港股GPU第一股或要来了，估值160亿

Core Insights - Wallran Technology plans to issue up to 372,458,000 overseas listed ordinary shares and convert 873,272,024 domestic unlisted shares into overseas listed shares for trading on the Hong Kong Stock Exchange [1] - The company is the only one among the "Four Little Dragons of Domestic GPUs" to choose the "full circulation" listing in Hong Kong, having previously filed for an IPO on the Sci-Tech Innovation Board in September 2024 [1] - Wallran Technology's valuation has reached 16 billion RMB according to the 2025 Hurun Unicorn List [2] Company Overview - Founded in 2019 by former SenseTime president Zhang Wen, Wallran Technology focuses on high-end cloud computing GPGPU technology [3] - The company released its first general-purpose GPU chip, BR100, in 2022, achieving over 1000T of 16-bit floating-point computing power and over 2000T of 8-bit fixed-point computing power, setting a global computing power record [3] - Wallran Technology is one of the first domestic GPU companies to commercialize Chiplet packaging technology, allowing for performance comparable to advanced process chips while reducing reliance on a single process [3] Product Development - The BR100 and its derivatives have led to the creation of the Wallli series of commercial hardware products, which include liquid-cooled and air-cooled OAM modules and AI acceleration cards [4] - The company has filed over 1,200 patents globally, ranking first among Chinese general-purpose GPU companies, with over 550 patents granted [4] Funding and Investment - Wallran Technology has raised over 5 billion RMB in D-round financing, with investors including Qiming Venture Partners, IDG Capital, Huaden China, Ping An Group, and Hillhouse Capital [5] Technological Advancements - The company is developing an OCS supernode project at the Shanghai Yidian Smart Computing Center, integrating silicon photonics technology with its high-performance general-purpose GPU liquid-cooled modules [7] - The supernode architecture aims to enhance computing density, communication density, and bandwidth density, addressing the challenges of traditional air-cooled AI servers [7]