今日 AI 领域资本与市场动作频繁：Anthropic 正式向 SEC 递交招股书拟纳斯达克上市，抢跑 OpenAI 冲刺“AI 第一股”；OpenAI 以 1.25 万亿美元投前估值筹资 300 亿美元并对 Anthropic 大客户放出最高 90% 折扣抢单。政策端，国办印发意见缓解 AI 就业替代，英国投入 1 亿英镑设立全球首个 AI 生物安全实验室。具身智能持续升温，SuperMimic 技术使人形机器人训练成本骤降 1000 倍，OpenAI 也发力人形机器人研发。
资本与市场
1. Anthropic 正式递交 IPO 招股书，拟纳斯达克挂牌抢跑“AI 第一股” 🔥
当地时间 10 月 26 日，Anthropic 向美国 SEC 公开递交 IPO 招股书，拟以代码「ANT」在纳斯达克上市。该公司凭借 Claude 系列大模型在企业服务市场快速崛起，选择在市场波动加大、秋季 IPO 队列收缩之际逆势推进，展现打造下一代变革性 AI 系统的长期野心。
来源：路透社/财闻
2. OpenAI 估值下调至 1.25 万亿美元，拟融资 300 亿美元 🔥
OpenAI 正以 1.25 万亿美元投前估值筹集至多 300 亿美元，较一个月前 1.4 万亿美元要价下调超 10%，由阿联酋投资者与 Coatue 领投，并将融资与算力获取捆绑。年化营收已从 8 月 400 亿美元跃至 9 月 500 亿美元，年底预期 700 亿美元。
来源：钛媒体
3. 华尔街据称推动 Anthropic 与 OpenAI 尽早上市
据市场消息，政治风险正促使华尔街加速布局，一些机构要求 Anthropic 和 OpenAI 尽早推出可能轰动的 IPO，不要因人工智能泡沫风险而瞻前顾后。大型投资者担忧民主党掌控国会的威胁可能挤压 AI 产业发展空间。
来源：格隆汇/新浪
4. DeepSeek 第二轮融资增至约 1000 亿元，宁德时代拟再投 100 亿 🔥
DeepSeek 即将完成第二轮融资，拟融资金额增至约 1000 亿元，投前估值 5000 亿元。宁德时代首轮出资约 50 亿元，本次拟再加码 100 亿元；腾讯等老股东继续参与，吉利、中芯聚源等产业资本入局。
来源：网易科技
模型与技术
5. OpenAI 针对 Anthropic 头部客户放出最高 90% 折扣，打起价格战抢客户
OpenAI 通过高额折扣定向挖掘 Anthropic 大客户，部分企业切换后 AI 使用成本最高下降 90%。在 Anthropic 增长放缓背景下，OpenAI 主动压缩利润率抢占市场份额，DeepSeek 等国内模型厂商也持续给全球 API 定价带来下行压力。
来源：CSDN
6. OpenAI 公布三项数学与科学新突破并发布成果分配计划
OpenAI 于 10 月 26 日宣布推出三项数学与科学研究成果，并发布将 AI 生成发现分配给学术界的新计划，推动 AI 在科学发现领域的落地。
来源：站长之家
7. SuperMimic 技术让人形机器人运动控制训练成本骤降 1000 倍 🔥
加州大学伯克利分校论文显示，SuperMimic 技术可将人形机器人全身运动控制器训练算力消耗压缩逾 1000 倍——原本需要 128 块 GPU 运行一周的任务，如今仅需少量资源即可完成，正在重塑人形机器人行业竞争格局。
来源：网易科技
8. 机构报告：OpenAI 发力人形机器人研发，将 AI 大脑作为核心突破口
10 月 26 日华龙证券发布人形机器人行业研报，指出 OpenAI 以 AI 大脑为核心突破口发力人形机器人研发。报告显示 10 月 19 至 23 日人形机器人指数涨幅 +11.48%，同期沪深300 下跌 1.57%。
来源：华龙证券/网易
产品与应用
9. 智元机器人 AI WAVE 全栈技术成果发布
智元机器人开启系列技术发布（10.22-30），陆续推出工业重载机器人、智能体框架、仿真平台、基座大模型、群体学习技术等具身智能核心成果，覆盖具身智能全栈。
来源：半两财经
10. 英伟达推出 AI 运维平台 Minerva，每小时分析数十亿数据点
英伟达推出 AI 运维平台 Minerva，整合基础设施与 AI 工作负载遥测数据，每小时可分析数十亿个数据点，提前发现故障、提高 GPU 运行时间并提升 AI 工厂产出效率。
来源：快科技
11. 全国首个 AI 应用风险综合保障方案推出，覆盖生成内容侵权与性能偏差
国内推出首款 AI 应用风险综合保障方案，覆盖生成内容侵权与性能偏差损失，为 AI 产品商业化落地提供风险兜底，填补了生成式 AI 应用保险保障的市场空白。
来源：CSDN
政策与生态
12. 英国将设全球首个 AI 生物安全实验室，OpenAI、Anthropic 与谷歌 DeepMind 参与支持
英国将投入 1 亿英镑设立国家级人工智能生物安全实验室，由 OpenAI、Anthropic 和谷歌 DeepMind 共同支持，旨在利用 AI 识别、监测新兴生物安全威胁，并获得最高价值 2 亿英镑的 AI 模型调用额度。
来源：英国政府公告/腾讯网
13. 国办印发意见缓解 AI 就业替代，推动就业友好型发展
国务院办公厅印发《关于适应人工智能技术发展助力构建就业友好型发展方式的意见》，提出增强 AI 创造就业、赋能传统岗位、提升就业质量的正面效应，有效减缓 AI 应用就业替代影响。
来源：百家号
14. 广告产业发展“十五五”规划：人工智能生成广告纳入监管
《广告产业发展“十五五”规划》正式发布，首次将人工智能生成广告纳入监管范畴，明确对 AI 生成内容的监管边界，为产业合规发展划清红线。
来源：微信公众号
算力与芯片
15. 工信部披露智能算力规模达 2592 EFLOPS，同比大幅增长
工信部信息通信管理局局长谢存在国新办发布会披露，截至 9 月底我国智能算力规模已达 2592 EFLOPS（每秒百亿亿次浮点运算），同比大幅增长，人工智能核心产业规模超 1.2 万亿元、企业超 6600 家。
来源：搜狐/工信部
