DeepSeek 或将加快军事 AI 规模化部署 20250210 摘要 Q&A 大模型 AI 技术的进步对我国军事 AI 规模化部署有哪些影响？ 大模型 AI 技术的进步对我国军事 AI 规模化部署具有深远影响。首先， DeepSeek 等大模型的推出及其低成本化和开源模式，为军工应用场景落地提供 了可能性，并引领作战模式向无人化、智能化方向演变。传统军事任务执行模 式 OODA（观察、判断、决策、行动）在人工智能技术的发展下，已经在观察和 判断环节引入了大量数据分析、图像识别和机器学习等先进技术。然而，自主 决策和行动环节仍依赖人力，特别是在复杂战场环境中。 随着大模型时代到来， 人工智能技术在真实战场中的应用逐渐增加。例如，有人机与无人机协同作战 中，希望未来无人机能够实现完全自主决策，而不仅仅是信息接收与传递。这 对军事 AI 提出更高要求，包括模型训练中的数据获取、安全性、鲁棒性以及决 • 人工智能技术正重塑军事任务执行模式，通过在 OODA（观察、判断、决策、 行动）循环中引入数据分析、图像识别和机器学习，提升决策效率，但自 主决策和行动环节仍依赖人力。 • 大模型 AI 技术在军事领域的应用日益 ...

Summary of DeepSeek Conference Call Company and Industry Overview - **Company**: DeepSeek - **Industry**: AI and Technology Key Points and Arguments 1. **Recent Advances in AI**: DeepSeek has made significant progress in AI, particularly in pre-training and post-training models, with the launch of the 136 model and RE model achieving breakthroughs in less than a month [2][4] 2. **Cost-Effective Model Training**: DeepSeek trained a base model comparable to GPT-3.5 using only $5 million and 2,400 H800 GPUs, challenging the high investment model prevalent in North America and prompting Wall Street to reassess high computing power demands [2][4] 3. **Open Source Approach**: The company adopts an open-source model similar to other projects, paving the way for future applications and development by other vendors, which may lead to irrational short-term computing investments but will ultimately promote long-term growth in total computing demand [2][5] 4. **Positive Market Response**: The DCC large language model's V3 version received a positive response in North America, with app downloads surpassing competitors and global traffic reaching one-third of GPT-3's within a week [2][8][9] 5. **Democratization of AI Technology**: The DCC open-source model lowers the barriers for SMEs and individual developers to commercialize AI technology, accelerating the democratization of AI and potentially reducing investor reliance on computing power and chips [2][10] 6. **Innovative Techniques in DPC Model**: The DPC large language model incorporates key technologies from OpenAI, new data labeling methods, and high-quality data cold starts, reducing costs and improving training efficiency [2][12] 7. **DPT V3 Version Innovations**: The DPT V3 version features significant innovations such as MLA, Deep CMOE, and Multi-task Prediction, enhancing training efficiency and reducing memory requirements, although it introduces potential hallucination issues due to multi-token predictions [2][15][18] 8. **Attention from Major Tech Companies**: Major companies like Meta and OpenAI are closely monitoring DPT model innovations, considering resource allocation for future explorations, although their primary goal is to enhance model performance rather than save on GPU costs [2][14][20] 9. **Impact on Financial Markets**: DeepSeek's low-cost, high-efficiency performance raises concerns on Wall Street regarding the necessity of previous large investments, as seen with the Stargate project aiming for $500 billion in funding [4][10] 10. **Future of AI Development**: The trend is shifting towards algorithmic innovation for efficiency rather than solely relying on hardware investments, indicating a sustained growth in overall computing resource demand but with more diverse and intelligent approaches [7][29] Other Important Insights 1. **Research and Development Efficiency**: The DPT team excels in engineering practices, effectively translating exploratory research into practical applications, which is crucial for maintaining efficiency with limited resources [19] 2. **Challenges in Pre-training**: Major companies face challenges in pre-training models due to limited high-quality data sources and stringent data regulations, which contrasts with the more flexible data acquisition strategies of Chinese firms [31][34] 3. **Multi-modal Data Training**: While multi-modal data training presents potential, it also faces challenges in efficiency and compatibility with text-based models, indicating that breakthroughs may be slower compared to pure text models [34] This summary encapsulates the key discussions and insights from the DeepSeek conference call, highlighting the company's innovative approaches and the broader implications for the AI industry.

