Group 1 - The core viewpoint of the article highlights that He Sheng New Materials (002290) has strategically invested in Yizhi Electronics, which, along with companies like SenseTime, Caiyue Xingchen, and Biran Technology, has launched the first batch of "Chip Model Intelligent" ecological co-construction solutions across various fields such as healthcare, securities, and intelligent manufacturing, promoting the integration of digital and physical realms [1] Group 2 - The solutions aim to empower multiple sectors, indicating a broad application potential and collaboration among leading technology firms [1] - The initiative reflects a growing trend in the industry towards digital transformation and smart solutions, which are increasingly important in today's market [1] - The collaboration among these companies signifies a strategic move to enhance innovation and competitiveness in the technology landscape [1]

Core Viewpoint - He Sheng New Materials has strategically invested in Yizhi Electronics, collaborating with companies like SenseTime, Caiyue Xingchen, and Biran Technology to launch the first batch of "Chip Model Intelligence" ecological co-construction solutions, which span multiple fields and empower the integration of digital and physical realms [1][2]. Group 1 - The collaboration focuses on sectors such as healthcare, securities, and intelligent manufacturing [1][2]. - The "Chip Model Intelligence" solutions represent a deep integration of models, chips, data, and intelligence within specific industries, showcasing a closed value loop [1][2].

Core Insights - The article discusses the challenges and advancements in the development of Agentic models, emphasizing that the main bottleneck is not the models themselves but the replication of real-world environments and stable access to external interfaces and databases [2][4][5] - It highlights the current limitations of general-purpose models in achieving industrial-level performance across various vertical agent scenarios, suggesting that tailored models for specific applications are more effective [5][12] - The article also explores the evolution of multi-modal models, indicating that while there have been significant advancements, a unified modeling approach for understanding and generating across modalities remains a key goal for the future [17][20] Group 1: Agentic Models - The primary focus is on enhancing models to perform effectively in various vertical agent scenarios, particularly in coding applications [4] - Current general-purpose models lack the capability to achieve stable generalization across diverse environments, necessitating the customization of models for specific applications [5] - The complexity of real-world environments, including external dependencies and interfaces, poses significant challenges for training agentic models [5][6] Group 2: Multi-Modal Models - The transition from single-modal to multi-modal models has introduced visual capabilities into language models, with a focus on aligning text and visual tokens [17][18] - Despite advancements, the industry faces challenges in scaling multi-modal models due to the difficulty in obtaining high-quality, aligned data [18] - Future directions include the pursuit of unified modeling that integrates generation and understanding capabilities, although current results indicate that separate optimization yields better performance [20][21][22] Group 3: Reinforcement Learning and Training Efficiency - The article emphasizes the importance of reinforcement learning systems for continuous model iteration in specific scenarios, with a focus on high efficiency and throughput [6][9] - The scaling of reinforcement learning has not yet reached a consensus in the industry, but there is recognition of its potential to enhance model capabilities significantly [10][11] - Efficient training processes, particularly in generating diverse paths for evaluation, are critical for the success of reinforcement learning in agentic models [9] Group 4: Future Trends and Directions - The article predicts that the development of agentic models with stable and accurate tool-calling capabilities will expand beyond coding applications to a broader range of real-world APIs [28] - The concept of "world models" is discussed, highlighting the evolution from language models to dynamic models that understand physical world operations [26] - The integration of tools into agent development is seen as a crucial pathway for enhancing model capabilities, reflecting the importance of tool usage in human intelligence evolution [25]

Core Viewpoint - Shanghai Pudong Development Bank (SPDB) is transforming its role from a "fund provider" to an "ecosystem partner," leveraging financial resources to support the development of an "AI City" and the broader technology finance landscape [7][31]. Group 1: AI Industry and IPO Trends - The AI industry is experiencing a rare "IPO wave" in early 2026, with several unicorns like MiniMax and others successfully listing on the Hong Kong Stock Exchange, marking a significant moment for AGI companies [3][4]. - In just 30 days, Shanghai has contributed five major AI and integrated circuit companies to the capital market, indicating a profound financial ecosystem transformation [4]. Group 2: SPDB's Role in Technology Finance - As of Q3 2025, SPDB's technology finance loan balance exceeded 1 trillion yuan, serving over 250,000 technology enterprises, with nearly 90% coverage of companies listed on the Sci-Tech Innovation Board [6][10]. - SPDB's transition to an "ecosystem partner" reflects a strategic shift towards building a comprehensive technology finance ecosystem, moving from initial trials to large-scale, systematic operations [10][11]. Group 3: Innovative Evaluation Systems - SPDB has developed the "Technology Five Forces Model" and "Technology Radar" evaluation systems to bridge the gap between traditional banking assessments and the unique needs of tech companies, focusing on innovation and future potential rather than historical financial metrics [14]. - The bank's proactive approach includes a "Technology Enterprise M&A Loan Action Plan," with over 10 billion yuan allocated to support mergers and acquisitions in the tech sector [14]. Group 4: Long-term Commitment to Technology Innovation - SPDB emphasizes the importance of patience and understanding in supporting technology innovation, positioning itself as a reliable partner throughout the growth cycles of tech companies [16]. - The bank has played a crucial role in supporting domestic GPU companies, providing tailored financial solutions that alleviate cash flow pressures during critical development phases [19]. Group 5: Ecosystem Development in Shanghai - SPDB's technology finance strategy aligns with Shanghai's ambition to become a global technology innovation center, with projections indicating that the AI industry in Shanghai will exceed 550 billion yuan by 2025 [26]. - The bank's comprehensive service model integrates financial support with local industry needs, fostering collaboration among various stakeholders in the technology ecosystem [29][30]. Group 6: Comprehensive Service Framework - SPDB's technology finance services have formed a "four-chain integration" system, connecting innovation, industry, funding, and talent to empower technology enterprises [32][33]. - The bank's approach goes beyond financing, focusing on integrating knowledge, technology, and resources to become a key enabler and builder of the industry ecosystem [34].

