敏捷机器人公司联合创始人谈纳斯达克上市、来自亚马逊的教训以及工作的未来

内容总结:
俄勒冈州塞勒姆市的人形机器人公司Agility Robotics于本周四登陆纳斯达克,成为美国首家专注于人形机器人的上市公司,既展现了该技术将如何改变工作的宏大愿景,也对实现这一目标所需的时间持谨慎态度。
“也许我们永远不会有超光速旅行,也许我们永远不会有《星际迷航》里的传送机,”该公司联合创始人兼首席机器人官乔纳森·赫斯特表示。“但我们绝对会拥有能够完成人类所有体力工作的机器。”不过他也警告说,“这不会在明年或五年后实现。这确实是一项巨大的挑战。”
Agility通过与由金融家迈克尔·克莱因领导的特殊目的收购公司Churchill Capital Corp XI合并实现上市。该交易在新投资前对Agility的估值为25亿美元,并带来了约4.65亿美元的总收益。其股票在纳斯达克以AGLT为代码交易。在最初跌破开盘价后,该股周四收于11.51美元,上涨2.2%,逆势上扬——当天纳斯达克指数下跌5.5%,创下18个月来最差单日表现。
公司机器人: Agility的主力产品是Digit,一款双足机器人,双臂安装在躯干顶部附近,使其能够高举料箱并从跌倒中恢复。它在仓库和工厂中搬运周转箱和集装箱。
最新版本Digit 5于9月发布。它身高5英尺11英寸,体重284磅,全身覆盖传感器,头部装有摄像头和天线,身体周围的安全防护壳内装有安全级传感器。Agility表示,它能举起50磅重物,触及7.2英尺高度,在24小时内工作超过20小时。但最大的进步在于,Digit 5被设计为无需隔离栏即可与人协同工作。早期版本必须与工人隔开,这限制了客户的使用场景。
Agility在亚马逊——该公司的早期投资者——身上深刻体会到了这一点的重要性。亚马逊曾在仓库中测试Digit早期版本的搬运和堆叠料箱功能。赫斯特说,Digit在那里完成了有用的工作,但当Agility询问扩大规模时,亚马逊表示无法部署这些机器人。亚马逊向Agility展示了其自身轮式机器人通过三到四年安全认证所需的努力,并表示不知道平衡型机器人如何才能达到同样的标准。
“这方面的安全论证难度超乎想象,”赫斯特说。“从来没有人部署过动态稳定平衡机器,这带来了新的风险。”让机器人走出隔离栏需要多年的工程和标准工作。Agility一直在联合编写该品类的ISO安全标准,并向保险公司和仓库运营商证明自己——这些方在将平衡机器放在人身旁之前需要数据支撑。Digit 5利用AI和多个传感器监测附近人员。根据不同情况,它会绕行、停止或坐下。视觉和音频提示会告知其即将进行的动作,当有人靠得太近时,安全控制器会触发响应。
赫斯特将来自亚马逊以及GXO、丰田和舍弗勒的经验教训归功于塑造了Digit 5的设计。“亚马逊在推动我们达到这一步的过程中发挥了巨大作用,”他说。根据Agility的 filings,亚马逊在2024年为105万美元购买五台机器人的交易支付了首付款。Agility在2025年从亚马逊录得约104万美元的净销售额,占其全年总销售额的58%。赫斯特表示,他无法透露亚马逊目前是否仍是客户。
凭借安全改进和其他进步,赫斯特称Digit 5是Agility的“规模化时刻”。“这就是那把钥匙,”他说。据其 filings,一位未具名的客户已承诺签订一份3亿美元、为期三年的合同,购买1,000台Digit 5机器人。这些机器人将以租赁而非出售的方式提供,承诺受里程碑和规格约束,客户获得随着机器人部署而逐步归属的认股权证。
赫斯特说,Agility已建造了约10台Digit 5机器人,首批原型刚刚下线。其塞勒姆工厂最终设计产能为每年多达10,000台机器人。
