Alibaba Makes Major Moves! AI Chips Experience Internal and External Resonance, Institutional Attention and Margin Buyer Accumulation Stocks List Announced

On the morning of March 24, the AI chip concept sector reversed in a V-shape, with domestic GPU giants soaring together. Among them, Muxi Co. surged about 17%, Moore Threads rose over 9%, and Cambrian also increased by 6.08%. Additionally, Youboxin rose nearly 15%, and stocks like Yuntian Lify and Yanshan Technology also showed upward momentum.

In terms of news, Alibaba Damo Academy’s latest chip product was officially unveiled. Overseas, tech giants are also heavily investing in chip development.

Alibaba Launches “World’s Highest-Performance RISC-V CPU”

On March 24, the “2026 Xuantie RISC-V Ecosystem Conference” hosted by Alibaba Damo Academy was held in Shanghai. During the event, Alibaba Damo Academy officially released the new generation flagship CPU, Xuantie C950.

Alibaba Damo Academy Chief Scientist Meng Jianyi stated that, as the “world’s highest-performance RISC-V CPU,” Xuantie C950 uses 5nm process technology, with a frequency of 3.2GHz. Its single-core general performance in the SPECint2006 benchmark exceeded 70 points, setting a new global record for RISC-V CPU performance. It is suitable for cloud computing, generative AI, high-end robotics, edge computing, and other fields, with overall performance more than three times that of the previous generation Xuantie C920.

Furthermore, Xuantie C950 leverages RISC-V’s open-source features and is equipped with self-developed AI acceleration engines, supporting large models like Qwen3 and DeepSeek V3 with hundreds of billions of parameters natively for the first time.

Notably, during Huawei’s China Partner Conference 2026 held a few days ago, Huawei announced the Atlas 350 AI training and inference acceleration card, equipped with the new Ascend 950PR processor. Compared to the previous Ascend chips, Ascend 950PR has significantly improved in low-precision data formats, vector computing power, interconnect bandwidth, and self-developed HBM.

Overseas Giants Bet on AI Chips

In the overseas market, recent developments in AI chips have also accelerated.

On March 22, local time, Elon Musk announced the launch of the Terafab project, jointly developed by Tesla, SpaceX, and xAI.

As the core of its compute power vertical integration strategy, this “largest chip manufacturing plant ever” plans to adopt 2nm process technology, serving electric vehicles, humanoid robots, and space AI infrastructure, with an annual output target of about 100 to 200 billion chips.

Meanwhile, to compete for dominance in the AI chip field, Samsung Electronics recently announced plans to invest over 110 trillion Korean won (about RMB 504.1 billion) in facilities and R&D by 2026, with most funds focused on the chip business’s equipment solutions (DS) division. This investment is a 21.7% increase over Samsung Electronics’ capital expenditure in 2025, setting a new company record.

Additionally, Jensen Huang recently projected revenue reaching one trillion dollars by 2027 at NVIDIA’s GTC conference, further boosting market confidence in the computing industry chain.

China Galaxy Securities pointed out that the demand for computing power and capital investment are resonating upward, entering a high-intensity expansion cycle of “demand explosion—price transmission—capital injection” in the AI industry.

Shanxi Securities stated that major domestic clients continue to increase procurement of domestic AI chips, and leading domestic manufacturers are accelerating in single-card performance, rack-level solutions, ecosystems, and capacity to catch up with NVIDIA. Their new flagship products are expected to quickly capture market share from NVIDIA chips. They remain optimistic about investment opportunities in the AI industry chain and recommend related stocks in the top tier of domestic AI computing chips.

Multiple Stocks Attract Attention from Institutions and Margin Traders

According to Eastmoney Concept Board, there are nearly 40 AI chip concept stocks in the A-share market, with a total market value exceeding RMB 2.3 trillion. Haiguang Information and Cambrian both have market caps over RMB 400 billion, with Moore Threads, Muxi Co., and Lankei Technology ranking third to fifth.

Since the beginning of this year, the performance of AI chip concept stocks has been mixed, with nearly half recording gains. Guoke Micro led the gains with a 59% increase, Hengshuo Co. rose nearly 40%, and Yanshan Technology and Hesheng New Materials both increased over 20% in the period.

Data from Eastmoney Choice shows that over the past month, seven AI chip concept stocks received institutional research, and some also saw increased leverage positions. Lankei Technology was heavily bought with RMB 466 million in financing, Jingchen Co. had a net financing purchase of RMB 156 million, and Guoxin Technology and Xingchen Technology had net financing inflows of RMB 70 million and RMB 46 million, respectively.

(Article source: Eastmoney Research Center)