Summary of Conference Call Notes Industry Overview - The automotive industry is experiencing a significant shift towards intelligence and automation, with expectations for a major upturn in the market by 2025, similar to the valuation increases seen in 2020 [2][15][16]. - The penetration rates for L3 autonomous driving are projected to rise from 10% to between 50% and 80% over the next three years, while electric vehicle penetration is expected to exceed 80% [5][15]. Key Companies and Recommendations - Recommended companies in the intelligent vehicle sector include Xiaopeng, Ideal, BYD, and Xiaomi, with a focus on those with strong intelligent attributes [2][3]. - In the components sector, companies benefiting from increased penetration of intelligent features include Horizon Robotics and Black Sesame [2][3]. Technological Developments - The DeepSeek model is expected to positively impact the optimization of intelligent driving algorithms, although it will not drastically change the competitive landscape in the short term [6][10]. - The VROOM model is anticipated to accelerate applications in vehicles, leading to a decrease in computational power requirements [4][10]. - The adoption of fast-slow system architectures by more automakers, such as Ideal's VTM architecture, is expected to enhance performance and efficiency [11][10]. Market Trends and Predictions - The automotive market is predicted to replicate the comprehensive uptrend seen in 2020, with overall vehicle PS valuations expected to rebound significantly by 2025 [16][18]. - The intelligent components sector is projected to see annual revenue growth rates between 30% and 70%, with leading companies like Desay SV and Huayang benefiting from this trend [29][28]. Investment Insights - The next three to five years are expected to bring significant changes in the automotive intelligence market, with a focus on fundamental performance rather than speculative narratives [15][22]. - The market for robo-taxis is anticipated to commercialize successfully between 2028 and 2030, providing substantial opportunities for growth [15]. Competitive Landscape - Domestic brands such as Huawei, Xiaopeng, and Ideal are positioned in the leading tier of autonomous driving, with expectations for significant improvements in their operational capabilities by 2025 [12][14]. - The competitive dynamics within the automotive sector are expected to intensify, with a potential consolidation of market share among the top players [21]. Component Industry Insights - The average selling price (ASP) of chips is rising, driven by increased demand for advanced technologies like lidar, which is crucial for the intelligent driving sector [26]. - Companies like Horizon Robotics and Black Sesame are highlighted as key players in the chip market, with expectations for substantial revenue growth by 2025 [30]. Conclusion - The automotive industry is on the brink of a transformative phase driven by intelligence and electrification, with significant investment opportunities emerging in both vehicle manufacturers and component suppliers. The focus on technological advancements and market dynamics will be critical for stakeholders in the coming years.

（原标题：DeepSeek入局搜索混战，10亿网民流量再分配开始了 | 海斌访谈） 成百上千亿的资金从搜索端口聚了又散。DeepSeek应用以史上最快速度达成3000万日活。 据调研机构QuestMobile，1月28日，DeepSeek日活跃用户数首次超越豆包，随后在2月1日突破3000万大 关。目前DeepSeek、字节旗下的豆包、月之暗面的Kimi智能助手以及百度的文小言等，是中国AIGC市 场的头部应用。对于它们，AI搜索功能是标准配置。 中国有10亿网民，成百上千亿的资金从搜索端口聚了又散。人工智能企业争先恐后入局搜索，参与这场 流量和利益再分配的游戏。DeepSeek骤然崛起，一切都只是开始。 人人做搜索 2月9日，第一财经记者分别向DeepSeek应用、豆包和Kimi智能助手检索了相同问题：蛟龙行动怎么 样？ 蛟龙行动是市场争议的一部春节档电影。DeepSeek、豆包以及Kimi智能助手都给出正反两方面的文字 总结。 DeepSeek并没有为泼天流量做好充足准备。相比豆包和Kimi智能助手，DeepSeek的反应时间更长而吐 字速度更慢。而且，DeepSeek目前还没有提供语音交互，需要文字输入 ...

一方面，节后首周市场成交量持续放大，春季躁动加速明显，抢筹指标和活跃私募仓位处于高位，外资 的流出和部分公募产品的赎回边际改善。强势行业集中在AI产业链方向，TMT板块成交占全A股总成交 比重快速提升至45%；风险偏好已经抬升，事件催化还在继续，预计科技行情还将演绎，但分化不可避 免，从普涨到分化到最后缩圈/轮动的过程也符合产业逻辑和过往规律。另一方面，中美科技股在2022 年底AI产业革命以来出现了明显的估值拉大，并在近期伴随中国技术突破出现变化，差距开始缩小。 从中美重点行业龙头分析来看，发现中国资产估值偏低的行业为互联网、通信硬件、智能车和智能驾 驶；估值偏高的为半导体、软件；估值接近的为消费电子。在配置上，预计以AI产业链为核心的科技 行情还会演绎，当前建议以产业逻辑为核心，关注景气度确定性高的端侧AI板块；预计后续低波动风 格将逐步体现出超额收益，建议突出非美出海主题+消费类和垄断类红利的杠铃策略。 校对：王蔚 盘面上看，零售、餐饮、传媒、旅游、食品饮料等板块拉升，地产、医药、建筑、半导体、酿酒等板块 上扬，DeepSeek概念持续活跃，算力、信创、AI应用概念等走强。 中信证券表示，技术变革驱动的 ...

