"我们看了那篇论文，心想：'天啊，怎么我们落后了？我们之前也在做这些事情啊，'结果就是，我们的工程主管决定 说：'取消春节假期，所有人都留在公司，加班睡办公室，我们要加速开发进度。'几周之内，我们就推出了自己的版本， 也就是Qwen系列模型。它非常具有竞争力"。 硬·AI 作者 | 赵雨荷 编辑 | 硬 AI 媒体报道，DeepSeek今年1月推出的低成本、功能强大的人工智能模型震惊了全球科技行业后，阿里巴巴 集团的工程师们取消了假期，在春节假期期间继续工作，奋起直追，连夜加班赶进度。 阿里巴巴董事会主席蔡崇信（Joe Tsai）周三在巴黎举行的VivaTech科技大会上表示，中国充满活力的消 费互联网环境，加上本地工程师之间激烈的竞争文化，正不断推动中国在AI领域的创新。 他透露，阿里巴巴是在DeepSeek发布R1模型之后，才意识到自己在AI领域已经落后了。他举例说明了中 国科技行业竞争的激烈程度： 蔡崇信还说，开源大型语言模型（LLMs）将带动人工智能应用的激增，并推动云计算需求的增长。目 前，阿里巴巴正经历"一个巨大考验的时期"之后，进一步聚焦自身庞大业务帝国的重心。 他说，阿里巴巴选择将Qwen系列 ...

甲骨文周三盘后公布财报显示，该公司第四财季业绩超预期，虽然云基建略微逊于预期，但管理层预计2026财年云基础 设施营收预计将增长超过70%，同时资本支出继上年猛增三倍后，新财年将继续增至250亿美元。甲骨文股价盘后一度跌 约3%，随后一度转涨8%。 硬·AI 作者 | 赵雨荷 编辑 | 硬 AI 甲骨文（Oracle）周三盘后公布财报显示，该公司第四财季业绩超预期，虽然云基建略微逊于预期，但管 理层预计2026财年云基础设施营收预计将增长超过70%，同时资本支出继上年猛增三倍后，新财年将继续 增至250亿美元。甲骨文股价盘后一度跌约3%，随后一度转涨8%。 以下是甲骨文第四财季财报要点： 主要财务数据： 第四财季营收： 甲骨文第四财季营收159.0亿美元，同比增长11%，分析师预期155.9亿美元。 第四财季运营利润： 甲骨文第四财季运营利润70.4亿美元，分析师预期68.5亿美元。 第四财季整体云收入： 甲骨文第四财季整体云收入（IaaS加上SaaS）同比增长27%，达到67亿美 元，占总营收的42% 首席执行官Safra Catz在分析师电话会议上表示，2026财年云基础设施营收预计将增长超过70%，相比 ...

Group 1 - OpenAI plans to utilize Google's cloud services to meet its increasing computing power demands, marking a strategic shift in its resource acquisition approach [1] - The collaboration with Google comes after a long-standing partnership with Microsoft, which was OpenAI's exclusive cloud service provider until January of this year [1] - This partnership is seen as a response to the exponential growth in computing power needs driven by advancements in AI technology, allowing OpenAI to diversify its computing sources [1] Group 2 - Morgan Stanley analyst Brian Nowak highlighted the significance of the partnership, suggesting it could accelerate growth for Google Cloud, which is currently undervalued at approximately 18 times the expected earnings for 2026 [2] - The collaboration signals Google's confidence in its long-term search position, indicating that if Google were genuinely concerned about threats to its search business, it would not provide critical computing resources to a potential competitor [2] - The report emphasizes that Google possesses a strong understanding of its business health, user base, data assets, and innovation pace, reflecting its confidence in the future [2]

Core Viewpoint - The Hong Kong stock technology sector has regained momentum after a significant pullback in April, driven by advancements in AI and cloud computing, policy support for Chinese companies returning to Hong Kong, and expectations of interest rate cuts by the Federal Reserve [1] Group 1: Index Comparison - The Hang Seng Technology Index consists of 30 selected Hong Kong-listed companies highly related to technology, with a weight limit of 8% for non-foreign companies and 4% for foreign companies, adjusted quarterly [2] - The Hang Seng Hong Kong Stock Connect Technology Theme Index reflects the performance of 30 Hong Kong-listed companies related to technology that can be traded through Stock Connect, with a weight limit of 10% per stock and adjusted semi-annually [2] Group 2: Industry Distribution - The Hang Seng Technology Index includes approximately 17% automotive weight and about 10% in tourism, home appliances, and pharmaceutical stocks, while the Hang Seng Hong Kong Stock Connect Technology Theme Index excludes these consumer sectors, focusing more on software services, information technology equipment, and semiconductors [4] Group 3: Concentration of Weighting Stocks - As of June 9, the top ten stocks in the Hang Seng Hong Kong Stock Connect Technology Theme Index accounted for 76% of its total weight, higher than the 71% concentration in the Hang Seng Technology Index, indicating greater potential for returns but also higher volatility risk [6][10] Group 4: QDII Quota Utilization - The Hang Seng Technology Index includes non-Stock Connect stocks, which may face quota restrictions when investing through QDII channels, while all stocks in the Hang Seng Hong Kong Stock Connect Technology Theme Index can be traded through Stock Connect, enhancing convenience [9] Group 5: Historical Performance - Over the past year, the Hang Seng Technology Index experienced a return of 44%, while the Hang Seng Hong Kong Stock Connect Technology Theme Index, due to its higher concentration, achieved a return of 51.4%, with both indices exhibiting high volatility around 40% [10]

