Agility Robotics
首个进入生产环境部署的商用人形机器人
Agility Robotics 做到了其他人形机器人公司还没做到的事:经验证、生产级、规模化商业部署。GXO Logistics 已搬运 100,000+ 个 tote,公司累计融资 $641M,并拥有目标年产 10,000 台的专用工厂,因此相对同行有 12–24 个月商业化领先。 核心投资问题是:Figure AI 估值 $39B、Tesla Optimus 拥有制造规模,竞争对手补上部署差距前,Agility 能否把先发优势转成持久市场位置。规模化单位经济尚未验证,客户集中度是真风险,协作安全认证仍是更广泛采用的技术闸门。
封面要素
公司概况
Agility Robotics 是一家位于 Oregon 州 Tangent 的机器人公司,设计、制造并部署 Digit——一款专为物流和仓储任务打造的双足人形机器人。公司 2015 年从 Oregon State University 分拆成立,开创了受食火鸡启发的仿生腿部设计,Digit 因此拥有标志性的节能步态。2024 年 6 月,公司与 GXO Logistics 签下全球首份人形机器人多年期商业机器人即服务(RaaS)协议,拿到关键行业里程碑。到 2025 年 11 月,Digit 在 GXO 位于 Georgia 州 Flowery Branch 的履约中心真实商业部署中已搬运超过 100,000 个周转箱,证明技术已达到生产级,而不只是展示。2025 年 3 月,Agility 完成 $400M Series C 融资,投后估值 $2.1B,投资方包括 WP Global Partners、SoftBank、Amazon Industrial Innovation Fund 和 NVIDIA 的 NVentures。公司正在 Oregon 州 Salem 建设 70,000-square-foot 制造设施 RoboFab,目标年产能 10,000 台。
- 成立时间
- 2015-01-01
- 创始人
- Jonathan Hurst, Damion Shelton
- 创立地点
- Corvallis, Oregon, USA
- 总部
- Tangent, Oregon, USA
- 产品
- Digit 是一款 5'9"、65 kg 的双足人形机器人,载荷 16 kg,电池续航 4+ 小时,并配有针对物流场景优化的末端执行器。它集成 3D LiDAR、立体摄像头和机载 AI,可在仓库环境中导航并搬运周转箱和包裹。Arc 云平台提供车队管理、AMR 集成(MiR、Zebra Technologies)以及 WMS/WES 连接。Digit 按功能安全(FSoE CAT1)标准运行,目前部署在安全围栏环境中,目标在 2026–2027 年拿到协作安全认证。
- 客户
- 一级第三方物流运营商(3PL)、电商履约中心和制造设施;这些客户在重复性物料搬运任务上面临用工短缺。
- 商业模式
- 混合模式:机器人即服务(RaaS)按每台 $10K–$15K/月计费,签多年期合同;直接资本销售约每台 $150K。Arc 云平台订阅收入(价格未披露)。
- 阶段
- Series C — commercial deployment phase
- 融资情况
- 已融资 $641M;2025 年 3 月 Series C $400M;投后估值 $2.1B。
执行摘要
主要优势
- 首个也是唯一已商业部署的人形机器人——在 GXO 实时履约流程中搬运 100,000+ 个 tote。
- 与 GXO Logistics(全球最大纯 3PL)签下多年 RaaS 协议,建立耐久锚定收入。
- 融资 $641M,战略投资人包括 Amazon IIF 和 NVIDIA NVentures,既给资金,也给生态入口。
- RoboFab 专用制造设施目标产能 10K 台 / 年,形成可防守的供应链。
- Cassie/Digit 仿生运动 IP 来自 OSU 十年研究;在物流环境中,步态功能更优。
- Arc 云平台提供 fleet management 和 AMR 集成,制造软件层切换成本。
主要风险
- 客户集中度:GXO 和 Amazon 占已知商业部署的大多数;任何一个锚定客户流失,都会造成重大损害。
- 单位经济在规模化下未验证:运营成本 $10–12/hr,对比目标 RaaS 价格 $30/hr;毛利率走到 50%+ 需要显著扩产。
- Tesla Optimus(目标价格 $20–30K)和 Figure AI(估值 $39B、Helix AI)构成竞争威胁——两者分别具备制造规模或 AI 优势。
- 协作安全认证延迟:拿到认证前,当前笼式隔离部署会限制可服务市场和客户类型。
- 所有收入和财务预测都是分析师估算;没有经审计财务;实际烧钱速度和现金跑道未披露。
未决问题
- 任何财年都没有经审计财务报表和可验证收入数据。
- RoboFab 单位成本结构和制造成本曲线未公开披露。
- GXO 和 Amazon 合同续约条款、价格、多站点扩张承诺未披露。
- 协作安全认证时间线和监管路径仍属推测。
- 专利组合范围,以及相对 Tesla、Figure AI、Boston Dynamics 的防御性 IP 位置尚未充分评估。
目录
01公司概况
1.1 身份与商业模式
Agility Robotics 为工业物流和仓库自动化设计、制造并部署双足人形机器人。公司的核心产品 Digit 被推向市场时定位为首个持续进入商业生产部署的人形机器人,这一定位有 GXO Logistics 和 Amazon 履约中心的已记录部署支撑。Agility 主要靠机器人即服务(RaaS)订阅模式销售 Digit 使用权,通常每台机器人每月收费 $10,000–$15,000,把硬件、维护、软件更新和车队管理打包在一起。公司也通过约每台 $150,000 的直接销售,以及 Agility Arc 云平台订阅获得收入;Arc 负责编排机器人车队,并与仓库管理系统集成。Agility 的公开使命是自动化仓库里“枯燥、肮脏、危险、重复”的任务,解决物流和供应链运营中长期存在的用工短缺。公司总部位于 Washington 州 Redmond,并运营位于 Oregon 州 Salem 的 RoboFab,这是一座 70,000 sq ft 制造设施。公司称 Digit 约 6,000 个零部件中约 80% 来自美国,体现其供应链韧性策略。截至 2026 年,Agility 称自己正在交付“经验证的成果”,并把 Digit 与 Agility Arc 软件平台一起定位为面向设施运营商的完整自动化方案。 [CO001, CO002, CO003, CO004, CO005, CO006]
1.2 创立历史与领导层
Agility Robotics 于 2015 年从 Oregon State University 的 Dynamic Robotics Laboratory 分拆成立,实验室由 Jonathan Hurst 教授领导,他多年研究腿式运动和双足机器人。联合创始人 Damion Shelton 自公司成立起担任 CEO 至 2024 年初,把公司从学术原型带到商业产品。2024 年 3 月,Microsoft 前总裁、Magic Leap 前 CEO Peggy Johnson 接任 CEO,为一家从研发转向大规模部署的公司带来企业级商业化经验。Hurst 现任首席机器人官,聚焦技术愿景;Shelton 则出任总裁兼董事长。其他关键高管包括 Melonee Wise(首席产品官,曾任 Fetch Robotics CTO/CEO)、Pras Velagapudi(CTO)、Aindrea Campbell(COO,曾任职 Apple 和 Zymergen)、Rich Bohne(首席商务官)和 Ana Lang(首席法务与人力官)。Hurst 与 Shelton 这对创始组合仍是公司身份的核心:深厚的学术机器人能力与商业野心结合在一起。截至 2026 年中,公开记录中没有有争议的创始人离职或治理纠纷,不过 2024 年的 CEO 交接是一次实质性组织变化。公司源自 OSU,使其在腿式运动研究上拥有被认可的研究传承。 [CO007, CO008, CO009, CO010, CO011, CO012]
| 姓名 | 职务 | 背景 / 曾任职位 | 创始人 | 关键人物依赖 |
|---|---|---|---|---|
| Peggy Johnson | CEO | 曾任 Microsoft 总裁;曾任 Magic Leap CEO | 否 | 高 – 外部引入的商业化支柱 |
| Damion Shelton | 总裁兼董事长 | 联合创始人兼 CEO,2015–2024 | 是 | 高 – 原始商业愿景承载者 |
| Jonathan Hurst | 首席机器人官 | OSU 教授;Dynamic Robotics Lab 创始人 | 是 | 高 – 技术 IP 与研究身份核心 |
| Melonee Wise | 首席产品官 | 曾任 Fetch Robotics CTO/CEO | 否 | 中 – 产品路线图负责人 |
| Pras Velagapudi | CTO | 机器人 AI 与部署专家 | 否 | 中 – AI / 软件平台负责人 |
| Aindrea Campbell | COO | Apple、Zymergen 运营背景 | 否 | 中 – 规模化运营执行 |
| Rich Bohne | 首席商务官 | 销售与 GTM 领导经验 | 否 | 中 – 收入增长 |
| Ana Lang | 首席法务与人事官 | 法务与 HR 领导经验 | 否 | 低 – 支持职能 |
创始人身份和关键人物依赖为研究判断;正式接班计划未公开披露。
[CO007, CO008, CO009, CO010, CO011]1.3 融资历史与投资方
Agility Robotics 多轮融资累计约 $641M。最重要的一轮是 2025 年 3 月完成的 $400M Series C,由私募股权机构旗下创投部门 WP Global Partners 领投,SoftBank Group 和 Amazon Industrial Innovation Fund 参投。本轮融资对应约 $1.75B 投前估值和约 $2.12B 投后估值。在 Series C 之前,公司于 2022 年完成 $150M Series B,估值约 $620M,由 Amazon Industrial Innovation Fund 领投,DCVC 和 Playground Global 参投。Amazon 的参与早于 2022 年:2021 年,它通过 Industrial Innovation Fund 投入一笔未披露金额,同时也是战略客户。Ford Motor Company 约在 2020 年较早的种子轮 / Series A 阶段投资约 $6M。Canon 和 Sony Innovation Fund 也被列为早期轮次投资方,反映硬件和影像公司对该赛道的战略兴趣。投资方组合值得注意:既有战略企业运营方(Amazon,同时也是大客户),也有大型综合基金(SoftBank、WP Global),还有使命契合的深科技投资方(DCVC、Playground Global)。SoftBank 参与 Series C 也有信号意义,因为此前有报道称 SoftBank 曾探索以超过 $900M 的估值收购 Agility。约 $641M 的已披露资本基本预先覆盖了 RoboFab 扩产和商业扩张,按当前烧钱估计,至少可支撑未来 18–24 个月。 [CO013, CO014, CO015, CO016, CO017, CO018]
| 投资方 / 利益相关方 | 角色 | 轮次 | 战略重要性 | 尽调问题 |
|---|---|---|---|---|
| WP Global Partners 投资方 | 领投方 | Series C 轮($400M,2025) | 主要财务支持方;私募股权关联方 | 核实持股比例和董事会权利 |
| SoftBank Group | 战略投资方 | Series C 轮(2025) | 大型科技投资方;曾探索收购 | 理解 SoftBank 的战略诉求和锁定风险 |
| Amazon Industrial Innovation Fund | 战略投资方兼客户 | Series B (2022)、Series C (2025) | 客户与投资人利益一致;关键部署合作伙伴 | 核查是否存在优惠条款、排他性或供给义务 |
| DCVC (Data Collective) | VC 投资人 | Series B (2022) | 深科技 VC;可带来组合协同 | 治理权利与按比例跟投权 |
| Playground Global | VC 投资人 | Series B (2022) | 硬件方向 VC | 董事会构成与投票权 |
| Ford Motor Company | 战略投资人 | Seed/Series A(约 2020) | 早期产业验证方;关注汽车场景 | 当前持股与持续参与度 |
| Canon / Sony Innovation Fund | 战略投资人 | 更早轮次 | 与影像和硬件行业方向契合 | 核实现有持股水平 |
持股比例未公开披露。Amazon 同时是投资人和主要客户,商业条款需要独立尽调。
[CO013, CO014, CO015, CO016, CO017]按时间梳理 Agility Robotics 从 2015 年创立到 2026 年 5 月的关键里程碑,覆盖创立、产品发布、融资事件、制造扩张和商业部署。
[CO007, CO013, CO015, CO022, CO028, CO030]展示 Agility Robotics 已知各轮累计融资,体现其从早期投资走到 2025 年 C 轮独角兽里程碑的过程。
早期 / 种子轮和 Amazon 2021 年金额为估计值或未披露。账面累计约 $647M;公司口径为累计融资约 $641M。
[CO013, CO014, CO015, CO016, CO017]1.4 封面指标与规模信号
截至 2026 年 5 月研究日,Agility Robotics 作为私人公司,大多数财务指标仍未披露。最可靠的已记录数字包括:投后估值约 $2.12B(2025 年 3 月 Series C);累计融资约 $641M;以及 2025 年估计年收入约 $35.5M,该数字来自第三方数据提供商(CompWorth、CBInsights),置信度较低。员工数估计在 250 到 500+ 之间,仅 Salem RoboFab 设施就以 500+ 为目标。公司未公开披露毛利率、ARR 或 RaaS 合同数量。公司称 Digit 已“进入生产部署”,且至少有两个具名企业客户(GXO Logistics 和 Amazon)。2025 年 11 月,GXO 部署搬运周转箱超过 100,000 次,形成一个有记录的运营吞吐里程碑。Agility 预计 RoboFab 完全建成后年产能为 10,000 台 Digit,但 2024 年和 2025 年的实际产量处在数百台而非数千台区间。考虑到制造扩张所需资本投入,公司收入预计尚未达到盈亏平衡。截至研究日,公开信息中没有披露债务融资或老股交易。 [CO020, CO021, CO022, CO023, CO024]
| 指标 | 数值 / 状态 | 日期 / 期间 | 置信度 | 缺口 / 备注 |
|---|---|---|---|---|
| 投后估值 | $2.12B | March 2025 | 高 | 据 The Information,Series C 轮条款清单 |
| 累计融资 | ~$641M | March 2025 | 高 | 各轮累计 |
| 估计年收入 | ~$35.5M | 2025 估计 | 低 | 第三方估计;未披露 |
| 员工数 | 250–500+ | 2025–2026 | 中 | 公司称 RoboFab 目标 500+ 人;总人数未披露 |
| 产能(RoboFab) | 10,000 units/yr | 目标 | 中 | 目标尚未达成 |
| 当前产量 | 数百台 | 2024–2025 | 中 | 爬坡阶段,未量产 |
| 已具名企业客户 | 2 (GXO, Amazon) | 2026 | 高 | 公开确认的商业部署 |
| GXO 周转箱里程碑 | 100,000+ totes | Nov 2025 | 高 | 公司新闻稿;GXO 合作 |
| 定价(RaaS) | $10K–$15K/mo per unit 价格 | 2025 | 中 | 据报道;未正式公布 |
| 定价(直接销售) | ~$150,000/unit | 2025 | 中 | 据报道;未正式公布 |
收入、员工数和产量估计来自第三方数据提供商(CBInsights、CompWorth)及行业报道;应按近似值处理。
[CO001, CO002, CO020, CO021, CO022, CO023]截至 2026 年 5 月,Agility Robotics 的关键快照指标,覆盖估值、融资、客户数、产能规模和产品部署状态。
收入和员工数为第三方估计。定价来自报道,Agility 未官方确认。
[CO001, CO013, CO020, CO022, CO023]1.5 里程碑与公司时间线
