[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fYVre6a9UNzFK7c6m4G8wOeT4sRq6DPQ6uxxfgEPj_ec":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1227,"人形机器人未来展望：技术瓶颈与市场潜力深度对话","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">在2025年世界机器人大会期间，NVIDIA Omniverse和仿真技术副总裁Rev Lebaredian、宇树科技创始人兼CEO王兴兴，以及北京大学助理教授、银河通用创始人兼CTO王鹤，共同出席了一场小型媒体会。会上，他们深入探讨了人形机器人大规模商用的难点、关键挑战、技术瓶颈以及未来趋势。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">NVIDIA作为全球领先的机器人基础设施供应商，其芯片和平台已在机器人企业中广泛应用。目前，NVIDIA正与阿里云、北京人形机器人创新中心等多家企业合作，共同推动中国在人形机器人和AI解决方案方面的发展。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">宇树科技作为国内知名的机器人独角兽企业，近期启动了IPO，并为其新型人形机器人R1部署了NVIDIA全栈机器人技术。R1售价3.99万元，预计年底前实现量产。宇树科技还自研了灵巧手，希望在未来一两年内实现自然交互。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">银河通用作为具身大模型通用机器人独角兽，成立两年来已完成超过20亿元融资。其G1 Premium是首批搭载NVIDIA Jetson Thor的人形机器人之一，采用轮式底盘，下一代人形机器人项目将采用纯双足设计。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">在媒体会上，各位嘉宾就人形机器人商用难点、关键技术瓶颈等问题进行了详细解答。王兴兴表示，成本和硬件并非人形机器人商业化的关键问题，具身智能模型的泛用性和实用性才是当前最大的挑战。王鹤则指出，机器人完成任务的能力有限，限制了其规模化部署。然而，一旦在有限技能范围内实现通用水平，人形机器人市场将具有千亿级规模。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">针对人形机器人在工厂车间的应用，王鹤提到，搬运和分拣是当前主要的应用场景。然而，分拣速度仍是机器人需要克服的难点。熟练工人的分拣速度极快，而机器人在模型和硬件层面尚难以达到这种效率。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">对于物理AI领域的能耗、热管理和体积限制挑战，Rev Lebaredian表示，NVIDIA致力于通过全栈优化提升每瓦性能和每美元性能。这包括芯片、算法、软件以及应用层面的整体优化，以实现显著的性能飞跃。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">关于人形和双足设计的必要性，王鹤认为，从长远来看，人形机器人必须能够融入人类生活。王兴兴也表示，人形机器人是通用机器人的重要载体，其结构相对简洁，未来有望像电脑一样普及。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">在谈到中国在物理AI和机器人领域的优势时，Rev Lebaredian指出，中国拥有顶尖的AI人才、电子与计算技术能力以及庞大的制造业基础，为物理AI和机器人产业的快速发展提供了坚实基础。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">对于未来机器人普及的场景，王兴兴认为，家用机器人普及的最大难题并非技术，而是伦理和安全要求极高。他预计，随着更强大、更通用的AI大模型的出现，机器人将在更多通用场景中表现更好。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">关于机器人数量对生育率的影响，Rev Lebaredian认为，人们是否选择生育与机器人数量无直接关联。然而，打造机器人以提升生产力，对于维持和推动社会经济发展至关重要。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">在机器人算力、模型和数据挑战方面，王鹤提到，多模态大模型目前仍处于发展阶段，数据不足是核心问题。仿真技术将是解决这一问题的有效途径。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Rev Lebaredian强调，仿真是构建安全机器人系统的唯一方式。他介绍了NVIDIA在提升仿真速度、精度以及构建代表现实世界数据方面的努力。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">NVIDIA为机器人打造了三台计算机：机器人本体计算机、AI工厂计算机和仿真计算机。其中，Jetson Thor是专为物理世界中的智能推理Agent打造的超级计算机，具有强大的计算能力和能效比。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">宇树科技展示了多款机器人产品，包括足式人形机器人、机器狗和轮式机器人等。王兴兴表示，宇树科技重视机器人上肢和手部的研发，希望在未来实现自然交互。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">银河通用则展示了\u003C\u002Fspan>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">其在无人药房、无人零售等领域的应用。王鹤提到，银河通用的机器人能够在复杂环境中高效导航，得益于其长期研发的大规模具身大模型VLA。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fce20c6d8bc4d46c6a0576023dd7240bc\u002FAA1KoXAT.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">最后，各位嘉宾一致认为，机器人是连接物理世界和信息世界的桥梁。通过全球协作，推动机器人与AI技术的发展，将让计算的力量进入更广阔的市场，为人类文明迈向新高度提供强大动力。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】ITBEAR科技资讯 \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.msn.cn\u002Fzh-cn\u002Fnews\u002Fother\u002F%E4%BA%BA%E5%BD%A2%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%9C%AA%E6%9D%A5%E5%B1%95%E6%9C%9B-%E6%8A%80%E6%9C%AF%E7%93%B6%E9%A2%88%E4%B8%8E%E5%B8%82%E5%9C%BA%E6%BD%9C%E5%8A%9B%E6%B7%B1%E5%BA%A6%E5%AF%B9%E8%AF%9D\u002Far-AA1KoVbR?ocid=msedgntp&pc=LCTS&cvid=689c595b41f04df5ae47439972661e3e&ei=12\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fu5a.cn\u002FPR121\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F32c2f1fc0f6b44149dd71735ecc3fd19\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fthumbs\u002F32c2f1fc0f6b44149dd71735ecc3fd19\u002FAI领域.jpg",0,1,207,"2025-08-15 16:58",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Aa0c2e07b-c88d-403b-8462-bf884737a2e0%3A0.wav?Expires=1755255247&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=3tLCo3Lnl4f%2Flk%2FQ9dIcVxpzThI%3D",8589040,"a0c2e07b-c88d-403b-8462-bf884737a2e0","2025-08-15 16:54","Future Outlook of Humanoid Robots: In-Depth Dialogue on Technical Bottlenecks and Market Potential","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">During the 2025 World Robot Conference, Rev Lebaredian, Vice President of NVIDIA Omniverse and Simulation Technology, Wang Xingxing, founder and CEO of Unitree, and Wang He, assistant professor at Peking University and founder and CTO of Galaxy General, participated in a small media event. They discussed in depth the difficulties, key challenges, technical bottlenecks, and future trends of large-scale commercialization of humanoid robots.