点击蓝字，关注我们 以云计算服务和数据库软件闻名全球的科技巨头 甲骨文(ORCL.US) 在周四美股收盘后公布业绩报 告，该公司预计， 本月开启的2026财年，在无比强劲的云端AI算力资源需求以及爆发式增长的实体 AI算力基础设施需求推动下，其云基础设施销售额将一举跃升逾70% ，这一展望大幅推动了投资者们 对于甲骨文备受关注云计算业务的乐观情绪，令投资者们愈发坚信在AI推动之下，甲骨文在云计算领 域的份额终将比肩亚马逊AWS与微软Azure。 最新业绩与展望指引凸显出长期以来以数据库软件闻名的甲骨文正在加速跻身云计算服务超级巨头， 重点吸引专注于对AI算力需求无比庞大的人工智能训练/推理系统的大客户。 今年早些时候，甲骨文 宣布与AI领域领军者OpenAI联合组建名为"Stargate"(即星际之门)的大型数据中心合资企业，为 OpenAI提供海量AI算力资源。 不仅如此，特斯拉CEO埃隆·马斯克所创立的 AI新势力xAI ，以及美国科技巨头——Facebook母公司 Meta Platforms(META.US) 等众多聚焦于AI基础设施建设的科技行业领军者 ，都已成为其云计算客 户。 甲骨文首席执行官萨 ...

华尔街见闻此前提及，有媒体爆料OpenAI已与谷歌在上月敲定合作协议，谷歌云为OpenAI训练和运行 模型提供算力，之前因OpenAI与微软的锁定协议而无法达成合作。而今年1月之前，微软一直是OpenAI 的独家云服务供应商。 虽然这项潜在合作的条款和规模尚未公布，但分析师预计随着生成式AI推理阶段的持续发展，OpenAI 将需要越来越多的容量。作为参考，今年早些时候OpenAI与CoreWeave签署了119亿美元的协议（包括 3.5亿美元的股权投资），最近又将该协议额外延长了40亿美元。 报告指出，对于谷歌而言，这一合作可能谷歌云业务的一个重要转折点。 目前，谷歌云的年化收入已接近500亿美元，过去两年保持了20%至30%的增长率，但相较于亚马逊的 AWS和微软的Azure，其市场份额和增速并未展现出持续的压倒性优势——而OpenAI的加入，若规模 可观，可能成为谷歌云加速增长的催化剂。 报告认为，以谷歌当前股价对应的2026年市盈率约18倍来看，市场尚未充分给予谷歌云应有的估值溢 价。 报告补充称，随着谷歌芯片产能限制在下半年有所缓解，这可能成为推动谷歌云更快、更持久增长的驱 动力。这对投资者而言，是一 ...

Core Viewpoint - The article emphasizes the importance of archival work in serving the party, the country, and the people, highlighting the need for high-quality development in Changsha's archival sector [1][4]. Group 1: Archival Work Importance - Archival work is described as a vital undertaking that carries civilization, records history, and serves the public, benefiting the nation and the people [4]. - The archival system in the city is urged to maintain a strong political attribute, ensuring that political considerations are integrated throughout all aspects of archival work [4]. Group 2: Development and Modernization - There is a call for enhanced efforts in the collection, management, and protection of red archival resources to preserve the urban memory of modern Chinese practices in Changsha [4]. - The article stresses the need for deep research and development of cultural archival resources to effectively tell the story of Changsha and promote Hunan culture [4]. - A push for digital transformation is highlighted, with an emphasis on the application of cloud computing, big data, and artificial intelligence to improve the modernization of archival work [4]. Group 3: Safety and Talent Development - The importance of maintaining archival security is underscored, with recommendations for improving personnel, physical, and technical defenses [4]. - There is a focus on strengthening the professional talent pool to support the high-quality development of the city's archival work [4].

