JAMES PATON2026年10月28日Neura Robotics is not a household name. It’s not even the most prominent business in Metzingen, a German town of 22,000 people that is also home to the fashion brand Hugo Boss.纽鼐机器人(Neura Robotics)并非家喻户晓的名字，甚至不是梅青根最显眼的企业——这座人口2万余人的德国小镇同时也是时尚品牌雨果博斯(Hugo Boss)的总部所在地。But in just a few months, Neura’s founder, David Reger, has attracted investment from some of the world’s most powerful technology companies, including Nvidia and Amazon, and received invitations to meet the leaders of Europe’s three largest economies.然而短短数月间，纽鼐创始人戴维·雷格尔(David Reger)已获得英伟达(Nvidia)、亚马逊(Amazon)等全球最具实力的科技公司投资，并受邀与欧洲三大经济体的领导人会面。The company’s rise underscores the importance of artificial intelligence and robotics, technologies that could help Europe transform its ailing manufacturing sector and take on Chinese competition. Neura is among companies at the forefront of Europe’s efforts to seize the potential economic gains as A.I. makes the leap from the digital world to the physical.这家公司的崛起凸显出人工智能与机器人技术的重要性：它们可能帮助欧洲改造萎靡不振的制造业，并应对来自中国的竞争。随着人工智能从数字世界跨入物理世界，欧洲正力图抓住这一潜在经济收益，而纽鼐充当着开路先锋的角色。“We are not efficient anymore,” said Mr. Reger, who raised as much as $1.4 billion in June. “But I do believe that this can be solved.”“我们已不再高效，”6月筹得高达14亿美元的雷格尔说。“但我确实相信，这个问题能够解决。”Whether A.I.-powered robots can live up to their promise remains unclear. Companies need to teach them to perform complex tasks, a challenge that requires the collection of vast amounts of data.人工智能驱动的机器人能否兑现承诺，眼下仍不明朗。企业必须教会它们执行复杂任务，而这一挑战需要收集海量数据。European robots are also far outnumbered.欧洲的机器人数量也远远落后。After rolling out 2.3 million industrial robots, almost half the global fleet, China is moving into the next frontier. Unlike the mostly one-armed machines that have been used for decades to perform repetitive tasks, the next generation of robots is expected to see, hear, feel and — most important — learn.在部署了230万台工业机器人、几乎占全球总量一半之后，中国正在迈向下一个前沿。过去数十年间用于执行重复性任务的机器大多只有单臂，而下一代机器人预计将具备视觉、听觉和触觉，还有最重要的学习能力。纽鼐正将人工智能应用于机器人技术，以帮助振兴欧洲的制造业。 Roderick Aichinger for The New York Times但它必须与拥有大量机器人的中国展开竞争。 Roderick Aichinger for The New York TimesOne of the keys to Europe’s strategy is a tech hub on the edge of Munich Airport. Inside the facility, as many as 100 humanoids and other robots have embarked on a 24-7 training mission. On a recent morning, a day after joining business leaders in Berlin for a discussion with Chancellor Friedrich Merz of Germany and Prime Minister Andy Burnham of Britain, Mr. Reger opened the “robogym” to a journalist for the first time. He declined to provide details of the meetings with the government leaders.欧洲战略的关键一环是慕尼黑机场边缘的一座技术中心。设施之内，多达100台人形机器人及其他机器人已投入全天候训练。不久前，雷格尔在柏林与商界领袖一道同德国总理弗里德里希·默茨和英国首相安迪·伯纳姆会谈。第二天上午，他首次向一名记者开放了这座“机器人健身房”。他拒绝透露与政府领导人会面的细节。Wearing a black blazer, a black shirt and black pants, Mr. Reger, 38, introduced the team. The robots picked up car parts and placed them into containers, stacked boxes and carried out other tasks, mimicking the movements of human trainers who guided them remotely. One robot practiced tightening its grip on a glass, learning how to respond when water is added.现年38岁、身着黑色外套、衬衫和长裤的雷格尔介绍了团队。机器人拿起汽车零件放入容器，堆叠箱体，并完成其他任务，模仿着遥控它们的人类训练员的动作。一台机器人反复练习握紧玻璃杯，学习杯中加水后该如何应对。