Core Insights - NVIDIA has announced the full production of the Vera Rubin platform, which weighs nearly 2 tons and integrates six new chips, significantly enhancing inference cost and training efficiency, achieving AI computations at a trillion operations per second, marking it as a true AI factory [2] - The company has also open-sourced its first inference VLA (Vision-Language-Action) model, Alpamayo 1, designed for vehicles to "think" and solve problems in unexpected situations, utilizing a 10 billion parameter architecture [3] - The first vehicles equipped with NVIDIA technology are set to hit the roads in the US in Q1, Europe in Q2, and Asia later in the year [4][17] Production and Performance - The new Rubin platform has a performance increase of 5 times compared to the previous Blackwell version, with training performance being 3.5 times better [5][8] - The Rubin platform can reduce inference token generation costs by up to 10 times and decrease the number of GPUs required for training mixture of experts models by 4 times [8] - The Vera CPU in the Rubin platform features 88 cores, providing double the performance of its predecessor, and is designed for agent inference, making it the most energy-efficient processor in large-scale AI factories [8] Ecosystem and Deployment - Major cloud providers, including Microsoft, are expected to be among the first to deploy the new hardware in the second half of the year, with Microsoft’s next-generation Fairwater AI super factory set to utilize NVIDIA's Vera Rubin NVL72 systems [6] - NVIDIA maintains a long-term bullish outlook, predicting the total market size could reach several trillion dollars despite concerns about increasing competition and sustainability of AI spending [7] Innovations and Technologies - The Rubin platform incorporates five innovative technologies, including the sixth-generation NVLink interconnect technology and a third-generation transformer engine, which provides 50 petaflops of NVFP4 computing power for AI inference [9] - The platform's modular design allows for faster assembly and maintenance, with an 18 times quicker process compared to Blackwell [9] Open Source Initiatives - NVIDIA has released the Alpamayo model as part of a complete open ecosystem for autonomous driving development, which includes simulation frameworks and datasets [14][21] - The Alpamayo model is designed for the autonomous driving research community, allowing developers to adapt it for vehicle development and as a foundational tool for autonomous driving technology [15][18] - NVIDIA has also launched various open-source models and tools across different sectors, including the Nemotron family for agent AI and the Cosmos platform for physical AI [26][27]

A股市场存储芯片股表现强势，其中，盈新发展、雷科防务涨停，国科微、普冉股份涨超8%，精测电 子、雅创电子、中微公司、雅克科技、恒烁股份涨超7%，长川科技、石英股份、兆易创新、安集科 技、澜起科技涨超5%。值得注意的是，长川科技、航天智装创历史新高。 | 代码 | 名称 | | 涨幅% ↓ | 总市值 | 年初至今涨幅%。 | | --- | --- | --- | --- | --- | --- | | 300821 | 东岳佳材 | 家 | 11.20 | 161亿 | 14.92 | | 600691 | 潞化科技 | | 10.14 | 72.23亿 | 10.14 | | 600618 | 氯碱化工 | 长 | 10.02 | 152亿 | 10.94 | | 600078 | 澄星股份 | | 10.01 | 74.38 Z | 14.60 | | 002324 | 普利特 | + | 10.01 | 215亿 | 21.01 | | 600955 | 维远股份 | -澳 | 10.00 | 98.61 乙 | 10.68 | | 001231 | 农心科技 | 服 | 9.99 | 30.15亿 ...

功能覆盖方面，DPNPU内置90余种神经网络算子，全面兼容CNN、RNN等主流神经网络架构，并支持 LSTM、GRU等RNN变体，通过RISC-V指令和硬件通用性设计可扩展支持更多新型算子，适配未来AI 模型发展需求。量化技术上，该产品支持训练后量化（PTQ）的四种主流方式，同时兼容INT8和FP16 混合精度量化，在保持模型精度的前提下，大幅降低计算资源与存储空间占用。此外，产品集成自研 PCD（Parameter Compression & Decompression）模块，通过参数压缩与硬件自动解压缩技术，有效缓 解数据吞吐瓶颈，提升系统整体性能。 为降低AI应用开发门槛，国芯科技围绕DPNPU构建了完整的软件生态——C*Core NPU Studio，涵盖工 具套件、驱动和运行时软件三大核心组成。其中，工具套件提供模型转换、预处理、量化、编译、仿真 等端到端模型部署能力；运行时软件包含推理框架软件和扩展软算子库；驱动则适配RISC-V等主流 CPU平台，支持Linux/RTOS/Bare-metal等多种应用环境，可显著缩短AI应用从开发到部署的周期，提升 落地效率。 公告指出，当前人工智能技术正加速 ...

