Core Insights - The seventh edition of "Company Secrets" is now available for pre-sale, focusing on in-depth financial analysis rather than superficial business advice [1] - The publication aims to uncover the underlying logic of companies' profitability and losses, featuring case studies from various companies [1] Summary by Topics Company Performance - NIO's L90 achieved a monthly sales figure of 15,000 units, attributed to a pricing strategy of 260,000 and platform-based production, marking a significant recovery from 5,000 monthly sales to over 40,000 [1] - Pop Mart's revenue reached 3.4 billion, with profits surpassing global toy giant Mattel, driven by the normalization of IP blockbuster products [1] - Bawang Tea's revenue increased by 167% within a year, supported by a high-end positioning strategy and expansion into lower-tier markets [1] Industry Trends - The publication also addresses challenges faced by Nintendo with the Switch, Hengrui's transformation in centralized procurement, and Cambrian's resurgence in AI computing power [1]

Market Performance - The A-share market experienced a day of volatility, with the Shanghai Composite Index falling by 0.64%, the Shenzhen Component Index down by 1.37%, the ChiNext Index decreasing by 1.96%, and the Sci-Tech Innovation Board Index dropping by 2.66% [1][2]. Sector Performance - The commercial aerospace sector saw a significant decline, with multiple stocks hitting the daily limit down [2][4]. - The computing hardware supply chain faced a downturn, particularly in the server and CPO segments [2]. - Conversely, sectors such as AI applications, innovative pharmaceuticals, medical services, and ultra-high voltage concepts showed strength [2]. Notable Stocks - Oil and gas stocks experienced a rally, with TBEA Co., Ltd. and Zhongyou Co. hitting the daily limit up, and Tongyuan Petroleum rising over 10% [2][3]. - Specific stocks that gained include: - Tongyuan Petroleum: +11.66% to 7.18 - Zhongyou Co.: +10.00% to 9.02 - CNOOC Services: +46.03% to 15.29 [3]. Declining Stocks - The commercial aerospace concept stocks collectively retreated, with companies like China Aerospace Science and Technology and China Satellite Communications hitting the daily limit down [2][4]. - Notable decliners include: - Aerospace Electronics: -10.01% to 28.40 - Putian Technology: -10.01% to 36.78 - Aerospace Long Peak: -10.00% to 26.45 [4]. Trading Volume - The total trading volume in the Shanghai and Shenzhen markets reached 3.65 trillion, setting a new historical high, with over 3,700 stocks declining [4]. Capital Flow - Main capital flows showed net inflows into sectors like medical devices, gaming, and energy metals, while there were net outflows from consumer electronics, aerospace, and telecommunications [6]. - Specific stocks with net inflows include: - TBEA Co., Ltd.: +1.846 billion - Haige Communication: +1.661 billion - Zhangqu Technology: +791 million [6]. - Stocks with significant net outflows include: - Goldwind Technology: -5.043 billion - Aerospace Electronics: -4.378 billion - Bluefocus Communication: -2.976 billion [6]. Institutional Insights - Guotai Junan Securities noted that market fluctuations do not alter the slow bull market trend, suggesting a focus on AI and cyclical opportunities [6]. - Jinyuan Securities indicated that the market is in a consolidation phase, recommending attention to undervalued sectors [6]. - Zhongtai Securities projected that the market may exhibit characteristics of bottom lifting and active main lines in the first quarter and beyond [6].

1月13日消息，由阿里钉钉前副总裁王铭创立的攀峰智能（K2 Lab）宣布完成数千万元的种子轮融资， 云时资本独家投资。 编辑｜瑜见 本轮融资将主要用于产品和AI能力打造、用户增长和AI Native团队的构建，推进面向超级个体的内容 电商Agent基建。首款产品能帮助达人完成选品推荐、脚本生成、多镜头视频制作、智能剪辑等。 记者｜刘晓洁 ...

1月13日，AI应用板块表现活跃，AI应用细分概念——"GEO"（生成式引擎优化）再走强，被市场称为 新"易中天"组合的——易点天下、中文在线、天龙集团均大涨，易点天下一度三连板。 据媒体报道称，马斯克当地时间1月10日在社交媒体平台X发文称，将在一周内正式开源X平台最新的内 容推荐算法，覆盖"所有用于决定向用户推荐自然内容和广告内容的代码"。马斯克表示，"此过程将每 四周重复一次"，同时附带开发者说明，标注算法和逻辑上的改动内容。 截至发稿，天龙集团20%涨停，易点天下、中文在线均涨超10%。浙文互联、引力传媒、利欧股份等涨 停。 编辑丨瑜见 | 代码 | 名称 | 米唱 | 息金额 | 息市值 | 现价 | | --- | --- | --- | --- | --- | --- | | 920021 流金科技 | | +27.23% | 13.22 Z | 37.39 Z | 12.05 | | e88382 | 光云科技 | +20.02% | 16.19 Z | 125.1 乙 | 29.38 | | 301 408 | 华人健康 | +20.00% | 13.06 Z | 110.4 Z | 27. ...