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">As a leading global supplier of robot infrastructure, NVIDIA's chips and platforms have been widely used in robot companies. Currently, NVIDIA is collaborating with Alibaba Cloud, Beijing Humanoid Robot Innovation Center and other companies to jointly promote the development of humanoid robots and AI solutions in China.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Unitree, a well-known robot unicorn company in China, has recently launched an IPO and deployed NVIDIA's full-stack robot technology for its new humanoid robot R1. The price of R1 is 39,900 yuan, and it is expected to achieve mass production by the end of this year. Unitree also independently developed a dexterous hand, hoping to achieve natural interaction within the next one or two years.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Galaxy General, a unicorn company of embodied large model general-purpose robots, has completed over 2 billion yuan in financing in two years. Its G1 Premium is one of the first humanoid robots equipped with NVIDIA Jetson Thor, using a wheeled chassis. The next generation humanoid robot project will use a purely bipedal design.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">During the media event, the guests provided detailed answers to issues such as the commercial challenges of humanoid robots and key technical bottlenecks. Wang Xingxing stated that cost and hardware are not the key issues for the commercialization of humanoid robots. The versatility and practicality of embodied intelligence models are the biggest challenges at present. Wang He pointed out that the ability of robots to complete tasks is limited, which restricts their large-scale deployment. However, once a general level is achieved within a limited skill range, the humanoid robot market will have a scale of hundreds of billions of yuan.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regarding the application of humanoid robots in factory workshops, Wang He mentioned that moving and sorting are the main application scenarios currently. However, the sorting speed remains a challenge for robots. The sorting speed of skilled workers is extremely fast, while robots still find it difficult to reach this efficiency at the model and hardware levels.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regarding the challenges in the physical AI field such as energy consumption, thermal management, and volume limitations, Rev Lebaredian stated that NVIDIA is committed to improving performance per watt and performance per dollar through full-stack optimization. This includes overall optimization at the chip, algorithm, software, and application levels to achieve significant performance leaps.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regarding the necessity of humanoid and bipedal designs, Wang He believes that in the long run, humanoid robots must be able to integrate into human life. Wang Xingxing also said that humanoid robots are an important carrier of general-purpose robots, and their structure is relatively simple. In the future, they are expected to become as common as computers.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">When talking about China's advantages in the fields of physical AI and robotics, Rev Lebaredian pointed out that China has top AI talents, electronic and computing technology capabilities, and a vast manufacturing base, providing a solid foundation for the rapid development of physical AI and the robot industry.