Summary of Conference Call on DeepSeek's Impact on ERP and OA Industry and Company Involved - The discussion centers around the impact of DeepSeek on the Enterprise Resource Planning (ERP) and Office Automation (OA) sectors, highlighting its influence on large model applications in these areas [1][3][5]. Core Points and Arguments 1. **Rapid Growth of DeepSeek**: DeepSeek has seen significant growth, with daily active users surpassing 30 million, making it the leading domestic large model in terms of user engagement [4]. 2. **Cost Reduction in Model Deployment**: DeepSeek's innovations in architecture and software collaboration have led to a substantial decrease in the costs associated with deploying large models, with costs for input and output tokens being notably low [2][4]. 3. **Transformation of Software Applications**: The introduction of DeepSeek is expected to revolutionize application software, particularly in ERP and OA, by enabling more efficient use of AI agents, which can automate many tasks previously handled by humans [3][8]. 4. **Integration with Existing Systems**: DeepSeek's standardized API allows for easy integration with mainstream ERP and OA systems, facilitating low-code development and reducing implementation time for enterprises [7]. 5. **Enhanced User Experience**: The use of AI agents is anticipated to simplify interactions with complex management software, improving user experience through natural language processing and reducing the need for extensive training on software functionalities [9][10]. 6. **Market Opportunities for SaaS Models**: The shift towards AI-driven solutions opens up new market opportunities for ERP and OA vendors, including the potential for flexible pricing models based on AI performance and efficiency improvements [12][13]. 7. **Competitive Landscape**: Companies like Owning and Kingdee are actively developing AI capabilities within their ERP solutions, showcasing the competitive nature of the market and the ongoing advancements in AI applications [14][15]. Other Important but Possibly Overlooked Content 1. **Focus on Accuracy in AI Applications**: There is a strong emphasis on the importance of accuracy in AI applications to avoid errors that could negatively impact business operations [17]. 2. **Future Market Potential**: The potential for AI applications to transform various business processes is highlighted, with a call for continued attention to how AI can be leveraged to meet specific enterprise needs [18]. 3. **Diverse Application Scenarios**: The discussion includes various application scenarios for AI in ERP and OA, such as contract management, financial reporting, and human resources, indicating a broad scope for AI integration [14][16]. This summary encapsulates the key insights from the conference call regarding the transformative impact of DeepSeek on the ERP and OA industries, emphasizing cost efficiency, user experience, and market opportunities.

一直致力在这种AI上面AI应用和Agent的一个创新的应用公司在2023年年初就投入了这一块的研发和一些商品化的一些落地和订单的一些逐步的转化时至今日其实已经过去了两年整的时间了我认为在这个过程中我们确实也走了一些弯路因为毕竟AI是一个新鲜的事物但是在这个过程中我们也取得了一个莫大的成绩 然后完成了一个是智能培训考试系统的一个情况当时接的是DeepSeq的V3未来上线可能会更新到R1这样的一个程度主要是在课程制作上和考试环节两大层面上去实现了DeepSeq的一个融合另外与DeepSeq上我们还做了数据分析 也是和用它的RE模型去做的大概是这两类的一个深度的应用和整合当然其实对于我们这种厂商来讲的这种因为我们自己不做这种通用大模型的训练那对于模型来讲的更多的其实是一个使用和复能的一个过程那这个过程中我们也去适配了多样化的一个厂商就是主要还是与客户的需求以及它这个成本 需求功能应用等等有比较大的一个关联性这就是目前我在弹幕型适配的一个情况我就在与大家之前公布的基础上更新这么多内容然后接下来我就介绍一下我们的今天的一个专家的嘉宾他是田总是我们的首席创新官是协同智能研究院的院长田总在咱们的B端的领域其实是有着非常深厚 ...