When one robot learns a new skill, they all get smarter as the data is fed into a cloud-based platform, or a shared brain. Neura has signed up dozens of automotive, logistics, pharmaceutical and other companies to carry out their own training in the gym, declining to identify them. More than 2,600 have applied for space in one of the sites, Neura’s chief said.当一台机器人掌握新技能，数据会被送入一个云端平台——或者说一个共享大脑——所有机器人便随之变得更聪明。纽鼐已签约数十家汽车、物流、制药等领域的公司，让它们在健身房中进行各自的训练，但拒绝披露这些公司的名称。纽鼐首席执行官称，已有超过2600家企业申请在其中一个场地获得空间。纽鼐首席执行官戴维·雷格尔通过自学掌握了工程技术，并在瑞士制造出了自己的第一批机器人。 Roderick Aichinger for The New York TimesThe training center, developed with the Technical University of Munich, is part of a wider plan. By the end of the year, Neura aims to have five gyms operating globally, including one already opened in Hangzhou, China. Ten more are expected to follow.这座由纽鼐与慕尼黑工业大学共同开发的培训中心是一项更大计划的一部分。纽鼐的目标是在年底前在全球运营五座健身房，其中一座已在中国杭州开设，预计还将新增十座。Unlike generative A.I. models trained on enormous amounts of internet data, physical A.I. depends on the costly collection of real-world data. Mr. Reger likened it to learning how to swim. You can’t simply watch videos; you have to get in the water.生成式人工智能模型依靠海量互联网数据训练，物理人工智能则不同，它依赖于成本高昂的现实世界数据采集。雷格尔将其比作学游泳：光看视频不够，你必须下水。China, however, is building most of the pools.然而，大多数的泳池正由中国建造。More than 90 humanoid data collection centers are running across China, according to Interact Analysis, a consulting firm. And the campaign, fueled by state support and a need to replace retiring workers, shows no signs of slowing down. At least 50 more sites are under development. Europe has fewer than 10 facilities.据咨询公司Interact Analysis统计，中国各地已有90余个人形机器人数据采集中心投入运行。在政府支持与替换退休工人的需求推动下，这场运动毫无放缓迹象，至少还有50个场所正在建设中。欧洲的设施则不足10个。It’s not just the scale of China’s operation that is making headlines. A Chinese humanoid in August broke Usain Bolt’s record in a 100-meter sprint, while dozens of others ran a half-marathon in Beijing. At the Paris Motor Show, robots developed by China’s Agibot danced to the Dua Lipa song “Levitating.”登上头条的不只是中国行动的规模。8月，一台中国人形机器人在百米短跑中打破了尤塞恩·博尔特的纪录，另有数十台在北京跑完半程马拉松。在巴黎车展上，中国智元机器人(Agibot)研发的机器人随着杜阿·利帕的歌曲《Levitating》起舞。纽鼐机器人通过云端系统共享新掌握的技能。 Roderick Aichinger for The New York Times已有超过2600家公司申请在纽鼐的健身房中训练机器人。 Roderick Aichinger for The New York Times“The skill is undeniable, but a lot of it, especially in the humanoid sector, is around what we’d call entertainment robots versus industrial robots,” said Arnaud Robert, the president of Hexagon Robotics, a Swiss company that is working on a pilot with BMW at one of the German automaker’s plants in Leipzig.“这种技能是毋庸置疑的，但其中许多——尤其是在人形机器人领域——属于我们所说的娱乐机器人，而非工业机器人，”海克斯康机器人(Hexagon Robotics)总裁阿诺·罗伯特说。这家瑞士公司正与宝马在其位于莱比锡的一家工厂开展试点合作。That leaves an opening for European companies, he said.他说，这给欧洲公司留下了空间。A.I. robots are hitting the factory floor in Europe for the first time, with Germany emerging as the main training ground. Robots are expected to move into production at the BMW site by the first quarter of next year. Agile Robots, in Munich, is working with partners including Google DeepMind. Another company, Humanoid, aims to integrate robots into the operations of Schaeffler, a German auto supplier, by the end of 2026.人工智能机器人正首次进入欧洲的工厂车间，德国正在成为主要训练场。预计到明年第一季度，机器人将在宝马基地投入生产。总部位于慕尼黑的思灵机器人(Agile Robots)正与谷歌DeepMind等伙伴合作。另一家公司Humanoid则计划在2026年底前将机器人整合进德国汽车供应商舍弗勒(Schaeffler)的运营。While Neura advances with the gyms, it is teaming up with another supplier, Bosch, and other companies, equipping more than 1,000 workers with sensor suits to record their movements and capture other data during their regular shifts.