Group 1: Technology Developments - NVIDIA launched a new generation of AI platform Rubin, featuring six new chips, including Rubin GPU with a reasoning power of 50 PFLOPS, five times that of Blackwell [2] - Qualcomm and Google announced an expansion of their decade-long partnership to create end-to-end automotive technology solutions, integrating Qualcomm's Snapdragon digital chassis with Google's automotive software [4] - NVIDIA introduced the Alpamayo series of open-source AI models and tools aimed at enhancing the development of safe and reliable autonomous driving systems [11] Group 2: Corporate Collaborations - Zhiyuan Robotics partnered with MiniMax to enhance voice interaction experiences through customized AI technology [5] - GAC Group signed a comprehensive cooperation framework agreement with Huawei to deepen collaboration on smart vehicle technologies [5] Group 3: Market Movements - Chinese aerospace company China Aerospace Science and Technology Corporation increased its registered capital from 1 billion to 1.396 billion yuan, a 77.6% increase [10] - Defu Technology announced a cooperation agreement with a leading domestic CCL company to supply high-end copper foil products, expected to positively impact its 2026 performance [9] Group 4: Product Launches and Innovations - Raybird Innovation completed over 1 billion yuan in financing and plans to showcase the world's first AR smart glasses with eSIM technology at CES 2026 [15] - ByteDance's Doubao denied rumors regarding the imminent shipment of AI glasses, clarifying that there are currently no sales plans [8]

Core Viewpoint - The report from Guotou Securities suggests that the A-share market is currently in a "non-action" phase, characterized by a lack of clear leading sectors and accelerated industry rotation, but this may build momentum for a "proactive" market in the first half of next year [1] Market Status - The previous "high-cut low" market trend has ended, leading to a phase of confusion with no clear leading sectors. The market's mainline clarity assessment indicator is around 48%, indicating a state of "mainline chaos" [2] - Historically, this "non-action" chaotic state occupies about 40% of the year, typically lasting 3-4 weeks, often corresponding to unclear macroeconomic expectations or significant policy gaps [2] Policy Orientation - The macro policy environment has shifted from pursuing short-term growth to focusing on "internal strengthening," emphasizing both stock and incremental improvements without aggressive stimulus measures [3] - The report highlights positive progress in risk resolution in key areas, which may alleviate constraints and signal a transition towards new growth drivers in the medium to long term [3] Future Outlook - Despite the current "non-action" state, there is potential for a focus on technology and overseas expansion as dual mainlines for the market. The technology sector, particularly the AI industry chain, and overseas expansion are expected to see a rising profit share, surpassing 30% [4] - The report suggests that the market may need to wait for confirmation of global industry trends or domestic economic data to emerge from the current chaos, with funds likely to concentrate on sectors with clear industry trends and solid fundamentals [4] Overseas Variables - The Federal Reserve's internal divisions and the complex external environment have increased uncertainty regarding future policy directions, affecting global capital flows and market risk appetite [4]

科技创新是高质量发展的核心驱动力。人工智能大模型加速迭代，芯片自主研发不断突破，天问二 号叩问深空，人形机器人、无人机走向应用前沿……一系列科技成果，勾勒出新质生产力加快生长的清 晰轨迹。创新正从单点突破走向系统跃升，新质生产力愈发厚积成势。 "十五五"规划建议将"科技自立自强水平大幅提高"列为"十五五"时期经济社会发展主要目标之一。 牵住科技创新这个"牛鼻子"，更多推动生产方式变革、赋能人民美好生活的新质生产力将加速汇聚，为 高质量发展提供澎湃动力。 ——编 者 天问二号VLBI测轨分系统总师黄勇 2026年元旦，我们负责的VLBI（甚长基线干涉测量）测轨分系统已经完成了52次观测任务。当天 问二号飞行在距离地球4300万公里的宇宙空间时，作为VLBI测定轨团队的成员，我时刻牢记"追星"对 我们意味着什么——那是一条必须系在探测器身上，看不见却至关重要的宇宙"风筝线"。 VLBI测轨分系统是中国科学院集全院相关力量构建的国之重器。天问二号的目标是探测小行星和 彗星，小行星采样返回段约2.5年，主带彗星转移段长达7年。轨道复杂、距离遥远，对系统长时间运行 可靠性、网络通信稳定性、资源与人力保障和应急响应处置等 ...

2026 年刚开年，AI 行业最重磅的信号，不是来自模型发布会，而是来自芯片 CEO 的一场产业宣言。 1月5日，拉斯维加斯 CES 现场，英伟达 CEO 黄仁勋一登台就说： 计算行业的每一层，都要被重写一次。 黄仁勋并没有把模型升级作为主角，而是强调：AI的真正跃迁，不靠单点突破，而要靠一整套工业化 能力。 什么是工业化能力？ 不是展示一个更强的Demo，而是让 AI 能被复制、能被部署、能被验收、最终能规模化。 这次发布会，英伟达展示的正是这套完整的工业化体系： 这意味着：AI 应用不再是堆个模型挂在原来的程序上，而是从写代码变成教会一个智能体怎么做事。 过去的应用程序，是提前写好一套流程、预先编译、部署到设备上运行。 硬件层：Rubin平台全面量产，训练速度提升4倍、成本降低10倍 应用层：Physical AI标准路径，从Cosmos模拟到 Alpamayo 自动驾驶，2026年 Q1上路 生态层：全栈开源工具链，从模型到数据到工具，全部向行业开放 而现在的 AI 应用，是实时生成、实时理解、实时回应，连每一帧画面、每一个词，都是现场生成的。 这背后，底层逻辑变了三件事： 黄仁勋说，机器人领域的"Ch ...