2026.01.13 本文字数：4945，阅读时长大约8分钟 来源 |健闻咨询 乔燕薇 2025年上半年，京津冀地区的某基层医院院长刘钢（化名）拍板，引入了一套业内评测靠前的医疗大模 型系统，希望能够重点解决电子病历生成的效率问题，顺带提供简单的疾病辅助诊断。 在医院内部讨论时，以刘钢为代表的医院管理层对此普遍持乐观态度。 在过去的一两年时间里，开发医疗大模型的主导权始终被牢牢地攥在头部医疗AI公司和大三甲医院的 手中。单纯从技术逻辑来说，大模型具备向下兼容的特性，就好比能解高等数学的人，一定精通加减乘 除。 所以，对于既不想错过这波AI技术福利，又拿不出更多经费和资源的基层医疗机构而言，择优引入已 经在头部三甲医院"跑通"的大模型产品，是一种隐性共识。 但项目的进展却并不顺利。产品落地一个多月后，刘钢发现，无论是在电子病历生成环节，还是辅助诊 断层面，模型都没有达到预期效果，甚至还有点"负作用"。 比如，这套模型似乎并不能识别当地老百姓的"方言"，导致生成的病历文书错乱百出，医生还得花更多 时间来手工校正；在给出辅助诊断时，由于基层医疗机构的检查检验数据没有完全打通，模型给出的结 果也远不如在头部医院里运行 ...

2026.01.13 本文字数：1213，阅读时长大约2分钟 作者 |第一财经 林志吟 1月13日上午，A股以及港股的医药股集体出现走强趋势。 截至上午10点45分之际，A股中的生物医药板块有四成多的个股出现上涨，包括康乐卫士 （920575.BJ）、泓博医药（301230.SZ）、荣昌生物（688331.SH）等个股涨幅超过15%；港股主板的 医疗保健板块，也有一半以上的个股出现上涨，药明巨诺（02126.HK）、基石药业（02616.HK）、瑞 科生物（02179.HK）、荣昌生物（09995.HK）等个股涨幅超过8%。 两地医药股活跃背后，一年一度的摩根大通全球医疗健康大会正在美国旧金山召开，该会议聚焦生物技 术、生物制药等六大领域。作为全球规模最大、信息最密集的医疗投融资与合作盛会，可以洞察全球医 药生物产业发展的最新动态与最新预期。 自去年开始，中国创新药逐渐成为全球生物制药行业的中坚力量，来自中国的声音也是大会的焦点之 一，多家中国药企也将在大会上进行演讲。 从以往看，该大会召开期间，全球创新药领域的并购、对外授权动作表现活跃，同样，在本次大会，市 场也在期待今年达成的新交易。 当前，已有重磅对外 ...

Core Insights - DeepSeek has released a new paper focusing on the conditional memory module of large models, suggesting it will be a core modeling primitive in the next generation of sparse large models [2][5][7] Group 1: Research and Development - The new paper, co-authored with Peking University, is titled "Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models" [5] - The research identifies two distinct tasks within large models: deep dynamic computation for combinatorial reasoning and static knowledge retrieval, highlighting inefficiencies in the current Transformer architecture [5][6] - DeepSeek introduces conditional memory as a supplementary sparse dimension to optimize the balance between neural computation (MoE) and static memory (Engram) [6][7] Group 2: Performance and Implications - The team discovered a U-shaped scaling law indicating that the mixed sparse capacity allocation between MoE experts and Engram memory significantly outperforms pure MoE baseline models [6] - The introduction of the memory module not only aids knowledge retrieval but also shows significant improvements in general reasoning, coding, and mathematical tasks [6][7] - The paper essentially proposes a "division of labor" optimization for large models, allowing specialized modules to handle specific tasks more efficiently [6][7] Group 3: Future Developments - Industry speculation suggests that the proposed conditional memory may be part of the technical architecture for DeepSeek's upcoming flagship model, DeepSeek V4, expected to be released around February [7] - Initial tests indicate that V4 may surpass other leading models in programming capabilities, with the previous V3 model having already outperformed OpenAI's GPT-5 and Google's Gemini 3.0 Pro in various benchmarks [7]