财务状况: Agility的业务规模仍然很小,去年收入为178万美元,今年前六个月约为17.4万美元。上半年亏损3.62亿美元,高于去年同期的7,180万美元。股票薪酬——员工股权奖励的会计费用——同期增加了约2.52亿美元。截至6月底交易完成前,Agility拥有1,470万美元现金,管理层在其中期财务报表中表示,对公司持续经营能力存在重大疑虑。
赫斯特称公开市场是他最大的不确定因素。“我对我们的市场路径和技术路径完全有信心,”他说。如果公开股东认同,Agility将能够“成功继续这一使命,并在需要时筹集资金”。如果股东对公司持怀疑态度,“那将使我们的日子艰难得多”。
Agility拥有大多数竞争对手所没有的——正在从事商业工作的机器人。截至5月,其Digit 4模型已在九个承诺客户部署中累计运行超过65,000小时,据其 filings 显示。自2024年6月以来,Digit 4机器人在物流公司GXO运营的佐治亚州设施中搬运了超过100,000个周转箱。
投资者对仍在主要进行试点工作的竞争对手给出了高得多的估值,包括Figure AI据报道达390亿美元,是Agility合并时估值的15倍以上。“Digit今天正在客户运营中做真正的工作,承担重复性、体力要求高的任务,”Agility首席执行官佩吉·约翰逊在一份声明中说。她表示,上市为公司提供了更多资源,将Digit带入更多工作场所。
约翰逊于2024年3月出任CEO,当时赫斯特从CTO转任首席机器人官。她此前经营Magic Leap,再之前在微软担任高管六年,在CEO萨提亚·纳德拉麾下负责业务开发和交易。在本周早些时候接受GeekWire采访时,赫斯特对上市持长远视角,称其是更漫长道路上的一个里程碑。他说,把上市当天当作成就,是金融人士的思维框架:“我不住在那个世界里。”
从CMU到OSU: 赫斯特在这条路上已经走了近三十年。1997年,他在匹兹堡卡内基梅隆大学读大一时建造了第一个行走机器人。他留在卡内基梅隆攻读机器人学博士学位,于2008年毕业,他说当时全世界获得该学位的人不到100人。那一年他加入科瓦利斯的俄勒冈州立大学,从一开始就计划最终将研究成果转化为公司。在俄勒冈州立大学的实验室里,赫斯特的团队建造了ATRIAS——一款双足机器人,外形不像任何动物,因为它基于行走和奔跑的数学模型。这是第一台复现人类行走动力学的机器。
2015年,他与卡内基梅隆同学达米恩·谢尔顿(后者担任CEO至2024年)以及他的俄勒冈州立大学学生米哈伊尔·琼斯一起,从实验室中创办了Agility。赫斯特说,身在俄勒冈从一开始就有帮助。Agility几乎是孤军奋战在尝试建造人形机器人,搬到俄勒冈加入公司的工程师们都致力于这一使命。公司几乎没有人员流动,不同于硅谷每六个月换一份工作的模式。但当员工超过400人时,它需要经验丰富的工程管理者,而俄勒冈没有足够大的机器人产业来培养这类人才。Agility一直在加州招聘这类人员并让他们飞往北方,于是在那里开设了办公室。如今Agility除俄勒冈外,在匹兹堡和加州弗里蒙特设有办公室。
对就业的影响: 目前,Agility将Digit定位为填补难以招人的岗位的方式。约翰逊本周告诉TechCrunch,公司专注于“人工的、高度重复性的工作,这些工作很难找到人来干”。赫斯特称该技术的首批商业应用是通往他所说梦想的“第一步”:“机器人成为人类日常生活的一部分,做人们不愿意做的事情。”当被问及被机器人取代的人会怎样时,赫斯特以农业为例说明自动化如何改善了人们的生活。“美国历史早期,90%的美国人从事农业,今天只有2%,”他说。“而这一切之所以成为现实,唯一的原因就是自动化。”
中文翻译:
位于俄勒冈州塞勒姆的Agility Robotics公司周四在纳斯达克上市,成为美国首家专注于人形机器人的上市公司——它对这项技术将如何改变工作提出了雄心勃勃的愿景,同时对实现这一愿景需要多长时间持谨慎态度。
“也许我们永远不会有超光速旅行。也许我们永远不会有《星际迷航》里的传送机,”该公司联合创始人兼首席机器人官乔纳森·赫斯特说。“但我们绝对会拥有能够完成人类所有体力工作的机器。”