记者丨陶力 倪雨晴 彭新 孔海丽 董静怡 骆轶琪 编辑丨骆一帆 科技的迭代速度，正愈发超越人们想象。 2025年，人工智能、量子计算、聚变能源、航天工程等关键领域，均迎来突破性进展，中国 与美国在核心赛道的引领与竞争，推动全球科技格局发生深刻变革。从开源AI模型打破算力 垄断到"人造太阳"刷新能源探索纪录，从量子计算逐步突破到商业航天技术成熟，一系列重 大科技事件不仅重塑了产业生态，也改写了技术发展的底层逻辑。 随着技术积累的持续深化与应用场景的不断拓展， 2026年成为科技从"实验室走向产业化"的 关键一年 ，AI与物理世界的深度融合、脑机接口探索商业化、低空经济的规模化发展等趋 势，将进一步推动科技与民生、产业的深度绑定，开启新一轮全球科技革命与产业变革。 21世纪经济报道科技团队在采访行业专家和资深从业人员的基础上，经过深入讨论和研究， 系统梳理出了2025年全球重大科技事件，观察和透视2026年行业发展新趋势，力图呈现科技 浪潮下的产业变革与未来图景。 2 0 2 5年全球科技重大事件回顾 1.DeepSeek引领全球开源AI模型 2025年春节期间，DeepSeek火爆出圈，迅速在全球应用市场霸榜， ...

Group 1 - The State Intellectual Property Office of China has granted a patent titled "Fine-tuning Method and Device for Point Cloud Classification" to Hangzhou Biran Technology Development Co., Ltd. and Shanghai Biran Technology Co., Ltd., with the announcement number CN116778133B and application date of June 2023 [1] - Hangzhou Biran Technology Development Co., Ltd. was established in 2021, located in Hangzhou, and primarily engages in technology promotion and application services, with a registered capital of 50 million RMB. The company has a total of 4 patents [1] - Shanghai Biran Technology Co., Ltd. was founded in 2019, based in Shanghai, and focuses on software and information technology services, with a registered capital of approximately 42.23 million RMB. The company has invested in 10 enterprises, holds 135 trademarks, and has 1,495 patents, along with 4 administrative licenses [1]

Core Insights - The National Venture Capital Guiding Fund, launched at the end of 2025, has a 20-year duration and aims to support hard technology sectors without regional investment return requirements, marking a significant shift in China's venture capital landscape [1][2]. Group 1: Fund Characteristics - The guiding fund is designed to leverage a substantial amount of fiscal resources, aiming to mobilize trillions in social capital, focusing on strategic emerging industries and early-stage innovative small and medium enterprises [2]. - The fund's long duration and flexible return requirements are seen as beneficial for fostering a unified national market and aligning with the developmental needs of technology enterprises [2][5]. Group 2: Investment Trends - There is a noticeable trend of state-owned capital funds increasing their investments in hard technology, with significant investments in semiconductor and AI sectors, indicating a long-term commitment to strengthening key industrial chains [2][6]. - The guiding fund's approach emphasizes early and small investments, which aligns with the growth trajectories of innovative companies, as evidenced by the involvement of state-owned funds in various successful startups [1][6]. Group 3: Changes in Investment Strategy - The investment strategy has evolved to allow longer fund durations and more flexible return mechanisms, moving away from rigid return requirements that previously distorted investment decisions [4][5]. - A focus on a more patient capital approach is emerging, with an emphasis on supporting the entire lifecycle of investments rather than prioritizing quick returns [4][6]. Group 4: Sector-Specific Investments - The state-owned venture capital "national team" is actively investing in critical sectors such as semiconductors, advanced manufacturing, artificial intelligence, and new materials, reflecting a strategic focus on enhancing domestic capabilities [6][7]. - Investments in aerospace and robotics are also highlighted, showcasing the diverse interests of the national team in fostering innovation across various high-tech fields [7].

