Jensen Huang Resigns as NVIDIA CEO: A Window of Opportunity for Domestic Compute?
The NVIDIA board of directors announced that founder Jensen Huang has resigned from the positions of President, CEO, and board member, effective immediately; CFO Colette Kress will serve as interim CEO, and the board has begun the search for a new CEO. After the news was announced, NVIDIA's U.S. premarket shares fell by more than 10% at one point, and the decline in early trading that day was about 6.88%. The article also cites IDC data for the first half of 2026, stating that China's data center AI accelerator shipments were about 2.008 million units, with non-GPU accelerators at 1.016 million units, up nearly 80% year over year, surpassing GPUs in shipment volume for the first time; NVIDIA shipped 678,000 units, for a share of 33.8%.
Why read this
SelectedAI score 90
The article places Jensen Huang's resignation alongside changes in the shipment structure of domestic compute, so readers can use it to judge the direction of adjustments to supply chains and procurement strategies.
Topics
Body · AI translation
On November 12, the NVIDIA board of directors announced that company founder Jensen Huang has resigned from his positions as president, chief executive officer, and board member, effective immediately. Current executive vice president and chief financial officer Colette Kress will serve as interim CEO while continuing as CFO until a permanent successor is determined; the board has simultaneously launched a search for a new CEO and will consider both internal and external candidates. The board also appointed Harvey C. Jones as its first board chairman.
Kress joined NVIDIA in September 2013 and previously spent nearly 25 years in finance roles at technology companies including Cisco and Microsoft; Jones has more than 30 years of experience as an entrepreneur, technology executive, and venture capitalist, and previously served as CEO and executive chairman of Synopsys. Jensen Huang founded NVIDIA in 1993 and led it for 33 years.
The news of Jensen Huang's resignation sent shockwaves through capital markets. In premarket trading on November 12, NVIDIA shares at one point fell more than 10%, a sharp pullback from the previous trading day's closing price of $173.73, and recovered slightly after the open; as of 10:15 a.m. Eastern Time, NVIDIA shares stood at $161.77, down 6.88%. The market is reassessing the operational and governance risks facing this AI chip giant after losing its founder.
"The first thing the new CEO may do is a compliance review, examining the supply chain and customers," a veteran industry figure told an IT Times reporter. "What the transitional management also needs to stabilize is customer and supply chain expectations—whether existing orders can be delivered on time and how next year's capacity will be scheduled are more important than telling new stories."
The official announcement said Jensen Huang's contributions were "profoundly meaningful" and that the board and the company expressed "sincere gratitude," but as of publication time, Jensen Huang himself had not spoken publicly. After his departure, what NVIDIA's downstream customers need to recalculate is no longer "whether NVIDIA will be okay," but "for the next capacity expansion, where will the chips come from, and how long will it take for them to arrive."
Certainty declines in Southeast Asia's computing power market
In recent years, leading global technology companies have all been optimistic about Southeast Asia, building campuses and renting racks there.
On November 4, Google announced a $15 billion investment in Malaysia to expand data centers and cloud operations, with new campuses located in Selangor and Negeri Sembilan; the investment period is 2027 to 2028, about 7.5 times its 2024 commitment of $2 billion. Microsoft's data center near Kuala Lumpur is under construction, with its Malaysia investment totaling $2.2 billion, and in May 2025 it opened its first cloud region in Indonesia; Amazon Web Services launched its Asia Pacific (Thailand) region in early 2025.
Some Chinese companies also rely on their own resources and connections to procure servers and accelerator cards through overseas channels. According to Caixin, of the approximately 1.5 GW of approved data center projects in Thailand, projects with Chinese backgrounds or mainly serving major Chinese corporate clients account for 60% to 70%.
Infrastructure is expanding, but the "predictability" of supply is declining.
When an IT Times reporter inquired about pricing from multiple computing power leasing channels as a buyer, some channel representatives said that high-end computing power that can currently be stably rented out is "increasingly difficult to lock down," and some channels "cannot guarantee" whether contracts can be renewed next year.
A person from a server manufacturer told the reporter that Southeast Asia, Europe, and other regions remain options for some customers seeking overseas computing power at this stage, but uncertainty on the supply side is increasing. In addition to rising prices, customers must also consider risks such as contract renewals, continued resource supply, and future changes in regulatory policy.
On the other hand, NVIDIA's high-end computing power remains in high demand in the leasing market. Some complex training and high-performance computing scenarios still prioritize NVIDIA GPUs, and they are difficult to fully replace in the short term. But price has become the most direct variable, and channel representatives have found that since the beginning of this year, the number of customers inquiring about Chinese computing power solutions has increased noticeably.