Agility Robotics 的历史可分为五个阶段:学术起源(2015–2017)、早期商业化(2018–2020)、资本形成(2021–2022)、制造建设(2023–2024)和商业扩张(2025–2026)。公司 2015 年创立于 Oregon State University;第一个重要外部机器人 Cassie 于 2017 年发布,并卖给研究实验室。Digit 平台约在 2019–2020 年首次推出,是一款能力更强、带手臂和操作能力的双足人形机器人。2022 年的 $150M Series B 支撑公司转向商业部署。2024 年 6 月,公司与 GXO Logistics 签下多年期 RaaS 协议,被称为行业首例;随后在 2025 年 11 月达成搬运 100,000 个周转箱的里程碑。2025 年 3 月的 $400M Series C 是公司最大单笔融资事件。位于 Salem 的 RoboFab 于 2023 年 9 月公开宣布,2024 年开业,并正在爬坡至满产。2024 年 3 月,Peggy Johnson 加入并出任 CEO。截至 2026 年,Agility 把“协作安全”版本 Digit 作为下一个重大产品里程碑,目标是让机器人无需安全围栏即可与人类协作;计划 2026–2027 年推出,并配备载荷 50-lb 的第五代机型。 [CO025, CO026, CO027, CO028, CO029, CO030]
| 日期 | 事件 | 类型 | 金额 / 估值 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2015 | Agility Robotics 作为 OSU 衍生公司成立 | 创立 | Jonathan Hurst, Damion Shelton | OSU Dynamic Robotics Lab 学术 IP 开始商业化 | |
| 2016 | CASSIE 双足机器人原型发布 | 产品 | 内部 | 首次展示仿鸟双足运动平台 | |
| 2019 | 第一代 Digit 发布 | 产品 | 内部 | 具备手臂和操作能力的双足人形机器人推出 | |
| 2020 | Ford Motor Company 投资约 $6M | 融资 | $6M | Ford Motor Company | 首个大型汽车战略投资人验证物流用例 |
| 2021 | Amazon Industrial Innovation Fund 投资 | 融资 | 未披露 | Amazon IIF | 建立投资人兼客户的双重关系 |
| 2022-04 | 完成 $150M Series B 融资 | 融资 | 以约 $620M 估值融资 $150M | Amazon IIF、DCVC、Playground Global 等投资方 | 大额资金注入,支撑商业产品开发 |
| 2022 | 第三代 Digit(带手臂)发布 | 产品 | 内部 | 增加灵巧操作能力;商业化准备度提升 | |
| 2023-09 | 宣布在 Salem, Oregon 建设 RoboFab 工厂 | 规模化 | 70,000 sq ft;10,000 units/yr 目标 | Agility Robotics | 全球首座人形机器人专用工厂 |
| 2023 | Amazon 在 Sumner, WA 履约中心试点部署 | 合作 | Amazon | 首个大型企业试点;周转箱整合用例 | |
| 2024-03 | Peggy Johnson 出任 CEO | 治理 | 内部 | 领导层切换为以商业化为重心的 CEO | |
| 2024 | RoboFab 工厂开业;开始生产 | 规模化 | 内部 | 商业化生产爬坡;向首批合作伙伴计划交付 | |
| 2024-06 | 签署 GXO Logistics 多年期 RaaS 协议 | 合作 | GXO Logistics | 行业内首个多年期人形机器人部署协议 | |
| 2025-03 | 完成 $400M Series C 融资 | 融资 | 以约 $2.12B 投后估值融资 $400M | WP Global Partners、SoftBank、Amazon IIF 等投资方 | 最大一轮融资;估值进入独角兽以上阶段 |
| 2025-11 | GXO 设施达到 100,000 个周转箱里程碑 | 规模化 | GXO Logistics | 在真实仓库环境验证工业级吞吐量 | |
| 2026 | 以协作安全版 Digit 为目标(无需围栏作业) | 产品 | 内部 | 下一代能力可支持人机并肩部署 |
日期来自公开公告和媒体报道。Ford $6M 投资金额来自报道;公司未公开确认。
[CO025, CO026, CO027, CO028, CO029, CO030]1.6 图表与证据
02市场分析
2.1 市场定义与边界
Agility Robotics 竞争所在市场是人形机器人,具体落在物流和仓库自动化子赛道。公司的产品 Digit 面向人类尺度环境中的体力重复任务:搬运周转箱、装载传送带、堆叠容器,并支持最后一段履约流程。因此 Agility 处在两个更大市场的交叉点:(1)全球人形机器人市场;(2)更广义的仓库自动化市场。仓库自动化市场指所有部署在履约设施内、用于自动化仓储、分拣、拣选、包装和运输的软硬件。ResearchAndMarkets 估计该市场 2025 年规模约 $29.9B,到 2030 年增长至 $63.4B,CAGR 为 16.2%。其中,物流机器人硬件(AMR、AGV、人形机器人、拣选机械臂)按 Grand View Research 口径,2024 年是一个 $14.5B 细分市场。对 Digit 更相关的竞争边界是人形机器人市场本身;MarketsandMarkets 估计 2025 年全球规模约 $2.9B,到 2030 年增长至约 $15.3B,CAGR 为 39.2%。关键替代方案和相邻方案包括自主移动机器人(AMR)、机器人拣选臂、固定传送系统、传统用工中介和外包仓库劳动力。人形机器人的差异在于,它们可在非结构化、按人类尺度设计的环境中运行,无需改造设施;对无法承担或无法论证现有仓库布局改造的客户来说,这是核心卖点。 [CM001, CM002, CM003, CM004]
| 市场层级 | 定义 | 2025 年规模估计 | 2030 年预测 | CAGR | 对 Agility 的意义 |
|---|---|---|---|---|---|
| 仓储自动化(广义) | 仓库存储、分拣、拣选、包装、运输所需的全部硬件 / 软件 | $29.9B | $63.4B | 16.2% | 外层边界;包括非人形自动化 |
| 物流机器人市场 | 用于物流的机器人硬件,包括 AMR、AGV、人形机器人 | $14.5B (2024) | $35.1B | 15.9% | 竞争集合包括 AMR 和专用机械臂 |
| 人形机器人市场(全球) | 各垂直行业的人形机器人硬件和服务 | $2.9B | $15.3B | 39.2% | 直接 TAM;Digit 在此竞争 |
| 仓储 / 物流机器人市场 | 专门用于仓储和物流的机器人系统 | $7.6B | ~$20B+ | 26%+ | 最接近 Agility 可商业触达市场 |
| 人形物流 / 仓储 SAM | 可触达地区面向物流 / 仓储的人形机器人 | $580M–$870M | $2.3B–$4.6B | ~40% | Agility 在人形机器人市场中的主要 SAM |
市场规模估计来自 MarketsandMarkets、Grand View Research、ResearchAndMarkets 和 BCC Research。多套方法给出一个区间;单项估计应按方向性参考。
[CM001, CM002, CM003, CM004, CM005]展示从广义仓库自动化 TAM 到 Agility Robotics 2025 年 SOM 的市场规模层级,说明每一层如何收窄可服务机会。
所有数值均为多家分析机构研究估计,置信区间很宽。金字塔层级并非严格相加的子集;它们代表不同定义边界。
[CM001, CM002, CM003, CM005, CM006, CM007]展示多家分析机构对 2025 年和 2030 年人形机器人市场规模的估计区间,反映预测圈分歧显著。
中点和区间来自不同方法的分析报告研究。数字应视为方向性判断,而非精确预测。
[CM001, CM002, CM003, CM004]2.2 TAM、SAM 与 SOM 测算
物流人形机器人的 TAM 最适合用三种口径理解。口径 1(自上而下的人形机器人市场):全球人形机器人市场 2025 年估计为 $2.9B(MarketsandMarkets),物流 / 仓储是两大主要垂直之一。若物流占人形机器人支出的 20–30%,则物流专用人形机器人 TAM 在 2025 年为 $580M–$870M,到 2030 年增长至 $2.3B–$4.6B。口径 2(劳动力替代):美国仓库劳动力市场雇用约 1.8M 名工人,每名工人每年全负荷成本约 $42K–$55K,形成约 $75B–$99B 年度劳动力支出。即便到 2030 年只有 5% 被人形机器人替代,也意味着 $3.7B–$5.0B 的年度市场机会。口径 3(仓库自动化自下而上):在 2025 年 $29.9B 的仓库自动化 TAM 中,人形机器人约占当前支出的 3–5%,但预计到 2030 年将升至 $63.4B 市场的 10–15%,意味着十年末人形仓库自动化机会为 $6.3B–$9.5B。Agility 2025 年的 SAM 明显更小:美国和西欧愿意签多年期 RaaS 合同、设施面积 ≥250,000 sq ft 的第三方物流(3PL)和大型电商履约运营商。该 SAM 估计为 $800M–$1.5B。Agility 2025 年的 SOM 受制于产能(数百到低千台)和活跃销售管线,按已披露部署和估计 RaaS 合同价值,大概率为 $20M–$100M。 [CM005, CM006, CM007, CM008, CM009, CM010]
| 视角 | 方法 | 2025 年估计 | 2030 年估计 | 关键假设 | 置信度 |
|---|---|---|---|---|---|
| 视角 1:自上而下的人形机器人市场 | 全球人形机器人 TAM × 物流占比(20–30%) | $580M–$870M | $2.3B–$4.6B | 物流 = 人形机器人支出的 20–30% | 中 |
| 视角 2:劳动力替代 | 美国仓储劳动力支出 × 人形机器人替代率 | $3.7B–$5.0B(5% 替代) | $7.5B–$15B(10–20% 替代) | 美国仓储劳动力年支出约 $75–99B | 低 |
| 视角 3:仓储自动化占比 | 仓储自动化 TAM × 人形机器人占比(3–15%) | $900M–$1.5B | $6.3B–$9.5B | 人形机器人占 WAM 的比例从 3% 增至 10–15% | 中 |
| Agility SAM(2025) | 愿意签署 RaaS 的美国 / 欧洲 3PL 与大型电商 | $800M–$1.5B | $4B–$8B | 最小设施规模、RaaS 意愿、地域 | 低 |
| Agility SOM(2025) | 活跃部署数 × RaaS 合同价值 | $20M–$100M | 取决于产能爬坡 | 受产能约束;基于已知客户 | 低 |
所有估计均来自研究汇总。视角 2(劳动力替代)是理论上限,不是短期可触达市场。Agility SOM 还受 RoboFab 产能爬坡进一步约束。
[CM005, CM006, CM007, CM008]2.3 买方与细分客户格局
Digit 的主要买方是大型物流运营商、电商履约公司和合同制造商中的运营 / 供应链高管。仓库自动化预算通常归运营或资本开支 / 自动化部门管理;涉及数百万美元、多年期 RaaS 承诺时,CFO 和高管层会参与。可以识别出三个主要买方细分。细分 1——一级 3PL 运营商(如 GXO、DHL、Ryder、XPO):拥有大型、标准化设施,周转箱 / 纸箱搬运流程劳动密集,采购动机是降低人工成本和提升吞吐,同时具备部署人形机器人的财务能力和运营成熟度。细分 2——大型电商运营商(如 Amazon、Walmart、Target):兼具高货量、极高劳动强度和内部自动化能力;它们是早期采用者,但也会自研竞争性机器人方案。细分 3——专业制造商和零售商(如服装、汽车零部件):在难以用固定设备自动化的人类尺度环境中处理高混合、低批量流程。RaaS 模式的付款方是设施运营商,使用者是一线仓库员工。企业采用周期较长(销售周期 12–18+ 个月,集成 3–6 个月),原因包括安全认证要求、WMS 集成复杂度,以及在扩大车队前必须证明 ROI。较小运营商(收入 <$50M)因最低合同经济性限制,基本不在 Agility 当前 SAM 内。 [CM011, CM012, CM013, CM014, CM015]
| 细分市场 | 买方类型 | 核心痛点 | 预算所有者 | 付费意愿 | 采用阶段 |
|---|---|---|---|---|---|
| Tier-1 3PL 运营商 | 运营 VP/SVP | 劳动力短缺 + 吞吐量 | COO / CFO | 高(多年期 RaaS) | 早期采用者 — GXO 于 2024 年签约 |
| 大型电商履约 | 自动化总监 / 供应链 VP | 订单量增长 + 劳动力成本 | 内部自动化团队 / CFO | 高(先试点,再扩张) | 试点阶段 — Amazon 自 2023 年起测试 |
| 专业制造商 | 工厂经理 / 制造 VP | 多品种、不规则任务 | 运营 / CapEx 预算 | 中(ROI 需要验证) | 早期采用偏后;需要证明用例成立 |
| 中型零售商 / 配送中心运营商 | 运营总监 | 用工可靠性 | CFO / 运营预算 | 中低(CapEx 受限) | 采用滞后;RaaS 有帮助,但销售周期长 |
| 政府 / 国防 | 设施经理 | 危险环境自动化 | 政府采购 | 中高(任务导向) | 萌芽期;不是 Agility 2025 年主要目标 |
细分市场分类基于公开可观察的部署模式和分析师研究。买方画像是概括性原型。
[CM011, CM012, CM013, CM014]展示物流运营商从认知到全面部署 Digit 机队经历的阶段,并突出压缩漏斗的关键摩擦点。
漏斗数量为研究估计,用于说明转化动态;并非基于 Agility 披露的销售管线数据。
[CM011, CM012, CM013, CM014, CM015]2.4 增长驱动与采用顺风
四个主要驱动支撑人形机器人物流市场增长。驱动 1——结构性用工短缺:美国仓库和物流就业长期面临结构性短缺;BLS 数据显示,2024–2025 年即便工资高于 $20/hr,仓库岗位空缺也经常超过 500,000。人口老龄化和替代就业机会增加,使可服务劳动力池持续收缩。驱动 2——电商量增长:美国电商销售在 2024–2025 年期间年增速约 10%,订单量和履约复杂度同步上升。当日达和次日达预期要求设施以更高吞吐、更高灵活性运行。驱动 3——资本可得性:截至 2025 年,全球人形机器人公司累计获得超过 $10B 私募融资,主要投资方包括 SoftBank、Microsoft、Amazon 和 NVIDIA。这些资本支撑持续 R&D 和制造扩张。驱动 4——技术成熟度拐点:强化学习、大型视觉-语言-动作(VLA)模型和储能进步,已在 2024–2025 年把人形机器人能力从实验室演示推到持续商业部署;Digit 在 GXO 搬运 100K+ 个周转箱就是证据。此外,监管顺风包括 OSHA 越来越认可仓库场景中的人体工学伤害成本,给体力要求高的任务自动化带来制度压力。 [CM016, CM017, CM018, CM019, CM020]
| 因素 | 类型 | 对 Agility 的影响 | 时间范围 | 证据质量 |
|---|---|---|---|---|
| 仓储劳动力结构性短缺 | 驱动因素 | 高 — Digit 核心需求逻辑 | 当前(2025–2030) | 高 — BLS 数据、运营商新闻稿 |
| 电商订单量增长(约 10% YoY) | 驱动因素 | 高 — 拉动吞吐量需求 | 当前(2025–2030) | 高 — US Commerce Dept. 数据 |
| 私募资本流入(累计 $10B+) | 驱动因素 | 中 — 也会赋能竞争对手 | 当前 | 高 — 融资数据库 |
| AI/ML 能力提升 | 驱动因素 | 高 — 支持更多任务类型 | 2025–2028 | 中 — 部署中可观察 |
| 单台成本高(RaaS $10K–$15K/mo) | 约束 | 高 — 限制 TAM 可触达性 | 当前 | 中 — 第三方定价数据 |
| 企业销售周期长(12–18 mo) | 约束 | 高 — 拖慢收入爬坡 | 当前 | 中 — 从交易时间线观察 |
| WMS 集成复杂 | 约束 | 中 — 部署摩擦 | 当前 | 中 — 客户反馈 |
| 安全围栏要求(尚无协作安全能力) | 约束 | 高 — 限制部署环境 | 2025–2027 | 高 — 公司表述 |
| 竞争对手能力快速提升 | 约束 | 高 — 价格 / 性能竞争 | 2025–2028 | 高 — 市场报告 |
| 监管 / OSHA 人体工学标准 | 驱动因素 | 中 — 创造自动化激励 | 中期 | 中 — 监管趋势数据 |
增长驱动因素和约束评估基于研究汇总与分析师报告。影响评级为定性评估。
[CM016, CM017, CM018, CM019, CM020, CM021]2.5 采用约束与市场风险
多个结构性约束限制物流场景中人形机器人的近期采用。约束 1——单位经济性:以每台 Digit $10,000–$15,000/月(RaaS)的价格,与人工(全负荷约 $22–28/hr)相比达到盈亏平衡,要求机器人接近连续运行且停机极少。早期商业部署中的硬件可靠性和服务成本仍不确定。Interact Analysis 估计,到 2027 年,物流人形机器人实际市场渗透率仍会非常有限,2028 年后才会出现有意义的规模。约束 2——集成复杂度:仓库管理系统(WMS)集成、安全认证、设施测绘和工人共存规程,需要 3–6 个月上线时间和大量技术资源。约束 3——安全与监管:截至 2025–2026 年,所有 Digit 部署都在围栏或隔离区域运行,因为机器人尚未取得可在无限制人机共作环境中使用的完整协作安全认证。取得“协作安全”状态,是大多数大规模部署的必要前提。约束 4——技术集中与切换成本:早期多年期 RaaS 合同会把运营商锁进特定平台技术,形成有利于现有厂商(Agility)的切换成本;但当技术被认为尚不成熟时,也可能阻碍采用。约束 5——竞争强度:资金充足的竞争对手(Tesla Optimus、Figure AI、Boston Dynamics、1X Technologies)正在开发竞争平台,可能通过更低价格或更强能力侵蚀 Agility 的先发优势。 [CM021, CM022, CM023, CM024, CM025]
