In 2025, China’s technology landscape is changing rapidly. Once heavily reliant on American chipmaker Nvidia for powering artificial intelligence (AI) applications, Chinese tech giants are now turning to homegrown solutions. This shift is not just about technology—it is a response to new export restrictions, growing national ambitions, and the urgent need for self-reliance.
The Impact of US Export Restrictions on Nvidia’s Market in China
The most significant driver of this change is the tightening of US export controls on advanced semiconductors. In April 2025, new US regulations required licenses for exporting high-end chips like Nvidia’s H20 and AMD’s MI308 to China. These rules followed earlier bans on Nvidia’s H100 chip, effectively cutting off Chinese companies from the latest American AI hardware. As a result, Nvidia’s share of China’s AI chip market has plummeted from 95% before 2022 to just 50% in 2025. This has forced Chinese cloud service providers and tech giants to seek local alternatives.
China’s Ambitious Push for Semiconductor Self-Sufficiency
China’s government has responded by doubling down on semiconductor self-sufficiency. President Xi Jinping has made technological independence a national priority, with the goal of reaching 50% self-sufficiency in semiconductors by 2025. The government has invested over $47 billion through its National IC Industry Investment Fund to boost domestic chip development. This strategy is already reshaping the industry: the share of imported AI chips in China’s AI server market is expected to fall from 63% in 2024 to just 41.5% in 2025.
Huawei and Cambricon: Leading the Charge with Local AI Chips
Two companies stand out in this new landscape: Huawei and Cambricon Technologies. Huawei is preparing to mass-produce its Ascend 910C AI chip, which offers double the computing power and memory of its previous generation. The 910C is already being shipped to Chinese customers and is seen as a primary choice for AI model training and deployment, especially after US restrictions on Nvidia’s products. Huawei is also developing the next-generation Ascend 910D chip, aiming to surpass Nvidia’s H100 in performance.
Cambricon, another key player, is investing heavily in R&D and raising nearly $700 million to develop chips for large language models and AI applications. The company has recently turned profitable and is working to diversify its client base to compete with Huawei. Morgan Stanley predicts that Chinese GPU makers, led by Huawei and Cambricon, could capture 70% of the domestic market by 2027, up from 30% in 2024.
Alibaba, Baidu, and the Rise of Custom AI Chips
China’s tech giants are not just buying local chips—they are designing their own. Alibaba and Baidu are actively developing application-specific integrated circuits (ASICs) to replace Nvidia GPUs in their data centers. These efforts are part of a broader trend: Chinese companies want more control over their supply chains and the ability to tailor chips to their unique AI workloads. This move towards custom chips is expected to accelerate as export controls tighten and local expertise grows.
Nvidia’s Response: Cheaper, Less Advanced Chips for China
Nvidia is not giving up on the Chinese market, which still represents 13% of its global sales. The company plans to launch a new, cheaper AI chip based on its Blackwell architecture for China, with mass production starting as early as June 2025. However, this chip will have less advanced features and lower performance than the restricted H20 and H100 models, making it less attractive for cutting-edge AI applications. Nvidia’s CEO has acknowledged that the company’s options in China are now “limited” due to US regulations.
Challenges and Opportunities for China’s Chip Industry
Despite rapid progress, China still faces challenges in semiconductor manufacturing. While local companies have achieved high self-sufficiency in some production processes, they remain dependent on foreign technology for advanced steps like lithography and ion implantation. Yield rates for advanced chips, such as those produced by Semiconductor Manufacturing International Corp. (SMIC), are still lower than global leaders. However, strong government support, rising demand for AI, and a growing ecosystem of chip designers are driving continuous improvement.
Looking Ahead: The Future of AI Chips in China
The shift from Nvidia to local chips is transforming China’s AI landscape. By 2025, domestic brands are expected to capture a much larger share of the market, reducing reliance on foreign suppliers and fostering innovation within China’s semiconductor ecosystem. As Chinese companies continue to invest in R&D and expand production, the country is well on its way to becoming a global leader in AI hardware.