在推进健身房业务的同时，纽鼐还与另一家供应商博世(Bosch)及其他公司合作，为1000多名工人配备传感器套件，在正常轮班中记录他们的动作并采集其他数据。European companies want to be judged on their ability to develop technology that translates into tangible results. But they will need to prove that their robots can operate safely and efficiently — and at a low enough cost to make them economically viable.欧洲公司希望外界关注的是它们将开发技术转化为切实成果的能力。但它们必须证明自己的机器人能够安全、高效地运行——而且成本足够低，具备经济可行性。纽鼐正与多家企业合作，为1000多名工人配备传感器套件。 Roderick Aichinger for The New York Times这些套件旨在利用现实世界的动作数据来训练机器人。 Roderick Aichinger for The New York TimesSmart robots in factories, warehouses, hospitals and hotels are forecast to play a critical role as countries cope with aging populations and shrinking work forces. The number of humanoids deployed worldwide could rise to one million in 2030 from 15,000 last year, if costs fall and the technology does not encounter regulatory headwinds, said Zornitsa Todorova, a researcher at Barclays.随着各国应对人口老龄化和劳动力萎缩，工厂、仓库、医院和酒店中的智能机器人预计将发挥关键作用。巴克莱(Barclays)研究员佐尔尼察·托多罗娃表示，如果成本下降且技术不遭遇监管阻力，全球部署的人形机器人数量可能从去年的1.5万台增至2030年的100万台。“The next 10 years will be the decade of the robot — and at the moment, it looks like that decade might belong to China,” Ms. Todorova said. Still, Europe’s focus on safety and expertise in hardware, such as actuators that enable robots to move, provide an edge, she said.“未来十年将是机器人的十年——而眼下，这个十年看起来可能属于中国，”托多罗娃说。不过她认为，欧洲对安全的重视以及在硬件领域的专长——例如让机器人得以运动的执行器——构成了一种优势。Neura raised funds this year from several prominent companies, including existing partners Nvidia and Amazon. Leading the investment was Tether, the cryptocurrency company, which is valued at more than $180 billion and has faced scrutiny from lawmakers and regulators. Neura said part of the $1.4 billion was tied to milestones that it was on track to achieve.纽鼐今年从多家知名企业筹得资金，包括现有合作伙伴英伟达和亚马逊。领投方是加密货币公司泰达(Tether)，这家估值超过1800亿美元的公司正受到立法者和监管机构的审视。纽鼐表示，14亿美元中的一部分是与它有望达成的里程碑挂钩的。在公司与慕尼黑工业大学合作的机器人健身房开幕式上，纽鼐机器人步入舞台。 Roderick Aichinger for The New York TimesGrowing up as the eldest of 11 siblings in Rottweil, 50 miles from Metzingen, Mr. Reger dreamed of living overseas, eventually heading to California. Instead of finding a tech job in Silicon Valley, however, he got his start working with a Christian organization to help homeless people in San Francisco.雷格尔在距梅青根50英里的罗特韦尔长大，是11个兄弟姐妹中的长子。他梦想移居海外，最终去了加利福尼亚。但他一开始并没有在硅谷找到科技工作，而是与一家基督教组织合作帮助旧金山的无家可归者。After returning home, he developed a deeper appreciation for Europe and its health and social systems. Rather than attending a university, he taught himself engineering and got a job in Switzerland, where he built his first robots.回国之后，他对欧洲及其医疗和社会保障体系有了更深的认同。他没有上大学，而是自学了工程，在瑞士找到一份工作，并在那里造出了自己的第一批机器人。Mr. Reger, who also met with France’s president, Emmanuel Macron, in September, said he aimed to become “Europe’s largest taxpayer,” combining A.I. and robotics to help create an economic engine for Europe that is even bigger than the car industry. Failing to reap the gains, he said, would be costly.雷格尔在9月与法国总统埃马纽埃尔·马克龙会面。他表示，自己的目标是成为“欧洲最大的纳税人”，把人工智能与机器人技术结合起来，为欧洲打造一个比汽车业更大的经济引擎。他说，如果不能收获这些收益，代价将十分高昂。“This is a competition,” Mr. Reger said. “If you see other countries get more efficient through robotics and we don’t, it’s going to be very difficult to come out of this situation.”“这是一场竞争，”雷格尔说。“如果其他国家通过机器人技术变得更高效，而我们没有，那么要走出这种局面将非常困难。”James Paton是时报国际商业和经济记者，常驻伦敦。翻译：经雷点击查看本文英文版。