未来智造局|抓住国产GPU从“跟跑”到“并跑”的窗口期——专访壁仞科技战略委员会主席李新荣
Xin Hua Cai Jing·2025-12-15 14:58

Core Viewpoint - The rise of generative artificial intelligence has led to a surge in global computing power demand, positioning computing power as a key indicator of national technological competitiveness. Domestic GPU manufacturers, such as Shanghai Birun Technology Co., Ltd., are accelerating their breakthroughs in high-end general-purpose intelligent computing chips, with their GPU products now in mass production [1][2]. Group 1: Company Developments - Shanghai Birun Technology plans to issue up to 373 million shares for overseas listing on the Hong Kong Stock Exchange, marking a significant step towards becoming the first GPU stock in Hong Kong [1]. - The company has successfully launched multiple GPU products that are widely used in the artificial intelligence industry, including large model training and various inference application scenarios [3][6]. - Birun Technology is collaborating with major enterprises like China Mobile and China Telecom to deploy its GPU products in AI data centers and other critical industries [6][7]. Group 2: Market Dynamics - The domestic smart computing chip market is projected to exceed 350 billion yuan, with NVIDIA's H20 product orders reaching 16 billion USD (approximately 112.9 billion yuan) in the first quarter of this year [2]. - The global GPU market is currently dominated by foreign tech giants, but geopolitical uncertainties are driving countries to accelerate their own technology development, creating opportunities for domestic GPU growth [2][3]. Group 3: Strategic Focus - Birun Technology's strategy emphasizes three core areas for the next five years: architectural innovation to overcome performance bottlenecks, ecosystem collaboration to reduce user migration costs, and scenario adaptation to address practical application pain points [3][4]. - The company is focusing on supernode technology to enhance cluster performance, which is seen as a critical method to compensate for the limitations of single-chip processes [4][5]. Group 4: Industry Outlook - The artificial intelligence industry in China is experiencing rapid growth, but the chip sector is still in its early stages, leading to an impending period of market consolidation where only a few companies are expected to survive [6][7]. - The rise of large models and multi-modal training is pushing computing clusters into a new era, with supernode architecture becoming a key technological path for this evolution [5][6].