本文由 AI 热点日报自动整理生成，新闻均来自公开网络信息，版权归原媒体所有。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Today, capital and market movements in the AI field are high: Anthropic officially submitted a book to the SEC for listing in NASDAQ, sprinting Open AI's “AI first share”; Open AI raised $30 billion in pre-investment valuations of $1.25 trillion and issued up to 90 per cent discounts on top of Anthropic's large clients. At the end of the policy process, the Office of the United States Secretary of State issued advice on the mitigation of AI employment substitution, and the United Kingdom invested Pound100 million to set up the first AI biosafety laboratory worldwide. With intelligence and constant warming, SuperMimic technology has dramatically reduced the cost of robotic training by 1,000 times, and Open AI has developed human robotics.
Capital and markets
Anthropic officially submits an IPO bid book for NASDAQ to mount the "AI 1"
On October 26th, local time, Anthropic publicly submitted an IPO book to the US SEC to be listed in NASDAQ with the code "ANT". With Claude series LLM, the company has chosen to advance in reverse at a time of increased market volatility and fall IPO contraction, demonstrating its long-term ambition to build a next generation of transformative AI systems.
Source: Reuters/fiscal
2. The Open AI valuation was revised downwards to $1.25 trillion for proposed financing $30 billion
Open AI is raising up to $30 billion in pre-investment valuations of $1.25 trillion, up by 10 per cent from 1.4 trillion a month ago, with the United Arab Emirates investors and Coatue taking the lead and tying the financing to the calculus. The annualization of revenues jumped from $40 billion in August to $50 billion in September, with $70 billion expected at the end of the year.
Source: Titanic Media
3. Wall Street is said to have promoted the early listing of Anthropic and Open AI.
According to market news, political risks are driving Wall Street to speed up the layout, and some agencies are asking Anthropic and Open AI to launch a potentially sensational IPO as soon as possible without looking forward to the risks of artificially intelligent foams. Large investors are concerned that the threat of the Democratic Party taking control of Congress may squeeze the space for AI industrial development.
Source: Groon/New Wave
DeepSeek's second round of financing has increased to about $100 billion, with a further investment of 10 billion in the Ningde era.
DeepSeek is about to complete the second round of financing, which will increase to about $100 billion, with a pre-investment valuation of $500 billion. The first round of the Ningde era was about $5 billion, and this time it is proposed to add another $10 billion; old shareholders such as Tencent continue to participate and industrial capital, such as gili and central core, enters the bureau.
