Investment Rating - The report maintains a positive outlook on the computer industry, recommending stocks such as Baidu, Huawei, and others, with an "Outperform" rating for the recommended targets [5][10][6]. Core Insights - Baidu and Huawei's open-source initiatives are expected to significantly promote the adoption of foundation models, lowering the barriers for R&D and application in AI technology [11][12]. - The State Administration for Market Regulation has approved 7 national standards aimed at standardizing AI and digital services, which will enhance competitiveness and market opportunities for companies in the computer industry [12][13]. - The China Internet Association has initiated standards for algorithm management, which will facilitate the governance of recommendation algorithms and promote compliance within internet platform enterprises [13]. Summary by Sections Open-Source Initiatives - On June 30, Baidu and Huawei announced the open-sourcing of various models, including Baidu's Wenxin 4.5 series and Huawei's Pangu models, which will lower the R&D and application thresholds for AI technologies [11][10]. National Standards - The report highlights the approval of 7 national standards by the State Administration for Market Regulation, which will support the digital economy and enhance the interconnectivity of the computer industry [12][10]. Algorithm Management Standards - The initiation of algorithm management standards by the China Internet Association is expected to advance platform algorithm governance, focusing on compliance and security measures [13][10].

Core Viewpoint - The one-machine market is facing challenges with a lack of new customers and returning clients, leading to a shift in focus towards mid-to-low-end models and a more cautious market approach [2][3][9]. Group 1: Market Dynamics - A recent tender from a state-owned enterprise in Guangdong highlights the unclear demand for one-machine products, with only a few repeat orders from existing clients [2]. - The initial hype around one-machine products has diminished significantly within four months, with many manufacturers struggling to achieve sales despite having products listed [2][3]. - The market consensus indicates that the high-end model market is largely saturated, and future competition will focus on mid-to-low-end models [3]. Group 2: Sales Performance - In the first half of the year, leading hardware manufacturers experienced significant sales, with some reporting nearly 10 billion in revenue and monthly sales of over a thousand units [5]. - The price of one-machine products has increased, with some models seeing price hikes of around 10% within weeks, reflecting a supply-demand imbalance [5][6]. - Companies like Inspur reported a 165.31% year-on-year increase in net profit, driven by strong server sales [5]. Group 3: Customer Relationships - The success of leading manufacturers is attributed to established customer relationships rather than technological advantages, as many large clients have fixed procurement lists [6][7]. - The deployment of one-machine products is primarily driven by strategic needs, with clients focusing on hardware rather than software solutions [7][8]. Group 4: Future Outlook - The second half of the year is expected to see a shift towards a "guerrilla warfare" approach in the mid-to-low-end market, with companies needing to adapt to changing customer budgets and requirements [9][10]. - The demand for high-end models has decreased, with clients now more interested in cost-effective solutions and specific business needs [11]. - The market for one-machine products is becoming increasingly competitive, with new entrants facing challenges in gaining traction [13]. Group 5: Implementation Challenges - Many existing one-machine deployments are not being fully utilized, with clients struggling to see a return on investment [15][16]. - The gap between model capabilities and actual application needs is causing disillusionment among users, leading to a reassessment of the value of one-machine products [16][20]. - Successful implementation often requires a combination of different models and a deep understanding of client data, which many current deployments lack [19][21].

GLM-4.1V-9B-Thinking标志着GLM系列视觉模型实现从感知走向认知的关键跃迁。 本文为IPO早知道原创 作者｜ Stone Jin 微信公众号｜ipozaozhidao 学科解题：支持对数学、物理、生物、化学等学科问题的看图解题，通过推理给出详细的思考过程； 据 IPO 早 知 道 消 息 ， GLM-4.1V-9B-Thinking 凭 借 9B 的 模 型 尺 寸 ， 日 前 成 功 登 顶 HuggingFace Trending第一 。 | Hugging Face | Q Search models, datasets, users ... | | | | | --- | --- | --- | --- | --- | | Models 1,851,126 | Filter by name | Full-text search | Add filters | 1J Sort: Trending | | THUDM/GLM-4.1V-9B-Thinking | | | | | | 15. Image-Text-to-Text . . . : 10B = Updated about 15 ...