从两个维度映射物流人形机器人的关键采用约束:对 Agility 增长的影响严重度,以及估计解决时间。
[CM021, CM022, CM023, CM024, CM025]2.6 图表与证据
03竞争格局
3.1 竞争格局概览
Agility Robotics 所处的商用人形机器人市场仍处早期,但已有六个有实质意义的竞争对手,覆盖不同融资规模、技术路线和商业野心。格局可分为三层。第一层是资金充足、短期内瞄准同一物流和制造垂直的竞争者:Figure AI(2025 年 9 月估值 $39B)、Apptronik Apollo(估值约 $5B,Samsung 投资)和 1X Technologies NEO(寻求 $10B 估值,OpenAI 支持)。第二层是拥有相邻技术优势的平台型现有玩家:Tesla Optimus(背靠 Tesla 的垂直一体化制造和 FSD AI 栈)以及 Boston Dynamics Atlas(归 Hyundai Motor Group 所有,56 DOF、IP67 等级)。第三层是低成本中国进入者:Unitree G1(起售价 $16K),有能力侵蚀研究和试点细分。关键差异事实是,截至 2026 年中,Agility 是唯一一家公开记录显示已有多年期、商业规模 RaaS 部署的竞争者。这一部署领先带来运营数据、客户背书和流程集成锁定,而竞争对手尚未复刻。不过商业窗口有限:Figure AI、Apptronik 和 1X 都把商业生产目标放在 2026–2027 年,Tesla 也释放信号称 Optimus 可能在 2026–2027 年时间段对外可用。Agility 已经建立的竞争护城河真实存在,但有时间边界;防御力高度取决于客户获取、产能爬坡和 AI 能力提升的执行速度。 [CP001, CP004, CP017, CP022, CP033]
用序数四象限图从两个维度比较人形机器人竞争对手:商业部署成熟度(x 轴,0–10)和 AI 能力广度(y 轴,0–10)。Agility 在商业部署上领先,Figure AI 在 AI 能力广度上领先。Tesla 和 Boston Dynamics 属于专项 / 自用玩家;1X 和 Apptronik 仍在冒头。
评分为序数估计,来自分析报告、产品文档和部署证据。商业部署成熟度:10=企业规模、持续多年商业 RaaS;0=无商业部署。AI 能力广度:10=通用多任务 VLA 模型;0=仅遥操作。
[CP001, CP003, CP007, CP009, CP017, CP022]3.2 直接竞争者画像
Figure AI(2022 年创立,位于 California 州 Sunnyvale)是知名度最高、资本最充足的纯人形机器人竞争者。其 B2 机器人高 5'6",重 60kg,载荷 20kg,步行速度 1.2 m/s,电池周期 5 小时。Figure AI 的 Helix 视觉-语言-动作模型代表了真正差异化的 AI 能力,可从人类示范中学习通用任务;Agility 当前 AI 栈在通用性上无法匹配。Figure 于 2025 年 9 月以 $39B 投后估值完成超过 $1B 的 Series C。不过,其 BMW 工厂试点——公司的第一个有实质意义的商业测试——在 2025 年初中止,原因是发现明显依赖遥操作;截至 2026 年中,Figure AI 尚未达到与 Agility 可比的商业规模。Tesla Optimus Gen 2 是战略变数:机器人高 5'8",重 57kg,载荷 20kg,一旦商业化,目标单价为 $20K–$30K。截至 2026 年,Tesla 仅在自有制造设施内部部署 Optimus,没有对外商业销售。Boston Dynamics electric Atlas(2024)是身体能力最强的平台,高 6'2",重 90kg,载荷 50kg,但单价 $150K+,重点是 Hyundai/Kia 汽车装配,而非物流。挪威创业公司 1X Technologies 的 NEO 提供 $20K 直接销售 / 每月 $499 订阅模式,承载能力 25kg;它完成由 OpenAI 和 EQT Ventures 支持的 $125M Series B,据报道正以 $10B 估值寻求 $1B 融资。Apptronik Apollo 是一款 5'9"、72kg、25kg 载荷的平台,拥有 NASA 合作、Samsung 投资,以及开放开发者平台路线。Unitree G1 起价约 $16K,远低于所有西方竞争者,主要瞄准研究和试点预算。 [CP001, CP002, CP003, CP005, CP006, CP007]
| 竞争对手 | 类别 | 融资 / 估值 | 目标细分市场 | 关键差异化 | 主要限制 |
|---|---|---|---|---|---|
| Agility Robotics (Digit) | 直接 — 商用人形机器人 | 已融资 $641M / 估值约 $2.1B | 物流、3PL、仓库 | 唯一商业规模 RaaS 部署;Arc WMS 集成;RoboFab 产能爬坡 | 载荷最低(16kg);仅限围栏隔离;AI 任务范围窄 |
| Figure AI (B2 + Helix) | 直接 — 商用人形机器人 | 已融资 $1B+ / 估值 $39B | 制造、通用工业 | 面向通用任务的 Helix VLA 模型;融资最多的纯人形机器人公司;五指手 | BMW 之后没有持续商业部署;BMW 试点因远程操作短板结束 |
| Tesla Optimus (Gen 2) | 直接 — 工业人形机器人 | 内部资金(Tesla 资产负债表);目标价格约 $20–30K | Tesla 内部工厂使用(尚未对外销售) | FSD AI 栈;Tesla 制造规模;最低目标单价 | 未对外商业供应;仅限 Tesla 工厂自用 |
| Boston Dynamics Atlas(电动版) | 相邻 — 汽车 / 工业人形机器人 | Hyundai 子公司(无独立融资) | 汽车装配(Hyundai/Kia) | 同类最佳敏捷性(56 DOF);50kg 载荷;IP67 等级 | 成本最高($150K+);不瞄准物流;锁定在 Hyundai 生态 |
| 1X Technologies (NEO) | 直接 — 家用 / 轻工业人形机器人 | 已融资 $125M Series B / 寻求约 $10B 估值 | 主要家用;工业合作次要 | 最低商业价格($20K / 每月 $499);OpenAI 支持;25kg 载荷 | 家用优先降低短期物流竞争;商业化早期 |
| Apptronik (Apollo) | 直接 — 制造 / 物流人形机器人 | 已融资约 $935M / 估值约 $5B | 制造、物流、太空 | 开放开发者平台;NASA 合作;Samsung 投资;电池快换设计 | 灵巧性不及 Figure AI;未披露商业定价;公开部署证据有限 |
| Unitree (G1/H1) | 相邻 — 研究 / 试点人形机器人 | 中国本土资本;未披露全球融资 | 研究、学术试点、轻工业 | 价格最低($16K);制造效率强;高产量潜力 | AI 成熟度较低;出口限制风险;主要为研究级 |
画像覆盖直接竞争对手(人形机器人)和相邻竞争对手(近人形机器人)。数据来自公司官网、融资公告和第三方分析师报告。估值数字为最近披露轮次的投后估值估计。
[CP001, CP002, CP005, CP006, CP007, CP008]| 能力 | Agility Robotics | Figure AI B2 | Tesla Optimus | Boston Dynamics Atlas | 1X NEO | Apptronik Apollo |
|---|---|---|---|---|---|---|
| 商业部署 | 生产部署(GXO、Amazon 多年 RaaS) | 试点结束(BMW 试用已终止) | 仅内部使用(Tesla 工厂) | 研发 / Hyundai 内部 | 商业化前 | 商业化前 |
| 载荷能力 | 16 kg | 20 kg | 20 kg | 50 kg(持续 30 kg) | 25 kg | 25 kg |
| 电池 / 运行时长 | 4+ hours | 5 hours | Unknown | 作业周期有限 | Unknown | 4 hours + 热插拔 |
| AI / 任务学习 | 特定任务(面向物流优化) | Helix VLA —— 通用多任务 | 源自 FSD 的 AI 栈 | 基于行为 + 手动编程 | 示范学习 | 第三方 AI 集成(开放平台) |
| 安全认证 | FSoE CAT1;围栏隔离 | 未知 —— 尚无商业认证 | 未知 —— 尚未商业部署 | IP67;敏捷度高但无商业安全认证 | Unknown | Unknown |
| 行走速度 | 1.5 m/s | 1.2 m/s | 8 km/h(约 2.2 m/s) | 敏捷度高;具体规格未披露 | Unknown | 4 km/h(约 1.1 m/s) |
能力评估基于公开产品规格、公司公告和独立分析师评论。标记为“未知”的单元格表示没有可靠公开证据。商业部署反映 mid-2026 研究日期时的状态。
[CP002, CP003, CP004, CP005, CP008, CP010]3.3 能力与功能差异化
从整个竞争场看,物流客户最看重的能力维度是载荷、电池续航、任务灵活性、安全认证和商业集成成熟度。Agility 的 Digit 载荷为 16kg,在主要竞争者中最低;Figure AI B2、Tesla Optimus、1X NEO 和 Apptronik Apollo 都把载荷目标定在 20–25kg。Digit 聚焦周转箱物流流程,而 16kg 已超过实际需求,这缓和了载荷差距,但规格弱点仍然存在。电池续航方面,Digit 的 4+ 小时与 1X NEO 的竞争性参数相当;Boston Dynamics Atlas 的运行周期最受限制。Agility 的关键能力差异来自部署栈:Agility Arc 平台、3D LiDAR + 立体摄像头传感器融合、WMS 集成能力,以及 FSoE CAT1 功能安全认证。没有竞争者在第三方物流环境中取得同等商业规模部署认证。AI 能力方面,Figure AI 的 Helix VLA 模型被认为技术上优于 Agility 当前方案,因为它可从示范中支持通用多任务操作,而 Digit 的 AI 栈针对范围更窄的物流任务优化。如果 Tesla 决定投入人形机器人专用训练管线,源自 FSD 的 AI 栈会给 Optimus 长期 AI 开发优势。Apptronik 的开放开发者平台差异化在生态扩张,而非核心 AI 能力。 [CP002, CP003, CP008, CP015, CP020, CP026]
矩阵比较六家人形机器人竞争对手的六项能力。绿色单元格表示竞争强项;琥珀色表示部分能力;红色或未知表示缺口。Agility 在商业部署和安全认证上领先;Figure AI 和 Tesla 在 AI 广度上领先;Boston Dynamics 在载荷和机动性上领先。
能力评级为定性评估,基于公开产品规格和分析报告。标为「未知」的单元格表示缺乏可靠公开证据,并不代表已确认没有该能力。
[CP002, CP003, CP004, CP008, CP013, CP015]3.4 定价模式与市场进入对比
截至 2026 年中,人形机器人市场有三类清晰的定价和市场进入范式。Agility 的 RaaS 模式($10K–$15K/月,或约 $30/hr)把硬件、软件、维护和车队管理打包在一起——为客户把资本开支转成运营开支,也为 Agility 创造经常性收入。该模式让 Agility 与企业客户激励一致,因为 Agility 保留硬件所有权,并在运行时间和可靠性上有财务利益。直接销售选项($150K/台)存在,但不是主要商业动作。截至 2026 年中,Figure AI 尚未公开披露商业定价模式,不过行业报告称其可能采用类似 RaaS 路径。Boston Dynamics 每台 $150K+ 的 Atlas 面向 Hyundai 内部资本预算,预计不会达到 RaaS 或大众市场价位。1X Technologies 的双轨定价——$20K 直接销售或 $499/月——显著低于 Agility,但瞄准不同客户细分(家庭 / 轻工业,而非企业物流)。Unitree G1 起价 $16K,从绝对价格上压低所有竞争者,瞄准学术和试点客户。Tesla Optimus 的目标单价为 $20K–$30K,但时间表尚未兑现。Apptronik 尚未公开发布价格。最关键的市场进入差异是,Agility 已证明能打穿一级运营商(GXO、Amazon)的企业销售周期,而竞争对手仍处于商业化前或早期商业阶段。 [CP006, CP009, CP010, CP018, CP022, CP024]
| 竞争对手 | 模式类型 | 单机 / 订阅价格 | 包含内容 | 折扣 / 未知项 | 战略含义 |
|---|---|---|---|---|---|
| Agility Robotics | RaaS(主);直销(次) | $10K–$15K/month 或 ~$30/hr;直销 $150K | 硬件、维护、软件更新、Arc 车队管理、WMS 集成支持 | 多机部署可能有批量折扣;未披露 | RaaS 模式带来经常性收入和留存;硬件所有权让激励更一致 |
| Figure AI | 未知 —— 未披露商业定价 | 未披露 | Unknown | Unknown | 高端 AI 定位可能支撑更高单机价格;暂无商业价格证据 |
| Tesla Optimus | 直销(目标) | $20K–$30K/台(目标;尚未供应) | 仅硬件(假设);软件来自 Tesla 生态 | 目前无商业销售;定价仍属目标 | 若实现,按 ~14–18 个月 TCO 口径会压低 Agility RaaS 价格 |
| Boston Dynamics Atlas | 直销 | $150K+/台 | 硬件;Spot 生态集成;Hyundai 专属定制 | 不面向外部物流运营商 | 对 Agility 不构成定价威胁;垂直场景和成本结构不同 |
| 1X Technologies NEO | 双模式:直销 + 订阅 | 直销 $20K 或 $499/month | 硬件;基础软件;生态有限 | 工业合作的批量价格未知 | 西方人形机器人中的价格低点;面向家庭 / 轻工业,而非 3PL |
| Unitree G1 | 直销 | 起价 ~$16K | 硬件;基础 SDK | 研究折扣常见 | 行业价格地板锚点;若客户用 Unitree 试点,Agility 在研究 / 试点环节有风险 |
定价数据来自公司官方页面、新闻稿和行业报告。多数竞争对手尚未披露商业定价。Agility 的 RaaS 定价是该领域公开资料最完整的样本。直销与订阅对比仅供示意。
[CP006, CP009, CP010, CP016, CP018, CP024]3.5 竞争护城河与替代风险
Agility 的竞争优势包括:(1)先发部署数据——GXO 超过 100,000 次周转箱搬运产生专有运营数据,可反哺 AI 和机械优化。(2)RaaS 软件锁定——Agility Arc 的 WMS 集成形成类似企业 SaaS 的切换成本;更换 Digit 需要把 WMS 与竞争平台重新集成、重新培训操作员,并重新设计安全区域。(3)生产基础设施——位于 Oregon 州 Salem、70,000 sq ft 的 RoboFab 目标年产 10,000 台,提供多数竞争者尚未建立的供应链规模。(4)安全认证——FSoE CAT1 与既有围栏隔离部署经验带来监管熟悉度和合规路径,竞争者还需要复刻。关键替代风险包括:(a)Figure AI 到 2027 年实现商业部署规模,凭更强通用 AI 能力和可比价格进入,这是考虑融资和人才后最可能的近期威胁。(b)Tesla Optimus 达成 $20–30K 目标单价,使直接购买在 12–18 个月总拥有成本口径下可与 Agility 的 RaaS 竞争。(c)中国制造商,尤其是 Unitree,通过价格压力侵蚀试点和早期采用者细分,迫使 Agility 下探到低于其企业定位的市场。(d)如果 Figure AI 或 Apptronik 在 Agility 之前拿下 GXO 竞争对手的企业合同,商业窗口会比 Agility 扩张速度更快收窄。IEEE 在 2025 年发布评估指出,全行业在 AI 可靠性、运行时间和安全集成等商业成熟度方面仍有显著缺口;这些问题也适用于 Agility 及其竞争者。 [CP017, CP019, CP021, CP023, CP024, CP028]
| 护城河 / 风险 | 类型 | 严重性 | 现有证据 | 缓释措施 / 尽调问题 |
|---|---|---|---|---|
| 部署数据领先(100K+ 次周转箱搬运,多年 RaaS) | 护城河 —— 数据优势 | 高 | GXO 部署有公开记录;唯一商业规模人形机器人 RaaS | 评估专有数据积累速度,并与 2026–2027 年进入市场的竞争对手对比 |
| 通过 Agility Arc 锁定 WMS 集成 | 护城河 —— 切换成本 | 中高 | Arc 平台已在 GXO 生产环境和 Amazon 试点中使用 | 核实 WMS 集成合同条款;若竞争对手做出兼容 API,评估可迁移风险 |
| Figure AI 到 2027 年实现商业物流规模化 | 风险 —— 能力替代 | 高 | Figure AI 估值 $39B,已融资 $1B+;BMW 项目已结束,但计划下一次部署 | 跟踪 Figure AI 商业部署公告;评估 Agility 能否在 Figure 扩张前锁定 2–3 家额外 Tier-1 运营商 |
