英伟达产业链ET 通信ETF华夏（515050）深度聚焦电子（芯片、PCB、消费电子）+通信（光模块、服务器）算力硬 件。前五大持仓股：新易盛、中际旭创、立讯精密、工业富联、兆易创新。场外联接(A类：008086；C 类：008087)。 3月12日早盘，光模块、CPO算力硬件板块盘初拉升，相关算力ETF方面，同指数规模最大通信ETF华夏 （515050）高开，截至9:58涨幅达0.61%，持仓股中，光迅科技迅速涨停，三安光电、紫光股份领涨， 网宿科技、亨通光电、太辰光跟涨。 消息面上，英伟达正在自主智能体基础设施竞争中发力，美东时间11日周三，英伟达宣布推出新一代开 源大语言模型Nemotron 3 Super，专为企业级多智能体系统设计，凭借全新的混合专家（MoE）架构， 将推理吞吐量提升至上一代模型的五倍以上。该模型的总参数量达1200亿，推理时仅激活120亿参数， 原生支持100万token上下文窗口。英伟达表示，Nemotron 3 Super在效率与开放性方面已登顶Artificial Analysis榜首，同等规模模型中准确率领先。 高盛认为，AI作为通用智能层的下一阶段应用已经显现。OpenC ...

Group 1 - The article reports that the Iranian Tasnim News Agency has listed U.S. companies with ties to Israel as "new targets" for Iran, indicating an expansion of the conflict to infrastructure warfare [1] - Companies mentioned include Google, Microsoft, Palantir, IBM, Nvidia, and Oracle, whose technologies have been utilized for military purposes in the region [1] - The report highlights that the rapid expansion of U.S. tech companies' data center networks in the Middle East has made these multi-billion dollar AI infrastructures targets in the ongoing conflict [1]

公众号记得加星标⭐️，第一时间看推送不会错过。 英伟达支持的法国初创公司Scintil Photonics周三表示，已开始向客户提供激光芯片进行测试。 Scintil是众多致力于研究如何利用光脉冲而非电信号在人工智能服务器（例如Nvidia和AMD生产的 服务器）内部传输信息的初创公司之一。这项技术有望简化将多个芯片连接起来组成一台大型计算机 的难度。分析师预计，Nvidia将在下周于硅谷举行的开发者大会上公布更多关于这项名为"共封装光 学"的技术计划。 所有光学系统都依赖激光芯片来产生承载信息的光束，而这些芯片由一种名为磷化铟的特殊材料制 成，主要用于长距离通信网络，但目前的产量不足以满足人工智能数据中心的需求。正是由于这种供 应状况，英伟达在本月初分别向两家最大的激光芯片制造商 Lumentum 和 Coherent 各投资了 20 亿 美元。 Scintil 去年从英伟达获得了 5800 万美元的融资，该公司找到了一种方法，可以将磷化铟激光器与光 通信所需的其他一些元件封装到单个芯片中，并与以色列的 Tower Semiconductor 合作，后者是其 制造合作伙伴。 Scintil公司首席执行官马特 ...

Group 1 - The core topic revolves around the rapid rise of OpenClaw, an open-source AI project, which has sparked significant interest and policy support across various regions in China [1][2] - Multiple cities, including Hefei and Shenzhen, have introduced measures to support the development of OpenClaw and one-person companies (OPC), with Hefei offering up to 10 million yuan in funding [1][2] - Concerns regarding the security of OpenClaw have been raised, highlighting its high-level permissions and the lack of strict security audits in the skill package market [2] Group 2 - Tencent has announced a series of products under the "lobster" brand, including WorkBuddy, which is fully compatible with OpenClaw and aims to enhance usability and security [3][4] - Nvidia plans to invest $26 billion over the next five years to develop open-source AI models, marking a strategic shift from being solely a chip manufacturer to becoming a leading AI laboratory [4][5][6] - The investment will cover model development, computational infrastructure, talent acquisition, and ecosystem building, with the first models expected to be released by late 2026 or early 2027 [6][8] Group 3 - Google has completed a $32 billion acquisition of Wiz, marking its largest acquisition to date, aimed at enhancing its cloud security capabilities [11][12] - The acquisition is part of Google's strategy to strengthen its position in the cloud computing and AI sectors, responding to the growing demand for secure AI infrastructure [11][12] - Wiz will continue to operate with a multi-cloud service strategy, providing its products through competitors' platforms like AWS and Azure [12] Group 4 - Xiaomi has launched its official car insurance, FaBa Tianxing, which offers competitive pricing and additional services compared to other insurance providers [18] - The insurance is currently available for renewal in Beijing, with a focus on providing enhanced services to customers [18][19] Group 5 - NIO's CEO Li Bin emphasized that the company's first quarterly profit is just a starting point, indicating a long-term competitive strategy in the automotive industry [20][21] - NIO plans to maintain a quarterly R&D investment of 2 to 2.5 billion yuan in 2026, focusing on improving efficiency and ensuring strong investment in key products and technologies [20][21]