Core Insights - DeepSeek has released a new paper focusing on the conditional memory module of large models, suggesting it will be a core modeling primitive in the next generation of sparse large models [1][4]. Group 1: Research Findings - The new paper, co-authored with Peking University, is titled "Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models" and highlights the need for a native knowledge retrieval mechanism in existing Transformer architectures [4]. - The research identifies two distinct tasks in large models: deep dynamic computation for combinatorial reasoning and static knowledge retrieval, indicating that current models inefficiently simulate retrieval processes [4][5]. - DeepSeek introduces conditional memory as a supplementary dimension of sparsity, optimizing the trade-off between mixture of experts (MoE) and static memory (Engram) [4][6]. Group 2: Performance Improvements - The team discovered a U-shaped scaling law, showing that the mixed sparse capacity allocation between MoE experts and Engram memory significantly outperforms pure MoE baseline models [5]. - The introduction of the memory module not only aids knowledge retrieval but also yields notable improvements in general reasoning, coding, and mathematical tasks [5][6]. - The paper essentially proposes a "division of labor" optimization for large models, allowing specialized modules to handle specific tasks, thereby enhancing efficiency and resource allocation [6]. Group 3: Future Developments - Industry speculation suggests that the proposed conditional memory may be integral to the architecture of DeepSeek's upcoming flagship model, DeepSeek V4, expected to be released around February [6]. - Initial tests indicate that V4 may surpass other leading models in programming capabilities, with the previous model, V3, having already outperformed OpenAI's GPT-5 and Google's Gemini 3.0 Pro in various benchmarks [6].

2026.01.13 使用人工智能模型来设计和发现新药物，是目前全球制药巨头加速布局的方向，此举目标是缩短新药研 发上市所需的时间。 本文字数：1275，阅读时长大约2分钟 作者 |第一财经 钱童心 当地时间1月12日，摩根大通医疗健康大会（JPM Conference）在旧金山正式召开。大会首日，全球市 值最大的科技公司英伟达与全球市值最大的制药公司礼来宣布，将在五年内斥资10亿美元在旧金山湾区 建立一个新的联合研究实验室，以加速AI药物研发进程。 截至发稿，英伟达市值为4.5万亿美元，位居全球市值最大公司榜首；礼来市值稳定在1万亿美元上方， 牢牢占据全球市值最大制药企业的地位。 据介绍，该实验室将使用英伟达最新一代AI芯片Vera Rubin。就在几个月前，礼来曾表示，正使用1000 多个英伟达当前一代AI芯片Grace Blackwell构建一台超级计算机。这也有望成为全球最强大的AI计算机 之一。 微信编辑| 七三 第一财经持续追踪财经热点。若您掌握公司动态、行业趋势、金融事件等有价值的线索，欢迎提供。专 用邮箱：bianjibu@yicai.com （注：我们会对线索进行核实。您的隐私将严格保密。） ...

2026.01.13 本文字数：518，阅读时长大约1分钟 作者 |第一财经 栾立 第一财经持续追踪财经热点。若您掌握公司动态、行业趋势、金融事件等有价值的线索，欢迎提供。专 用邮箱：bianjibu@yicai.com （注：我们会对线索进行核实。您的隐私将严格保密。） i茅台"秒空"仍在继续，今日（1月13日）标价4199元的53%vol 500ml陈年贵州茅台酒（15）首日在i茅台 app上线也告秒空。 官方消息显示，2026年1月13日起，陈年贵州茅台酒（15）也正式登录i茅台APP抢购，陈年贵州茅台酒 （15）也是茅台的"塔尖"产品。记者注意到，9点09分刚到，陈年贵州茅台酒（15）已经被抢购一空。 据酒商介绍，陈年贵州茅台酒（15）平时多用于礼赠，需求并不是特别多。其也属于2025年市场价格倒 挂比较严重的产品，近期批发价一般在3900元/瓶左右。尚不清楚i茅台每日的陈年贵州茅台酒（15）供 给量是多少，因此并不好做判断，而且也不排除有个人投机者在其中操作。 根据独立白酒评论员肖竹青介绍，近期茅台刚刚调整过陈年贵州茅台酒（15）的价格，出厂价从5399 元/瓶下调至3409元/瓶，零售价从5999 ...