当前，驱动科学研究的人工智能（AI for Science，AI4S）在单点取得了可观的进展，实现了工具层面的革新，然而要成为 「 革命的工具 」 ，需要采用 「 通专融 合 AGI 」 方式。大模型的突破性能力逐步改变科学研究的模式，而大模型在科学领域的深度应用亟需科学的评测支撑。 现有科学评测面临着两大痛点：现有测试多聚焦知识记忆，而真实科研需要从原始科学数据感知到复杂推理的全链条能力；天文、地球、生命和材料等领域存在 大量未开发的多模态数据分析需求。 为此，上海人工智能实验室 AI4S 团队推出了 Scientists' First Exam（以下简称 SFE）—— 系统评估多模态大模型（MLLMs）多学科、高难度的科学专业领域认 知能力的评测基准 。 SFE 首创 「 信号感知 - 属性理解 - 对比推理 」 三级评估体系，涵盖 五大科学领域的 66 项高价值任务 ，采用原始科学数据和中英双语问答形式。测试表明，尽管 主流模型在传统基准表现优异，但在 SFE 高阶科学任务上仍面临显著挑战（ SOTA 大模型综合得分仅为 30 左右 ）。SFE 通过系统全面地评测大模型在科学任务 上的能力短板，为科学 ...

Group 1 - The forum focused on national strategic needs and technological frontiers in the fields of pattern recognition and artificial intelligence, gathering nearly 20 experts and representatives from renowned universities, research institutes, and leading enterprises in China [1][3] - The event aimed to foster the cultivation of new productive forces and interdisciplinary integration, injecting new momentum into scientific research innovation and the collaborative development of academic journals [1] Group 2 - Various professors presented specialized reports, including topics such as "3D/4D content creation for arbitrary sparse data," "embodied intelligent robots with emotional intelligence," and "visual perception in unmanned systems" [5][7][11] - A roundtable discussion was held, focusing on new trends and challenges in multimodal large models and generative artificial intelligence, addressing the transformation of research paradigms and talent cultivation in the era of large models [15]

佐力药业（300181）公告，公司与浙江大学签署了《浙江大学-佐力药业智赋食药联合研发中心共建协 议》和《技术开发(委托)合同》，计划投入金额共计2000万元。其中，共建协议涉及1000万元，合同涉 及1000万元。合作旨在利用AI技术研发食药同源垂类大模型与智能体，赋能乌灵菌功能食品/日化用品 等配方和工艺。此次合作有利于强化公司技术护城河，丰富产品矩阵，拓宽C端市场空间，提升产品竞 争力，为公司大健康产业发展注入科技驱动力。 ...

Core Viewpoint - The investment in the plush toy brand "Super Vitality Factory" by the reading group signifies a growing interest in the collectible toy market, driven by the success of Bubble Mart and the potential of IP development in this sector [1][3][5]. Group 1: Investment and Market Dynamics - "Super Vitality Factory" has secured a strategic investment from the reading group, acquiring a 10% stake, marking a significant event in the plush toy industry [1][3]. - The investment reflects a broader trend where major companies are rapidly entering the collectible toy market, following the success of Bubble Mart, which has a market capitalization exceeding 300 billion [5][12]. - The collectible toy market is increasingly seen as a new narrative driven by emotional value, with companies exploring various avenues such as AI companions and virtual stars [3][5]. Group 2: IP Development and Brand Strategy - The success of Bubble Mart is closely tied to its ability to create and manage IP, with a reported revenue of 13.038 billion in 2024, a year-on-year increase of 106.92% [5][12]. - The reading group, with its extensive IP resources, is positioned to leverage its existing content to enhance the collectible toy market, potentially creating a closed-loop system where toys activate consumer engagement [13][28]. - The competitive landscape is shifting towards IP creation, with companies like Lehua Entertainment and its WAKUKU brand aiming to replicate the success of Bubble Mart through strategic partnerships and celebrity endorsements [9][25]. Group 3: Future Trends and Challenges - The future of the collectible toy market may not follow a single path but will likely balance emotional resonance and long-term value, with various models emerging, including traditional IP development, AI-driven companionship, and content-based strategies [29][30]. - The challenge remains in effectively translating literary IP into appealing visual symbols for the younger demographic, ensuring that the products resonate with both fans and casual consumers [28][30]. - The evolving landscape suggests that the next successful brand will be one that can deeply connect with users, transcending mere collectibles to become meaningful emotional companions [29][30].

