Core Insights - The humanoid robot industry faced a harsh elimination race in 2025, with notable companies collapsing and the gap between leading and lagging firms widening significantly [4][5][11] - The industry is transitioning from a phase of hype and funding to one focused on actual production and commercial viability, highlighting the need for sustainable business models [9][12] Industry Overview - In 2025, the humanoid robot sector saw 610 financing events, nearly tripling from the previous year, with total funding exceeding 57 billion yuan, indicating a stark divide between top-tier and struggling companies [16] - The customer base shifted from academic institutions to industrial enterprises, reflecting a growing recognition of humanoid robots in production settings [16] Company Failures - K-Scale Labs, a once-prominent startup, collapsed just before mass production due to cash flow issues, revealing the challenges faced by many startups in the U.S. lacking a robust local supply chain [6][11] - Rethink Robotics, a pioneer in collaborative robots, faced its second closure in 2025 after failing to meet market demands, illustrating the rapid pace of technological evolution and competition [7][13] - iRobot, known for its Roomba, filed for bankruptcy in December 2025, struggling against low-cost competitors and regulatory hurdles, marking a significant decline for a once-leading company [8][11] Key Challenges - Many humanoid robot companies are experiencing funding shortages, with a lack of follow-up financing leading to operational difficulties [11] - Over 50% of humanoid robot orders are primarily for public relations and data collection rather than genuine productivity enhancement, indicating a failure in commercial viability [12] - Product homogeneity has intensified competition, with many companies producing similar low-barrier products, making it difficult to establish a competitive edge [12][13] Future Outlook - The survival of companies in the humanoid robot sector will depend on securing continuous funding and achieving repeat orders in industrial applications [16][17] - Companies must differentiate themselves, establish real commercial cycles, and leverage data to enhance AI capabilities to thrive in the evolving market [17] - The industry is expected to see a significant increase in humanoid robot shipments in 2026, contingent on technological advancements [16]

2017年7月17日，罗德尼·布鲁克斯（Rodney Brooks）在麻省理工学院克雷斯吉礼堂举行的TechCrunch机器人技术会议 上发言。图片来源：Paul Marotta—Getty Images for TechCrunch 当投资者正忙于向人形机器人领域投入数十亿美元时，一位拥有三十年机器人研发经验的麻省理工学院 机器人专家却指出，此类投资不过是徒劳之举。 扫地机器人Roomba制造商iRobot联合创始人罗德尼·布鲁克斯指出，埃隆·马斯克（Elon Musk）设想 的"人形机器人化身万能助手"的未来图景"纯属空想"，部分原因在于，机器人在动作协调性上存在难以 突破的瓶颈。 布鲁克斯在博文中写道："尽管风险投资机构和大型科技企业投入数亿甚至数十亿美元用于训练，当今 的人形机器人仍无法掌握灵巧操作技能。" 触觉是人体最复杂的感知系统之一。人类的手部拥有1.7万个低阈值机械感受器，用于感知轻微触碰， 且在指尖末端的分布更为密集。这些感受器能对多种刺激（如压力）做出反应，并与15类神经元同步振 动。人类试图在机器人中复刻的正是这套复杂机制。 布鲁克斯指出，尽管人工智能已通过大量语音识别和图像处理数据完成 ...