证券研究报告 | 行业策略 gszqdatemark 2026 01 29 年 月 日 计算机 2026 年度策略：锚定 AI 未来，共启科技新篇 金融科技：新工具+新货币，重塑金融新范式。1）新工具：牛市先锋，优 选炒股软件。2026 年全球流动性合理充裕，国内保持适度宽松基调；同 时，险资兜住下行风险，存款搬家也有望成为重要增量。2026 年我们看好 指数继续上攻，炒股软件是牛市先锋，成交活跃下选择肌肉记忆票，牛市 有望持续受益，建议关注同花顺、东方财富、九方智投、指南针等。2）新 货币：数字人民币与人民币国际化双轮驱动。数字人民币 2.0 在 1 月 1 日 正式实施，从 M0 到 M1 的货币属性跃迁。同时，人民币国际化持续推进， 《人民币跨境支付系统业务规则》将自 2026 年 2 月 1 日起施行，加大拓 展 CIPS 网络力度是工作重点。数字人民币建议关注收单方、清算机构以 及银行做数币支付系统厂商，如新大陆、新国都、拉卡拉等；人民币国际 化核心受益方包括 CIPS 相关厂商。 AI 应用：C 端看巨头起舞，B 端是最后一公里的赛跑。1）在即将到来的 2026，模型端的进步仍然值得期待：采用 s ...

Group 1 - The core viewpoint of the news highlights the performance of the big data industry ETF (516700), which showed a decline of 0.52% in intraday trading, while certain constituent stocks like Qimingxingchen, Guotou Intelligent, and Qi Anxin performed strongly with increases of 2.17%, 1.49%, and 0.98% respectively [1] - The article discusses the recent developments in the AI sector, particularly the launch of the new GPU architecture "Huagang" by Moore Threads, which supports full precision computing and high-performance AI training chips, achieving international mainstream levels in several metrics [3] - The focus of the AI industry is shifting from "fast thinking" to "slow thinking," emphasizing algorithm and training paradigm optimization, with a transition from benchmark testing to real-world task execution [3] Group 2 - The big data industry ETF (516700) passively tracks the CSI Big Data Industry Index, heavily investing in sectors such as data centers, cloud computing, and big data processing, with key stocks including Zhongke Shuguang, Keda Xunfei, and Unisplendour [4] - There are three main catalysts for investors interested in the direction of technological self-reliance: the call for "technology to lead," the top-level design of Digital China to activate digital productivity, and the acceleration of the domestic substitution process [4] - The article notes that the trend of "信创" (Xinchuang) is expected to accelerate, indicating a broad outlook for technological self-reliance [4]

Market Performance - A-shares showed mixed performance from December 15 to December 19, with the Shanghai Composite Index closing at 3890.45 points, up 0.03%, while the Shenzhen Component Index fell 0.89% to 13140.21 points, and the ChiNext Index dropped 2.26% to 3122.24 points [1][2] - Over 54% of stocks in the A-share market rose during the week, with 107 stocks gaining over 15% and 31 stocks declining more than 15% [1] International Market Trends - The New York stock market indices rose on December 19, with the Dow Jones Industrial Average increasing by 183.04 points to 48134.89 points (up 0.38%), the S&P 500 rising by 59.74 points to 6834.50 points (up 0.88%), and the Nasdaq Composite gaining 301.26 points to 23307.62 points (up 1.31%) [3][4] - European stock indices also saw gains, with the FTSE 100 closing at 9897.42 points (up 0.61%), the CAC 40 at 8151.38 points (up 0.01%), and the DAX at 24288.40 points (up 0.37%) [3][4] - International oil prices experienced slight increases, with WTI crude oil futures rising by 51 cents to $56.66 per barrel (up 0.91%) and Brent crude oil futures increasing by 65 cents to $60.47 per barrel (up 1.09%) [3][4] Regulatory Developments - The National Development and Reform Commission, along with other regulatory bodies, released the "Internet Platform Pricing Behavior Rules" to enhance price regulation and protect consumer rights in the platform economy [5] - The National Internet Information Office and the China Securities Regulatory Commission are intensifying efforts to combat false information in the capital market, warning that spreading such information will lead to legal consequences [6] Automotive Industry Advancements - Changan Automobile received the first official license plate for L3-level autonomous driving, marking a significant milestone in the development of intelligent connected vehicles in China [6][7] - The Ministry of Industry and Information Technology granted the first batch of L3-level conditional autonomous driving vehicle licenses, allowing companies like Changan and BAIC to conduct road trials in designated areas [6][7] Technology Sector Developments - Moore Threads unveiled its new GPU architecture "Huagang," which supports full precision computing and is designed for large-scale computing clusters, with plans to launch high-performance AI chips in the coming year [8] - The AI infrastructure investment is expected to grow, with the global computing chip market expanding, benefiting hardware suppliers in the industry [8] Financial Sector Initiatives - The People's Bank of China and other financial authorities in Fujian Province are implementing a plan to enhance financial support for technology innovation and strategic emerging industries, aiming for a 10% annual growth in technology-related loans by 2025-2027 [9] Scientific Breakthroughs - Researchers at Shanghai Jiao Tong University achieved a breakthrough in the field of optical computing chips, which could support large-scale semantic media generation models, addressing the performance gap in traditional chip architectures [10] Commodity Market Insights - International silver prices reached a new high, with a year-to-date increase of 132.11%, significantly outpacing gold prices [11] Investment Strategies - Tianfeng Securities suggests focusing on three main investment directions: breakthroughs in AI technology, economic recovery, and undervalued assets benefiting from AI trends [12] - Huaxi Securities indicates that conditions for a "spring rally" are accumulating, recommending investments in sectors supported by industrial policies and consumer stimulus [13]