Core Insights - The National Venture Capital Guiding Fund, launched at the end of 2025, has a 20-year duration and aims to support hard technology sectors without regional reinvestment requirements, marking a significant shift in China's venture capital landscape [1][2] Group 1: Fund Characteristics - The fund emphasizes a focus on hard technology and strategic emerging industries, primarily targeting seed, early, and mid-stage innovative small and micro enterprises [2] - The fund's long duration and flexible reinvestment requirements are designed to align with the growth cycles of hard technology companies, promoting a more market-oriented investment approach [3][4] Group 2: Investment Trends - State-owned capital is increasingly investing in hard technology, with significant contributions from central enterprises and local government funds, such as the 18.32 billion USD investment from the National Integrated Circuit Fund into SMIC [2][3] - Local government funds are also expanding, with Beijing establishing a 100 billion yuan fund focused on AI and robotics, and Chengdu launching a fund covering similar future industries [2] Group 3: Investment Strategy Evolution - The investment strategy has shifted towards longer fund durations and more flexible reinvestment mechanisms, with 53% of new guiding funds allowing for durations over 10 years [3][4] - The focus has moved from short-term gains to respecting industry growth patterns, reducing the pressure for forced exits during market downturns [3][4] Group 4: Key Sectors and Companies - The state venture capital "national team" is actively investing in critical sectors such as semiconductors, advanced manufacturing, AI, and new materials, with notable investments in companies like Moer Technology and Longxin Technology [1][6] - In the semiconductor sector, investments have been made in key areas such as core equipment and materials, while in AI, the focus includes hardware and software solutions [6][7] Group 5: Future Industry Focus - The national team is also targeting emerging fields like embodied intelligence and robotics, with investments in companies that facilitate faster technological validation and industrial scaling [7] - The collaboration between the guiding fund and various capital sources is expected to accelerate the formation of a mature patient capital ecosystem, driving China's high-quality economic development [7]

Core Insights - MiniMax, a company founded just four years ago, successfully went public on the Hong Kong Stock Exchange, with its stock price surging by 109.09% on the first day, resulting in a market capitalization of HKD 105.4 billion [2] - The company has gained significant investor interest due to its C-end AI applications and over 70% of its revenue coming from overseas, highlighting a narrative of globalization and high growth [2] - Founder Yan Junjie, who transitioned to AGI entrepreneurship in 2022, has led the company from inception to IPO in a remarkably short time, showcasing impressive growth [2] Investment Background - MiniMax's early shareholder roster prominently features investors from Shanghai Jiao Tong University, with notable figures like Liu Wei, co-founder of miHoYo, who has been involved since the company's inception [3] - Liu Wei's relationship with Yan Junjie dates back years, and miHoYo's investment strategy aimed to leverage MiniMax's capabilities to enhance its own game AI development [3] - Other early investors included Hillhouse Capital, IDG, and Ming Shi, with three out of five angel investors being affiliated with Shanghai Jiao Tong University [4] Funding and Support - Cloud Capital's partner Chen Yu was one of Yan Junjie's earliest investor acquaintances, recognizing the potential of foundational models before the AI boom [4] - Ming Shi Venture Capital was one of the earliest investors in MiniMax, participating in six consecutive funding rounds leading up to the IPO [5] - Red Dot Capital, led by alumni from Shanghai Jiao Tong University, has also significantly invested in MiniMax, with their involvement marking one of the largest investments in the large model sector [6] Strategic Partnerships - The company has attracted strategic investments from state-owned funds, enhancing its growth trajectory and providing essential resources for expansion [7] - Investors from Shanghai Jiao Tong University have combined their technical expertise with strategic insights from state-owned funds to support MiniMax's development [7] - The IPO of MiniMax reflects the investment logic of Jiao Tong alumni, emphasizing long-term technological innovation over short-term gains [8] Industry Impact - The success of MiniMax is part of a broader trend in China's hard tech sector, with several AI companies going public and achieving significant valuations [8] - Jiao Tong alumni have played a crucial role in this wave of listings, demonstrating a strong focus on technology-driven investments in sectors like AI and semiconductors [8] - The Jiao Tong alumni network has effectively integrated capital with technological advancements, contributing to the growth of China's hard tech industry [11]

正帆科技回复称，正帆科技深耕集成电路行业，主要为芯片产业链中的晶圆制造商及核心工艺设备厂 商，提供关键装备、核心零组件及材料，以及专业服务。正帆科技对长鑫存储有少量间接投资，是长鑫 存储的长期合作伙伴和重要供应商。 格隆汇1月15日｜有投资者在互动平台向正帆科技提问：正帆科技是否给长江存储、长鑫科技、沐曦集 成、燧原科技和壁仞科技供应生产存储芯片和GPU的配套材料？正帆科技是否间接投资了长鑫科技？ ...