【九点特供】合肥助力! 安徽百信完成DeepSeek-V2版本在 产服务器上的模型适配,这家公司刚刚亲承和百信合作共 同推广销售异腾服务器产品;东风、长安两大汽车央企均筹 划童年 九点特供 2025.02.10 08:04 星期一 【市场主线】 上周五市场全天冲高回落,创业板指大涨超2.5%。盘面上,DeepSeek和比亚迪智驾概念股延续强势,光伏板块同样展开反弹, 仅银行等少数板块下跌。指数层面,两市成交额近2万亿目仅1只个股跌停,连续第二天实现了"指数放量大涨+跌停消灭"的目 标，而之所以能够大涨，就好比一个人在黑夜里摸着石头过河，等发现自己已经上岸了，便开始加速奔跑。如图所示，周五上证 冲了稻体上沿后回落，创业板指更是接近收复了年初的跌幅，但现在毕竟不像去年春节后（当时A股是从前一年国庆假期后一路 跌到2月），去年9月底之前也是连续跌了近4个月，因此经过连续两日的放量后，指数大概率会盘整一番，乐观预期下，等下方 的均线跟上后,有望突破箱子C并回到上方的箱体B开启新的区间震荡,关键还是看量能,量能看什么,看消息驱动,比如2月的 海外宏观(本周即将宣布"对等关税"措施）,3月的全国两会,4月的财报季。 题材 ...

Summary of Conference Call Records Industry Overview - The conference call primarily discusses the automotive industry, focusing on advancements in autonomous driving technology and the impact of AI on vehicle performance and market dynamics [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47]. Key Points and Arguments 1. **Market Positioning**: The top companies in the autonomous driving sector include Xiaofeng, Songqi, Sais, Linxiang, Biyading, and Xiaomi, with a focus on companies benefiting from the penetration of smart technology in downstream applications [1]. 2. **Technological Advancements**: The call emphasizes the importance of computational power in both cloud and vehicle environments, with a noted decline in demand for traditional algorithms, leading to the adoption of new models like VRM and DeepSeek [2][3]. 3. **Growth Projections**: Predictions indicate that the penetration rate of L3 autonomous driving will increase from 10% to 50% over the next three years, with electric vehicle penetration expected to rise from 50% to 80% [3]. 4. **DeepSeek's Impact**: DeepSeek is highlighted as a significant player in optimizing driving algorithms, enhancing the performance of autonomous vehicles through improved data utilization and model training [4][5][6][7][8]. 5. **Industry Dynamics**: The call suggests that the automotive industry is entering a growth phase after a prolonged introduction period, with a shift towards more competitive dynamics among leading companies [3][4][5][6][7][8][9][10]. 6. **Investment Opportunities**: The discussion includes insights on potential investment opportunities in companies that are well-positioned to leverage advancements in AI and autonomous driving technologies [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47]. 7. **Performance Metrics**: The call mentions that the response time for inference has decreased by 30% due to the integration of DeepSeek technologies, indicating improved efficiency in vehicle systems [15][16][17][18][19][20]. 8. **Valuation Trends**: The anticipated recovery of industry valuations is discussed, with expectations for a return to higher price-to-earnings ratios as the market recognizes the potential of smart technologies in vehicles [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47]. Additional Important Insights - **Competitive Landscape**: The call notes that the competitive landscape is evolving, with a potential for significant consolidation among automotive companies, particularly as they strive to enhance their autonomous driving capabilities [30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47]. - **Technological Integration**: The integration of AI and machine learning into vehicle systems is expected to enhance user experience and operational efficiency, particularly in mid-range and low-end models [37][38][39][40][41][42][43][44][45][46][47]. - **Future Outlook**: The overall sentiment is optimistic regarding the future of the automotive industry, with expectations for rapid advancements in technology and increased market penetration of electric and autonomous vehicles [22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44][45][46][47].

● 本报记者 杨洁近日，工信部发布2025年春节假期通信业运行情况显示，国内移动数据流量实现较快 增长，国际出入境漫游业务表现亮眼，AI赋能成为创新服务的重要方向。 中国电信借助数字人高效交互能力，为客户带来便捷的全新体验。 运用AI、5G技术助力企业数字化能力提升，春节期间提供百亿次高精定位服务调用，助力车企提供智 驾车辆春运导航；提供春节语音质检服务，提升金融企业对疑似涉诈等的检测能力；打造云电脑，支持 企业随时随地高效办公。 运用AI、大数据技术助力政府治理能力提升，中国联通开展春运、景区客流的实时监测与动态研判分 析，对重点设施等进行全天候智能监控，对春节假期消费情况和节后复工复产开展分析，深度赋能2025 年春运、文旅、安全、经济等多项工作。 数据显示，2025年春节期间（1月28日0时至2月4日24时共8天）累计移动互联网用户接入流量达660.3万 TB，按可比口径较2024年春节增长9.9%。 三家基础电信企业力推最新AI技术普惠应用，均全面接入DeepSeek开源大模型，实现在多场景、多产 品中应用。 视频类App持续活跃工信部数据显示，2025年春节期间，5G移动互联网用户接入流量保持快速增 ...