Core Insights - Serve Robotics Inc. reported third-quarter 2025 results with adjusted earnings missing estimates while revenues exceeded expectations, indicating a growth in top line but a decline in bottom line compared to the previous year [1][4][9] Financial Performance - The company reported an adjusted loss per share of 54 cents, wider than the Zacks Consensus Estimate of 37 cents, compared to a loss of 20 cents in the same quarter last year [4][9] - Revenues reached $0.69 million, surpassing the consensus estimate by 0.3%, and increased from $0.22 million in the prior-year quarter [4][9] Operational Highlights - Serve Robotics achieved significant milestones, deploying over 1,000 robots and recording a 713% year-over-year increase in daily supply hours [2][5] - Delivery volume rose 66% sequentially, expanding market coverage to over 3 million people and 1 million households in cities like Chicago, Dallas, Miami, and Los Angeles [2] Segment Performance - Fleet services generated revenues of $0.43 million, a 136.6% increase from $0.18 million in the previous year, driven by a 31% sequential rise in delivery revenues and a 120% increase in branding revenues [5] - Software services revenues reached $0.25 million, up 551.3% from $0.04 million in the prior-year period [5] Cost and Loss Analysis - The company reported a gross loss of $4.4 million, compared to a loss of $0.16 million in the same quarter last year [6] - General and administrative expenses surged 564.3% year over year to $13.2 million, while research and development expenses rose 167.9% to $13.4 million [6][7] Cash Position - At the end of the quarter, Serve Robotics had cash and cash equivalents of $116.8 million, down from $123.3 million as of December 31, 2024 [10] Future Guidance - The company anticipates generating revenues of $2.5 million in 2025 and expects around 10x revenue growth for 2026, with plans to deploy its 2,000th robot ahead of schedule by mid-December 2025 [11]

Elon Musk's bid to become the world's first trillionaire increasingly rests on a single proposition: that Tesla can evolve from an electric-vehicle company into a robotics powerhouse. Musk has made the case repeatedly that the company's future valuation depends not on cars but on full autonomy and an "army" of humanoid robots called Optimus that could one day perform the work of millions. It is a bold vision and a huge bet — one many technologists believe is both technologically plausible and financially tr ...

"We're entering an era where robots don't just assist humans—they extend human reach," said Andrew Ashur, CEO of Lucid Bots. "With Lavo AI, we're helping industries scale faster, protect workers, and unlock productivity that was never possible before." Post this With Lavo AI, operators can clean up to 6,000 square feet per hour with precision and consistency, with no manual operation required. Built with Click-and-Clean™ technology powered by the Lucid OS™, the robot allows operators to easily set up, save ...

智通财经APP讯，开创电气(301448.SZ)发布公告，公司("丙方")于2025年11月12日召开了第三届董事会 第六次会议，审议通过了《关于拟与清华大学签署建立多模态具身智能联合研究中心合作协议书暨对外 担保的议案》，同意公司与控股子公司燕谷(北京)机器人智能技术有限公司(以下简称"燕谷智能"或"甲 方")、清华大学(具体承办单位：交叉信息研究院，以下简称"乙方")签署《燕谷(北京)机器人智能技术有 限公司、浙江开创电气股份有限公司、清华大学联合建立清华大学-浙江开创电气股份有限公司多模态 具身智能联合研究中心合作协议书》；同时自愿为甲方在本协议项下的全部义务，向乙方提供不可撤销 的连带责任保证。 本次合作是公司实现技术外延与产业升级的战略决策。本协议的签署将能够充分利用清华大学在人工智 能和机器人相关领域强大的科研实力，促使公司与清华大学在多模态具身智能领域内就相关技术研发与 成果转化建立长远的战略合作关系， 以"产学研"相结合的方式进行多模态具身智能领域的研发，有利 于公司切入高壁垒的具身智能赛道，巩固和提升现有主业的智能化优势，战略性布局未来智能装备产 业，为公司长期发展注入新动能，提升公司研发能 ...

Key Takeaways Tesla plans to scale Optimus robot output, targeting 1M units yearly by late 2026.A new Texas Gigafactory line is expected to boost capacity to 10M Optimus units.Optimus aims to handle repetitive tasks, marking Tesla's expansion into robotics.Tesla (TSLA) is taking steps to scale up the production of its humanoid robot, Optimus. CEO Elon Musk has great expectations for Optimus and thinks that it has the potential to become the biggest product of the company.Per Teslarati, Tesla plans to expand ...

融中财经 . 中国领先的股权投资与产业投资媒体平台。聚焦报道中国新经济发展和创新投资全产业链。通过全媒体资讯平台、品牌活动、研究服务、专家咨询、投资 顾问等业务，为政府、企业、投资机构提供一站式专业服务。 以下文章来源于融中财经 ，作者吕敬之 来源 | 融中财经 作者 | 吕敬之 编辑 | 吾人 人形机器人被视作"下一个万亿终端"，但 2025 年的赛道先传来倒闭声：硅谷明星 K-Scale Labs 烧光数万美元后黯然关门， 100 多台预购订单被 迫退款。 融资窗口收紧、供应链缺位，让机器人"量产元年"看上去更像"生死元年"。 技术端，关节能耗、感知误差与数据荒形成隐形天花板；成本端， 40 万元 BOM 是马斯克 2 万美元目标的整整两倍；场景端，工业产线狭窄、家庭 付费意愿不足， ROI 算不过来。资本市场开始区分 PPT 与现金流，谁能先突破性价比阈值，谁才有资格谈" iPhone 时刻"。 01 人形机器人，开始倒闭 K-Scale Labs ，这家成立仅一年的类人机器人初创公司正在关闭。 该消息来自《 The Information 》看到的由 CEO Ben Bolte 发给投资者的一封邮件。 ...