Industry Overview - The global multiphysics software industry is at the intersection of technological innovation and industrial transformation, with a market size projected to reach approximately $4.1 billion in 2024, representing a year-on-year growth of 7.89% [1][13] - The integration of cloud computing, artificial intelligence, and quantum computing is emerging as a major trend in the technology sector [1][13] - ANSYS Cloud has reduced fluid-structure coupling simulation time by 40% through GPU acceleration technology, while Siemens Simcenter has achieved a 50% reduction in design iteration cycles for wind turbine blades by leveraging digital twin technology [1][13] Industry Development History - The Chinese multiphysics software industry has undergone four main stages: 1. The initial stage before 2010, where core technologies relied on international vendors, and domestic applications were limited to high-end manufacturing [4] 2. The technology accumulation period from 2010 to 2015, marked by national emphasis on industrial software and the establishment of special funds to support simulation software development [4] 3. The self-innovation breakthrough period from 2016 to 2020, where domestic platforms achieved commercial viability and expanded applications beyond the power equipment sector [5] 4. The ecosystem construction period from 2021 to present, characterized by explosive market growth and a focus on technological innovation and ecosystem collaboration [6] Industry Chain - The upstream of the multiphysics software industry includes hardware devices, basic software, and technical resources, while the midstream involves software development and solution integration [8] - The downstream applications span various sectors, including industrial manufacturing, energy, electronics, academia, and research institutions [9] Market Size - The global multiphysics software market is projected to reach approximately $4.1 billion in 2024, with a year-on-year growth of 7.89% [1][13] - The industry is witnessing significant advancements in technology, with ANSYS Cloud experiencing a 47% annual growth rate [15] Key Companies' Performance - Domestic companies like Cloud Road Intelligent Manufacturing and Chipray Microelectronics are breaking through technological barriers through independent research and development [15] - ANSYS and COMSOL dominate the high-end market, with ANSYS holding over 60% market share in automotive collision safety after acquiring LS-DYNA [15][17] - The market is seeing a shift towards cloud-based solutions, with ANSYS Cloud's subscription model lowering the entry barrier for small and medium enterprises [15] Industry Development Trends 1. Accelerated technological integration is driving breakthroughs in simulation accuracy and efficiency, with cloud computing and edge computing reshaping simulation paradigms [21] 2. The renewable energy revolution is becoming the largest incremental market, with significant applications in battery thermal management and wind turbine design [22] 3. The restructuring of ecosystems is accelerating, with policies supporting the development of domestic software and increasing the procurement ratio of domestic software in leading enterprises [23]

公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容 编译自 pcisig 。 PCI-SIG今日宣布向会员正式发布PCI Express ( PCIe ) 7.0规范，该规范速度达到128.0 GT/s。PCIe 7.0规范面向数据驱动型应用，例如人工智能/机器学习 (AI/ML)、800G 以太网、云计算和量子计 算。PCIe 8.0规范的探索工作已在进行中，以继续支持行业在PCIe技术生态系统中的投资和产品路线 图。 PCIe 7.0 规范特性： PCI-SIG 总裁兼主席 Al Yanes 表示："PCIe 技术在过去二十多年一直是高带宽、低延迟 IO 互连的 首选。我们很高兴宣布发布 PCIe 7.0 规范，这延续了我们每三年将 IO 带宽翻一番的悠久传统。随 着人工智能应用的持续快速扩展，下一代 PCIe 技术将满足部署 AI 的数据密集型市场的带宽需求， 包括超大规模数据中心、高性能计算 (HPC)、汽车以及军事/航空航天。" 通过 x16 配置提供 128.0 GT/s 原始比特率和高达 512 GB/s 的双向速率 采用 PAM4（4 级脉冲幅度调制）信令和基于 Flit 的编码 提供 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容 编译自 pcisig 。 PCI-SIG今日宣布向会员正式发布PCI Express ( PCIe ) 7.0规范，该规范速度达到128.0 GT/s。PCIe 7.0规范面向数据驱动型应用，例如人工智能/机器学习 (AI/ML)、800G 以太网、云计算和量子计 算。PCIe 8.0规范的探索工作已在进行中，以继续支持行业在PCIe技术生态系统中的投资和产品路线 图。 PCIe 7.0 规范特性： 通过 x16 配置提供 128.0 GT/s 原始比特率和高达 512 GB/s 的双向速率 采用 PAM4（4 级脉冲幅度调制）信令和基于 Flit 的编码 提供更高的电源效率 保持与前几代 PCIe 技术的向后兼容性 | Revision ** | Max Data Rate | Encoding | Signaling | | --- | --- | --- | --- | | PCle 8.0 (TBD) | Pathfinding | TBD | TBD | | PCle 7.0 (2025) | 128.0 GT/s | 1b/1b (Fli ...