Market Overview - Major U.S. indices ended mixed, with the Dow Jones Industrial Average falling 0.61% to 47,417.27, the S&P 500 slipping 0.08% to 6,775.80, and the Nasdaq edging up 0.08% to 22,716.13 [1] UiPath Inc (NYSE:PATH) - UiPath's stock surged by 6.82%, closing at $12.38, with an intraday high of $12.42 and a low of $11.65, while the stock fell 7.67% to $11.43 in after-hours trading [1] - The company's fourth-quarter earnings exceeded expectations, reporting revenue of $481.11 million against estimates of $464.49 million, and adjusted earnings per share of 30 cents, surpassing the expected 26 cents [2] - For fiscal 2027, UiPath expects revenue of $1.754 billion to $1.759 billion, above the $1.744 billion estimate, and forecasts first-quarter revenue of $395 million to $400 million versus the $394.03 million consensus [3] Nebius Group N.V. (NASDAQ:NBIS) - Nebius shares climbed 16.15%, closing at $112, with a high of $114.08 and a low of $104.49, while slipping 2.67% to $109.01 in after-hours trading [4] - The increase followed a strategic partnership announcement with Nvidia, which will invest $2 billion into Nebius, showcasing confidence in their AI technology capabilities [4] Navitas Semiconductor Corp (NASDAQ:NVTS) - Navitas Semiconductor saw a 24.88% increase, closing at $10.84, with an intraday high of $10.88 and a low of $8.72, while dropping 4.05% to $10.40 in extended trading [5] - The rise came after the company launched new products in its 5th generation GeneSiC silicon carbide power semiconductor platform, targeting AI data centers and energy infrastructure [5] Bumble Inc - Bumble's stock increased by 1.07%, closing at $2.84, with a high of $2.99 and a low of $2.82, while jumping 21.83% to $3.46 in after-hours trading [6] - The company reported fourth-quarter revenue of $224.17 million, beating the $221.3 million consensus estimate, while posting a loss of $4.06 per share versus expectations for earnings of 24 cents [7] - For the first quarter, Bumble expects revenue of $209 million to $213 million, roughly in line with the $210.89 million estimate [8] Papa John's International Inc (NASDAQ:PZZA) - Papa John's stock jumped 19.42%, closing at $38.86, with a high of $39.95 and a low of $31.92, following reports of a takeover bid by a Qatari-backed investment fund, valuing the company at $47 per share [9]

寿司郎回应消费者吃出金枪鱼寄生虫卵 往期推荐 多家黄金品牌宣布：调价！ 周三， 英伟达 宣布推出新一代开源大语言模型Nemotron 3 Super，专为企业级多智能体系统设计，凭借全新的混合专家（MoE）架构， 将推理吞吐量提升至上一代模型的五倍以上。 来源：第一财经 关注" 新华网财经 "视频号 更多财经资讯等你来看 ...

编辑 | 七三 周三， 英伟达 宣布推出新一代开源大语言模型Nemotron 3 Super，专为企业级多智能体系统设 计，凭借全新的混合专家（MoE）架构，将推理吞吐量提升至上一代模型的五倍以上。 ...