Group 1 - Ukraine's President Zelensky announced constructive progress in negotiations with the U.S. regarding the end of the Russia-Ukraine conflict, security guarantees for Ukraine, and post-war reconstruction [2] - Israel's Foreign Minister stated that Gaza can only be rebuilt if Hamas disarms, emphasizing that Hamas is currently trying to consolidate its power in Gaza [3] - The U.S. Starlink satellite experienced an anomaly, losing contact and producing debris, with monitoring efforts underway to assess potential threats to other satellites [4] Group 2 - China's Ministry of Commerce held a meeting to summarize and plan for business work, focusing on the implementation of the 14th Five-Year Plan and preparing for the 15th Five-Year Plan [5] - New regulations were issued for internet platform pricing, prohibiting platform operators from selling below cost to eliminate competition, with exceptions for certain types of goods [6] - China successfully launched the Communication Technology Test Satellite No. 23, which will be used for multi-band, high-speed satellite communication technology verification [7] Group 3 - China's animated film box office exceeded 25 billion yuan in 2025, marking the highest year in the history of Chinese animated films [8] - Chinese robots won a championship at the second Future Sports Games held in Abu Dhabi, showcasing advancements in robotics [9] Group 4 - The Russian Foreign Ministry denied any secret discussions with South Korea regarding North Korea's nuclear issues, asserting that the meeting was misrepresented by the media [10] - The Russian Defense Ministry reported advances in Ukraine, with Russian forces occupying more advantageous positions and targeting Ukrainian military infrastructure [11] - Thailand's Prime Minister announced that the military has controlled nearly all target areas, emphasizing adherence to international law and the aim to prevent further conflict escalation [12] Group 5 - Venezuelan President Maduro stated that the country has faced 25 weeks of aggression but remains committed to defending national peace and stability [13] - A shooting incident in South Africa resulted in 10 deaths and 10 injuries, highlighting ongoing security challenges in the region [14][15] Group 6 - Moore Threads announced a new GPU architecture that supports large-scale intelligent computing clusters, improving computing density by 50% and energy efficiency by 10 times [16] - Changan Automobile received the first official license plate for L3-level autonomous driving, marking a significant milestone in the autonomous vehicle industry [17] Group 7 - Nearly 190 billion yuan worth of restricted stocks will be unlocked this week, with several companies having significant unlock values, including Shouchuang Securities and Sany Heavy Energy [18][19][20] - Three new stocks are set to be issued this week, with specific details on issuance prices and subscription limits provided [22][23]

