Core Viewpoint - The collaboration between Shanghai Artificial Intelligence Laboratory's OpenDataLab team, DeepLink team, and domestic chip manufacturers has successfully adapted over ten mainstream domestic computing power solutions for the MinerU AI document parsing tool, enhancing its ecosystem compatibility and adaptability [1] Group 1: Collaboration and Development - The OpenDataLab team and DeepLink team have partnered with more than ten domestic chip manufacturers, including Ascend, PingTouGe, MuXi, HaiGuang, SuiYuan, MoErThread, TianShuZhiXin, HanWuJi, KunLunChip, TaiChuYuanQi, and BiRan [1] - The collaboration aims to implement a full-stack optimization strategy through hardware and software synergy [1] Group 2: Product Features - MinerU, developed by the Shanghai Artificial Intelligence Laboratory, is an AI document parsing tool that utilizes a self-developed VLM model [1] - The tool achieves an impressive accuracy rate of 99% in capturing elements from PDFs and complex web pages [1]

值得一提的是，据智通财经报道，2月12日，智谱新发布的GLM-5已完成与华为昇腾、摩尔线程、寒武 纪、昆仑芯、沐曦、燧原、海光等主流国产芯片平台的深度推理适配与算子级优化，能够在国产算力集 群上实现高吞吐、低延迟的稳定运行。据报道，这背后的意义不止是一次工程适配，更关乎 AI算力供 应链的自主可控，叠加国产算力平台的可控供给，使得开源Agentic Engineering级生产力工具第一次具 备可大规模交付的现实条件。 智谱(02513)涨幅扩大逾26%，截至发稿，涨24.01%，报388港元，成交额9.25亿港元。 消息面上，2月11日，智谱正式发布新一代旗舰模型GLM-5。在Coding与Agent能力上，GLM-5取得开源 SOTA表现，在真实编程场景的使用体感逼近Claude Opus4.5，擅长复杂系统工程与长程Agent任务。目 前，GLM-5模型已上线智谱开发平台BigModel.cn、Z.ai及智谱清言(App/网页端)，可供用户体验。需要 注意的是，由于算力限制，当前仅Max用户套餐中包含GLM-5调用额度，官方表示将逐步扩大开放范 围。 ...

值得一提的是，据智通财经报道，2月12日，智谱新发布的GLM-5已完成与华为昇腾、摩尔线程、寒武 纪、昆仑芯、沐曦、燧原、海光等主流国产芯片平台的深度推理适配与算子级优化，能够在国产算力集 群上实现高吞吐、低延迟的稳定运行。据报道，这背后的意义不止是一次工程适配，更关乎 AI算力供 应链的自主可控，叠加国产算力平台的可控供给，使得开源Agentic Engineering级生产力工具第一次具 备可大规模交付的现实条件。 智通财经APP获悉，智谱(02513)涨幅扩大逾26%，截至发稿，涨24.01%，报388港元，成交额9.25亿港 元。 消息面上，2月11日，智谱正式发布新一代旗舰模型GLM-5。在Coding与Agent能力上，GLM-5取得开源 SOTA表现，在真实编程场景的使用体感逼近Claude Opus4.5，擅长复杂系统工程与长程Agent任务。目 前，GLM-5模型已上线智谱开发平台BigModel.cn、Z.ai及智谱清言(App/网页端)，可供用户体验。需要 注意的是，由于算力限制，当前仅Max用户套餐中包含GLM-5调用额度，官方表示将逐步扩大开放范 围。 ...

不仅如此，摩尔线程还首次公布了MTT S5000的算力参数。该卡基于第四代MUSA架构"平湖"打造。其 单卡AI算力最高可达1000 TFLOPS，配备80GB显存，显存带宽达到1.6TB/s，卡间互联带宽为 784GB/s，完整支持从FP8到FP64的全精度计算。 随后有消息传出，目前智谱GLM-5已完成与华为昇腾、摩尔线程、寒武纪、昆仑芯、沐曦、燧原、海 光等主流国产芯片平台的深度推理适配与算子级优化，能够在国产算力集群上实现高吞吐、低延迟的稳 定运行。 2月11日，智谱AI上线并开源GLM-5，随后多款国产芯片完成对GLM-5的Day0适配。 海光信息方面表示，在本次Day0适配合作中，海光DCU团队深度协同智谱AI，凭借DTK等自研软件栈 优势，重点优化底层算子与硬件加速，使GLM-5在海光DCU上实现高吞吐、低延迟的稳定运行，充分 发挥"国产算力+国产大模型"价值。 摩尔线程表示，基于SGLang推理框架，该公司在旗舰级AI训推一体全功能GPU MTT S5000上，Day-0 完成了全流程适配与验证。凭借MUSA架构算子覆盖与生态兼容能力，摩尔线程打通了模型推理全链 路，并深度释放MTT S500 ...

