In 2025, the Chinese technology giant was already testing the Ascend 910D, a high-end AI chip designed to compete with Nvidia’s processors. Huawei planned to send samples to Chinese technology companies for testing, although the chip was still in early development and faced significant testing and production challenges.
Huawei’s latest push to challenge Nvidia is not its first, and that effort has now moved into another phase. Huawei’s rotating chairman, David Wang, said the company plans to launch the 960DT in the first quarter of 2027 and the Ascend 960PR in the third quarter as it expands its AI chip lineup. Huawei is also developing UnifiedBus, a technology that allows large numbers of AI processors to work together, and says it has already built systems capable of connecting up to 1 million AI processors.
🔺Huawei plans to launch two new AI chips in 2027 as it looks to challenge Nvidia in the AI chip market. https://t.co/RrNlKwguXF
— Indian Trends Hub (@IndianTrendsHub) September 17, 2026
Is Huawei the only Chinese company building Nvidia alternatives?
Huawei is already the biggest name in China’s domestic AI-chip push, but it is not alone. Cambricon, Biren, Moore Threads and MetaX are also developing processors aimed at AI training and inference.
Biren is one example of how quickly the domestic market is developing. The company reported $183.9 million in revenue for the first half of 2026, up almost 2,000% from a year earlier. But it still recorded a $56.2 million loss, showing the gap between finding demand for AI chips and building a profitable chip business. Biren is also developing its own software stack to compete with Nvidia’s CUDA platform.
Cambricon is another important player, while Moore Threads is pursuing a bigger GPU strategy that covers both graphics and AI. This means Huawei’s 2027 chips will enter a Chinese market where several companies are already trying to capture demand created by restrictions on Nvidia and other foreign suppliers.
You may also like: California Authorities Crack Down on Crypto Scams
China has tried to build an Nvidia rival before
Huawei’s earlier attempt was to build its own AI processors that Chinese companies could use instead of Nvidia’s chips. The company developed the Ascend 910C as an alternative to Nvidia’s H100, with the chips designed to handle AI training and inference. Huawei also began testing the processor with Chinese technology companies to see how it performed in real-world AI workloads.
Huawei then tried to make up for the performance gap by connecting many Ascend chips together. Its CloudMatrix 384 system, for example, links 384 Ascend 910C processors so they can work as one large AI computing system. Huawei has also developed its own networking and software tools to help those chips communicate and run AI models at scale.
The results have been mixed. DeepSeek’s research found that the Ascend 910C delivered around 60% of Nvidia’s H100 inference performance. Nvidia has since moved beyond the H100 to newer generations such as Blackwell, while Huawei still faces limits around advanced manufacturing, memory, packaging and software.
China has a reason to keep buying Huawei chips
US restrictions have turned AI chips into a strategic issue for Chinese technology companies. Instead of competing only on price or performance, Huawei is increasingly benefiting from China’s push to build a domestic alternative to Nvidia.
There are signs that this is already changing the Chinese market. A Bernstein estimate by the Washington Post put Nvidia and Huawei at roughly 40% each of China’s AI chip market in 2025, but projected Huawei’s share could reach about 50% in 2026, while Nvidia’s could fall to around 8%. Those are estimates rather than reported market-share figures, but they show how quickly the competitive balance could change inside China.
The gap with Nvidia, however, remains significant. Epoch AI estimates Huawei will produce less than 4% of Nvidia’s AI compute output in 2026, while differences in chip performance and access to high-bandwidth memory remain major constraints.
If it can provide Chinese companies with a workable domestic stack of chips, networking and software, U.S. restrictions could end up creating a large protected market for Huawei at home. The harder question is whether that technology can eventually compete with Nvidia outside China, where customers have more freedom to choose.
What are investors and users saying about Huawei’s AI ambitions?
Reaction to Huawei’s latest AI chip plans shows that investors remain divided over how far China can close the gap with Nvidia. Some users questioned whether Huawei’s progress is being overstated, with Night suggesting that Western media often swings between underestimating and overestimating China’s chip capabilities, creating volatility in Nvidia’s share price. He argued that the bigger issue is not the chip itself, but the supply chain behind it.
Derrick pointed to materials such as gallium, germanium and rare earths, where China remains a dominant global supplier. Those materials are essential for semiconductors, power systems and advanced electronics, meaning the AI competition extends beyond processors to control of the resources needed to manufacture them.
@business Export controls on Huawei are the supply-chain version of sanctions. The real question is what materials the alternative manufacturer requires — gallium, germanium, and rare earths are the hidden inputs that China still controls 80%+ of globally.
— Derrick Dao (@derrick_dao) September 17, 2026
China imposed export restrictions on gallium and germanium in 2023, highlighting how critical minerals have become part of the broader technology rivalry.
Kyle focused on another concern of scale. Every market Nvidia loses in China could give Huawei more customers, more software developers and more real-world experience deploying AI systems. That matters because Nvidia’s advantage is not only its chips, but also its CUDA software ecosystem and developer base.
Enjoyed this? Bookmark DeFi Planet, explore related topics, and follow us on Twitter, LinkedIn, Facebook, Instagram, Threads and CoinMarketCap Community for seamless access to high-quality industry insights
Take control of your crypto portfolio with DEFI PLANET PRO, DeFi Planet’s suite of analytics.