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regarding the scenario for the popularization of future robots, Wang Xingxing believes that the biggest problem for the popularization of home robots is not technology, but the high ethical and safety requirements. He expects that as more powerful and general AI large models emerge, robots will perform better in more general scenarios.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regarding the impact of robot numbers on fertility rates, Rev Lebaredian believes that whether people choose to have children is not directly related to the number of robots. However, building robots to enhance productivity is crucial for maintaining and promoting social and economic development.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regarding the challenges of robot computing power, models, and data, Wang He mentioned that multi-modal large models are still in the development stage, and insufficient data is the core issue. Simulation technology will be an effective way to solve this problem.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Rev Lebaredian emphasized that simulation is the only way to build a safe robot system. He introduced NVIDIA's efforts to improve simulation speed, accuracy, and build representative real-world data.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">NVIDIA has created three computers for robots: the robot body computer, the AI factory computer, and the simulation computer. Among them, Jetson Thor is a supercomputer specifically designed for intelligent reasoning agents in the physical world, with strong computing power and energy efficiency ratio.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Unitree demonstrated a variety of robot products, including legged humanoid robots, robotic dogs, and wheeled robots. Wang Xingxing stated that Unitree values the research and development of robot upper limbs and hands, hoping to achieve natural interaction in the future.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Galaxy General, on the other hand, demonstrated its applications in areas such as unmanned pharmacies and unmanned retail. Wang He mentioned that Galaxy General's robots can efficiently navigate in complex environments, thanks to its long-term developed large-scale embodied large model VLA.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fce20c6d8bc4d46c6a0576023dd7240bc\u002FAA1KoXAT.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Finally, all the guests agreed that robots are the bridge connecting the physical world and the information world. Through global collaboration, promoting the development of robot and AI technologies will bring the power of computing into broader markets and provide strong momentum for humanity's advancement to a new height.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source] ITBEAR Tech News \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.msn.cn\u002Fzh-cn\u002Fnews\u002Fother\u002F%E4%BA%BA%E5%BD%A2%E6%9C%BA%E5%99%A8%E4%BA%BA%E6%9C%AA%E6%9D%A5%E5%B1%95%E6%9C%9B-%E6%8A%80%E6%9C%AF%E7%93%B6%E9%A2%88%E4%B8%8E%E5%B8%82%E5%9C%BA%E6%BD%9C%E5%8A%9B%E6%B7%B1%E5%BA%A6%E5%AF%B9%E8%AF%9D\u002Far-AA1KoVbR?ocid=msedgntp&pc=LCTS&cvid=689c595b41f04df5ae47439972661e3e&ei=12\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fu5a.cn\u002FPR121\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reprinted by this website to provide readers with more information and news. The content does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views of the article are not the views of this website and are for reference only.）\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A80fa5321-253d-410a-ac70-d524c8c7dc10%3A0.wav?Expires=1774838495&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=TONdsr%2BYBmz0qFVQbaXpAo3Ccwk%3D","80fa5321-253d-410a-ac70-d524c8c7dc10",10866586]