不过,他告诫说,“这不会在明年或五年后实现。这确实是一项巨大的挑战。”
Agility通过与丘吉尔资本公司XI合并实现上市,这是一家由金融家迈克尔·克莱因领导的特殊目的收购公司(SPAC)。该交易在新投资之前对Agility的估值为25亿美元,并带来了约4.65亿美元的总收益。其股票在纳斯达克交易,代码为AGLT。
在最初跌破开盘价后,该股周四收于11.51美元,上涨2.2%,逆势而上——当天纳斯达克指数下跌5.5%,创下18个月来最差单日表现。
Agility的机器人:该公司的主角是Digit,一款双腿机器人,躯干顶部附近安装了两只手臂,使其能够高举料箱并从跌倒中恢复。它能在仓库和工厂中搬运周转箱和集装箱。
最新版本Digit 5于9月亮相。它身高5英尺11英寸(约1.8米),体重284磅(约129公斤),全身覆盖传感器,头部装有摄像头和天线,身体周围的安全防护壳内装有安全级传感器。Agility表示,它能举起50磅(约23公斤)的重物,触及7.2英尺(约2.2米)的高度,在24小时内工作超过20小时。
但最大的进步在于,Digit 5的设计目标是无需隔离屏障即可与人协同工作。早期版本必须与工人隔离开来,这限制了客户的使用场景。
Agility在亚马逊那里深刻认识到了这一点的重要性。亚马逊是该公司的早期投资者,曾测试Digit的先前版本在其仓库中搬运和堆叠料箱。
赫斯特说,Digit在那里做了有用的工作,但当Agility询问扩大规模时,亚马逊表示无法部署这些机器人。亚马逊向Agility展示了其自己的轮式机器人通过三到四年安全认证所需的条件,并表示不知道一款平衡机器人如何才能达到那个标准。
“这方面的安全论证难度超乎想象,”赫斯特说。“从来没有人部署过动态稳定平衡机器,这带来了新的风险。”
让机器人走出隔离屏障花了多年的工程和标准制定工作。Agility一直在联合编写该品类的ISO安全标准,并向保险公司和仓库运营商进行论证——这些人希望在将平衡机器放在人旁边之前看到数据证明。
Digit 5利用人工智能和多个传感器来监测附近的人员。根据具体情况,它会绕行、停下或坐下。视觉和音频提示会告知它即将做什么,当有人靠得太近时,安全控制器会触发其响应。
赫斯特将来自亚马逊的经验教训,以及来自GXO、丰田和舍弗勒的经验,归功于塑造了Digit 5的设计。
“亚马逊是推动我们走到那一步的极其重要的部分,”他说。
根据Agility的申报文件,亚马逊在2024年支付了一笔首付款,用于购买五台机器人,总价105万美元。Agility在2025年从亚马逊录得约104万美元的净销售额,占其全年总销售额的58%。赫斯特表示,他无法评论亚马逊目前是否仍是客户。
凭借安全性能的改进和其他进步,赫斯特称Digit 5是Agility的“规模化时刻”。“这就是那把钥匙,”他说。
根据其申报文件,一位Agility未透露名称的客户已承诺签订一份3亿美元、为期三年的合同,购买1,000台Digit 5机器人。这些机器人将以租赁而非出售的方式提供,该承诺受里程碑和规格条件约束,客户获得随着机器人部署而逐步归属的认股权证。
赫斯特说,Agility已经制造了约10台Digit 5机器人,首批原型机刚刚下线。其塞勒姆工厂的最终设计产能为每年多达10,000台机器人。
财务状况:Agility的业务规模仍然很小,去年收入为178万美元,今年前六个月约为17.4万美元。上半年亏损3.62亿美元,而去年同期亏损7,180万美元。股票薪酬——员工股权奖励的会计费用——同期增加了约2.52亿美元。
截至6月底交易完成前,Agility拥有1,470万美元现金,管理层在其中期财务报表中表示,对公司持续经营能力存在重大疑虑。
赫斯特称公开市场是他最大的不确定因素。