Group 1: Moore Threads Technology Roadmap - Moore Threads has unveiled its new generation full-featured GPU architecture "Huagang," which boasts a 50% increase in computing density and a 10-fold improvement in energy efficiency, supporting full precision calculations from FP4 to FP64 and capable of supporting over 100,000 card intelligent computing clusters [1] - The company is set to release the "Huashan" AI training and inference integrated chip and the "Lushan" high-performance graphics rendering GPU, with a computing power of 10 EFLOPS for the Wan Card intelligent computing cluster, and the S5000 single card inference sets a new record for domestic GPU performance [1] - The AI computing book MTT AIBOOK, equipped with the "Yangtze River" SoC chip, offers 50 TOPS heterogeneous AI computing power and can locally run large models with up to 30 billion parameters, now available for pre-sale on JD.com [1] Group 2: OpenAI's GPT-5.2-Codex Launch - OpenAI has launched GPT-5.2-Codex, which is considered the most advanced intelligent coding model to date, achieving state-of-the-art performance in SWE-Bench Pro and Terminal-Bench 2.0 benchmark tests [2] - Compared to GPT-5.2, it has improved instruction-following capabilities, long context understanding, and network security features, with better performance in Windows environments and significant improvements in token efficiency at mid-high inference levels [2] - The model is now available to paid ChatGPT users across all Codex platforms, with plans to open access to API users in the coming weeks and provide more lenient access for defensive cybersecurity professionals [2] Group 3: Google's Gemma Models - Google has open-sourced two models from the Gemma 3 family, T5Gemma 2 and FunctionGemma, with T5Gemma 2 being the first multi-modal long-context encoder-decoder model, available in sizes of 270M-270M, 1B-1B, and 4B-4B [3] - FunctionGemma is optimized for function calls, running on just 270 million parameters, suitable for mobile and browser devices, and supports precise structured data output for external API calls, making it ideal for edge AI agent applications [3] - T5Gemma 2 returns to the classic Encoder-Decoder architecture, surpassing similarly sized Gemma 3 models in multi-modal performance, code reasoning, and long context capabilities, while FunctionGemma can be reduced to 135MB for operation through quantization [3] Group 4: NVIDIA's NitroGen Model - NVIDIA has open-sourced the NitroGen foundational model, designed to play over 1,000 games, using game video frames as input to output real controller operation signals, and supports rapid adaptation to new games through post-training [4] - The model is based on the GR00T N1.5 architecture and utilizes 500 million parameters, trained by automatically extracting action labels from 40,000 hours of publicly available game videos, covering various game types including RPGs, platformers, and racing [4] - It can accomplish non-trivial tasks without fine-tuning, achieving a task success rate improvement of up to 52% compared to models trained from scratch, and the dataset, evaluation suite, and model weights have been made open-source [4] Group 5: OpenAI's Codex Agent Skills Support - OpenAI has announced that Codex now fully supports Agent Skills, integrating with industry-standard specifications led by Anthropic, which include markdown commands and optional script resources [5] - It allows for explicit calls (via /skills command or $selection) and implicit calls (automatically matching descriptions based on tasks), with skill storage prioritized from the current working directory to the user's personal directory [5] - Built-in tools like $skill-creator and $skill-installer are provided to automatically generate skill frameworks or install skills from third-party repositories like GitHub, with an official Skill library released by OpenAI [5] Group 6: Luma AI's Ray3 Modify - Luma AI has launched the Ray3 Modify feature, emphasizing a "real person first, AI follows" approach to video production, where actor performances and camera movements serve as the foundational input for AI processing [6] - It supports keyframe control (start and end frames), character reference capabilities, and retains the integrity of performances, allowing the same performance to be placed in different scenes for various content versions without reshooting [6] - Integrated into the Dream Machine platform, it targets film production, advertising creativity, and post-production processes, enabling creators to maintain control without the need for repeated filming [6] Group 7: METR Report on Claude Opus 4.5 - The METR report indicates that Claude Opus 4.5 can sustain coding for approximately 4 hours and 49 minutes, marking the longest time span reported to date, surpassing GPT-5.1-Codex-Max's 2 hours and 53 minutes [9] - The task duration for AI coding agents is showing exponential growth, doubling every 7 months from 2019 to 2024, and expected to double every 4 months from 2024 to 2025, with predictions that AI will complete a full workday's tasks by April 2026 [9] - The industry views long-term memory as the final challenge towards achieving AGI, as current models rely on retrieval tools and context compression, lacking true self-learning and persistent memory capabilities [9] Group 8: Google AI's Success Story - Josh Woodward, the head of Google AI products, has driven the Gemini application’s monthly active users from 350 million in March to 650 million in October, surpassing ChatGPT to top the App Store rankings [10] - At 42 years old and from Oklahoma, he joined Google through an internship in 2009, contributing to Chromebook development, founding the NBU initiative, and leading the expansion of Google Pay, before taking over as Gemini application head in April 2025 [10] - He has promoted the NotebookLM project to break Google's traditional practices by utilizing Discord for community engagement, establishing a "Block" ticketing system to eliminate bureaucratic obstacles, and initiating the "Papercuts" plan to address minor issues, emphasizing the balance between AI innovation and social responsibility [10]

