Hello! I’m Yuan Hehe. Please follow me for more exciting content. Nvidia China AI chip market share

Nvidia once held the vast majority of the Chinese AI chip market share, but this advantage is rapidly disappearing after export restrictions were implemented. In my view, this change is not just about a single company’s declining sales, but rather a complete reshaping of the supply landscape of China’s AI computing power market.

Since 2022, the United States has been tightening restrictions on the export of advanced AI chips to China, significantly limiting the products Nvidia can sell and the projects it can participate in in China. Meanwhile, domestic companies are accelerating their shift towards domestically developed solutions, and Huawei has used this opportunity to expand the market influence of its Ascend chips.

Bernstein predicts that Nvidia’s share of the Chinese AI chip market may fall to about 8% by 2026, while Huawei is expected to reach about 50%. Jensen Huang has also previously stated that under existing export restrictions, Nvidia is practically finding it difficult to compete in the Chinese data center computing market.

In less than five years, Nvidia went from holding an absolute advantage to becoming a marginal player in the Chinese market. More noteworthy is that the demand it left behind hasn’t disappeared; instead, it’s being rapidly absorbed by Huawei and other domestic chip manufacturers.

How Nvidia lost control of the Chinese market

Jensen Huang remains very popular in China. His recent visit to Beijing, where he was seen eating Zhajiangmian (noodles with soybean paste), attracted considerable attention, demonstrating that Nvidia’s influence among Chinese developers and tech companies hasn’t completely disappeared. However, the problem is that brand appeal is no longer easily translated into orders.

What truly changed the situation was the increasingly stringent export restrictions on AI chips imposed by the United States. Nvidia could no longer provide its most advanced products to Chinese customers as it had in the past, and even though some restrictions were later adjusted, domestic companies have begun to accelerate their shift towards domestically produced chip solutions.

In my view, this change has given Huawei the most direct opportunity. Nvidia originally held an absolute advantage in the Chinese AI chip market, but with limited product supply, customers had to look for alternatives, and Huawei took the opportunity to expand the influence of its Ascend ecosystem.

Huawei’s rise was not achieved overnight.

When Huawei was placed on the U.S. trade blacklist in 2019, it was widely believed that this would weaken its competitiveness in the chip and communications fields in the long term. But in retrospect, the restrictions did create enormous pressure and forced Huawei and the entire Chinese semiconductor industry to find alternatives more quickly.

Over the past few years, domestic chip companies have continuously increased their R&D investment, and Huawei has gradually become one of the most watched. The continuous advancement of the Ascend processor, supporting software ecosystem, and large-scale computing clusters shows that it is not just making single chips, but trying to complete the entire capability from hardware to system.

Industry analysts believe that Huawei’s new-generation Ascend processors are beginning to perform close to Nvidia’s high-end products in some commercial workloads. Huawei’s showcase of its domestically produced chip cluster also demonstrates that China is attempting to bridge the gap in single-chip performance and advanced manufacturing processes through greater chip collaboration.

In my view, Huawei cannot yet claim to have fully caught up with Nvidia, but it has moved from passively seeking alternatives to being able to provide complete solutions. Huawei executive He Tingbo also remains cautious: “Who will move faster is still unanswered; ultimately, it depends on technological iteration, the software ecosystem, and the ability to implement solutions in practice.”

Nvidia still maintains a significant advantage.

Huawei has indeed made rapid progress, but for now, Nvidia remains the unavoidable benchmark in the high-end AI computing field. Its advantages extend beyond the chips themselves, encompassing a mature software ecosystem, development tools, and a global supply chain—capabilities that cannot be fully replicated in a short period.

Nvidia is responsible for chip design, while advanced products are mainly manufactured by TSMC, which in turn relies heavily on ASML’s high-end lithography equipment. Domestic companies currently lack access to the most advanced EUV lithography technology, meaning that domestic solutions still face a real gap in cutting-edge processes and single-chip performance.

Therefore, as long as institutions can legally obtain NVIDIA chips, they will generally prioritize CUDA for training cutting-edge models, conducting university research, and deploying large-scale AI systems. The reason is not complicated: the CUDA ecosystem is mature, existing software compatibility is more complete, and migration costs are lower.

More importantly, losing part of the Chinese market has not slowed Nvidia’s global growth. The company’s revenue reached approximately $81.6 billion in the most recent quarter, and is projected to increase to approximately $91 billion in the next quarter; this guidance does not include its data center business in China. Its fiscal year 2026 revenue is close to $216 billion, significantly higher than Huawei’s total revenue of approximately $126 billion during the same period.

In my view, Huawei is narrowing the gap in hardware and system capabilities, but what Nvidia truly cannot replace is its overall advantage in chips, software, and the developer ecosystem. The landscape of the Chinese market has changed, but this does not mean that Nvidia’s leading position in the global AI industry has been shaken.

China’s AI strategy is becoming clearer

What truly deserves attention now is no longer whether Huawei can immediately replace Nvidia, but rather that China is gradually establishing an AI system that is less dependent on overseas technology.

DeepSeek’s latest V4 model has been adapted for Huawei’s Ascend platform, indicating that the collaboration between domestically developed large-scale models and domestically produced chips is moving from simple compatibility to deeper joint optimization. How to design the model and how to maximize chip performance will likely be collaboratively addressed from the R&D stage in the future.

Some analysts believe that DeepSeek and Huawei have already engaged in extensive technical cooperation and may continue to explore using domestically produced hardware to train subsequent models. In my view, this shift is more important than simply comparing the performance of a particular chip, because it relates to whether software, hardware, and development tools can truly form their own ecosystem.

Of course, this process won’t be completed overnight. Nvidia’s advantages in performance and software ecosystem remain significant, and domestic solutions still have many issues to resolve. However, judging from the collaboration between DeepSeek and Huawei, the direction of China’s AI industry in reducing its reliance on overseas chips is becoming increasingly clear.

The AI ​​chip race has entered a new phase.

In the past, discussions about AI chip competition focused primarily on who had stronger computing power and faster speeds. However, the real determinants now include not only chip specifications but also manufacturing capabilities, supply chains, software ecosystems, and industrial policies.

For China, Huawei’s significance has long surpassed that of a chip manufacturer. It is taking on a greater role in domestically developed computing infrastructure and has become a crucial part of China’s efforts to reduce its dependence on foreign technology.

Nvidia remains the global leader in the AI ​​hardware market, a position unlikely to change in the short term. However, in the Chinese market, its near-monopoly has been broken, with Huawei and other domestic manufacturers rapidly filling the gap.

In my view, the competition between Nvidia and Huawei is no longer just about two companies vying for orders, but a long-term contest between two different technological systems, supply chains, and industrial approaches. The ultimate outcome will not only reshape the Chinese AI chip market but also influence the trajectory of the global artificial intelligence industry in the coming years.

If you enjoyed this article, please consider giving a tip! Also, don’t forget to share, like, comment, and follow!