“我对我们的市场路径和技术路径完全有信心,”他说。如果公开市场股东认同,Agility将能够“成功地继续这一使命,并在需要时筹集资金”。如果他们怀疑公司,“那将让我们的日子难过得多”。
Agility相对于大多数竞争对手的优势在于,它的机器人正在做商业性工作。根据其申报文件,截至5月,其Digit 4型号已在九个已承诺的客户部署中累计运行超过65,000小时。自2024年6月以来,Digit 4机器人在物流公司GXO运营的佐治亚州设施中已搬运超过100,000个周转箱。
投资者对仍在主要进行试点工作的竞争对手给出了高得多的估值,包括Figure AI据报道达390亿美元,是Agility在合并中估值的15倍以上。
“Digit今天正在客户运营中做真正的工作,承担重复性、体力要求高的任务,”Agility首席执行官佩吉·约翰逊在一份声明中说。她说,上市为公司提供了更多资源,将Digit带入更多工作场所。
约翰逊于2024年3月出任CEO,当时赫斯特从CTO转任首席机器人官。她此前经营Magic Leap,再之前在微软担任高管六年,在CEO萨提亚·纳德拉手下负责业务开发和交易。
在本周早些时候接受GeekWire采访时,赫斯特对股票上市持长远眼光,称这是更长道路上的一座里程碑。他说,把上市当天当作成就,是金融人士的思维框架:“我只是不活在那个世界里。”
从CMU到OSU:赫斯特在这条路上已经走了近三十年。
1997年,他在匹兹堡卡内基梅隆大学读大一时就制造了了他的第一个步行机器人。他留在卡内基梅隆攻读机器人学博士学位,2008年毕业,他说当时全世界获得该学位的人不到100人。
那一年他加入了位于科瓦利斯的俄勒冈州立大学,从一开始就计划最终从自己的研究中剥离出一家公司。在俄勒冈州立大学的实验室里,赫斯特的团队建造了ATRIAS——一款双腿机器人,看起来不像任何动物,因为它是基于步行和奔跑的数学模型设计的。这是第一台再现人类行走动力学的机器。
2015年,他与卡内基梅隆同学、直至2024年担任CEO的达米恩·谢尔顿,以及他在俄勒冈州立大学的学生之一米哈伊尔·琼斯一起,从那里的实验室创立了Agility。
赫斯特说,从一开始在俄勒冈州就有所帮助。Agility几乎是唯一一家尝试建造人形机器人的公司,搬到俄勒冈州加入它的工程师们都致力于这一使命。公司几乎没有人员流动,不像硅谷那种每六个月就跳槽的模式。
但当员工超过400人时,它需要经验丰富的工程经理,而俄勒冈州没有足够大的机器人产业来培养这类人才。Agility一直在加州招聘这类人员并让他们飞往北方工作,因此在那里开设了办公室。如今Agility除了俄勒冈州,还在匹兹堡和加州弗里蒙特设有办公室。
对就业的影响:目前,Agility将Digit定位为填补难以配备人员的岗位的方式。约翰逊本周告诉TechCrunch,公司专注于“手动、高度重复、很难找到人来干的工作”。
赫斯特称这项技术的首批商业应用是迈向他所描述的理想——“机器人成为人类日常生活的一部分,做人们不愿做的事情”——的“第一步”。
当被问及被机器人取代工作的人会怎样时,赫斯特以农业为例,说明自动化如何改善了人们的生活。
“在美国历史早期,90%的美国人从事农业,而今天只有2%,”他说。“而这一切之所以成为现实,唯一的原因就是自动化。”
英文来源:
Agility Robotics of Salem, Ore., made its Nasdaq debut Thursday as the first publicly traded U.S. company devoted solely to humanoid robots — offering an ambitious vision for how the technology will change work, and a cautious view of how long it will take.