Core Insights - The first MUSA Developer Conference (MDC2025) was held by Moore Threads in Beijing, emphasizing the advanced and leading technology of the MUSA architecture as the foundation for full-function GPUs [1] - The company aims to enhance its research and development investments to tackle core technological challenges from hardware to software, focusing on building a self-reliant domestic computing industry ecosystem [1] Group 1 - The MUSA architecture supports full precision computing from FP4 to FP64, with a 50% increase in density and a 10-fold improvement in efficiency, capable of supporting computing clusters exceeding 100,000 cards [1] - Moore Threads plans to launch two high-performance chips based on the MUSA architecture: the "Hua Shan" chip for AI training and inference, and the "Lu Shan" chip for high-performance graphics rendering, both set to be mass-produced next year [1] Group 2 - The company showcased the "Kua E Wan Ka" intelligent computing cluster, demonstrating its capability to support trillion-parameter model training with reliability and achieving international mainstream levels in several key precision metrics [2] - Moore Threads shared its planning for the MTT C256 super-node architecture aimed at the next generation of ultra-large-scale intelligent computing centers [2] - The company also released the MTTAIBOOK, an AI computing device equipped with the intelligent SoC chip "Chang Jiang" [2]

Core Insights - The company unveiled its next-generation GPU architecture "Huagang" at the first MUSA Developer Conference (MDC2025) in Beijing, showcasing advancements in AI training clusters and various technologies [1][2] - The new architecture supports full precision computing from FP4 to FP64, with a 50% increase in computing density and a 10x improvement in energy efficiency [1] - The company plans to launch the "Huashan" chip focused on AI training and inference, and the "Lushan" chip aimed at graphics rendering [1] Architecture and Performance - The "Huagang" architecture enhances training cluster capabilities with the "Kua'e" 10,000-card intelligent computing cluster, achieving 60% training utilization on dense models and 40% on mixture of experts models, with a linear scaling efficiency of 95% [1] - In inference, the company collaborated with Silicon-based Flow to achieve a single card Prefill throughput exceeding 4000 tokens/s and Decode throughput over 1000 tokens/s on the DeepSeek R1671B model [1] Software Ecosystem - The MUSA 5.0 version optimizes programming models, computing libraries, and compilers, with core computing library muDNN's GEMM and FlashAttention efficiency exceeding 98% and communication efficiency reaching 97% [1] - The company plans to gradually open-source some core components, including computing acceleration libraries and system management frameworks [1] Graphics and AI Integration - The new architecture integrates hardware ray tracing acceleration engines and supports self-developed AI generative rendering technology [2] - The company introduced the MTLambda simulation training platform and the MTT AIBOOK based on the "Yangtze" SoC, focusing on cutting-edge fields like embodied intelligence and AI for Science [2] Future Infrastructure - The company announced the MTTC256 super-node architecture design for next-generation large-scale intelligent computing centers, emphasizing high-density hardware and energy efficiency optimization [2] - The comprehensive technology layout from chip architecture to cluster infrastructure and edge devices aims to support the development of the domestic AI computing ecosystem [2] - Industry experts believe the company is positioning itself to compete directly with Nvidia by releasing its architecture early to boost confidence in its software ecosystem [2]