| Tesla Optimus 对外价格达到 $20K–$30K | 风险 —— 价格压缩 | 高 | Tesla 提出 $20K–$30K 目标;内部部署已确认 | 在不同价格点建模比较 Agility RaaS TCO 与 Tesla 直销;找出 RaaS 保持竞争力的价格地板 |
| Unitree 侵蚀试点 / 研究市场 | 风险 —— 商品化 | 中 | G1 以 $16K 切入,低于所有西方平台的试点预算 | 评估客户若用 Unitree 而非 Digit 试点,试点转商业的转化率是否变化 |
| IEEE 指出全行业商业就绪度缺口 | 风险 —— 可信度 | 中 | IEEE 2025 年评估指出,整个行业存在 AI 可靠性和运行时长担忧 | 获取 GXO 的 Digit 运行时长数据,用具体性能证据对冲全行业可信度风险 |
护城河耐久性和竞争风险评估,是基于公开证据的定性综合判断。严重性评级反映未来 3 年内对 Agility 竞争位置的估计影响。尽调问题是给投资人的建议后续动作。
[CP017, CP022, CP023, CP028, CP029, CP030]汇总 Agility Robotics 相对行业的竞争耐久指标。这些指标反映截至 2026 年年中,公司当前护城河深度和就绪度姿态,并给出方向性的定性判断。
KPI 数值为研究估计和定性判断。数字项(如料箱搬运次数、估值倍数)来自可得证据;方向性判断基于分析师综合研判。
[CP017, CP021, CP022, CP023, CP028, CP029]3.6 图表与证据
04财务情况
4.1 商业模式与收入结构
Agility Robotics 采用混合收入模式,结合经常性订阅收入、一次性硬件销售和仍处早期的软件平台收入。主要渠道是机器人即服务(RaaS):企业客户估计每台机器人每月支付 $10,000 到 $15,000,获得硬件使用权、维护、软件更新,以及通过 Agility Arc 平台提供的车队管理支持。该订阅结构把物流运营商的资本开支决策转成运营开支,降低大规模部署的采用门槛。公司公开的长期目标是每机器人小时 $30——与美国仓库环境中全负荷人工成本相当——因此 RaaS 经济性本质上是直接替代劳动力。第二条收入流是约每台 Digit $150,000 的直接硬件销售,可能服务偏好资产所有权或研究采购的客户。第三,Arc 云平台提供车队编排和仓库管理系统集成;Arc 独立订阅定价尚未公开披露。由于公司为私人公司,这三条收入流的结构未知。第三方对 2025 年年收入估计从 $10M 到 $80M 不等,最常被引用的中点接近 $35M(Growjo)。公司未披露官方收入数字或经审计财务报表。收入确认方法——尤其是 RaaS 硬件组成部分如何按 ASC 606 或 IFRS 15 处理——尚未公开说明,任何交易前都需要尽调审查。 [CI001, CI002, CI003, CI004, CI005, CI006]
| 收入来源 | 机制 | 单位 | 当前数值 / 状态 | 收入质量 | 尽调问题 |
|---|---|---|---|---|---|
| 机器人直销 | 一次性硬件采购 | 每台 | ASP ~$150,000(公司引用区间) | 公司披露的标价;批量折扣未知 | 确认扣除批量折扣、考虑地域组合后的实际 ASP |
| RaaS 订阅 | 月度订阅,捆绑硬件、维护和软件 | 每台机器人 / 月 | ~$10,000–$15,000/month(每台 ~$120K–$180K/yr) | 第三方报道;未发布官方价目表 | 索取当前 Digit 价格表和合同条款 |
| Arc 云平台 | 用于车队编排和 WMS 集成的云 SaaS 订阅 | 按车队或站点(未披露) | 定价未公开披露;可能并入 RaaS 或单独销售 | 未验证;公司网站确认产品存在 | 获取 Arc 单独定价、合同结构和续约条款 |
| 专业服务 / 部署支持 | 实施、集成和现场就绪服务 | 按项目或工时材料 | 未单独披露;可能打包进合同金额 | 根据典型企业机器人项目推断 | 厘清专业服务是独立收入线还是成本中心 |
所有收入数字均为第三方估计;Agility Robotics 不公开披露收入、合同数量或收入结构。RaaS 和直销价格来自分析师及媒体来源。Arc 平台定价尚未披露。
[CI001, CI002, CI003, CI004]| 定价层级 | 标价 | 合同期限 | 折扣 / 未知项 | 隐含单机年收入 | 来源 |
|---|---|---|---|---|---|
| Digit 直销 | ~$150,000/台 | 一次性 | 假设有批量折扣但未披露;没有可用的多机价目表 | ~$150,000(一次性) | 第三方分析师报告(Growjo、Owler) |
| RaaS 月度订阅 | ~$10,000–$15,000/robot/month 价格 | 多年期(期限未披露) | 可能按车队规模分层折扣;未发布官方分层列表 | ~$120,000–$180,000/robot/yr 收入 | 媒体报道、部署跟踪来源 |
| RaaS 小时等效 | 按月费 ÷ ~720 hrs/month 作业小时数推算为 ~$14–$17/hr | 内嵌于 RaaS 合同 | 与 $30/hr 目标的差距,反映当前定价偏低或利用率有限 | ~$120,000–$180,000/robot/yr 收入 | 由已披露月费和作业小时假设估算 |
| Arc 平台单独销售 | 未披露 | 推测为年度 SaaS 订阅 | 定价模型、层级和打包逻辑尚未公开 | Unknown | 公司产品页;无定价数据 |
所有价格均为标价,来自第三方分析师报告和媒体报道;包含企业批量折扣的实际成交价没有公开资料。小时等效费率按每年 8,760 可用小时、~60% 利用率计算(实际利用率未披露)。
[CI002, CI003, CI005, CI006]按收入流拆分 2025 年收入估计,展示直接整机销售、RaaS 订阅和 Arc 平台费如何拼出总收入估计。所有数值均为分析师估计;Agility Robotics 未披露官方收入。
所有数字均为分析师估计,依据公开价格、第三方收入追踪机构(Growjo、Owler)和部署数据。公司没有可用的审计财务报表。Arc 平台收入为剩余项估计,即在总收入和已知收入流代理数据之间倒推出的数值。
[CI001, CI002, CI003, CI004]4.2 融资历史与资本结构
Agility Robotics 多轮累计融资约 $641M,最高峰是 2025 年 3 月完成的 $400M Series C,由 WP Global Partners 领投,SoftBank Group、Amazon Industrial Innovation Fund 和 NVIDIA 的 NVentures 参投。Series C 确立了约 $1.75B 投前估值和接近 $2.1B 的投后估值。上一轮 Series B($150M,2022 年)由 Amazon Industrial Innovation Fund 领投,DCVC 和 Playground Global 参投。更早资本包括 Ford Motor Company 投资(约 $6M,约 2020 年)以及 Amazon Industrial Innovation Fund 未披露金额的种子期股权(2021 年)。投资方组合值得注意,既包括战略运营方 / 客户(Amazon),也包括大型综合基金(SoftBank、WP Global)、深科技专门机构(DCVC、Playground Global)和半导体伙伴(NVIDIA)。公开文件中没有发现已备案债务工具、信贷额度或项目融资安排。公司似乎也没有发行可转债或收入分成融资。Series C 的董事会构成和治理条款——包括清算优先权、反稀释条款和按比例跟投权——未公开披露。Series D 门槛尚未宣布;如果到约 2027 年单位经济性仍未达到商业化自我维持门槛,公司预计将寻求 Series D 或 IPO 前轮融资。任何正式交易中,都应对照 Series C 条款清单和股东协议核验资本结构信息。 [CI008, CI009, CI010, CI011, CI012, CI013]
4.3 单位经济性与成本结构
Digit 的单位经济性仍处于早期、盈亏平衡前阶段,这是一代硬件即服务业务的典型状态。按当前产量估计,每台 Digit 的物料清单和组装成本(COGS)约为 $80,000–$100,000;相对于约 $150,000 的直接销售 ASP,意味着毛利率约 30–40%。该估算来自可比工业机器人制造商和第三方部件成本分析;Agility 未披露实际 COGS。RaaS 模式下,每机器人小时的全负荷运营成本——包括用电、计划维护、软件授权和摊销资本成本——估计为 $10–$12/hr。当前 RaaS 计费约 $14–$17/hr(由每月 $10K–$15K 除以典型运行小时推算),因此毛利率从负到略正不等。公司公开目标 $30/hr,既体现匹配人工成本的意图,也体现规模化后随着硬件 COGS 随制造量下降、软件成本摊到更大车队上而释放经营杠杆的野心。管理层曾提到成熟规模下目标毛利率为 60%+,与遵循软件利润模型的机器人即服务业务一致。RoboFab 的 70,000 sq ft 设施目标年产 10,000 台;达到规模后,组件采购量提升、制造良率改善和面向制造的设计迭代,预计会压缩 COGS 并推高毛利。公司未公开分享客户层面的回本或 ROI 数据,不过按 $30/hr 对比美国仓库典型全包人工 $25–$35/hr,如果满负荷使用,客户理论上可在 12–24 个月内达到盈亏平衡。 [CI014, CI015, CI016, CI017, CI018, CI019]
| 指标 | 数值 / 估计 | 置信度 | 重要性 | 尽调问题 |
|---|---|---|---|---|
| ASP —— 直销 | ~$150,000/unit | 中(公司经媒体引用) | 设定硬件收入线的毛利率上限 | 确认扣除批量折扣及任何政府 / 研发定价后的实际 ASP |
| 单机 COGS(硬件) | ~$80,000–$100,000(估计) | 低(分析师推断) | 决定毛利率和规模经济;若要达到目标毛利率,COGS 必须下降 | 向管理层获取 BOM 成本拆分和制造良率数据 |
| 毛利率 —— 直销 | ~30%–40%(按当前产量估计) | 低(推导估计) | 显示硬件盈利路径;60%+ 目标需要压低 COGS | 通过经审计财务报表或管理账确认毛利率 |
| 每机器人小时运营成本 | ~$10–$12/hr(估计) | 低(分析师估计) | RaaS 毛利率的关键驱动;必须相对 $30/hr 目标计费率下降 | 索取实际电力、维护、折旧摊销和支持成本表 |
| RaaS 毛利率(当前) | 负值至 ~20%(估计) | 低(推导) | 判断当前 RaaS 订阅是否已产生正向单机经济性 | 获取 RaaS P&L 按队列和合同年份的贡献 |
| 规模化目标毛利率 | 60%+(公司表述方向) | 中(公司叙述) | 未来 robotics-as-software 毛利模型的基准 | 验证推动毛利率改善的假设(产量、自动化、软件组合) |
所有单机经济性数值均为分析师估计,或依据可比硬件业务推断;Agility Robotics 未发布官方毛利率、COGS 或单机运营成本数据。置信度反映来源质量。所有空字段在尽调场景下都需要管理层直接提供数据。
[CI014, CI015, CI016, CI017, CI018, CI019]从机器人 ASP 到运营成本再到估计毛利,定性拆解单机经济性流向,展示当前已实现毛利率与目标毛利率之间的差距。
所有数值均为估计,依据分析师报告和可比硬件即服务公司。Agility Robotics 未披露 COGS、运营成本表或毛利率。节点代表大致当前状态;目标状态节点反映公司声明的方向性目标。
[CI007, CI014, CI015, CI016, CI017, CI018]4.4 资本充足性与现金跑道分析
2025 年 3 月 $400M Series C 交割后,扣除交易成本和短期义务,Agility Robotics 估计持有 $300M–$400M 现金及现金等价物。月度现金消耗估计为 $4–$7M,包括 250–500 名员工约 $3–$5M 的工资成本,以及 RoboFab 资本开支、零部件库存建设、下一代 Digit 迭代和协作安全系统 R&D 支出。按约 $5.5M/月的中位烧钱速度,从 2025 年 3 月起,现金跑道延伸至约 55–70 个月(约 4.5–6 年),意味着现金可支撑到 2029 年末或 2031 年。在高烧钱情景下($7M/月,含制造扩张成本),跑道缩短至约 43 个月(2028 年中)。公司未披露 Series C 资金用途拆分。资本配置优先项推断包括:RoboFab 爬坡至目标产能、下一代 Digit 与 Arc AI 功能 R&D、面向企业销售周期的商业团队扩张,以及零部件库存所需营运资金(每台约 6,000 个零部件,80% 美国来源)。公开文件或媒体报道中没有发现已披露债务、循环信贷或项目融资安排。公司下一次重大资本事件预计是 Series D 或 IPO 前轮融资,触发条件可能是 RaaS 单位经济性接近盈亏平衡,或车队规模跨过需要额外制造投入的门槛。现金跑道充足性被评为近期足够,但对制造爬坡速度和毛利率兑现敏感。 [CI021, CI022, CI023, CI024, CI025, CI026]
| 项目 | 估计 / 状态 | 置信度 | 来源依据 | 尽调动作 |
|---|---|---|---|---|
| 在手现金(Series C 后) | ~$300M–$400M(扣除交易成本后) | 中(由 $400M 融资减去近期义务推导) | Series C 交割公告;未披露资产负债表 | 核实资产负债表中的现金余额,以及受限 / 非受限现金拆分 |
| 月度现金消耗 | ~$4M–$7M/month(估计) | 低(由员工数、资本开支、R&D 信号推断) | 可比公司基准;未披露烧钱速度 | 向管理层索取月度现金流量表 |
| 估计现金跑道 | 自 2025 年 3 月起 ~36–72 个月(3–6 年,取决于烧钱速度) | 低(烧钱速度不确定,区间很宽) | 由估计现金和烧钱速度区间推导 | 获取董事会批准的财务模型和 18 个月现金预测 |
| 债务 / 项目融资义务 | 公开文件或媒体报道中未发现 | 中(未见证据) | SEC EDGAR Form D 文件;公开媒体报道;无信贷额度公告 | 在尽调中确认不存在债务、可转债、收入分成融资 |
| 下一次融资触发点 | 若单机经济性到 ~2027 年仍未改善,可能需要 Series D 或 pre-IPO 融资 | 低(分析师推断) | 可比公司融资周期分析;IPO 市场环境 | 了解董事会资本计划和 Series D 条款预期 |
现金余额和烧钱速度均为估计;没有公开可用的审计资产负债表。现金跑道按估计现金 ÷ 估计月度烧钱速度计算。公开来源未发现债务和项目融资义务;正式尽调应对照公司财务报表核实。
[CI021, CI022, CI023, CI024, CI025]以熊市、基准、牛市假设给关键财务指标做情景区间估计,展示可能结果范围。所有数值均为分析师估计;没有审计数据。
所有数值为第三方分析师估计,或由公开数据推导。熊市情景假设商业爬坡有限且成本高;牛市情景假设机队快速部署且成本早期压缩。没有审计财务数据可锚定这些估计。
[CI004, CI020, CI021, CI022, CI031]展示 Series C 资金主要用途,说明在软件研发和商业运营并行推进时,扩大人形机器人制造需要大量资本。
资本按项目分配未公开披露。以下数字为比例权重估计,依据已披露设施面积、员工人数估计和可比硬件公司的资本结构。
[CI021, CI022, CI023, CI024, CI025, CI026]4.5 财务风险与盈利路径
Agility Robotics 面临多项财务风险,可能实质影响其盈利路径。第一,每机器人小时运营成本($10–$12/hr)仍低于目标计费价格 $30/hr;差距必须靠硬件降本和软件杠杆共同收窄,但两者都没有保证。如果当前现金耗尽时成本曲线仍未改善,公司将需要追加资本,而届时 RaaS 模式可行性仍未被最终证明。第二,收入集中是实质风险:两个公开已知客户 Amazon 和 GXO Logistics 可能占当前合同价值的绝大多数;任何部署放缓或合同重谈都会对收入造成不成比例影响。第三,直接销售 $150K ASP 和 RaaS 每月 $10K–$15K 的价格,可能在 Boston Dynamics、1X Technologies、Figure AI 等竞争者推出竞争产品后承受下行压力,在规模尚未形成前压缩利润率。第四,RoboFab 资本强度高——70,000 sq ft,目标 10,000 台 / 年——需要持续高利用率才能获得单台成本杠杆;低生产良率或爬坡延迟会推高实际 COGS。第五,隐含收入倍数为 15×–60×($2.1B 估值对应估计 $10M–$80M 收入),对近期商业执行提出极高要求。持怀疑态度的行业分析质疑,人形机器人能否在下一轮融资周期前达到足以支撑当前定价和估值溢价的可靠性和利用率。缺少经审计财务数据、且公司为私人公司,使得在没有正式尽调权限的情况下独立评估这些风险较为困难。 [CI027, CI028, CI029, CI030, CI031, CI032]
| 缺失指标 | 重要性 | 尽调路径 | 对分析的影响 |
|---|---|---|---|
| 经审计收入和收入结构 | 没有确认收入数据,无法评估商业模式可行性或做预测 | 索取 FY2024–2025 经审计 P&L 或管理账 | 高:收入估计区间($10M–$80M)太宽,难以支撑有意义的承销 |