Core Insights - First Tellurium Corp. is launching research and development for thermoelectric cooling devices aimed at the robotics industry, providing solid-state cooling for sensitive electronics in harsh environments [2][3] - The global robotics market is projected to grow significantly, with estimates ranging from US$33.9 billion in 2024 to US$60.5 billion by 2030, reflecting a CAGR of 9.9% from 2025 to 2030, while another estimate suggests a market value of US$47.8 billion in 2024, reaching US$211.1 billion by 2034, with a CAGR of 16.6% from 2025 to 2034 [3][4] - The company’s proprietary manufacturing process, Capillary Casting, allows for lower production costs compared to existing devices, positioning it favorably in the market [4][5] Company Strategy - First Tellurium aims to capitalize on the growing robotics market and is exploring additional markets such as high-payload drones, AI data centers, and combustion engines [5] - The company’s business model focuses on generating revenue through the development of tellurium-based technologies, mineral discovery, and project development [6]

本周，在莫斯科的一场科技展示活动上，俄罗斯机器人公司 Idol 重磅推出了其号称"首个 AI 驱动的人形机器人"——Aldol。活动现场配 上电影《洛奇》（Rocky）的经典主题曲，灯光闪烁、气氛热烈，媒体摄影机齐刷刷对准舞台。 然而，万众期待的场面瞬间变成了"事故现场"：当 Aldol 抬起手、准备向观众挥手致意时，它突然失去平衡，向前一倒，重重摔在舞台 上，外壳碎裂、零件四散。 主办方赶紧冲上前想把机器人扶起来，最后只能用一块黑布将"事故现场"遮盖起来，却反而让布料缠在机器人身上，而机器人仍在地上 不断抽搐乱动…… 短短几分钟，原本被视为展示"俄罗斯 AI 实力"的科技盛会，突然变成了一场全球围观的事故。 一场"高光时刻"，变成了当众"扑街" 据官方介绍，Aldol 的设计初衷并不简单。它由 Idol 公司开发，号称是俄罗斯首个基于"具身人工智能（Embodied AI）"的人形机器人， 结合传感器、算法与动作控制，使其能"像人一样"感知世界并做出反应。 研发团队宣称，Aldol 具备三大关键能力： Idol 公司表示，Aldol 搭载了 19 个伺服电机，能够实现十几种基础表情、上百种微表情；其硅胶"仿 ...

Core Insights - Knightscope reported a 24% year-over-year revenue growth for Q3 2025, with revenue reaching $3.1 million compared to $2.5 million in Q3 2024, primarily driven by a $0.6 million increase in product revenue due to improved delivery execution [2][4] Financial Performance - The company experienced a gross loss of $1.6 million in Q3 2025, compared to a gross loss of $0.5 million in the same quarter last year, attributed to a non-cash inventory write-off of approximately $0.6 million during a comprehensive review of inventory and manufacturing processes [3] - Operating expenses increased by 10% year-over-year to $7.9 million, mainly due to higher research and development expenses as the company invests in new product development [4] - The net loss for Q3 2025 was $9.5 million, or $(0.98) per share, an improvement from a net loss of $10.9 million, or $(3.58) per share, in Q3 2024 [4] - Cash and cash equivalents stood at $20.4 million as of September 30, 2025, reflecting a $15.2 million improvement from the previous year, supported by disciplined cost controls and equity investments [5] Business Developments - Knightscope completed its move to a new headquarters in Sunnyvale, California, in August 2025, more than doubling its footprint at lower-than-market rates [6] - The company is advancing its next-generation K7 platform, autonomy, AI-powered analytics, and advanced emergency communication device solutions [6] - Operational efficiency initiatives include a fully staffed second production shift, which reduces overtime costs and improves capacity utilization [7] Management Commentary - The CEO highlighted solid top-line growth and successful order fulfillment, emphasizing the transition to the new headquarters as a significant factor in enhancing production capabilities and operational efficiency [8] - The CFO noted that disciplined cost management and a strengthened balance sheet are key priorities, with over $20 million in cash available for strategic growth initiatives [8] Market Outlook - Knightscope anticipates strong interest in its autonomous security and emergency communication product lines, expecting to maintain revenue momentum in Q4 as production efficiency improves [9] - The company is optimistic about its proactive procurement strategy and inventory controls supporting continued delivery execution despite uneven supply chain conditions [9]

DemoHLM团队 投稿 量子位 | 公众号 QbitAI 近日，来自北京大学与BeingBeyond的研究团队提出 DemoHLM 框架，为人形机器人移动操作（loco-manipulation）领域提供一种新思 路——仅需1次仿真环境中的人类演示，即可自动生成海量训练数据，实现真实人形机器人在多任务场景下的泛化操作，有效解决了传统方 法依赖硬编码、真实数据成本高、跨场景泛化差的核心痛点。 核心挑战：人形机器人移动操作的"三重困境" 移动操作是人形机器人融入人类环境的核心能力（如搬箱子、开门、递物等），但长期受限于三大难题： 现有方案要么局限于仿真场景，要么需消耗数百小时真实遥操作数据，难以满足家庭、工业等复杂场景的实用需求。 DemoHLM：分层架构与数据生成革新，破解三重困境 DemoHLM的核心创新在于"分层控制+单演示数据生成"双引擎，既保证了全身运动的稳定性，又实现了极低数据成本下的泛化学习。 数据效率低 ：传统方法需采集大量真实机器人遥操作数据，成本极高且难以规模化； 任务泛化差 ：依赖任务特定的硬编码设计（如预定义子任务、专属奖励函数），换任务需重新开发； Sim-to-Real迁移难 ：基于 ...