Group 1 - The AI inference token consumption in China is projected to increase from approximately 10 trillion in 2025 to about 3,900 trillion by 2030, representing a growth of around 370 times over five years [1] - The explosive growth is driven by two main factors: increased penetration of AI in consumer and enterprise workloads, and the expansion of application scenarios from simple conversational AI to complex intelligent agents and multimodal outputs [1] - As of December 2025, the daily token consumption of ByteDance's Doubao series models has exceeded 50 trillion, showing a year-on-year increase of over 10 times [1] Group 2 - Domestic models are entering a release window as the Spring Festival approaches, with rapid iterations expected to capture user AI interaction entry points, leading to an accelerated demand for inference-side computing power [1] - 2026 is anticipated to be a pivotal year for the domestic super node market, with numerous domestic manufacturers releasing next-generation super node solutions, including Huawei's Atlas 950/960 and others [1] - The industry chain is expected to reach a significant growth point as both supply and demand sides are actively engaged [1] Group 3 - Kehua Data has signed an ecological cooperation agreement with Wallen Technology and Digital China to jointly build a four-in-one industrial ecosystem encompassing "chip-server-data center-computing power services" [2]

Group 1 - The Shanghai Stock Exchange Sci-Tech Innovation Board Semiconductor Materials and Equipment Theme Index rose by 0.27% as of February 10, 2026, with notable gains from Huafeng Measurement Control (up 8.05%) and Aisen Co., Ltd. (up 7.48%) [1] - The ChiNext Semiconductor Materials and Equipment Theme Index fell by 0.10%, with Huafeng Measurement Control leading the gains at 8.76%, while Xidian Co., Ltd. saw the largest decline at 3.97% [1] - The Sci-Tech Semiconductor ETF (588170) experienced a turnover of 3.4% with a transaction volume of 274 million yuan, while the Semiconductor Equipment ETF Huaxia (562590) had a turnover of 1.65% and a transaction volume of 45.47 million yuan [1] Group 2 - The latest net outflow for the Sci-Tech Semiconductor ETF was 63.71 million yuan, while the Huaxia Semiconductor Equipment ETF saw a net inflow of 3.83 million yuan over the last 21 trading days, with a total inflow of 1.269 billion yuan [2] - Zhongke Shuguang announced plans to raise up to 8 billion yuan through convertible bonds for projects related to advanced computing power systems for artificial intelligence [2] - The semiconductor equipment and materials industry is identified as a key area for domestic substitution, benefiting from the AI revolution and ongoing technological advancements [3]

Core Viewpoint - The article highlights the significant growth in the AI computing sector, driven by advancements in domestic models and government initiatives aimed at enhancing computing resource efficiency and service levels [3][4]. Group 1: Market Performance - Major indices saw an increase, with AI computing and application sectors leading the gains [1]. - Cloud computing ETFs (159890) and software leader ETFs (159899) both rose over 3% during trading [1]. - Notable stock performances included 360 Technology hitting the daily limit, and Oriental Guoxin rising over 18% [1]. Group 2: Government Initiatives - A recent government notification aims to promote the "1+M+N" national computing interconnection node system, which is expected to enhance the quality of the computing industry and improve the efficiency of public computing resource usage [3]. - The total number of operational computing center racks in the country has increased from 5.2 million to 12.5 million over the past five years, with an annual growth rate of 30% [3]. Group 3: Model Development and Investment - The period leading up to the Spring Festival has seen a surge in the release of domestic AI models, including significant updates from companies like DeepSeek, Alibaba, and Baidu [4]. - ByteDance plans to invest 160 billion yuan in capital expenditures by 2026, while Alibaba is advancing a 3-year plan for 380 billion yuan in AI infrastructure [4]. - The acceleration in model iteration is expected to drive rapid growth in demand for inference-side computing power, with 2026 anticipated to be a pivotal year for domestic super node deployment [5]. Group 4: Industry Outlook - The rapid iteration of domestic models is likely to significantly increase demand for inference-side computing power, benefiting the domestic computing industry [5]. - The cloud computing ETF (159890) tracks major players in the AI infrastructure and application sectors, while the software leader ETF (159899) focuses on key software development companies, both expected to benefit from the ongoing industry trends [5].