| 按收入线拆分毛利率(硬件 vs. RaaS vs. Arc) | 混合毛利率会掩盖 RaaS 经济性;硬件毛利看似可接受时,RaaS 可能深度为负 | 索取分部级 P&L 或贡献毛利桥 | 高:负 RaaS 毛利率可能持续到资本无法承受 |
| 单机 COGS 和 BOM 拆分 | 没有组件成本数据,无法验证毛利率改善路径或制造杠杆逻辑 | 向运营团队索取物料清单、组装成本和良率数据 | 高:$80K–$100K COGS 估计来自分析师推导,可能大幅偏离 |
| 生产量(出货单位) | 车队规模和产能爬坡速度决定收入轨迹和 RoboFab 利用率逻辑 | 按季度向管理层索取生产和部署数据 | 高:公开资料没有披露已制造或规模部署了多少台 Digit |
| 客户合同条款和续约率 | 收入质量取决于合同期限、续约率和流失率;两家客户集中度会放大这一风险 | 审查客户合同、续约历史和管线数据 | 高:收入集中在 Amazon 和 GXO,且未披露续约指标 |
本表列出 Agility Robotics 尚未公开披露、但对承销分析重要的财务指标。所有项目都是标准尽调数据点,正式交易场景下应可获得,但公开来源不可得。
[CI027, CI028, CI029, CI030, CI033]05产品与技术
5.1 产品概览与 Digit 硬件架构
Digit 是 Agility Robotics 的商用双足人形机器人,从一开始就为物流和仓库自动化而设计。当前一代 Digit 高 175 cm(5'9"),重约 65 kg,可处理最高 16 kg(35 lbs)载荷。其运动系统采用受食火鸡启发的反向弯曲腿,源自 Oregon State University Dynamic Robotics Laboratory 在 Jonathan Hurst 教授领导下十余年的生物力学研究。机器人通过一套感知组件在仓库中自主导航,包括 3D LiDAR、立体 RGB 摄像头、惯性测量单元(IMU)、关节编码器、接近传感器和力传感器。机载计算运行双 Intel i7 CPU,模块化开放载荷舱可容纳 NVIDIA Jetson 或 Intel NUC 加速器,用于 AI 推理任务。Digit 最高步行速度 1.5 m/s,可上下楼梯和路缘,并能运行 4+ 小时后自主对接充电站。连接方式包括 WiFi 和 4G 蜂窝网络。每台机器人约由 6,000 个零部件组成,其中约 80% 来自美国供应商,体现了与本土制造优先级一致的供应链韧性策略。末端执行器针对物流周转箱搬运而非通用五指操作优化,这是一项有意的设计取舍:用更简单、更可靠的操作能力,服务一个范围较窄但已被商业验证的任务集。 [CE001, CE002, CE003, CE004, CE005, CE006]
| 模块 / 资产 | 描述 | 成熟度 / 状态 | 主要用户 | 差异化 | 尽调缺口 |
|---|---|---|---|---|---|
| Digit(当前代) | 175cm、65kg 双足人形机器人;16kg 载荷;4hr 电池;3D LiDAR | 商业化 —— 围栏隔离 | 仓库运营商 / 物流 3PL | 仿生 OSU 腿部 IP;面向任务优化的末端执行器;美国制造 | 各来源对具体 DOF 规格说法不一致;实际现场 MTBM 未披露 |
| Arc Cloud Platform 云平台 | 车队管理、远程诊断、WMS/WES/MES/PLC 与 AMR 集成 | 商业化 —— 活跃部署 | 仓库 IT / 系统集成商 | 异构机器人编排(Digit + MiR + Zebra AMR) | SDK 深度、API 文档未完全公开;定价未披露 |
| 自动充电桩 | 用于自主充电的对接站;无需人工干预 | 商业化 | 设施运营商 | 省去电池更换人工;延长作业窗口 | 大规模部署时,多机位充电坞容量受限;负载下换电时间未公布 |
| 模块化载荷舱 | 开放式舱位,可容纳 NVIDIA Jetson / Intel NUC AI 计算加速器 | 可用——可选加装件 | 企业 AI / 解决方案集成商 | 让机器人端 AI 可升级,无需更换平台 | 第三方加速器认证状态未披露 |
| 任务优化末端执行器 | 面向物流的专用夹爪,针对料箱和包裹搬运优化 | 商用 | 3PL / 履约运营商 | 在狭窄任务域内比 5 指机械手更可靠;按用途工程化设计 | 不是通用方案;任务超出料箱搬运后,切换成本高 |
| Cassie(旧款,已退役) | 2017-2019 年售给高校实验室的研究型双足机器人;Digit 的前身 | 已退役 / 传承产品 | 研究机构 | Agility 基础运动控制 IP 和学术论文记录的基础 | 无活跃商业化;仅有研究价值 |
截至 2026 年 5 月,状态反映公开可知的商业部署进展。“成熟度”评估基于有记录的 GXO/Amazon 部署和公司沟通,而非独立审计。末端执行器的通用性、Arc API 文档深度, 仍是集成伙伴的开放尽调项。
[CE001, CE002, CE003, CE008, CE009, CE010]5.2 运动研究基础与 AI 感知技术
Digit 的运动架构可追溯到 Oregon State University 2015 年开始开发的 Cassie 双足机器人;这条技术谱系代表了能效型腿足运动领域最严谨的一批学术成果。 仿生设计把动态稳定性放在准静态平衡之前,因此 Digit 面对地面扰动和外力偏差时, 比静态稳定方案更能顶住并恢复。腿部几何受到食火鸡关节构型启发,行走时能被动回收能量, 降低执行器负荷并延长电池续航。Agility 将强化学习用于抓取和操控任务, 训练机载 AI 模型在不同周转箱形状和输送线高度之间泛化。感知管线把 3D LiDAR 点云 与双目摄像头画面结合起来,在机载端跑神经网络推理,不靠人工遥操作也能自主规划路径和避障。 IEEE 等同行评审刊物上围绕 Cassie 和 Digit 平台发表的学术论文, 给出了一条有文献支撑的技术谱系;依赖更新、公开发表更少研究的竞争者不容易复制。 不过,Agility 尚未公布商业部署中的任务成功率、操控延迟或故障恢复时间等具体基准结果, 限制了外部对其所称性能的独立技术验证。开放式载荷舱允许接入第三方 AI 计算模块, 硬件代际升级时,平台可以提升推理能力,而不必整机更换。 [CE012, CE013, CE019, CE028, CE033, CE034]
5.3 Arc 云平台与企业集成架构
Arc 云平台是 Agility 的软件即服务层,负责编排 Digit 机队、接入企业仓储系统, 并提供远程监控和诊断。Arc 通过 webhook 和自定义 API 直接连接仓库管理系统(WMS)、 仓库执行系统(WES)、制造执行系统(MES)和 PLC,让运营人员能在既有运营工具里 给 Digit 分配任务,不必再搭一套并行的人机界面。2025 年 3 月宣布的关键差异化能力 是 AMR 直连集成:Arc 能在同一工作流里调度 MiR 和 Zebra Technologies 的 自主移动机器人,让 Digit 在全自动周转箱搬运链路中装卸 AMR。Arc 因此不只是 Digit 机队管理器,而是面向已经运营 AMR 机队的物流设施的异构机器人统一编排平台。 Arc 还支持安全机队监控和远程停机能力,这对需要集中控制室监督的企业部署至关重要。 平台在机器人内部控制通信中使用 EtherCAT 现场总线,并由安全 PLC 协调符合 FSoE 的 安全指令。Arc API 架构让 WMS 集成商和系统集成商能搭建横跨 Digit 与第三方自动化设备的 自定义工作流自动化;集成库越成熟,这条能力越可能形成扩张护城河。截至 2026 年, 具体 API 文档、SDK 可用性和公开开发者资源仍不充分透明,对评估集成深度的技术伙伴来说 是一个缺口。 [CE009, CE010, CE011, CE027, CE032]
| 用户任务 / 作业 | 当前工作流(人工) | Agility 方案 | 宣称的可衡量收益 | 限制 |
|---|---|---|---|---|
| 输送线旁料箱归集 | 人工拣选员走到输送线,搬起料箱并码放,供分拣使用 | Digit 从入站输送线抓取料箱,并码放到出站点 | GXO 有 100,000+ 个料箱记录;成本约 $3/hr,对比人工 $17–25/hr | 需要结构化环境;仅限同尺寸料箱格式 |
| AMR 装载 / 卸载 | 人工在站点交接处为 AMR 小车装卸货 | Arc 调度 Digit 装卸 MiR 或 Zebra AMR 载荷 | 省去 AMR 交接站点人工;打开无人值守作业窗口 | 截至 2025 年,集成仅限 MiR 和 Zebra;定制 AMR 需要工程投入 |
| 场内库存搬运 | 人工拣选员长距离步行,在区域间搬运料箱 | Digit 或 Digit + AMR 串联自主完成搬运 | 减少拣选员步行时间;把人力转到增值任务 | 仅适用于结构化巷道布局;需要地面建图和站点设置 |
| 重复分拣劳动力 | 夜班工人分拣包裹;岗位难招且流动率高 | Digit 在低峰和夜间班次无人值守运行 | 缓解夜间物流岗位招聘难;吞吐更稳定 | 当前围栏要求降低开放式设施的部署灵活性 |
| 机队监控与告警 | 主管手动检查机器人状态;故障响应慢 | Arc 提供实时机队仪表盘、告警和远程 E-stop | 降低主管与机器人配比;支持远程运营监管 | Arc 监控深度随集成而异;4G 下远程控制延迟未公开基准测试 |
收益主张来自公司沟通和 GXO 新闻稿。$3/hr 运营成本是公司引用数字,未经过独立审计。 截至 2026 年 5 月,所有部署都需要结构化环境。Agility 未公开围栏成本和安装时间。
[CE009, CE010, CE011, CE020, CE027, CE032]| 层 / 组件 | 作用 | 关键依赖 | 风险 |
|---|---|---|---|
| Arc 云平台 | 机队编排、WMS/WES 集成、远程诊断、AMR 调度 | 云基础设施(AWS 或同等);WMS 厂商 API 稳定性 | 与传统 WMS 集成复杂;API 版本管理未公布 |
| 安全 PLC + FSoE EtherCAT | 实时执行安全指令;CAT1 停止;E-stop 继电器 | EtherCAT 现场总线厂商;IEC 61508 合规 | FSoE 是专有协议;仅在非协作配置下达到 SIL 等级 |
| 板载 AI 计算(Intel i7 + Jetson) | 传感器融合、任务推理、基于 RL 的抓取、路径规划 | Intel CPU 供应;可选 NVIDIA Jetson GPU 供货 | 更高吞吐 AI 模型可能让计算成为瓶颈;持续高负载下有散热上限 |
| 3D LiDAR + 双目视觉 | 环境建图、避障、料箱检测与定位 | LiDAR 传感器供应商;全生命周期校准稳定性 | 高粉尘、低光或反光表面环境中的表现未公开基准测试 |
| 自主充电坞 | 免人工充电;通过板载传感器定位充电坞 | 充电坞硬件可靠性;地面标记或信标基础设施 | 充电坞故障会成为班次连续性的单点故障 |
| WiFi + 4G 连接 | Arc 遥测、远程指令、诊断流传输 | 场馆 WiFi 基础设施;4G 运营商可用性 | 负载下 4G 指令延迟;弱连接环境中远程 E-stop 时序未记录 |
架构信息来自公司产品页面、IEEE 论文和第三方技术媒体。EtherCAT 现场总线和 FSoE 细节已由多个技术来源确认。 计算规格反映当前一代 Digit;未来代际可能不同。
[CE006, CE009, CE011, CE014, CE015, CE033]Agility Robotics 产品栈的分层视图:从云端编排,到 AI/感知、安全与控制,再到底层 Digit 硬件平台,并列出每层关键集成和组件。
[CE009, CE011, CE014, CE033]端到端客户工作流,展示仓库运营商如何通过 Arc 平台部署 Digit:从 WMS 下发任务,到机器人执行,再到报告回传。
[CE009, CE010, CE011, CE032]5.4 制造、供应链与知识产权
Agility Robotics 运营 RoboFab,这是一座位于 Oregon 州 Salem 的专用人形机器人制造设施, 2023 年 9 月宣布,2024 年投产。该设施面积 70,000 sq ft,满产后的产能目标为 每年 10,000 台 Digit,是全球最大的专用人形机器人生产基地之一。截至 2026 年初, 约 100 台已在 GXO 和 Amazon 商业部署,说明该设施仍处在产能爬坡阶段。单台约 6,000 个零件的 BOM 中,约 80% 来自美国供应商;相比更依赖中国来源组件的竞争者, 这降低了单一地理供应链风险。Agility 的知识产权组合覆盖双足运动(源自 Cassie 研究谱系)、 操控与任务优化末端执行器设计、纳入 FSoE 协议的安全与控制系统架构,以及 Arc 平台中的 多机器人编排。确切专利数量未公开披露。关键基础 IP 来自 OSU 的 Dynamic Robotics Laboratory, Jonathan Hurst 关于被动动态行走和能效型双足运动的研究被转化为商业专利申请。 RoboFab 的专用产能、美国来源供应链和学术 IP 谱系叠加,形成制造与 IP 护城河; 新进入者要复制会耗费多年时间。不过,10,000 台设计产能与当前约 100 台部署之间的落差, 是重大的执行与规模化风险,投资者和伙伴应在 2026–2027 年密切跟踪。 [CE016, CE017, CE018, CE019, CE029, CE037]
一张有向无环图,展示影响 Digit 生产、部署和平台运行的关键外部依赖,覆盖零部件供应商、云基础设施、监管机构和企业集成伙伴。
[CE018, CE033, CE009, CE016, CE020]5.5 安全认证、合规与产品路线图
Digit 当前的安全架构包括 Functional Safety over EtherCAT(FSoE)CAT1 停止能力、 安全 PLC、机身和无线急停系统,以及 Arc 的远程机队停机。尽管有这套硬件安全基础设施, 截至 2026 年,Digit 仍在围栏隔离环境中运行,机器人主动作业时由物理屏障与人类工人隔开。 公司尚未取得符合 ISO/TS 15066 的协作安全认证;该认证将允许 Digit 在无屏障环境中 直接与人并肩工作。CEO Peggy Johnson 公开把 2026 年定为协作安全认证目标, 这将是 Digit 商业部署潜力中最大的阶跃式能力提升,可打开不适合或无法承受围栏安装成本的设施。 IEEE Spectrum 和 Fortune 的独立分析师与批评者指出,该认证时间表已较早前承诺推迟; 物流环境中的完整无围栏人机协作还涉及监管审查,可能延伸到 2027 年相当晚的阶段。 已宣布的产品路线图包括:2025 年提升续航、扩展 AMR 集成并改进安全功能;2026 年取得 协作认证;2027 年及以后推出第五代 Digit,目标是完全协作式工作流和更高载荷 (50 lbs,而当前为 35 lbs)。更长期看,Agility 称希望从物流扩展到轻制造、医疗和零售, 但这些垂直领域尚无公开确认的具体产品变体或客户合作。可能限制这类扩张的关键技术瓶颈包括: 末端执行器按任务设计而非通用型、当前 AI 模型内置了结构化环境假设,以及电池容量不足以在 没有自主充电基础设施的情况下支撑连续多班次部署。 [CE014, CE015, CE020, CE021, CE022, CE023]
| 控制 / 认证 | 当前状态 | 范围 | 缺口 / 开放问题 |
|---|---|---|---|
| FSoE CAT1 紧急停止 | 已实施并部署 | 机器人端硬件 + 无线 E-stop;EtherCAT 现场总线 | CAT1 认证仅覆盖围栏隔离运行;尚未获得协作认证 |
| 安全 PLC(IEC 61508) | 已实施 | Digit 硬件控制系统 | 当前配置达到的 SIL 等级未公开披露 |
| ISO/TS 15066 协作安全 | 尚未获得 | 无隔离屏障的人机协作运行 | 目标为 2026 年;独立分析师提示可能滑至 2027 年;未点名认证机构 |
| CE / FCC 标志 | 状态未公开披露 | 在美国和欧盟商售所需的产品安全 / EMC 合规 | 商售必需;缺少公开披露构成尽调缺口 |
| Arc 平台安全 / SOC 2 | 状态未公开披露 | 云平台数据安全;企业机队数据 | 未公开 SOC 2 认证;企业采购可能要求确认 |
合规状态基于公司产品沟通和第三方报道。截至 2026 年 5 月,Agility 尚未发布公开的信任 / 合规页面。 多个独立来源确认 ISO/TS 15066 协作安全尚未获得。CE/FCC 和 SOC 2 缺口是已识别的尽调项,并非已确认缺陷。
[CE014, CE015, CE020, CE023, CE035, CE036]| 阶段 / 日期 | 功能 / 里程碑 | 状态 | 影响 | 来源 |
|---|---|---|---|---|
| 2025(已完成) | 改进续航,AMR 集成(MiR/Zebra),增强安全功能 | 已按 2025 年 3 月公告交付 | 提升单台机器人部署价值;扩展可服务设施类型 | Agility Robotics Arc 平台公告;Wired;IEEE Spectrum |
| 2026(目标) | 协作安全认证——无围栏人机共存 | 推进中;尚未获得 | 最关键里程碑:解锁绝大多数仓库地面配置 | CEO Peggy Johnson 公开表态;The Verge;IEEE 反向分析 |
| 2027+(规划) | Digit v5:更高载荷(目标 50 lbs),完全协作工作流 | 已公布路线图;未确认里程碑日期 | 扩大物流和轻制造可服务市场 | 公司沟通;MIT Technology Review |
| 2027+(规划) | Arc 平台扩展至 MiR/Zebra 之外的更多 AMR / 机器人品牌 | 已宣布意向;未披露厂商名单 | Arc 要成为真正的多机器人编排标准,此项至关重要 | Agility 官方沟通;Supply Chain Brain |