Core Viewpoint - The A-share market is experiencing a significant rebound, driven by substantial inflows of medium to long-term capital, indicating a potential start of a bullish trend as of December 17, 2025 [2][4]. Market Performance - The A-share market saw a strong performance with a total of 2,977 stocks rising during the week, marking it as the best-performing week in December [5]. - The market indices showed a mixed performance, with most major indices rebounding after initial declines earlier in the week [8]. Capital Inflows - The A500 ETF has seen a remarkable increase in scale, reaching 245.935 billion yuan, with a net inflow of 32.7 billion yuan in the past week, accounting for nearly 70% of total net inflows into stock ETFs [2]. - The Huatai-PB A500 ETF has become the first ETF tracking this index to exceed 40 billion yuan in scale, indicating strong investor interest [2]. Policy and Economic Environment - Short-term policies are expected to remain supportive, with potential interest rate cuts and measures to boost consumption likely to be implemented [11]. - External risks are perceived to be limited, with stable Sino-U.S. relations and the Bank of Japan's recent interest rate hike contributing to a more favorable outlook [12]. Sector Focus - Analysts suggest focusing on growth and consumer sectors, as these areas are likely to benefit from supportive policies and have been relatively underperforming this year [16]. - The technology, media, and telecommunications (TMT) sectors, along with advanced manufacturing, are highlighted for their potential during the upcoming "spring rally" [16]. Market Sentiment - Recent events have reduced external uncertainties, contributing to a warming market sentiment [9]. - However, caution is advised as some high-performing stocks have shown signs of volatility, indicating potential shifts in risk appetite among investors [17].

Core Viewpoint - The article highlights the significant advancements made by domestic GPU manufacturer Moore Threads with the launch of their new GPU architecture "Huagang," aimed at enhancing local AI computing power supply capabilities [1]. Group 1: New GPU Architecture - Moore Threads announced the mass production of the "Huagang" architecture-based series chips, including the Huashan and Lushan series, set to launch next year [1]. - The new architecture features a next-generation instruction set, achieving a 50% increase in computing density and a 10-fold improvement in energy efficiency, supporting large-scale intelligent computing clusters of over 100,000 cards [1][2]. - The Huashan GPU chip has surpassed international mainstream chips in floating-point computing power, memory bandwidth, memory capacity, and high-speed interconnect bandwidth, marking a technological advancement in high-performance computing for domestic GPUs [2]. Group 2: Performance Enhancements - The Lushan high-performance graphics rendering chip, based on the "Huagang" architecture, is expected to deliver a 15-fold improvement in 3A game rendering, a 64-fold increase in AI performance, and a 50-fold enhancement in ray tracing performance [2]. - The architecture also includes a new first-generation AI generative rendering architecture (AGR) and a second-generation ray tracing hardware acceleration engine [5]. Group 3: Software Ecosystem Upgrade - Alongside hardware upgrades, Moore Threads announced a comprehensive software upgrade for its self-developed MUSA architecture, which supports a full-function GPU computing ecosystem with a complete software stack and broad ecological compatibility [7]. - The unified software stack covers the entire product range, emphasizing the importance of full-function GPUs in supporting increasingly complex AI models and diverse application scenarios [7]. Group 4: AI Computing Notebook Launch - Moore Threads introduced the MTT AIBOOK, the first AI computing notebook equipped with the self-developed intelligent SoC chip "Changjiang," featuring a high-performance full-core CPU and a full-function GPU [9][10]. - The notebook supports the MUSA unified architecture and offers heterogeneous AI computing capabilities of 50 TOPS, targeting AI developers and professional users [10]. - The MTT AIBOOK is available for pre-sale at a price of 9,999 yuan for the 32GB RAM and 1TB storage version [10].

Core Viewpoint - Moore Threads has launched its next-generation GPU architecture "Huagang" at the MUSA Developer Conference, showcasing significant advancements in computing power and energy efficiency [1][2]. Group 1: Product and Technology Developments - The "Huagang" architecture features a 50% increase in computing density and a 10-fold improvement in energy efficiency, supporting large-scale intelligent computing clusters of over 100,000 cards [1][2]. - Moore Threads has introduced a full-function GPU architecture and the "Kua'a" intelligent computing cluster, with plans to release high-performance AI training and inference chip "Huashan" and graphics rendering chip "Lushan" based on the new architecture [2]. - The company has also shared its MTT C256 super-node architecture for next-generation large-scale intelligent computing centers and launched the MTT AIBOOK, an AI-powered laptop with a self-developed SoC chip "Changjiang," available for pre-sale at a price of 9,999 yuan [2][3]. Group 2: Market Position and Financial Performance - Moore Threads is recognized as "China's version of NVIDIA" and has recently gone public, experiencing a significant stock price increase of over 400% on its first trading day, although it has since seen a decline to 664.1 yuan per share [4]. - For the first nine months of 2025, the company reported a revenue of 785 million yuan and a net loss of 724 million yuan, with projected losses for the full year ranging from 730 million to 1.168 billion yuan [6].