Supply-side variables push domestic chips into the spotlight
Uncertainty comes from the extension of regulatory expectations.
On January 12, 2026, the U.S. House of Representatives passed the Remote Access Security Act by a vote of 369 to 22. The bill seeks to fill the regulatory gap in current export controls regarding remote access to advanced computing resources, with the core aim of extending the jurisdiction of export controls from "items leaving the country" to "remote access."
As of November, the bill had not yet entered the next legislative stage and had not taken effect, but for companies obtaining high-end computing power through overseas data centers, whether the regulatory boundary will further extend to the cloud has become a new risk variable.
Supply-side variables have pushed domestic chips into the spotlight.
IDC data for the first half of 2026 show that China's data center AI accelerator shipments totaled about 2.008 million units, including 991,000 GPUs, down 26% year on year, and 1.016 million non-GPU accelerators, including inference chips, up nearly 80% year on year, surpassing GPUs in shipments for the first time. NVIDIA shipped 678,000 units for a 33.8% share, down 44% year on year; Huawei shipped 572,000 units for a 28.5% share; Alibaba's T-Head shipped 133,000 units, about 7%; Cambricon shipped 92,000 units, about 5%; Kunlunxin about 4%; Hygon about 3%; Moore Threads shipped 56,000 units, about 3%, ranking first among domestic GPU manufacturers.
Several industry insiders told the IT Times reporter that the tiers among domestic chips are now fairly clear. The first tier consists of Huawei Ascend and T-Head: the former leads in scale and has entered stable deployment in government and enterprise, telecom operators, and cloud systems; the latter's production capacity first serves Alibaba's own cloud and AI systems, with limited incremental volume for external customers. An industry insider told the reporter that some older T-Head models flowed into the market last year, but this year its cards are basically unavailable on the market, "it's very hard to get stock"; however, based on actual usage feedback, its product quality enjoys a solid reputation in the industry.
On November 10, Alibaba announced that it had submitted a spin-off plan to list T-Head separately on the Main Board of the Hong Kong Stock Exchange, and that T-Head had submitted an A1 form confidentially on September 11. If the spin-off listing ultimately goes through, T-Head is expected to gain more independent financing channels and a higher capital market valuation. But whether T-Head's Zhenwu chips can shift from mainly serving Alibaba internally to larger external supply still depends on production capacity, customer structure, and business strategy.
MetaX, Iluvatar CoreX, Moore Threads, Enflame, and Biren Technology are all in the second tier, with MetaX and Iluvatar CoreX at the top of that tier, relatively ahead in overall scale, ecosystem maturity, and system-level delivery capability.
MetaX's Xiyun C600 is equipped with 144GB of HBM3e memory and will enter mass production in May 2026, achieving a closed domestic supply chain from design and manufacturing to packaging and software stack, and in the first half of the year it was among the first to pass the national Security and Reliability Evaluation. Iluvatar CoreX's Tiangai 300, in third-party testing completed in July this year, achieved end-to-end throughput per card equivalent to 1.05 times that of mainstream international high-end GPUs under a typical PD-disaggregated deployment; according to its disclosure, it is the first domestic general-purpose GPU to reach this level in that dimension.
Kunlunxin holds about a 4% share by shipment volume, still in the second tier, and in January this year it confidentially submitted a listing application to the Hong Kong Stock Exchange, in May launched listing guidance for the STAR Market, advancing on a dual A+H track; Baidu responded that market rumors of a "delay in Kunlunxin's listing timetable" are untrue.
Moore Threads' semi-annual report released on August 9 shows first-half revenue of 1.736 billion yuan, up 147.42% year on year, already exceeding full-year 2025 revenue; net loss attributable to shareholders was 11.5631 million yuan, a narrower loss year on year. Its shipment volume ranks first among domestic GPU manufacturers, while ecosystem maturity and system-level delivery experience are still being accumulated.
However, NVIDIA's business fundamentals have not weakened. Its fiscal 2027 second quarter (ended July 26, 2026) revenue was $96.22 billion, up 106% year on year, with data center business revenue of $89 billion, up 117% year on year. An official announcement said third-quarter results will still be released as originally scheduled on November 17.
Predictability of supply is declining. The question enterprises need to answer now is no longer "are there cards available" but "who holds the cards, and can the next capacity expansion be delivered on time." How wide the window for domestic computing power can open depends on whether the first tier releases production capacity and whether the second tier can turn tested figures into sustained delivery.
This article is from the WeChat public account "IT Times" (ID: vittimes), author: Mao Yu, editors: Hao Junhui, Sun Yan, published by 36Kr with authorization.