| 2028+(推测) | 从物流扩展到医疗、轻制造、零售垂直领域 | 未公开确认产品变体或 LOI | 若实现将显著扩大市场;没有垂直特定证据时仍属推测 | 公司博客;由 CEO 公开表态推断 |
路线图里程碑基于公司公开表态和第三方报道。公司未公开披露 2025 年后功能或 Digit v5 的签约客户承诺。 协作安全认证时间线最关键也最不确定;分析师认为监管审查复杂性是风险因素。
[CE023, CE024, CE031, CE036]按成熟度评估 Digit 核心技术和产品模块,对照竞争基线给出当前商业状态,并识别关键缺口。
[CE008, CE012, CE013, CE020, CE021, CE030]06客户情况
6.1 客户群细分
Agility Robotics 瞄准的是仓储与物流市场中一块狭窄但高价值的需求:合约物流运营商 (3PL)和运营高吞吐、结构化履约环境的大型企业货主。主要买方画像是一线或二线 3PL、 或全渠道零售商的运营副总裁 / 首席供应链官,他们希望对冲仓库劳动力每年 30–40% 的流失率 和不断上升的工资压力。用户是仓库一线操作员;付款方是动用 RaaS 预算的物流业务单元。 次级客群包括电商纯玩家和汽车零部件分销商,它们同样有结构化、重复性任务。 Digit 专为具备三类特征的任务设计:结构化物理环境(通道、输送线、固定拣选 / 放置工位)、 中等载荷下的高重复动作(最高 35 lbs 的周转箱),以及大型轮式自动化难以进入的空间约束。 这让可触达客户基数小于通用人形机器人,但提升了已具名客户的任务成功率和部署信心。 Agility 当前商业销售明确避开非结构化或变化大的环境,把足迹限制在拥有明确输送线基础设施的 中大型配送中心(DC)。 各细分客群的收入和战略价值差异很大。一个 24/7 运转的 3PL DC,在 RaaS 模式下每台机器人 每年可贡献 $120K–$180K;若单站点部署 5–50 台机队,年合同额可达 $600K–$9M。 Amazon 这类大型电商运营商看重的,可能不只是即时财务回报,而是共同开发以及提前试点 新兴劳动力替代平台的战略价值——这也解释了 Amazon Industrial Innovation Fund 的持股, 并不要求 Amazon 在现阶段成为主要收入贡献方。 [CU001, CU002, CU003, CU004, CU005, CU006]
| 细分市场 | 买方 / 用户 / 付款方 | 用例 | 规模 | 收入 / 战略价值 | 缺口 |
|---|---|---|---|---|---|
| 合同物流(3PL) | 运营 VP / 一线操作员 / BU P&L | 料箱回收、输送线装载、拖车卸货 | 全面铺开后,每站点 50–500 个机器人年 | 主要收入细分;$120K–$180K/robot/yr RaaS | 部署周期长;每个设施需 OSHA 现场批准 |
| 大型企业发货方(电商) | 供应链 VP / 运营 / 公司 CapEx | 高吞吐分拣、入站料箱处理 | 目标规模下,大型 DC 每个 100+ 台 | 高销量潜力;Amazon 是锚定账户 | Amazon 有自有机器人部门;合作持久性不确定 |
| 汽车 / 工业分销 | 工厂经理 / 一线 / 制造预算 | 零件配套、料箱码放、结构化取放 | 单站点机队更小;站点更多 | 相邻市场;尚无具名客户 | 无确认部署;较重工业零件受载荷上限约束 |
| 零售全渠道履约 | DC 负责人 / 一线 / 零售运营预算 | 履约中心料箱处理、退货处理 | 大型 DC 每个 10–50 台 | 劳动力流动率高,ROI 叙事强 | SPANX(经 GXO)是首个零售相邻部署;复制数据有限 |
细分定义和规模估算来自公开部署数据和分析师报道推断;Agility 未发布正式市场分层研究。 单台机器人收入基于公开引用的 RaaS 定价 $10K–$15K/month;实际合同条款未披露。
[CU001, CU002, CU003, CU004]6.2 已具名客户部署与证据
截至 2026 年初,GXO Logistics 是 Agility Robotics 唯一确认的商业生产客户。 双方关系始于 2024 年 6 月宣布的多年 RaaS 协议,Digit 被部署在 Georgia 州 Lithia Springs 由 GXO 运营的 SPANX 履约中心。到 2025 年底,Digit 已在真实生产中完成超过 100,000 次 周转箱搬运;Robotics & Automation News 于 2025 年 11 月独立报道了这一里程碑, 并称其为全球首个正式商业人形机器人部署。该部署在围栏隔离环境中运行,并取得 OSHA NRTL 站点特定批准;这是关键安全资质步骤,使 GXO 能在 OSHA 管辖下让 Digit 与人类工人同场运行。 GXO 按收入计也是全球最大的纯合约物流公司,因此是可信度很高的标杆客户。 Amazon 确认参与 Agility Partner Program——这是 Agility 面向企业客户的结构化商业接入路径—— 并已在 Amazon 履约设施的探索性试点场景中部署 Digit。Amazon Industrial Innovation Fund 参与了 Series B 和 Series C 两轮融资,既提供资本,也传递了隐性的产品市场验证信号。 不过,Amazon 尚未被确认为生产部署客户,且公司自身运营 Amazon Robotics 机器人研发项目, 因而长期独家依赖 Agility 的可能性不高。 法国全球物流运营商、全球前十 3PL GEODIS 确认参与 Agility Partner Program, Digit 试点聚焦仓库周转箱搬运和拖车卸货任务。截至报告日期,GEODIS 尚未公开宣布生产部署 或多年协议。 据报道,未具名零售商和电商运营商也在贸易展演示(NRF 2025、MODEX 2024)中表现出 试点阶段兴趣,但尚无其他已具名生产客户获得公开确认。 [CU007, CU008, CU009, CU010, CU011, CU012]
| 客户 | 细分市场 | 部署与用例 | 生产与试点 | 结果 | 限制 |
|---|---|---|---|---|---|
| GXO Logistics | 合同物流(3PL) | 料箱回收和输送线装载,SPANX DC,Lithia Springs GA | 生产——多年期 RaaS 协议 | 已搬运 100,000+ 个料箱;OSHA NRTL 单站点批准;全球首个商用人形机器人部署 | 机队规模未披露;单一站点;未发布吞吐基准或正常运行时间数据 |
| Amazon | 电商 / 履约 | Amazon 履约设施中的 Digit 试点;Partner Program 成员 | 试点 / 探索——未确认生产部署 | Amazon Industrial Innovation Fund 投资 B 轮和 C 轮;试点验证了硬件安全性 | 未宣布生产部署;Amazon Robotics 是内部竞争能力 |
| GEODIS | 合同物流(3PL)——全球 | 用于料箱处理和拖车卸货的 Digit 试点;Partner Program 成员 | 试点——未宣布生产协议 | 通过新闻稿公开确认试点;战略上验证 3PL 细分需求 | 试点范围、台数和结果指标未披露;无多年期协议 |
具名客户名单仅基于公开新闻稿和媒体报道;Agility 未发布完整客户名单。 生产 / 试点分类使用公开可得证据。Amazon 和 GEODIS 的结果指标未公开披露。
[CU007, CU008, CU009, CU010, CU011, CU012]按关键验证维度评估具名客户关系:部署阶段、OSHA 批准、多年期合同、机队规模是否已知、公开背书状态和战略投资人关系。
[CU007, CU010, CU012, CU013, CU014, CU016]6.3 采用轨迹与商业动能
GXO 的 100,000 个周转箱里程碑,是目前最具体的公开采用指标。Agility 尚未披露 GXO SPANX 站点部署了多少台 Digit、每个周转箱的循环时间或正常运行率;这些指标本可用于估算机队规模 和吞吐效率。行业估计基于报道的周转箱量和运行小时数,认为该部署约有 4–8 台 Digit, 但这是推断,不是确认数据。 多年 RaaS 协议结构是强留存和收入信号:不同于硬件销售,多年订阅带来经常性收入和合同锁定, 降低合同期内的流失风险。Agility Partner Program 对符合条件的企业客户开放, 设置了明确的准入、试点和生产路径,形成结构化商业管线。除 GXO 外,Amazon 和 GEODIS 是仅有的公开具名 Partner Program 参与方。 采用受三类因素约束:(1)部署复杂度高,需要 OSHA 站点特定批准、设施改造(围栏基础设施、 充电站、传感器标记)和 WMS/Arc API 集成,每一项都会让从部署到收入确认增加 3–9 个月; (2)RoboFab 产能仍在从低得多的基数爬坡至每年 10,000 台,限制了同时服务客户数量; (3)客户在承诺前会测试多个人形机器人平台,竞争评估周期较长。GXO SPANX 取得的 OSHA NRTL 批准是站点特定的,每个新设施都要重新办理,形成快速扩张的监管瓶颈。 积极的一面是,仓库劳动力市场动态强烈支持采用:美国仓库和物料搬运劳动力的全负担成本平均为 $18–$22/hour,30–40% 的年流失率带来每次招聘和培训 $2,000–$5,000 的成本。 在 $10K–$15K/robot/month 的 RaaS 价格下,ROI 要打平,每台机器人每班需要替代约 1.5–2.0 个 FTE;高吞吐 DC 能达到这一门槛。 [CU017, CU018, CU019, CU020, CU021, CU022]
| 指标 | 数值 | 日期 | 来源 | 置信度 | 影响 | 缺失分母 |
|---|---|---|---|---|---|---|
| 生产环境搬运料箱数(GXO SPANX) | 100,000+ | 2025-11 | SU003 | 高 | 证明生产环境可持续运行;全球首个商用人形机器人里程碑 | 机队规模、运行小时数、正常运行时间 %、单箱周期时间未披露 |
| 具名多年期 RaaS 协议 | 1(GXO) | 2024-06 | SU001 | 高 | 单一锚定账户;集中度风险 | 合同年限、ACV、续约条款未公开 |
| 具名 Partner Program 参与者 | 3(GXO、Amazon、GEODIS) | 2025-Q1 | SU022, SU013 | 中 | 潜在生产账户管线;不等同于收入 | Partner Program 总名单;未确认未具名参与者 |
| 商业部署中 Digit 台数估计 | ~4-8(GXO 站点,推断) | 2025-Q4 | SU007, SU009 | 低 | 商业机队极早期;按当前台数,收入可能低于 $5M ARR | Agility 或 GXO 未披露实际机队规模 |
| RoboFab 产能(设计目标) | 10,000 units/yr | 2025 | SU005 | 中 | 若需求兑现,规模化时供给不会成为硬约束 | 当前实际生产速率未披露 |
| Amazon 试点站点 | 数量未说明 | 2023-2025 | SU012, SU013 | 中 | Amazon 正在验证技术,但尚未进入生产部署 | 台数、设施数量、试点范围未公开 |
料箱数量是唯一经独立验证的生产指标;机队规模和吞吐指标为推断估算。 Partner Program 参与者数量仅基于公开公告;可能存在未披露参与者。RoboFab 产能是设计目标,不是确认生产速率。
[CU008, CU009, CU017, CU018, CU019, CU020]企业仓储客户从初次认知到多年期生产部署的端到端旅程,突出关键决策门、延迟点,以及 Agility 在每一阶段的角色。
[CU019, CU020, CU021, CU022]估算从企业入站兴趣,到 Partner Program 接洽、活跃试点、生产部署的漏斗,基于公开数据和推断管线深度展示各阶段转化。
阶段数量为估计,依据公开 Partner Program 公告和行业情报;Agility 未公布管线数据。生产部署数量 1(GXO)已确认;其他阶段均为推断。
[CU017, CU018, CU019, CU022]6.4 留存与客户耐久性
Agility Robotics 的客户留存证据目前更多是结构性信号,而非已观察到的数据,因为客户基数小、 商业历史短。与 GXO 的多年 RaaS 协议是主要留存证据:Agility 未披露合同期限, 但多年意味着至少两年;GXO 持续运行并完成 100,000 个周转箱里程碑,说明关系仍活跃且在扩展。 GXO 未公开宣布终止合同、缩小范围或重新竞标,这是一个弱正面信号。 结构性锁定机制包括:(1)Arc 平台与 GXO 的 WMS 集成,若改用替代供应商,需要重新做 API 集成, 形成切换成本;(2)设施改造投入(围栏基础设施、地面传感器、充电站)已经沉没, 换供应商时需要替换;(3)操作员围绕 Digit 专属工作流形成的培训和肌肉记忆;(4)OSHA NRTL 站点特定批准绑定 Digit,同一设施若换成任何替代系统都要重新取得批准。 Agility 或 GXO 均未公开 NPS、CSAT 或运营满意度数据。曾与 GXO 运营人员交流的行业分析师 (截至 2025 年底见于行业媒体引用)称,该部署达到或超过吞吐目标,但这属于轶事证据。 Amazon 和 GEODIS 尚未公开评论试点结果。GXO 协议签署 18 个月后仍没有第二个具名生产客户, 是一个警示信号——它可能反映正常企业销售周期、RoboFab 产能约束,或其他潜在客户未披露的 试点失败。 [CU025, CU026, CU027, CU028, CU029, CU030]
| 指标 | 数值或空值 | 细分市场 | 置信度 | 尽调询问 |
|---|---|---|---|---|
| GXO 合同续约迹象 | 未公开披露 | 3PL 生产 | 低 | 索取 GXO 合同条款清单;确认续约意向或扩展条款 |
| 客户 NPS / CSAT | 空值——未发布 | 全部 | 低 | 索取操作员调查数据,或与 GXO DC 经理做客户访谈 |
| 试点转生产转化率 | 空值——3 个具名试点中仅 1 个生产客户 | 全部 | 低 | 确认试点总管线;询问转化时间线和阻碍 |
| Digit 在 GXO 的正常运行时间 / 可用性 | 空值——未披露 | 3PL 生产 | 低 | 索取 GXO 部署中的 Arc 平台正常运行时间日志,以及计划内 / 计划外停机数据 |
| 站点重复扩张(GXO 多站点) | 未公开确认 | 3PL 生产 | 低 | 询问 GXO 是否已扩展到 SPANX DC 之外;审阅合同中的多站点选择权条款 |
| 结构性切换成本 | 高(推断)——Arc WMS 集成、OSHA 重新批准、已沉没的设施改造 | 全部生产部署 | 中 | 量化客户切换到竞争性人形机器人时的估计重新部署成本 |
所有留存指标均为空值或结构性推断;截至 2026 年 5 月,Agility Robotics 没有公开 NPS、CSAT、流失率或续约率数据。 切换成本估计是定性判断,未量化。
[CU025, CU026, CU027, CU028, CU029]Agility 生产部署客户的留存队列估计,基于合同结构和转换成本分析,展示部署随时间延续的假设比例。GXO 是唯一真实数据点;其余数值为情景估计。
[CU025, CU026, CU027, CU028]6.5 扩张与集中度风险
客户集中度风险很高。GXO Logistics 是唯一公开确认的生产客户;如果 GXO 决定不续约或不扩大部署, Agility 的商业可信度和近期收入都会受到实质性损害。Amazon 的参与在财务和战略上都很重要, 但它是投资方加试点关系,而非已确认的生产收入来源。如果 Amazon 决定利用自有机器人能力, 或部署竞争性人形平台(Boston Dynamics Spot、Figure AI、1X Technologies), Agility 市场定位的下游影响可能很严重。 扩张驱动因素可信,但取决于执行。先落地再扩张的模式——先部署在大型 3PL 网络内一个站点, 再推向更多设施——是主要增长路径。GXO 在全球运营数百个 DC;即便只是再增加 10–20 个 GXO 站点、每站点 5–10 台机器人,也意味着相对当前推断基数扩大 10–20 倍。向 GEODIS、Amazon 和新的 Partner Program 进入者扩张遵循同一模式。瓶颈在部署周期(OSHA 批准、设施改造、 WMS 集成),而不是需求。 RaaS 模式放大了收入集中度风险:多年合同能降低季度间波动,但若大客户不续约, 收入会台阶式下滑,而不是逐步衰减。尽调应确认 GXO 的续约意向、处于主动部署阶段的客户管线 (而不只是 Partner Program 准入),以及是否有任何试点被终止或转为“无决定”。 [CU031, CU032, CU033, CU034, CU035, CU036]
| 扩张驱动因素 | 集中度风险 | 影响 | 尽调路径 |
|---|---|---|---|
| 在 GXO 网络内落地扩张(全球数百个 DC) | 单一生产客户;GXO 不续约将终止生产收入 | 高——若 GXO 不扩张或续约,收入将断崖式下滑 | 确认 GXO 扩张路线图;正在积极评估的站点数量 |
| Amazon 试点转为生产部署 | Amazon 有内部竞争机器人能力(Amazon Robotics) | 高——若 Amazon 退出,最大潜在需求信号消失 | 评估 Amazon Robotics 范围与 Agility 合作关系;确认试点结果 |
