A-shares Subscription | One of the "Four Little Dragons of Domestic GPUs," Suiruan Technology (688801.SH), has opened for subscription today.

On September 2, Suiruan Technology (688801.SH) opened for subscription, with an issuing price of 142.18 yuan per share and a subscription limit of 6,500 shares. It is listed on the Shanghai Stock Exchange, with CITIC Securities as its sponsor. Suiruan Technology, along with Moor Threads, Muxi Co., and BIREN TECH, is collectively referred to as the "Four Little Dragons of Domestic GPUs." Previously, Moor Threads, Muxi Co., and BIREN TECH saw closing gains of 425.46%, 692.95%, and 75.82%, respectively, on their first day of trading. According to the prospectus, the company is one of the leading enterprises in China's cloud AI chip sector, committed to becoming a "leading company in general artificial intelligence infrastructure." The company adheres to an innovation-driven, self-research technology path, building a core competitive advantage and moat for sustainable long-term development. Over the past eight years, the company has independently developed four generations of architecture comprising five cloud AI chips, establishing a complete product system that covers AI chips, AI accelerator cards and modules, intelligent computing systems and clusters, as well as AI computing and programming software platforms. In core technology areas, after years of accumulation, the company has formed a three-category core technology system encompassing chips and hardware, software and programming platforms, and computing cluster solutions. Regarding underlying hardware, the company has independently developed an instruction set aligned with Nvidia's TensorCore accelerated computing units and NVLink interconnection technology, creating original architectures such as the GCUCARE accelerated computing unit and GCULARE high-speed interconnect technology between chips. This architecture not only offers programming flexibility but also provides deep support for high-parallelism accelerated computing in large AI models. On the software platform side, the company did not follow the CUDA ecosystem led by Nvidia and instead developed a full-stack AI computing and programming software platform called "Yusuan TopsRider," which includes drivers, programming languages and compilers, operator libraries, and toolchains to link the companys hardware with AI applications. This significantly reduces the programming development difficulty and migration costs of mainstream AI models based on the companys hardware, allowing its hardware products to better unleash performance in application scenarios. In terms of computing clusters, during the reporting period, the company has already generated income from its thousand-k and ten-thousand-k intelligent computing center projects. Currently, the company is jointly developing an ultra-node solution with customers and collaboratively building high-speed interconnected clusters with commercial value. In terms of ecosystem construction, benefitting from years of deep collaboration on hardware and software customization with large internet companies, multiple generations of the companys products have been widely commercially used in various internet AI scenarios, continuously providing AI computing power support for national-level internet applications transitioning from traditional AI models to large AI models. The companys continuous iterative development ability and product competitiveness have been validated and recognized by the market, gradually achieving a critical leap from a "technology product closed loop" to a "commercial value closed loop." Furthermore, the company is fully leveraging its advantages to build ecosystems with a broader range of downstream partners. In addition to participating in the national "Eastern Data and Western Computing" hub node's intelligent computing center project, the company is actively deepening cooperation with domestic telecom operators and exploring multiple business opportunities in various vertical industries, providing inclusive computing power support for AI empowerment across a wide range of sectors. Additionally, to achieve sustainable development, Suiruan Technology places high importance on long-term strategic cooperation with partners across the entire industry chain, including EDA/IP, wafer manufacturing, packaging and testing, and system components, ensuring stable product development and supply delivery. The domestic cloud AI chip industry is in its early development stage, with international player Nvidia holding a significant market share in China. However, local companies continue to break through technological barriers, capturing a certain market share. Domestic cloud AI chip manufacturers include DSA architecture companies represented by Huawei HiSilicon, Cambricon, and the company itself, as well as GPGPU architecture firms represented by Moor Threads, Muxi Co., ILUVATAR COREX, and BIREN TECH. According to IDC data and the companys sales figures, the overall shipment scale of AI accelerator cards in China is expected to reach approximately 4 million units by 2025, with Nvidia accounting for about 2.2 million units and approximately 55% of the market share. The companys AI accelerator card and module sales are projected to reach 66,000 units that year, corresponding to a market share of about 1.7% in the Chinese AI accelerator card market, ranking among the top domestic AI chip manufacturers. In the future, Nvidia's sales in the domestic market are subject to certain uncertainties, making it likely that local AI chip manufacturers will continue to increase their market share in China. In terms of finance, for the years 2023, 2024, and 2025, the company is expected to achieve operating revenues of approximately 301 million yuan, 722 million yuan, and 990 million yuan respectively; during the same period, net losses are estimated to be approximately 1.665 billion yuan, 1.510 billion yuan, and 1.164 billion yuan respectively.