Group 1 - The establishment of "Zhaofu Technology" marks a significant entry into the L4 autonomous driving sector, with a registered capital of 1.288 billion yuan, focusing on technology development, safety applications, and commercialization [2][3] - The three founding shareholders, Ant Group, Haibo, and CATL, have collectively invested over 3 billion yuan in the initial phase, indicating strong financial backing for the venture [3][4] - The collaboration leverages the strengths of each partner: Haibo's experience in shared mobility, Ant Group's advancements in AI and safety, and CATL's expertise in battery technology [4][5] Group 2 - The local government in Minhang District expresses high expectations for the project, aligning it with the region's focus on integrated circuits and software services, which will enhance technological transformation capabilities [5] - The Robotaxi market is projected to be a trillion-dollar opportunity, attracting various players, including Haibo, which aims to create a multi-layered, adaptable Robotaxi operating platform [7][8] - The market for Robotaxi in China is expected to reach a total scale of 47 billion dollars by 2035, with profitability anticipated starting in 2026 [7][8] Group 3 - Haibo's entry into the Robotaxi market is seen as a strategic move to diversify its offerings, building on its established presence in shared mobility [10][11] - The company aims to implement a differentiated competitive strategy, focusing on domestic markets initially while planning for international expansion [11][12] - The successful commercialization of Robotaxi will depend on reducing operational costs and enhancing consumer acceptance, with Haibo's operational expertise providing a competitive edge [11][12]

Core Insights - The article discusses the current state and trends of the large model API market in 2025, highlighting significant growth and shifts in market share among key players [1][2][25]. Token Usage Growth - In Q1 2025, the total token usage for AI models increased nearly fourfold compared to the previous quarter, stabilizing at around 2 trillion tokens per week thereafter [7][25]. - The top models by token usage include Gemini-2.0-Flash, Claude-Sonnet-4, and Gemini-2.5-Flash-Preview-0520, with Gemini-2.0-Flash maintaining a strong position due to its low pricing and high performance [2][7]. Market Share Distribution - Google holds a dominant market share of 43.1%, followed by DeepSeek at 19.6% and Anthropic at 18.4% [8][25]. - OpenAI's models show significant volatility in usage, with GPT-4o-mini experiencing notable fluctuations, particularly in May [8][25]. Segment-Specific Insights - In the programming domain, Claude-Sonnet-4 leads with a 44.5% market share, while Gemini-2.5-Pro follows [12]. - For translation tasks, Gemini-2.0-Flash dominates with a 45.7% share, indicating its widespread integration into translation software [17]. - The role-playing model market is fragmented, with small models collectively holding 26.6% of the share, while DeepSeek leads in this area [21]. API Usage Trends - The most utilized APIs on OpenRouter are primarily for code writing, with Cline and RooCode leading the way [25]. - The overall trend indicates a strong preference for tools that facilitate coding and application development [25]. Competitive Landscape - DeepSeek's V3 model has shown strong user retention and is favored over its predecessor, likely due to faster processing times [25]. - Meta's Llama series is declining in popularity, while Mistral AI has captured approximately 3% of the market, primarily among users interested in fine-tuning open-source models [25]. - X-AI's Grok series is still establishing its market position, and the Qwen series holds a modest 1.6% share, indicating room for growth [25].

宏观日报 | 2025-07-09 关注美国关税对上游商品影响 中观事件总览 生产行业：美关税8月1日施行。 1）日前，国务院办公厅印发《关于健全"高效办成一件事"重点事项常态化推进机 制的意见》，对进一步优化政务服务、提升行政效能，健全常态化推进机制作出部署。《意见》指出，坚持集约高 效、整体规划，加强资产调剂共享，避免低效闲置、铺张浪费，坚决防治和纠正政务服务中的"面子工程"。强化 统筹规划，在确保安全的前提下稳妥有序推进人工智能大模型等新技术在政务服务领域应用。持续完善全国一体 化在线政务服务平台，强化有关业务系统集约整合和互联互通，提高安全防护能力水平。2）当地时间7月8日，美 国总统特朗普在社交媒体平台"真实社交"上发文表示，根据7日发给各国的信函，以及接下来一段时间内将要发送 的信函，关税将于2025年8月1日开始实施，"该日期没有变化，以后也不会改变"。前一天（7日）特朗普曾表示， 将从8月1日起分别对来自日本、韩国等14个国家的进口产品征收25%至40%不等的关税。对此，外交部发言人表示， 在关税问题上，中方的立场是一贯的，也是非常明确的，关税战、贸易战没有赢家，搞保护主义损害各方利益。 服务 ...