“Maybe we’ll never have faster-than-light travel. Maybe we’ll never have transporters from Star Trek,” said Jonathan Hurst, the company’s co-founder and chief robot officer. “But we absolutely will have machines that can do all of the physical work that a person can do.”
However, he cautioned, “It’s not going to be next year or five years from now. It’s really a grand challenge.”
Agility went public through a merger with Churchill Capital Corp XI, a special purpose acquisition company, or SPAC, led by financier Michael Klein. The deal valued Agility at $2.5 billion before the new investment and brought in about $465 million in gross proceeds. Its shares trade on the Nasdaq under the ticker AGLT.
After initially dipping below its opening price, the stock closed Thursday at $11.51, up 2.2%, bucking the trend on a day when the Nasdaq fell 5.5%, its worst day in 18 months.
Agility’s robot: The company’s main character is Digit, a two-legged robot with two arms mounted near the top of its torso that let it lift bins high and recover from falls. It moves totes and containers in warehouses and factories.
Digit 5, the latest version, was unveiled in September. Standing 5-foot-11 and weighing 284 pounds, it’s covered in sensors, with cameras and antennas in its head and safety-rated sensors in protective housings around its body. Agility says it can lift 50 pounds, reach 7.2 feet high and work more than 20 hours in a 24-hour day.
But the big advance is that Digit 5 was designed to work alongside people without a barrier. Earlier versions had to be fenced off from workers, which limited where customers could use them.
Agility learned how much that mattered at Amazon, an early investor in the company, which tested prior versions of Digit for moving and stacking bins in its warehouses.
Digit did useful work in tests there, Hurst said, but when Agility asked about scaling up, Amazon said it couldn’t deploy the robots. It showed Agility what it had taken to get its own wheeled robot through three to four years of safety certification, and said it didn’t know how a balancing robot would ever get there.
“The safety case to do this is unbelievably hard,” Hurst said. “Nobody has ever deployed a dynamically stable balancing machine, and it brings new risks.”
Getting a robot out from behind the barrier took years of engineering and standards work. Agility has been co-authoring ISO safety standards for the category and making its case to insurance companies and warehouse operators who want proof in numbers before putting a balancing machine next to a person.
Digit 5 uses AI and multiple sensors to watch for nearby people. Depending on the situation, it steers around them, stops or sits down. Visual and audio cues signal what it’s about to do, and a safety controller triggers its responses when someone gets too close.
Hurst credits the lessons from Amazon, along with ones from GXO, Toyota and Schaeffler, with shaping the design of Digit 5.
“Amazon was a huge, huge part of getting us to that point,” he said.
Amazon made an initial deposit in 2024 toward a $1.05 million purchase of five robots, according to Agility’s filings. Agility recorded about $1.04 million in net sales from Amazon in 2025, 58% of its total for the year. Hurst said he couldn’t speak to whether Amazon is a customer now.
With the safety improvements and other advancements, Hurst called Digit 5 “the scaling moment” for Agility. “That is the unlock,” he said.
One customer, which Agility has not named, has committed to a $300 million, three-year contract for 1,000 Digit 5 robots, according to its filings. The robots would be rented rather than sold, the commitment is subject to milestones and specifications, and the customer gets warrants that vest as the robots are deployed.
Agility has built about 10 Digit 5 robots, Hurst said, with the first prototypes just coming off the line. Its factory in Salem was built to eventually produce as many as 10,000 robots a year.
