NVIDIA CEO Jensen Huang's visits to China and Taiwan reflect the importance of US-Asia linkages amidst a critical period in the company’s semiconductor and AI development programs.
Jensen Huang, CEO of California-based tech giant NVIDIA, paid visits to the company’s offices in China ahead of Chinese New Year and then met with his TSMC (Taiwan Semiconductor Manufacturing Company) counterpart C. C. Wei in Taipei. These trips come at a time when NVIDIA is facing both pressure from the US government to reduce chip exports to China as well as mounting challenges to their AI (artificial intelligence) development programs from rival companies in the United States.
Uncertainty in NVIDIA's Chinese chip market:
Huang's travel to China coincides with the US government's continued efforts to limit foreign tech investment, which has put NVIDIA's relationships with its Chinese clients at a crossroads. In October 2022, the US Department of Commerce's Bureau of Industry and Security instituted export controls on American semiconductor chip exports to China, prompting NVIDIA to modify two chips intended for use in China’s AI industry: the company created alternative versions of the banned-for-China A100 and H100 chips – the A800 and H800, respectively – in compliance with the restrictions.
More recently, in October 2023, the Bureau further expanded the scope of the regulations, which now includes the H800 among the list of newly banned exports. In November 2023, NVIDIA responded by unveiling three new compliant chips – the H20, L20, and L2 – but later announced production delays until the second quarter of 2024. Although NVIDIA has captured more than 90% of China's $7 billion AI chip market, these sanctions and resulting delays have enabled rival Chinese tech companies like Huawei Technologies to challenge NVIDIA's market share and a significant source of its revenue, a fifth of which comes from its sales to China.
Huang pays a visit to NVIDIA's Chinese offices:
Given this uncertainty with NVIDIA’s future in China, Huang embarked on a “goodwill trip” to key cities from January 15th to 19th, 2024. According to Chinese news outlet Yicai Global, the main purpose of the visit was to “happily spend a good time” with some of NVIDIA’s nearly 3,000 Chinese employees and express gratitude for their hard work. A spokesperson for the company claimed that the primary reason for Huang’s trip was to “resume his annual tradition of celebrating Chinese New Year with [NVIDIA’s] local employees.”
This was reflected in media coverage of his trip, which mainly focused on New Year’s celebrations such as drawing the winning number from a lottery box and performing a traditional northeastern Chinese 秧歌 (yānggē) dance on stage alongside his employees while adorned in a red-and-green floral vest. While NVIDIA has not released an official itinerary, leaving questions about Huang’s meetings with executive and government officials up in the air, the CEO reportedly attended meetings and inspected marketing, sales, and R&D teams across offices in Beijing (January 15th), Shanghai (January 17th), and Shenzhen (January 19th), ostensibly to keep in good graces with NVIDIA's Chinese clientele.
Huang meets with TSMC in Taipei:
Huang, who was born in Taiwan, traveled to Taipei on January 20th to meet with TSMC CEO C. C. Wei as well as TSMC founder and industry leader Morris Chang. At a January 25th press event held after the trip, Huang discussed the challenges posed to NVIDIA’s AI ambitions by the restrictions on chip supplies: “The single greatest challenge in AI, of course, is scaling the capacity of AI.” He went on to explain TSMC’s importance to NVIDIA’s business: “We’re working very hard, TSMC, all of our supply chain partners here, are working very hard to keep up with the demand.” NVIDIA relies heavily on TSMC to manufacture the chips that power most AI generative technology worldwide: in the past, Huang has openly spoken about how “perfectly safe” he feels about his company’s reliance on TSMC’s Taiwanese chip fabrication plants (fabs).
Huang has been to Taiwan three times in the past year, participating in Computex Taipei 2023, Hon Hai Tech Day (hosted by Taiwan-based Foxconn, the largest contract manufacturer of electronics in the world), and a ceremony where Chang was honored with the first annual Li Guoding Award, named after the so-called “Godfather of Taiwanese Technology.”
Implications for NVIDIA:
Huang’s meetings with both NVIDIA’s Chinese divisions and its Taiwanese chip manufacturer underscore the significance of US-Asia linkages in the global semiconductor and AI industries. With the Biden administration recently announcing billions of dollars in subsidies for companies like Intel, Micron, and TSMC to construct fabs in America through the $52.7 billion CHIPS (Creating Helpful Incentives for Producing Semiconductors in America) Act, many other key industry players will soon have access to in-house chip production. This presents challenges for NVIDIA, a fab-less company, whose close partnership with TSMC is the linchpin for its preeminent leadership in AI development. Having partnered with Reliance, Infosys, and Tata in October 2023 to jumpstart an AI platform for India, NVIDIA's alliances across the Indo-Pacific are shaping up to be its greatest asset in the ever-evolving cross-national competition for technological superiority.
Vincent Zhang is a Spring 2024 Young Professional at the East-West Center in Washington. He is a senior at the Edmund A. Walsh School of Foreign Service at Georgetown University, pursuing a B.S.F.S. in International Politics with a concentration in Foreign Policy.