Market Overview - The major indices showed mixed performance with the North Stock 50 up by 0.90%, while the Shanghai Composite Index decreased by 0.25%, the Shenzhen Component Index fell by 0.33%, the Sci-Tech Innovation 50 dropped by 0.71%, and the ChiNext Index declined by 0.73% [1] - The TMT sector led the gains with SW Communication Cables and Accessories up by 4.40%, SW Robotics increasing by 2.24%, and SW Printed Circuit Boards rising by 1.58% [1] - Conversely, the TMT sector also saw declines with SW LED down by 3.07%, SW Marketing Agencies falling by 2.74%, and SW Communication Network Equipment and Devices decreasing by 2.69% [1] Domestic News - TSMC plans to notify the Japanese government about changing the production focus of its JASM second wafer factory from 6nm logic semiconductors to 3nm, which the Japanese government believes will not compete with the ongoing 2nm project by Rapidus [2] - The Chinese government procurement website announced that China Mobile Guangdong Company won a bid for a project worth 155 million yuan, which includes Huawei's temperature-controlled data backup storage equipment priced at 44.5986 million yuan [2] - A national supercomputing internet core node has officially started trial operations, deploying three sets of the Sunway TaihuLight scaleX 30,000-card supercluster system, becoming the largest domestic AI computing pool in operation [2] - ASE Group indicated that its revenue growth in 2026 will be driven by advanced solutions, AI proliferation, and semiconductor market recovery, with revenue expected to double from 1.6 billion USD in 2025 to 3.2 billion USD [2] International News - Forge Nano, Inc. successfully demonstrated high-speed, defect-free atomic layer deposition (ALD) coatings on semiconductor structures with a 1000:1 aspect ratio, maintaining good conformality at production scale [3] - Amazon plans to invest approximately 200 billion USD by 2026, driven by strong demand for its existing products and services, as well as opportunities in AI, chips, robotics, and near-Earth orbit satellites [3] - Infineon announced the acquisition of ams Osram's non-optical analog/mixed-signal sensor product portfolio, enhancing its offerings in automotive and industrial sensors and expanding its product matrix in healthcare applications [3] - Sunic System signed a contract worth approximately 97.786 million yuan with Hongxi Technology for the supply of evaporation equipment for mass production of OLEDoS displays [3] AI Insights - The world's first South China Sea air-sea coupled intelligent model "Feiyu-1.0" has been released, expected to play a significant role in marine environmental protection, resource development, and climate change response [4] - The MiniCPM-o 4.5 has been released, featuring a native full-duplex architecture that allows for parallel input and output without blocking, capable of capturing audio and visuals simultaneously [4] - Anthropic released the Claude Opus 4.6 model, introducing an "Adaptive Thinking" mechanism that improves its ability to solve complex problems, achieving a score of 53.1% in the Humanity's Last Exam (HLE) test, significantly higher than previous versions [4] - Pony.ai has received a commercial operation license for robotaxi services across Shenzhen, marking it as the first major city globally to open L4 autonomous driving commercial operations [4] Industry Tracking - China successfully launched a reusable experimental spacecraft using the Long March 2F rocket from the Jiuquan Satellite Launch Center, aimed at providing technical verification for peaceful space utilization [5] - Dongyi Technology released PHYBOTC2, a robot capable of carrying 5kg with a movement speed of 4m/s, applicable in various scenarios such as commercial guidance and education [5] - A Chinese team has made significant breakthroughs in developing a three-dimensional conical carbon-based flexible brain cortex electrode array, addressing key challenges in the industry and completing the first international space environment technology verification [5] - Baosteel's large roller cladding flux-cored wire has successfully launched, providing quality material support for cost reduction and efficiency improvement in industries like metallurgy and mining [5]

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]

2月5日，国家超算互联网核心节点在河南郑州上线试运行。该节点是国家超算互联网平台上线以来接入 的最大单体国产AI算力资源池，可提供超过3万卡的国产AI算力，满足大规模AI计算需求。 据介绍，该节点的算力资源由中科曙光scaleX万卡超集群系统支撑，基于AI计算开放架构，兼容主流软 件生态，支持多品牌国产加速卡混合部署，并已完成400多个大模型的适配优化。在国家超算互联网平 台一体化算力调度下，可接入上千款应用。 截至2025年底，国家超算互联网平台已服务100多万用户，上架超7300个应用商城商品。 来源：新华社 作者: 温竞华 杨琳 ...

Core Insights - The article discusses the advancements in the domestic GPU industry, highlighting the launch of the "Huagang" architecture by Moore Threads and the "scaleX" supercluster system by Inspur, indicating a shift in focus from individual GPU performance to building scalable systems capable of handling massive computational tasks [2][6]. Group 1: Moore Threads Developments - Moore Threads unveiled its latest "Huagang" architecture, which boasts a 50% increase in computing density and a 10-fold improvement in efficiency compared to the previous generation [3]. - The "Huagang" architecture supports full precision calculations from FP4 to FP64 and introduces new support for MTFP6, MTFP4, and mixed low precision [3]. - Future chip plans include "Huashan," aimed at AI training and inference, and "Lushan," focused on high-performance graphics rendering, with "Lushan" showing a 64-fold increase in AI computing performance and a 50% improvement in ray tracing performance [4]. Group 2: Inspur Developments - Inspur's "scaleX" supercluster system, which publicly debuted, consists of 16 scaleX640 supernodes interconnected via the scaleFabric high-speed network, capable of deploying 10,240 AI accelerator cards [10]. - The scaleX system employs immersion phase change liquid cooling technology to address heat dissipation challenges, achieving a 20-fold increase in computing density per rack and a PUE (Power Usage Effectiveness) of 1.04 [11][12]. - The system supports multi-brand accelerator cards and has optimized compatibility with over 400 mainstream large models, reflecting a strategy to provide a versatile platform for various domestic computing resources [14]. Group 3: Industry Challenges and Solutions - The industry faces challenges in scaling up computational power, particularly in managing heat, power supply, and physical space limitations when deploying thousands of high-power chips in data centers [8][9]. - Both companies are addressing communication delays in distributed computing, with Moore Threads integrating a new asynchronous programming model and self-developed MTLink technology to support clusters exceeding 100,000 cards, while Inspur's scaleFabric network achieves 400 Gb/s bandwidth and sub-microsecond communication latency [12][13]. Group 4: Software Ecosystem and Compatibility - As the hardware specifications approach international standards, the focus is shifting towards optimizing the software stack, with Moore Threads announcing an upgrade to its MUSA unified architecture and achieving over 98% efficiency in core computing libraries [13]. - Inspur emphasizes the compatibility of its systems with various brands of accelerator cards, promoting an open architecture strategy that allows for coexistence of multiple chips [14].