Core Viewpoint - Nvidia is aggressively positioning itself as a leading provider of enterprise AI agents, leveraging its dominance in AI hardware and expanding its software capabilities [1][10]. Group 1: Nvidia's Market Position - Nvidia is becoming the operating system of the AI economy, dominating AI compute with its advanced GPU chips and servers [1]. - The stock has shown resilience, avoiding a drop below its 200-day simple moving average (SMA) of approximately $178, and has increased by 5% over the past week to $186, surpassing its 50-day SMA of around $183 [2]. Group 2: Enterprise AI Agents - Enterprise AI agents are autonomous systems capable of understanding goals, making decisions, and executing multi-step actions without constant human oversight, functioning as digital coworkers [3]. - Current enterprise AI agents primarily operate on cloud platforms like Azure, AWS, and Google, or through AI model providers such as OpenAI and Anthropic [4]. Group 3: Nvidia's NemoClaw Platform - Nvidia is launching an open-source enterprise AI agent platform named NemoClaw, designed to be chip-agnostic, allowing deployment across various hardware without licensing fees [5]. - NemoClaw will create an open software layer that runs on Nvidia's AI accelerators, providing enterprises with a secure and customizable foundation for AI agents [7]. - The platform is expected to include enterprise-grade security, privacy tools, governance, and auditability, addressing concerns about data exposure to external clouds [9]. Group 4: Strategic Implications - The open-source nature of NemoClaw is anticipated to drive demand for Nvidia's hardware and services, reinforcing its ecosystem lock-in strategy [10][11]. - Analysts suggest that Nvidia's approach may mirror Microsoft's strategy of using software to dominate larger markets, with Nvidia's hardware already leading to a market cap exceeding $4.48 trillion [12]. - The bullish sentiment surrounding Nvidia is supported by its recent performance, with expectations for continued growth as it expands its enterprise AI footprint through NemoClaw [14][15]. Group 5: Investment Outlook - Nvidia's stock is currently rated as a Zacks Rank 2 (Buy), trading near decade lows in forward P/E valuation at 23X, indicating a potentially attractive entry point for investors [16].

以下文章来源于投中网 ，作者曹玮钰 投中网 . 投中网是领先的创新经济信息服务平台，拥有立体化传播矩阵，为创新经济人群提供深入、独到的智识 和洞见，在私募股权投资行业和创新商业领域拥有权威影响力。官网：www.chinaventure.com.cn 本文来自微信公众号： 投中网 ，作者：曹玮钰，题图来自：视觉中国 段sir悄悄投AI了。 段永平的持仓有日子没变了，很长一段时间，他主要做做增减，或者玩一玩金融工具，比如看涨期权 （Call） 和看跌期权 （Put） 。 根据披露，2025年Q4段永平管理的H&H International Investment一次性建仓三支AI新股。和过往的 区别在于，新标的不是传统意义的平台巨头，而是颗粒度很细的三家AI垂直企业——AI算力租赁商 CoreWeave，高速互联芯片Credo，AI医疗Tempus AI。 三家仓位很轻盈，合计仅0.28%，折3.4亿人民币 （4831万美元） ，放在1200亿的总持仓显得九牛一 毛，是典型的观察仓。不过段永平一向的理念是"不懂不投"，一小步足以说明不少问题。 另一值得注意的是，段永平大幅增持了英伟达，一把翻11倍 （增幅1110 ...

Core Viewpoint - Meta Platforms plans to launch four generations of self-developed AI chips by the end of 2027 to support its growing AI computing needs and reduce reliance on external chip suppliers [1][2]. Group 1: Chip Development Plans - The company will introduce the MTIA 300, MTIA 400, MTIA 450, and MTIA 500 chips, part of the "Meta Training and Inference Accelerator (MTIA)" series, aimed at internal AI training and inference tasks [1]. - The MTIA 300 is already in mass production for content ranking and recommendation model training, while the MTIA 400 has completed lab testing and is being gradually deployed [1]. - The MTIA 450 and MTIA 500 are expected to achieve large-scale deployment by 2027, with the MTIA 450 anticipated to launch in early 2027 and the MTIA 500 approximately six months later [1][2]. Group 2: AI Development and Chip Procurement - The rapid pace of AI development has necessitated an accelerated chip development timeline, with Meta currently relying heavily on external chip suppliers like NVIDIA and AMD, with procurement agreements reaching hundreds of billions [2]. - Meta has made significant investments in AI to build competitive large models and AI products, leading to unprecedented computational demands [2]. Group 3: Internal Chip Development Strategy - Self-developed chips will be optimized for specific applications, enhancing efficiency and reducing costs by eliminating unnecessary features [3]. - The chip development process is costly and lengthy, typically requiring billions in investment and several years from design to mass production [3]. - Despite challenges, including the cancellation of the advanced AI training chip project "Olympus" due to design difficulties, Meta remains committed to developing AI processors for training large AI models [3].