澜起科技主营业务包括内存接口芯片、PCIe Retimer芯片、津逮CPU等。2025年前三季度，公司实现营 业总收入40.58亿元，同比增长57.83%；归母净利润16.32亿元，同比增长66.89%。 澜起科技同日公告，公司发行H股并在香港联交所主板上市的计划已取得关键进展。公司已于2025年7 月11日向港交所递交上市申请，并于2025年12月9日获得中国证监会出具的境外发行上市备案通知书。 2025年12月15日，香港联交所上市委员会已举行聆讯并审议通过公司本次发行上市的申请。 中证智能财讯澜起科技（688008）1月6日公告2025年第二次股份回购方案实施进展。截至2025年12月31 日，公司已累计回购A股股份166.20万股，占总股本0.14%，累计支付金额约2.20亿元。公司股票1月5日 收盘价为127.75元/股，上涨8.45%。 该回购方案已于2025年7月获股东大会通过，计划使用自有资金2亿元至4亿元，以集中竞价交易方式回 购股份用于减少注册资本。最初回购价格上限为不超过118.00元/股，后因权益分派调整为117.80元/股。 2025年10月，公司董事会审议通过将回购价格上限大幅上调 ...

英伟达(NVDA.US)宣布推出包括汽车大模型平台Alpamayo在内的一系列人工智能模型和工具，旨在扩 大其技术的应用范围、加速自动驾驶汽车的开发并为新一代机器人提供动力。英伟达公司备受期待的全 新 Rubin 数据中心产品将于今年发布，客户很快就能试用该技术，从而帮助加快人工智能的开发。 首席执行官黄仁勋周一在拉斯维加斯举行的 CES 贸易展上发表讲话时宣布，该公司推出了一款名为 Alpamayo 的车辆平台，该平台可以让汽车在现实世界中进行"推理"。英伟达表示，潜在用户可以自行 获取 Alpamayo 模型并进行重新训练。这项免费服务旨在打造能够自主应对突发情况(例如交通信号灯故 障)的车辆。 车载计算机会分析来自摄像头和其他传感器的输入数据，将其分解成多个步骤，并提出解决方案。英伟 达正与梅赛德斯-奔驰合作，致力于开发能够在高速公路上实现自动驾驶，同时也能在城市中自主导航 的车辆。 黄仁勋表示，首款搭载英伟达芯片的汽车将于第一季度在美国上路。欧洲市场将于第二季度跟进，亚洲 市场则将在下半年推出。他说："我们设想，未来有一天，路上行驶的十亿辆汽车都将是自动驾驶汽 车。" 该公司还推出了用于机器人的AI模 ...

Core Viewpoint - The Chinese stock market has experienced a significant rally, with the Shanghai Composite Index achieving a record twelve consecutive days of gains, surpassing the 4000-point mark for the first time in 33 years since 1993, indicating a strong market sentiment and potential for a cross-year rally [1][27]. Group 1: Market Performance - The total trading volume in the two markets surged to 2.55 trillion yuan, with over 4100 stocks rising, reflecting heightened market enthusiasm and capital inflow [3][27]. - Goldman Sachs has released a report predicting that the Chinese stock market will rise by 15%-20% annually in 2026 and 2027, recommending an overweight position in Chinese stocks [3][27]. Group 2: ETF Performance - Hong Kong ETFs have seen substantial gains, particularly in the innovative drug, internet, and chip sectors, which have been the main drivers of the market's upward movement [4][28]. - The top-performing Hong Kong ETFs include the Hong Kong Innovative Drug ETF (520880) with a 5.4% increase, the Hong Kong Internet ETF (513770) with a 4.4% increase, and the Hong Kong Information Technology ETF (159131) with a 3.6% increase [4][29]. Group 3: Sector Highlights - The innovative drug sector is boosted by the implementation of a new national medical insurance drug list and a commercial health insurance list for innovative drugs starting January 1, 2026, which is expected to enhance industry confidence [6][30]. - In the internet sector, Baidu's Kunlun Chip has submitted a listing application to the Hong Kong Stock Exchange, raising expectations for its parent company's asset value [6][30]. - The chip sector is experiencing a rally due to a significant increase in global memory chip prices and the acceptance of Changxin Storage's IPO application on the Sci-Tech Innovation Board, with leading stocks like Hua Hong Semiconductor and SMIC showing strong performance [6][31]. Group 4: ETF Trends in 2025 - The top-performing ETFs in 2025 include the Entrepreneurial Board Artificial Intelligence ETF (159363) with a 105% increase, and the Nonferrous Metals ETF (159876) with over a 92% increase, indicating strong growth in technology and materials sectors [7][32]. - Other notable sectors include technology, electronics, and chemicals, which have also shown impressive performance, attracting significant investor interest [7][33]. Group 5: Fund Inflows - The top three ETFs by net inflow are the Broker ETF (512000) with 159.3 million yuan, the Hong Kong Internet ETF (513770) with 91.7 million yuan, and the Financial Technology ETF (159851) with 49.7 million yuan, highlighting investor preference for these sectors [9][35].