| Partner Program 转化管线(GEODIS 及未具名成员) | 除 GXO 外,没有具名 Partner Program 成员转入生产 | 中——管线存在,但转化率未验证 | 索取 Partner Program 成员数量和阶段分布 |
| 跨设施复用的 OSHA NRTL 批准模板 | 逐站点批准会把每个设施都变成瓶颈 | 中 — 即便客户需求已确认,也会拖慢收入落地 | 确认 OSHA 是否允许重复部署使用批准模板或加速流程 |
| 新垂直行业:汽车、零售、医药分销 | 仓储 3PL 之外尚无生产环境验证 | 短期低、长期高 — 多元化有上行空间 | 要求披露仓储 / 3PL 之外正在推进的试点 |
影响评估为定性判断,反映分析师基于公开部署数据和市场结构的推断。 Agility 未公开披露任何管线或预测数据。
[CU031, CU032, CU033, CU034, CU035]07风险
7.1 监管与法律风险
Agility Robotics 所处监管环境快速变化,覆盖工作场所安全标准、AI 治理框架、劳动法、 知识产权和出口管制。最直接的监管敞口是商业仓库机器人部署中的 OSHA 合规。根据一般责任条款 和 OSHA 1910.212 机器防护标准,部署 Digit 的雇主必须确保充分防护并消除危险。 当前所有 Digit 部署都采用围栏隔离配置,满足 ISO 10218-1/2 工业机器人安全规范。 但若按 ISO/TS 15066 转向协作式无围栏运行,就需要一套明显不同的风险评估框架、 新的 OSHA 合规文件,并可能重新谈判客户部署协议。协作认证前一旦发生任何工作场所事件, 都可能带来 OSHA 传票、媒体反噬和客户流失,成为公司商业轨迹的门槛性风险。 2024 年 8 月生效的 EU AI Act 将部署在安全关键工作场所场景中的 AI 系统归为高风险, 要求合格评定、技术文件、人类监督和透明度。共享仓库环境中的人形机器人可能落入 Annex III 高风险类别。美国关于 AI 安全的行政令以及新出现的两党机器人立法,显示联邦审查正在加强。 日本、英国、韩国和欧盟正在为自主系统制定不同监管框架,给 Agility 计划中的国际扩张 带来碎片化合规格局。 IP 风险中等但真实存在。Tesla、Boston Dynamics、Figure AI 和 Apptronik 已提交大量覆盖 双足运动、操控和机队管理的专利组合。随着 Digit 部署扩大、Agility 商业能见度提升, 专利侵权索赔的概率会上升。美国 Bureau of Industry and Security 管理的出口管制适用于 先进机器人技术;Agility 80% 的美国组件来源提供了一定缓冲,但 20% 的国际组件敞口、 以及 Digit AI 系统的军民两用潜力,仍带来出口许可风险。美国多个州和欧盟提出的机器人税立法, 可能提高 Digit 客户的总拥有成本,压低需求增长。 [CR001, CR002, CR003, CR004, CR005, CR006]
| 风险 / 事项 | 适用辖区 | 当前状态 | 发生概率 | 严重性 | 缓释措施 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| 客户设施发生工作场所安全事故并触发 OSHA 调查 | 美国联邦 / 州 OSHA | 潜在 — 尚无已知事故;围栏隔离部署降低但不能消除风险 | 中 | 严重 | ISO 10218 围栏合规;Arc 远程 E-stop;操作员培训规程 | 高 — 取得协作认证前,任何事故都可能引发执法和媒体反噬 | 要求 Agility 提供客户现场事故日志、险肇报告和 OSHA 处罚历史 |
| ISO/TS 15066 协作安全未合规 — 时间线滑坡 | 全球 ISO 和 OSHA | 尚未认证;目标为 2026 年;独立工程师评估可能滑至 2027-2028 年 | 高 | 高 | 认证项目已启动;CEO 明确列为优先事项;分阶段部署路径 | 高 — 滑坡会推迟无围栏部署,并限制可覆盖的设施类型 | 要求披露已聘认证机构名称和当前已完成的审计里程碑 |
| EU AI Act 将仓库人形机器人定为高风险 AI 系统 | 欧盟和 EEA | 安全关键场景中的人形机器人可能按 EU AI Act 附件 III 归为高风险 | 中 | 高 | 跟踪 EU AI Act 指引;提前准备符合性评估文件 | 中 — 进入欧盟需要完整符合性评估,并重新评估 CE 标志 | 聘请欧盟监管法律顾问;按附件 III 标准映射 Arc 和 Digit |
| 机器人税和自动化替代劳工立法影响客户 TCO | 美国州议会和欧盟委员会 | 多个辖区已有提案;截至 2026 年尚无法律生效 | 低-中 | 中 | 参与行业游说;用 ROI 文件支撑劳动力再配置叙事 | 低-中 — 若税收落地,会抬高客户 TCO,并压低需求管线 | 跟踪立法动态;在 CA、NY、WA 等高风险州聘请政策顾问 |
| 拥有更大 IP 组合的竞争对手提出专利侵权索赔 | 美国及 PCT 国际 | 尚无已知在审诉讼;行业内竞争对手专利组合正快速扩大 | 中 | 中 | Agility 扩大 IP 组合;开展 FTO 分析;为争议领域预留许可预算 | 中 — 商业能见度上升后,外部专利主张更可能找上门 | 要求 Agility 提供专利清单和 FTO 分析;审阅待决争议或警告函 |
| 先进机器人和 AI 部件受到出口管制限制 | 美国 BIS、EAR 和 ITAR | 尚无已知执法行动;自 2023 年起,全行业都在复核军民两用分类 | 低-中 | 中 | 80% 部件来自美国;对 Digit 和 Arc 的出口分类开展法律审查 | 低 — 中国人形机器人竞争对手提升该行业的国家安全敏感度后,风险会上升 | 与出口管制法律顾问确认 Digit 部件的 EAR 分类 |
| 工会和劳工组织反对客户设施部署 Digit | 美国和欧盟物流、仓储工会 | 物流领域工会已主动反对人形机器人;尚无已知部署被阻断 | 中 | 中 | 管理客户现场劳资关系;强化劳动力再配置叙事;准备 ROI 文件 | 中 — 有工会的客户场地部署摩擦更高,可能拖慢规模化 | 审阅关键客户场地的 CBA 条款,查找自动化限制条款 |
本清单仅基于截至 2026 年 5 月的公开监管披露、行业报道和已知立法活动。各行按严重性从严重到中排序。Agility 尚未发布监管风险清单或法律敞口披露;实际监管面可能明显更宽。所有严重性和发生概率评估均为分析师估计,投资交割前需要正式法律审查。EU AI Act 和 OSHA General Duty Clause 两项,是近期合规尽调优先级最高的事项。
[CR001, CR002, CR003, CR004, CR005, CR006]二维风险热力图,将 Agility Robotics 主要风险类别按发生可能性和影响严重性排列。每个单元格标出对应可能性-影响象限中的风险,便于在不同风险域中排序缓释资源和监控重点。
[CR001, CR011, CR021, CR030, CR037]7.2 运营与安全风险
Agility Robotics 的运营风险集中在五类故障模式:安全事故、电池与可靠性限制、网络安全漏洞、 供应链中断和制造爬坡失败。与 Digit 相关的工作场所伤害是严重性最高的运营风险。 这类事件会触发 OSHA 调查,引发大量媒体报道,促使客户暂停部署,并可能对更广泛的人形机器人行业 造成永久性声誉损害。当前过渡阶段风险更高:围栏隔离部署降低但不能消除人机接触事件, 尤其是在维护、充电坞交互,或需要在运行机器人附近干预的紧急情况下。 电池与可靠性限制是持续存在的运营约束。Digit 当前 4 小时电池续航要求多班次运营配置多个充电坞, 提高部署总拥有成本,并在充电坞处形成单点故障。Agility 未公开现场可靠性数据, 包括平均维护间隔时间、故障率和电池衰减曲线,因此外部难以独立评估运营风险。 在 GXO 和 Amazon 的商业部署中,任何现场故障模式都会削弱支撑企业采用的 RaaS 单位经济性。 Arc 云平台引入了网络安全攻击面。Digit 通过 WiFi 和 4G 蜂窝网络与 Arc 通信; 机队级命令注入或拒绝服务攻击可能同时瘫痪所有客户站点的商业运营。公司尚未公布 SOC 2 认证、 渗透测试结果或公开安全披露计划。Dark Reading 和行业安全分析师已将机器人机队连接性 标记为新兴攻击向量。 制造爬坡失败是最可量化的运营风险。RoboFab 设计产能为每年 10,000 台;当前商业部署约 100 台。这个 99 倍差距代表根本性执行挑战。供应链依赖包括 20% 国际组件来源和来自 Intel 的 半导体供应,进一步带来产速风险。若半导体短缺长期持续,或美国主要供应商发生重大故障, 都可能延迟支撑 Agility 收入和估值增长逻辑的生产爬坡。 [CR011, CR012, CR013, CR014, CR015, CR016]
| 失效模式 | 发生概率 | 严重性 | 缓释成熟度 | 剩余敞口 | 未解决缺口 |
|---|---|---|---|---|---|
| 客户现场发生 Digit 相关工作场所伤害 | 低-中(围栏降低概率) | 严重 | 部分 — 仅围栏合规;协作认证尚未取得 | 高 — OSHA 调查、媒体冲击、客户流失、行业声誉受损 | 没有公开现场事故数据;险肇报告机制未披露 |
| 电池耗尽和充电桩故障导致班次中途生产停机 | 中(4 小时上限已知) | 高 | 部分 — 有自主充电桩;已有多桩规划指引 | 中 — 会实质影响 RaaS 正常运行时间承诺和客户 ROI 模型 | 没有公开 MTBM 或电池衰减数据;缺少充电桩失效模式文件 |
| Arc 平台遭网络安全入侵,或机器人机队被注入指令 | 低-中(联网工业机队风险更高) | 高 | 低 — 没有公开 SOC 2;没有安全披露计划;没有发布渗透测试 | 高 — 机队级运营停摆;若 E-stop 通道被攻破,可能造成人身伤害 | 没有 SOC 2 认证;没有公开渗透测试结果;没有漏洞披露政策 |
| 半导体和部件供应中断,阻断 RoboFab 产能爬坡 | 中(历史上周期性发生;有地缘政治敞口) | 高 | 部分 — 80% 来自美国;多来源韧性尚未在规模化下证明 | 中 — 产能爬坡已远低于设计产能,任何中断都会叠加延误 | 20% 国际部件敞口未逐项列出;未披露备份供应商策略 |
| 制造爬坡失败 — RoboFab 无法向每年 10,000 台扩产 | 高(当前与设计产能相差 99x) | 高 | 低 — 工厂已建成,但任何有意义规模下的生产执行都未被证明 | 高 — 限制收入规模,加深客户集中,削弱估值逻辑 | 未向投资者公开生产排期或爬坡里程碑 |
失效模式按严重性和发生概率综合排序。缓释成熟度评级仅反映公开证据;Agility 可能有未公开的内部控制。Arc 是机队安全控制平面,网络安全项应立即尽调;远程 E-stop 命令通道一旦被攻破,就是安全关键失效,会同时带来人身和商业后果。所有未解决缺口都应列为正式投资流程中的优先尽调问题。
[CR011, CR012, CR013, CR014, CR015, CR016]有向图展示 Agility Robotics 的主要风险事件如何沿因果链传导,进入影响收入、客户、毛利、运营和投资人信心的下游结果。每条边代表风险事件之间的因果或条件依赖。
[CR011, CR014, CR021, CR030, CR037, CR038]7.3 伙伴与依赖风险
伙伴与依赖风险集中在三组关系:Amazon 投资方兼客户的双重关系、GXO 锚定客户依赖, 以及横跨 Intel、NVIDIA 和国内组件供应商的技术供应链依赖。Amazon 既可能是 Agility 最大客户,也是通过 Amazon Industrial Innovation Fund 进入的财务投资方。这种利益一致 同时是战略资产和结构性风险:Amazon 有能力、资本和机器人野心去开发自有人形替代方案。 有关 Amazon 内部机器人投入和收购历史的报道说明,这一风险并非理论问题。若 Amazon 决定 减少或终止 Digit 部署,转向内部开发方案,Agility 会同时失去一个标志性客户背书, 也可能失去一个财务支持方。 GXO Logistics 是唯一公开确认的多年 RaaS 商业合同。不续签 GXO 协议将实质性损害 Agility 的商业牵引叙事和 RaaS 收入。GXO 受到 3PL 行业宏观压力影响;若物流市场持续下行, GXO 的技术投资预算可能减少,并影响续约概率。SoftBank 和 WP Global Partners 的投资者关系 引入财务依赖风险:如果任一投资方面临内部资本约束或战略调整,后续融资支持可能在关键制造爬坡阶段延后。 Intel 和 NVIDIA 对 Digit 计算栈的供应依赖构成外部技术风险。若出现类似 2021–2022 年 全球芯片短缺的重大半导体供应中断,RoboFab 生产爬坡会直接放慢。Arc 云平台依赖公有云基础设施, 带来运营可用性风险:任何主要云区域故障都会同时中断所有客户站点的机队遥测、远程安全监控和任务调度。 这些平台级依赖未公开披露,多区域故障转移架构的程度也未知。 [CR021, CR022, CR023, CR024, CR025, CR026]
| 依赖 | 对手方 | 角色 | 集中度 | 失效情景 | 严重性 | 缓释措施 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| 客户和投资者 | Amazon (AIIF) | 最大潜在客户和 Series B/C 投资者;旗舰部署案例 | 严重 — 可能占商业部署机队的大多数;唯一确认的大型试点 | Amazon 终止 Digit 部署,并推出竞争性的内部人形机器人项目 | 严重 | RaaS 合同条款;AIIF 投资形成利益一致;客户侧 Arc 集成带来切换成本 | 高 — 尚未发现规模或战略可信度相当的替代客户 |
| 商业锚定合同 | GXO Logistics | 唯一公开确认的多年期 RaaS 协议;投资者最核心的商业验证点 | 高 — 若不续约,将消除主要可公开核验的商业牵引证据 | GXO 合同到期后不续约;宏观下行压缩 3PL 技术资本开支 | 高 | 多年期 RaaS 结构;Arc WMS 和工作流集成带来高切换成本 | 中 — GXO 品牌价值高,但合同独占性和续约条款未确认 |
| 财务支持方 | SoftBank | Series B 领投方;后续融资轮的战略可信度锚点 | 高 — 尚未发现规模或战略画像相当的替代投资者 | SoftBank Vision Fund 面临赎回压力,或战略重心转离机器人行业 | 高 | WP Global 和 Amazon AIIF 等多名共同投资者降低单一投资者集中度 | 中 — SoftBank 过往曾对被投公司突然战略转向 |
| 计算供应链 | Intel 与 NVIDIA | Digit 机载计算使用双 Intel i7;开放载荷舱可选 NVIDIA Jetson GPU | 中 — Intel 是主要计算供应商;NVIDIA 是可选载荷配置中的次要供应商 | 半导体长期短缺推迟 RoboFab 生产建设和交付排期 | 高 | 开放载荷舱架构允许替换为其他供应商的计算模块 | 中 — Intel i7 是标准商用芯片,有多条分销供应渠道 |
| 制造生产 | RoboFab(Agility 自有) | 位于 Oregon Salem 的唯一生产设施;设计产能为每年 10,000 台 | 严重 — 未披露备份制造基地或合同制造安排 | 工厂火灾、洪水、劳工行动或监管关停会让所有商业生产停摆 | 高 | 自有工厂降低供应商风险;Oregon 建筑和安全规范适用于运营 | 高 — 单一设施集中,且未披露应急制造计划 |
对手方按集中度和严重性综合排序。Amazon 同时是投资者和客户,是最复杂的依赖:近期利益一致性强,但若 Amazon 大规模推进内部人形机器人研发,利益会分化。人形机器人分析中常低估 RoboFab 单一设施风险;投资者应优先索取业务连续性和灾难恢复计划文件。
[CR021, CR022, CR023, CR024, CR025, CR026]这张有向依赖关系图展示 Agility Robotics 的关键外部依赖:客户、投资人、制造、供应链、平台基础设施和监管认证路径。 这些依赖共同决定商业执行能否跑通。
[CR021, CR022, CR023, CR024, CR025, CR026]7.4 人才与执行风险
Agility Robotics 的人才风险集中在高级领导层和技术联合创始人层面。CEO Peggy Johnson 于 2024 年 3 月加入,背景来自 Microsoft、Qualcomm 和 Verizon 的商业与业务发展, 而非机器人工程。Johnson 在完成 $400M Series C 融资和签下 GXO RaaS 协议上展现出 很强的商业直觉,但她在机器人制造运营上的经验有限,在 RoboFab 产能爬坡关键阶段会带来执行风险。 董事会和投资方群体尚未公开说明 CEO 以下的接班计划或运营梯队深度。 Damion Shelton(总裁兼联合创始人)是最关键的技术与战略锚点。作为拥有深厚产品和业务发展经验的 联合创始人,Shelton 连接了公司的技术与商业两端。他若离开,将显著冲击投资者信心和运营连续性。 Jonathan Hurst(Chief Robotics Officer 兼联合创始人)发明了 Digit 底层受食火鸡启发的 双足运动系统,并深耕腿足机器人动力学。Hurst 若流失,将拖慢运动控制研发;鉴于他在 OSU 共同开发 IP 中的基础角色,也会带来专利连续性风险。 AI 和机器人工程师的人才竞争极其激烈。Google DeepMind、Tesla、Meta AI,以及包括 Figure AI 和 Physical Intelligence 在内的资金充足创业公司,都在激烈争夺同一批强化学习、 运动控制和嵌入式系统工程师。Agility 作为一家商业部署约 100 台机器人规模的公司, 能否留住并吸引这类人才,高度依赖股权激励结构和技术使命吸引力。董事会一致性风险也存在: Amazon、SoftBank 和 WP Global Partners 可能有不同的退出时间偏好和风险容忍度, 在下一次战略拐点上形成治理摩擦。 [CR030, CR031, CR032, CR033, CR034, CR035]