Financial picture: Agility’s business is still small, with revenue of $1.78 million last year and about $174,000 in the first six months of this year. It lost $362 million in the first half, up from $71.8 million a year earlier. Stock-based compensation, an accounting charge for employee equity awards, rose by about $252 million over the same period.
Agility had $14.7 million in cash at the end of June, before the deal closed, and management said in its interim financial statements that there was substantial doubt about the company’s ability to keep operating.
Hurst called the public markets his biggest unknown.
“I’m completely confident in our market path, our technological path,” he said. If public shareholders agree, he said, Agility will be able to “continue on this mission successfully and raise money when we need to.” If they doubt the company, “that’s going to make our life a lot harder.”
What Agility has that most rivals don’t is robots doing commercial work. As of May, its Digit 4 model had logged more than 65,000 hours across nine committed customer deployments, according to its filings. Digit 4 robots have moved more than 100,000 totes since June 2024 at a Georgia facility run by the logistics company GXO.
Investors have put far higher values on rivals still working largely in pilots, including Figure AI at a reported $39 billion, more than 15 times Agility’s valuation in the merger.
“Digit is doing real work in customer operations today, taking on repetitive, physically demanding tasks,” said Peggy Johnson, Agility’s CEO, in a statement. Going public, she said, gives the company more resources to put Digit into more workplaces.
Johnson became CEO in March 2024, when Hurst moved from CTO to chief robot officer. She previously ran Magic Leap, and before that spent six years as a Microsoft executive, leading business development and dealmaking under CEO Satya Nadella.
In an interview with GeekWire earlier in the week, Hurst took the long view on the stock listing, calling it a milestone on a much longer path. Treating a listing day as the achievement, he said, is a finance person’s frame: “I just don’t live in that world.”
From CMU to OSU: Hurst has been on that path for almost three decades.
He built his first walking robot in 1997, as a freshman at Carnegie Mellon University in Pittsburgh. He stayed at Carnegie Mellon for a Ph.D. in robotics, finishing in 2008, when fewer than 100 people in the world had earned one, he said.
That year he joined Oregon State University in Corvallis, planning from the start to eventually spin a company out of his research. In his Oregon State lab, Hurst’s team built ATRIAS, a two-legged robot that didn’t look like any animal because it was based on a mathematical model of walking and running. It was the first machine to reproduce the dynamics of human walking.
He started Agility in 2015 with Damion Shelton, a Carnegie Mellon classmate who was CEO until 2024, and Mikhail Jones, one of his Oregon State students, out of his lab there.
Being in Oregon helped from the start, Hurst said. Agility was almost alone in trying to build humanoid robots, and the engineers who moved to Oregon to join it were committed to that mission. The company had almost no turnover, unlike the Silicon Valley pattern of people jumping to a new job every six months.
But as it passed 400 employees, it needed experienced engineering managers, and Oregon had no robotics industry large enough to have produced them. Agility had been hiring those people in California and flying them north, so it opened an office there. Today Agility has offices in Pittsburgh and Fremont, Calif., in addition to Oregon.
Impact on jobs: For the time being, Agility pitches Digit as a way to fill jobs that are difficult to staff. Johnson told TechCrunch this week that the company has focused on “manual, very repetitive jobs, that it’s hard to find humans to do.”
Hurst calls the first commercial uses of the technology “step one” toward what he describes as the dream: “robots as part of everyday human life, doing the things people would rather not do.”
Asked what happens to the people whose work the robots replace, Hurst pointed to farming as an example of automation improving people’s lives.
“Early in American history, 90% of Americans worked in agriculture, and today 2% do,” he said. “And the only reason that’s true is because of automation.”
文章标题:敏捷机器人公司联合创始人谈纳斯达克上市、来自亚马逊的教训以及工作的未来
文章链接:https://news.qimuai.cn/?post=5495
本站文章均为原创,未经授权请勿用于任何商业用途