received purchase orders, and we're in the process of restarting our manufacturing," he told reporters at the company's GTC conference. The development followed policy adjustments under the Trump administration, which initially imposed licensing requirements leading to a $5.5 billion charge for Nvidia but later permitted shipments of the H200 with conditions, including a 25 percent US government share of sales revenue. As of late February, the company had yet to record revenue from such sales, citing ongoing security reviews.
China previously represented at least one-fifth of Nvidia's data center revenue before tighter controls took effect. Despite the constraints, Nvidia reported 73 percent revenue growth in its most recent quarter and projected 77 percent growth in the current period, assuming no contribution from China in its guidance. Huang has lobbied in Washington and visited China earlier this year as part of efforts to address the situation.
US enforcement actions have highlighted risks of circumvention. In March, federal prosecutors charged three individuals linked to server manufacturer Super Micro Computer with conspiring to divert approximately $2.5 billion worth of servers containing Nvidia AI chips to China in violation of export controls. The defendants allegedly used a Southeast Asia-based pass-through company to place orders, fabricated documents, and employed dummy replica servers to evade audits. One individual used hair dryers to alter labels and serial numbers on boxes and dummy equipment, according to surveillance evidence cited in court documents. Super Micro Computer, which was not named as a defendant, stated that it placed two of the individuals on administrative leave, terminated its relationship with the third, and is cooperating with investigators. The company emphasized that the alleged conduct violated its compliance policies.
A separate case involved two Chinese nationals charged with smuggling millions of dollars in restricted Nvidia chips to China via shipping routes through Malaysia and Singapore. Nvidia has stated that it does not provide support for unlawfully diverted systems and works with customers and government authorities on compliance. US officials have described such diversion schemes as direct threats to national security.
These developments occur as the US government weighs the interplay between export controls, technological competition and economic interests. Huang has advocated for regulations that allow American companies to compete globally without unnecessarily surrendering market share, arguing that all layers of the AI industry—including applications—benefit from sustained leadership across energy, chips, infrastructure, models and software. He questioned approaches that might sacrifice one segment to protect another.
The tensions reflect broader US-China strategic rivalry in semiconductors and artificial intelligence. While export controls aim to safeguard sensitive technologies, critics within industry warn that overly restrictive measures could hasten the decoupling of technology ecosystems, potentially diminishing the influence of US standards and innovation. Chinese progress in domestic alternatives, evidenced by rising market shares for local producers, suggests that restrictions have spurred investment and development in China, even if gaps remain in the most advanced capabilities.
Huang's position centers on the view that keeping global AI developers engaged with American technology stacks, including through open-source contributions, serves long-term US interests. He has rejected zero-sum framings of the competition, stressing the need for continuous innovation to maintain advantages in a field where ecosystems prove resilient but not immutable. As orders for compliant chips begin to move forward and enforcement actions continue, the trajectory of US policy will likely shape not only Nvidia's business outlook but also the structure of the global AI landscape for years ahead.
Questions persist about the effectiveness of current licensing regimes, the pace of Chinese technological catch-up, and the potential for further policy adjustments. Industry leaders, policymakers and analysts continue to debate the optimal balance between security imperatives and the commercial dynamics that have driven rapid advances in artificial intelligence. The coming months may bring additional clarity as shipments proceed under approved terms and market data further illuminate shifts in competitive positioning.

IndraStra Global is now available on
Apple News, Google News, Feedly, Flipboard, and  WhatsApp Channel

COPYRIGHT: This article is published under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. https://creativecommons.org/licenses/by-nc-nd/4.0/

REPUBLISH: Republish our articles online or in print for free if you follow these guidelines. https://www.indrastra.com/p/republish-us.html