Source: Web-enabled technologies.
Models and technologies
5. Open AI has released up to 90% discounts on the Anthropic head client and raised the price to rob the client.
Open AI directs the excavation of Anthropic large clients by a high discount, with some businesses switching with AI using costs reduced by up to 90%. Against the backdrop of slowing growth in Anthropic, Open AI actively compressed its profit margin to capture market shares, and domestic model manufacturers like DeepSeek continue to put downward pressure on global API pricing.
Source: CSDN.
6. Open AI Publication of three new mathematical and scientific breakthroughs and publish distribution plan
On October 26th, Open AI announced the launch of three mathematical and scientific research results and publish generated a new plan for the discovery of AI to be distributed to academia to promote AI's landing in the field of scientific discovery.
Source: Station Chief ' s Home.
7. SuperMimic technology for robotics control training by 1,000 times
According to a paper from the University of California at Berkeley, SuperMimic technology can compress the calculator-to-manufacturance controlr of the body of human robots by more than 1,000 times -- a mission that would have required 128 GPUs to run for a week, which can now be accomplished with only a small amount of resources and is reshaping the competitive pattern of the human robot industry.
Source: Web-enabled technologies.
8. Institutional report: Open AI Research and Development of Human Dynamics, using AI brain as a core breakthrough mouth
October 26th, the publish human robotics industry of the Zhualong securities reported that Open AI had developed an AI brain at its core to break through the oral human robot. The report shows an increase of +11.48 per cent in the human robot index from 19 to 23 October, with a drop of 1.57 per cent in the depth of 300 in the same period.
Source: Walloon Securities/Web Easy
Products and applications
9. Elite robot AI WAVE Total technology publish
The intellectual robots open a series of technologies publish (10.22-30) to launch the industrial heavy-load robots, the AI agent framework, the simulation platform, the base LLM, the group learning technology, among other smart cores, which cover the entire body of intelligence.
Source: Half of the financial resources.
10. In Weida launched the AI Transport Platform Minerva, analysing billions of data points per hour
In Weida launched the AI platform Minerva, which integrates infrastructure and AI task load telemetry data and analyses billions of data points per hour, detects failures in advance, improves GPU running time and improves AI plant output efficiency.
Source: Fast technology.
11. The first AI application of the comprehensive safeguards programme in the country covers content abuse and performance deviations.
The introduction of the initial AI application comprehensive safeguards programme to cover the generation of content torts and performance deviations, to provide a bottom for the commercialization of AI products and to fill market gaps in the generation AI application insurance.
Source: CSDN.
Policy and ecology
The United Kingdom will host the first AI Biosafety Laboratory worldwide, Open AI, Anthropic and Google DeepMind
The UK will invest Pound100 million to establish a national artificial intelligence biosafety laboratory, supported by Open AI, Anthropic and Google DeepMind, aimed at using AI to identify, monitor emerging biosafety threats and obtain the highest Pound200 million AI model to call quota.
Source: British Government Bulletin/Tencent Network
AI Employment Alternatives for Employment-Friendly Development
The Cabinet Office of the State Council issued the Opinions on Adapting to Artificial and Intelligent Technology Development for Employment-Friendly Development, which propose positive effects of enhancing AI ' s ability to create jobs, empower traditional jobs and enhance the quality of employment, and effectively mitigate the impact of AI ' s employment substitution.
Source: 100 family names.
The “Fifty Five” programme for the development of the advertising industry: the regulation of artificially intelligent advertising
The “Fifty Five” Programme for the Development of the Advertising Industry, which includes, for the first time, artificial intelligence-generated advertising in its regulatory framework, clearly defines the regulatory boundaries of the content generated by AI and clears the red line for industrial compliance development.
Source: WeChatWeChat Official Account
Calculator and Chip
15. The Ministry of Industry and Communications disclosed a significant increase in the volume of intelligent computing of 2592 EFLOPS over the same period.
The Director-General of the Information and Communication Management Department of the Ministry of Industry and Communications, Mr. publish, would like to acknowledge that at the end of September, the country ' s intelligent calculus had reached 2592 EFLOPS (billion floats per second), a significant increase over the same period, with artificial intelligence core industries over $1.2 trillion and enterprises over $6,600.
Source: Fox Search / Ministry of Industry & Communications
This post was created automatically by the Al Hotspot Daily News, with news coming from open online information and copyrighted in the original media.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.