计算机行业12月投资策略展望：国产开源大模型持续迭代，重塑AI竞争格局

Industry Overview - The report highlights the continuous iteration of domestic open-source large models, reshaping the AI competitive landscape, with DeepSeek V3.2 achieving inference capabilities comparable to GPT-5 [1][20][21] - The establishment of a "Space Data Center" in Beijing, with the first computational test satellite "Chengguang No. 1" set to launch, is noted as a significant development in AI computing infrastructure [2][22] Industry Data - In October 2025, the Producer Price Index (PPI) for the computer industry increased by 1.0% month-on-month but decreased by 0.3% year-on-year [5][23] - From January to October 2025, the software industry generated a total revenue of 125,104.00 billion yuan, reflecting a year-on-year growth of 13.2%, with total profits amounting to 15,721.00 billion yuan, up by 7.7% [5][28] - Software exports reached 51.09 billion USD, marking a year-on-year increase of 6.7% [5][28] - The cumulative production of electronic computers from January to October 2025 was 28,750.20 million units, a decrease of 1.2% year-on-year, while the notebook computer export volume in October was 10.29 million units, down by 17.9% [5][24] Company Announcements - Pingao Co., Ltd. announced an investment framework agreement to invest 4 billion yuan in Jiangyuan Technology, acquiring a 15.4182% stake, focusing on AI computing hardware [7][38] - Jiahua Technology is planning a major asset restructuring, intending to acquire a controlling stake in Shudun Information Technology [7][42] Market Review - The report notes a decline of 5.26% in the Shenwan Computer Industry Index from November 1 to November 30, 2025, with all sub-sectors experiencing downturns [9][43] - As of November 28, 2025, the price-to-earnings (P/E) ratio for the Shenwan Computer Industry was 209.42 times, with a valuation premium of 1477.89% relative to the CSI 300 [9][45] Monthly Strategy - The report emphasizes the ongoing iteration of AI large models by leading domestic manufacturers, which is expected to accelerate technological innovation and application deployment [11][50] - Domestic open-source models like DeepSeek, Tongyi Qianwen, and Kimi are narrowing the performance gap with top international closed-source models, potentially gaining a competitive edge due to better cost-performance ratios [11][50] - The report suggests a focus on AI application sectors driven by rapid technological implementation and demand release, recommending attention to leading companies with strong AI technology deployment capabilities [11][50]