Core Viewpoint - The Ministry of Industry and Information Technology (MIIT) has announced the construction of a national computing power interconnection node system, which is expected to enhance the efficiency and service level of public computing resources and promote high-quality development in the computing power industry [2][3]. Group 1: National Computing Power Interconnection Node System - The MIIT's notification outlines a "1+M+N" system for national computing power interconnection nodes, aiming to standardize and facilitate the efficient flow of computing resources across different regions and sectors [2][3]. - The construction goals include establishing a supply-demand matching system for computing power in regions and key industries, enhancing overall computing power levels [2][3]. - Regional nodes will provide a unified service platform for interconnection, including services such as resource aggregation, selection, and monitoring [2][3]. Group 2: Industry Trends and Demand - The domestic computing power demand is expected to accelerate, driven by rapid iterations of domestic AI models and the evolving interaction methods that are shaping the demand for inference-side computing power [4][5]. - Major AI models from companies like Alibaba and Baidu are being released, which will further drive the commercialization of these models and increase the demand for computing power [5][6]. - The infrastructure for AI computing power is undergoing systematic upgrades to support the explosive demand for training and inference, transitioning from vertical scaling to distributed cluster expansion and cross-domain collaboration [6][7]. Group 3: Hardware Innovations - Key hardware components such as optical modules and switches are innovating around high speed, low power consumption, and low cost, with a focus on co-packaged optics (CPO) and linear drive pluggable optics (LPO) [7]. - Domestic optical module manufacturers are gaining a leading position in the global market, which positions them to benefit significantly from the expansion of AI computing infrastructure [7].

Group 1 - The China Securities Information Technology Application Innovation Industry Index decreased by 0.05% as of February 6, 2026, with mixed performance among constituent stocks [1] - Geer Software led the gains with an increase of 8.45%, while ZTE Information fell by 2.81%, indicating volatility in the sector [1] - The Xinchuang ETF (562570) experienced a decline of 0.27%, with the latest price at 1.47 yuan [1] Group 2 - The Xinchuang ETF saw a significant growth in scale, increasing by 24.46 million yuan over the past week, leading its peers [2] - The ETF has recorded continuous net inflows over the past four days, with a peak single-day net inflow of 30.01 million yuan, totaling 52.30 million yuan and an average daily net inflow of 13.08 million yuan [1] - Financial analysts believe that the rapid iteration of domestic models is capturing user AI interaction points, which is expected to accelerate the demand for reasoning-side computing power [2] Group 3 - 2026 is projected to be a pivotal year for the deployment of domestic super nodes, with numerous domestic manufacturers releasing next-generation super node solutions [2] - The Xinchuang ETF closely tracks the China Securities Xinchuang Index, emphasizing domestic full-stack replacement and integrating key components such as storage chips, CPUs, and AIPC [2] - The ETF is currently the largest in its category, reflecting strong policy certainty and a comprehensive approach from hardware to industry intelligence [2]

Core Insights - The article discusses the increasing activity of state-owned capital (国资) in direct investments, highlighting a shift from traditional fund investments to direct project investments, with expectations for continued growth in this area by 2026 [3][9][21]. Group 1: Current Trends in State-Owned Capital - State-owned capital has become a dominant force in the primary market, with over 60% of private equity and venture capital funds managed by state-owned entities by the end of 2025 [6]. - In 2025, state-owned institutions participated in over 4,100 sub-funds, marking a 4% increase year-on-year, and directly invested in 5,444 events, covering 4,989 companies with an investment amount exceeding 600 billion yuan, reflecting a 23% and 28% increase in transaction frequency and investment amount respectively [8][21]. - The trend indicates a significant shift towards direct investment by state-owned entities, which is becoming a necessary strategy rather than an optional one [9][21]. Group 2: Impact of Dollar Funds Decline - The decline of dollar funds has led to a drastic reduction in large transactions, with the number of investment events dropping from 871 to 225 and investment amounts plummeting from 532.9 billion yuan to 82.7 billion yuan from 2021 to 2025, a decrease of over 84% [5][6]. - State-owned capital has stepped in to fill the void left by dollar funds, becoming crucial for the financing of unicorn companies, especially in their later funding rounds [11][19]. Group 3: Investment Patterns and Success Stories - Notable companies like Muxi and Moer have seen significant backing from state-owned entities, with state capital contributing approximately 80% of Muxi's total financing of 11.35 billion yuan during its reporting period [12][14]. - The financing patterns show that state-owned capital is increasingly leading or independently investing in projects, with many instances of state-owned entities coordinating with sub-funds to invest in promising projects [20][21]. - The article emphasizes that state-owned capital has proven to be as effective as market-oriented funds in achieving successful IPOs, with over 60 state-owned platforms achieving IPO exits in 2025 [21]. Group 4: Challenges and Future Outlook - Despite the successes, state-owned capital faces challenges related to talent, incentives, and accountability, with some direct investment departments reportedly becoming inactive [24]. - The article suggests that the evolution of state-owned capital from mere fund providers to active industry organizers and ecosystem builders is a natural progression, indicating a long-term shift in the investment landscape [24].