Cambricon Technologies is poised to play a significant role in China’s ambitious plans to establish itself as a leader in artificial intelligence (AI) . While it may not be as widely recognized as tech giants like Huawei or Moore Threads , Cambricon is among the leading firms specializing in the design of GPUs for AI applications , boasting substantial growth potential. Recently, the company secured the approval to raise $560 million on the Shanghai Stock Exchange, earmarked for the development of four new chips aimed at training and inference of AI models. Additionally, they plan to work on a robust alternative to NVIDIA’s CUDA programming platform.
Cambricon is riding a wave of success; in the past year, its stock value has tripled. The strategic emphasis on AI within China, bolstered by its ongoing technological and commercial rivalry with the United States , has catalyzed significant investments in hardware firms focusing on AI and large language models. As a result, Cambricon Technologies and other Chinese firms specializing in integrated circuit design are positioned for short-term and medium-term growth.
As previously discussed, the Chinese government has mandated that state-owned data centers must utilize at least 50% Chinese integrated circuits in their servers . This tactic represents one of the most effective strategies to promote innovation while reducing China’s dependence on foreign technologies. It’s likely that this initiative will be complemented by further measures encouraging the use of domestic chips across both public and private data centers throughout the country.
Huawei and Moore Threads: A Triumvirate for China’s AI Ambitions
Moore Threads, a rising player in the field, was added to the US government trade blacklist in October 2023. This designation prevents companies aligned with US interests from selling software or advanced hardware to Moore Threads. Despite its relatively young age—founded in 2020—its founder, Zhang Jianzhong , previously headed NVIDIA’s Chinese subsidiary, giving the company insights that are invaluable in a competitive landscape.
Moore Threads was blacklisted by the US in October 2023.
Moore Threads has developed several GPUs tailored for AI applications that aim to compete directly with products from NVIDIA, AMD, and Huawei. Their most notable offerings currently are the MTT S4000 and MTT S3000 cards . Interestingly, the MTT S80 card is marketed toward gaming and content creation, boasting a performance capability of 14.4 TFlops in single-precision floating point operations.
Although Moore Threads currently lacks the market presence to compete with NVIDIA on a global scale, it has attracted significant investment—around $800 million—over the last four years from various stakeholders, including Chinese tech powerhouses ByteDance and Tencent . This financial backing positions Moore Threads for a promising future in the fiercely competitive AI hardware sector.
Another vital player in China’s chip industry for AI is Huawei . The company’s flagship offering, the Ascend 910D chip , is designed to surpass the performance of NVIDIA’s H100 GPU . Additionally, Huawei has recently unveiled the Ascend 920 chip , aimed squarely at filling the void left by NVIDIA’s future plans for the H200 GPU in the Chinese market. Set for large-scale production in the latter half of 2025, the Ascend 920 will employ cutting-edge 6nm integration technology , reportedly developed in collaboration with SMIC (Semiconductor Manufacturing International Corporation).
For more information, refer to sources like SCMP.
As the landscape in AI technologies continues to evolve, these firms exemplify China’s commitment to self-sufficiency and dominance in the global tech arena. Cambricon’s innovative endeavors, alongside the strategic moves by Huawei and Moore Threads, showcase a determined effort to create an ecosystem that can rival international giants in AI chip development. The coming decade will likely reveal how effective these companies are in their quest to reshape the future of artificial intelligence.