| 角色 / 职能 | 依赖或缺口 | 中断概率 | 若流失的严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| CEO — Peggy Johnson(2024 年 3 月入职) | 机器人工程背景有限;关键制造爬坡期需要商业执行力 | 低-中(新任高管任期仍脆弱;此前无机器人运营经验) | 高 | 董事会支持强;在 Microsoft、Qualcomm、Verizon 有商业履历;已完成 Series C | 访谈 Johnson 的制造爬坡策略;要求提供 RoboFab 90 天运营计划 |
| 总裁兼联合创始人 — Damion Shelton | 深厚的产品和业务架构知识;公司最主要的组织记忆锚点 | 低(联合创始人股权绑定强;个人身份与公司使命绑定) | 严重 | 联合创始人股权绑定;组织知识嵌入产品和团队文化 | 梳理 Shelton 下方的汇报关系和决策权限;识别 VP 级继任者 |
| CRO 兼联合创始人 — Jonathan Hurst | 食火鸡式运动 IP 的核心发明人;运动控制 R&D 的技术可信度锚点 | 低-中(OSU 学术身份带来兼职状态风险) | 高 | OSU 专利转让;运动控制博士团队接受 Hurst 方法和路径训练 | 确认 OSU 研究的 IP 转让;核验 Hurst 是全职还是顾问身份 |
| AI 和机器人工程人才池 | 大型科技公司、Figure AI、Physical Intelligence、Tesla Optimus 对 RL 和 AI 人才竞争激烈 | 高(由外部市场条件驱动;Agility 无法单独控制) | 高 | 股权激励;技术使命吸引力;Oregon 生活成本低于 Bay Area | 要求提供过去 12 个月工程团队流失率;审阅 Glassdoor 和 Levels.fyi 数据 |
| 董事会和投资者一致性 | Amazon、SoftBank、WP Global 的回报周期、风险容忍度和战略利益不同 | 中(下一轮融资或 M&A 评估时摩擦会加大) | 中 | 结构化董事会治理;独立董事;带保护条款的投资者权利协议 | 审阅董事会构成;确认独立董事人数和资质;审阅关键投资者条款 |
联合创始人同时承担技术和战略角色时,人员风险更高。CEO 来自机器人行业外部,在制造爬坡期带来特定执行缺口风险——这也是 Agility 历史上运营复杂度最高的阶段。Jonathan Hurst 仍保留 OSU 关联,因此 CRO 职务需要厘清是全职还是顾问身份。人才市场竞争可部分通过招聘速度和 Glassdoor 数据观察;这些信号应与官方披露一并按季度跟踪。
[CR030, CR031, CR032, CR033, CR034, CR035]7.5 风险缓释与监控
Agility Robotics 已针对主要风险类别实施早期缓释措施,但大多数措施仍处于商业化前阶段, 尚未在规模化环境下验证。监管与安全风险的近期缓释,是围栏隔离部署方式和符合 ISO 10218 的 安全架构。若在 2026 年或 2027 年取得 ISO/TS 15066 协作安全认证, 将实质性降低监管暴露,并解锁更广泛的设施部署形态。运营风险中,Arc 的实时远程急停能力和 机队监控提供了部分网络安全缓释;但未公布 SOC 2 认证和安全披露计划,是一个监控缺口。 双来源组件认证和优先采用美国国内供应商,提供了部分供应链韧性。 投资者层面的风险监控应跟踪五个代表投资逻辑破裂触发点的门槛事件:第一,Digit 在客户站点发生 任何 OSHA 可报告事件;第二,GXO 合同不续约或缩小范围;第三,Amazon 宣布自研人形机器人 或逐步淘汰 Digit;第四,协作安全认证滑到 2027 年 Q4 之后;第五,Damion Shelton 或 Jonathan Hurst 离开高管或顾问角色。这些事件单独或叠加发生,都需要实质性重估投资逻辑。 能降低风险的正面监控指标包括:到 2026 年底已签多年 RaaS 合同达到三个或更多; 到 2026 年底商业场景中每月部署 Digit 达到 200 台或更多;以及协作认证进展里程碑, 包括具名认证机构和独立审计启动。学术与实务文献中的风险管理框架表明, 现阶段人形机器人部署应纳入季度运营 KPI 复盘、独立安全审计,以及针对重大监管变化预先谈妥的 合同升级条款。Agility 当前公开披露未确认这些机制存在,它们也是任何正式投资流程中的尽调问题。 [CR037, CR038, CR039, CR040]
| 风险 | 可监测触发信号 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 安全事故 — 客户设施发生涉及 Digit 机器人的工作场所伤害 | OSHA 可记录事故,或客户以安全担忧为由宣布暂停部署 | 任何需列入 OSHA 300 日志的 Digit 相关伤害;或客户以 Digit 安全为由主动叫停 | 立即重估投资逻辑;要求根因分析和整改计划;调查结果明确前,考虑减仓 |
| GXO 合同不续约,或范围较初始承诺大幅缩减 | GXO 新闻稿或 SEC 文件显示不续约;合同周年时 Agility 对续约保持沉默 | 确认 GXO 不续约,或范围缩减至原始 Digit 机队承诺的 50% 以下 | 从投资逻辑中移除主要商业验证点;下调收入增长假设;加快索取替代管线客户证据 |
| Amazon 竞争威胁 — 宣布内部人形机器人项目 | Amazon 发布内部人形机器人产品新闻稿或专利申请;Amazon 在 Sumner 设施逐步淘汰 Digit | 确认 Amazon 内部人形机器人开发项目进入生产测试阶段 | 重大重估 TAM 可触达性和商业独立性;评估 Amazon 从客户变成直接竞争对手的情景 |
| 协作安全认证滑至 2027 年底以后 | 到 2027 Q1 仍未披露认证机构名称;到 2027 Q3 仍未宣布审计启动里程碑 | CEO 或投资者沟通确认认证目标推迟到 2027 年 12 月之后 | 修订部署 TAM 和时间线假设;下调估值模型中赋予无围栏部署可选性的溢价 |
| 关键高管离职 — 总裁 Damion Shelton 或 CRO Jonathan Hurst | LinkedIn 或媒体发布离职消息;不再出现在投资者沟通中;出现职位空缺招聘 | 确认 Damion Shelton 辞任或被免去总裁职务,或 Jonathan Hurst 辞任或被免去 CRO 职务 | 高优先级响应:立即要求董事会继任计划;评估领导层连续性风险;考虑把保护性董事席位列为投资条件 |
| 制造爬坡停滞,低于最低可行商业门槛 | 季度生产指标;投资者更新中披露 RoboFab 产能利用率 | 到 2026 日历年底商业部署少于 300 台,即低于设计产能的 3% | 修订收入爬坡模型;提高扩产假设对应的折现率;要求提供包含已承诺供应商协议的详细生产计划 |
终止标准指需要重大重估投资逻辑的阈值事件,并不必然等于自动退出。每个触发信号都应通过公开披露、合作伙伴沟通和行业新闻按季度跟踪。安全事故触发信号优先级最高,因为它可能给 Agility 和整个人形机器人行业造成不可逆声誉损害。拥有董事会席位的投资者应确保所有触发信号纳入正式投后监控机制。
[CR037, CR038, CR039, CR040]08估值
8.1 投资逻辑与反向逻辑
截至 2026 年初,Agility Robotics 是商业化进展最靠前的双足人形机器人公司。 其投资逻辑建立在五个相互强化的支柱上。第一,公司做到了竞争对手尚未做到的事: 一个具名企业客户(GXO Logistics)在持续多月生产中运行商业 RaaS 合同,并在 SPANX 履约中心搬运超过 100,000 个周转箱。第二,公司拿下 Amazon 作为战略投资方和试点伙伴, 既获得资本,也获得接触全球最大仓库运营商的入口。第三,RoboFab 这个位于 Oregon 州 Salem 的 专用制造设施,代表了 $70M 资本投入,目标是建设每年 10,000 台的可扩展生产基础设施; 若达成,Agility 将成为第一个达到有意义工业规模的人形机器人制造商。第四,公司在 Oregon State University 十余年开发出的双足运动 IP,为人形机器人最难的工程问题提供了可防守护城河。第五, $400M Series C 带来 24–36 个月运营现金跑道,降低近期融资风险。 反向逻辑同样清楚。Agility 当前收入基础很薄:只有一个具名生产客户,估计 ARR 低于 $10M, $1.75B 估值意味着 150 倍或更高收入倍数——只有当 RaaS 规模化模型以强单位经济性跑通时, 这个倍数才站得住。Figure AI 在 Microsoft、OpenAI 等支持下以 $2.6B 估值融资, 其 Helix 通用 AI 模型可能让它更快打入广泛市场。Tesla Optimus 在 Tesla 既有 gigafactory 中制造,可能具备 Agility 无法匹敌的成本优势。Unitree G1 以每台 $16,000 进入市场—— 比 Digit 隐含的 RaaS 成本基准便宜三个数量级——瞄准开发者和研究市场,可能在企业市场成熟前 建立平台锁定。专业执行器的供应链集中度带来规模化生产风险。支撑 Agility 市场的劳动力替代叙事 也有政治风险:工会反对、自动化征税提案,以及 OSHA 按站点审批要求,都可能拖慢企业采用。
| 维度 | 评估 | 置信度 | 理由 | 行动含义 |
|---|---|---|---|---|
| 总体建议 | 观察 — 有条件买入 | 中 | 先发商业验证存在;收入规模未确认;按当前 ARR 看估值偏高 | 跟踪 RoboFab 产量、第二个客户、GXO 续约;在 Series D 或战略 M&A 事件时重新评估 |
| 风险评级 | 高 | 中 | 收入尚未规模化、生产客户单一、竞争对手资本强度高、供应链依赖 | 仓位应反映高风险属性;规模验证前避免超配 |
| 估值立场 | 偏高 / 合理 | 中 | Series C 隐含估值约 $1.75B,对应估计低于 $10M 的 ARR;只有 RaaS 按乐观情景放大才说得通 | 没有新催化前,老股交易不要支付高于 Series C 的价格;跟踪 Series D 锚定价格 |
| 置信水平 | 中 | 中 | 商业验证强于任何同业;财务规模尚未经过压力测试;竞争风险真实存在,但 Agility 有差异化 | 若 RoboFab 达到 1,000 台/年且签下第二个客户,上调为高信心买入 |
| 退出周期 | 3–5 年才可能出现有意义退出 | 中 | M&A 窗口为 2026–2028 年;IPO 准备度最早在 2028–2030 年 | 需要耐心;当前阶段流动性有限 |
建议仅反映公开证据。股权结构表细节、实际 ARR 和合同续约条款均未公开。立场反映 2026 Q1 的信息状态。
[CV001, CV002, CV003, CV004]| 主题 | 缺失证据 | 重要性 | 负责人 / 尽调路径 |
|---|---|---|---|
| GXO 合同续约状态 | GXO 是否已承诺续签初始多年期 RaaS 协议;合同 ACV 和剩余期限 | 唯一量产客户;不续约会打掉商业化证据;投资逻辑以这份合同为锚 | CFO/CRO 直接尽调;GXO 运营副总裁访谈;合同条款法律审阅 |
| RoboFab 产能爬坡验证 | Salem 工厂投产以来的实际季度产量;良率;供应链瓶颈 | 乐观 / 基准情景的分界完全取决于 1,000–5,000 台/年爬坡;未验证产量是财务模型的首要风险 | 运营审计;现场走访;供应商访谈;Agility CFO/COO 直访 |
| C 轮优先股堆叠与股权结构表 | 清算优先权结构;反稀释条款;$1.5B–$2.5B 并购情景下的优先权压力 | 在低于 $2B 的悲观 / 基准并购退出中,优先权压力可能吃掉普通股回报 | 公司律师审阅条款清单;ARCH Venture Partners;Delaware 公司注册证书 |
| 第二个具名量产客户 | 除 GXO 外,已签署 RaaS 协议的第二个具名企业客户 | 证明获客可复制;降低单一客户集中风险;验证定价模型 | 业务拓展渠道核查;物流行业访谈;供应链会议情报 |
| 当前机队规模下的毛利率轨迹 | GXO RaaS 合同的实际毛利率;每台 Digit 硬件成本;服务 / 支持成本结构 | 当前规模毛利率为负在预期内,但轨迹决定 D 轮价格 | CFO 直接尽调;供应链成本建模;参考工业自动化 RaaS 毛利率基准 |
尽调问题按对投资决策的重要性排序。第 1–3 项是上调至买入的阻断项。第 4–5 项对观察仓位重要但不阻断。
[CV022, CV023, CV024, CV025]8.2 融资背景与估值框架
Agility Robotics 已通过多轮股权融资累计募集约 $1.2B。融资历史始于 Oregon State 获 DARPA 资助的早期研究,随后是 2020 年 Series A、2022 年由 Amazon Industrial Innovation Fund 领投的约 $150M Series B,以及 2025 年 3 月由 ARCH Venture Partners 领投、NEA、DCVC 和 Amazon 参与的 $400M Series C。Bloomberg 和 CB Insights 报道的 Series C 投后估值约为 $1.75B,但精确股权结构——包括优先股堆叠、清算优先权和反稀释条款—— 尚未公开披露。投资者应假设存在标准 VC 清算优先权;若为 2x 参与型,在低于 $2.5B 的中等退出情景中, 普通股回报会被显著压低。 适用于 Agility 的估值框架是按情景加权的折现里程碑模型,而不是收入倍数。在 ARR 低于 $10M 时, 传统 SaaS 倍数没有意义。正确锚点是:RaaS 模型在 1,000+ 台机器人部署时的概率加权预期价值, 折现到今天是多少?在基准情景下,到 2028 年部署 2,000 台机器人、每台每年 RaaS 收入 $120K, 年经常性收入(ARR)约达 $240M;按工业软件可比公司 7–12 倍的远期 EV/ARR 倍数计算, 企业价值为 $1.7B–$2.9B。Series C 价格在这一进入点是合理的,但不便宜: 上行空间要求 RoboFab 执行到位、多站点客户扩张,以及毛利率在机队规模下从当前为负提升至 40–50%+。Forge 和 EquityZen 的老股交易市场数据表明流动性很薄,买卖价差相对 Series C 价格有 20–35% 折价,反映规模化不确定性。
| 论点 | 证据 | 权重 | 何种变化会改变这一判断 |
|---|---|---|---|
| 正向:唯一已达到商业化人形机器人生产里程碑 | GXO SPANX:搬运 100,000+ 个周转箱;多年期 RaaS 协议;OSHA NRTL 站点批准 | 强 | 若 GXO 不续约或公开披露运营失败,判断反转 |
| 正向:Amazon 战略投资并担任试点伙伴 | Amazon Industrial Innovation Fund;Agility Partner Program;Amazon 设施持续开展部署测试 | 强 | 若 Amazon 推出达到生产规模的内部人形机器人项目,论点削弱 |
| 正向:RoboFab 制造基础设施 | 资本开支 $70M+;Salem OR 设施;10,000 台/年设计产能;Agility 保留制造控制权 | 中 | 若 RoboFab 产能爬坡晚于 2025 Q4 承诺,论点明显削弱 |
| 反向:估值相对已披露收入偏高 | ARR 估计低于 $10M;投后隐含估值 $1.75B;收入倍数 >150x | 强 | 若 2026 年靠机队扩张把 ARR 推到 $100M+,该反向论点被中和 |
| 反向:竞争对手资本强度和覆盖广度 | Figure AI($2.6B)、Tesla Optimus(超级工厂支撑)、Unitree G1(每台 $16K)、Apptronik(Samsung 支撑) | 强 | 若 Agility 在 500+ 台机器人规模证明 RaaS 毛利率 >40%,该反向论点减弱 |
证据评级反映每个论点的来源深度和独立性。所有主张均基于公开信息。
[CV005, CV006, CV007, CV008, CV009]8.3 可比公司与市场基准
收入前阶段的人形机器人公司没有直接上市公司可比。最有用的可比对象横跨三类: 人形机器人私募轮次、相邻工业自动化公司,以及战略并购先例。 在人形机器人私募轮次中,Figure AI 是最相关的同业。Figure AI 在 2024 年 2 月一轮融资中 募集 $675M,投后估值 $2.6B,投资方包括 Microsoft、Bezos Expeditions 和 Intel Capital。 其 BMW 工厂试点和 OpenAI 集成显示出不同的 GTM(汽车 / 通用型,相对于 Agility 的物流专用型)。 Apptronik 2024 年从 Samsung 等投资方募集 $350M,估值未披露。1X Technologies 2024 年 在 OpenAI 支持下以约 $400M 估值募集 $100M。Tesla Optimus 未独立估值,但受益于 Gigafactory 制造集成。同业区间说明,收入前人形机器人的估值可在 $400M–$2.6B 之间波动, 取决于制造验证、战略支持和伙伴生态深度。Agility 的 $1.75B 处于该区间上沿, 其商业生产里程碑可以支撑这一位置,但相对未公开披露收入的同业仍显得偏紧。 在工业自动化领域,Zebra Technologies(工业移动机器人,约 $15B)、Cognex(机器视觉, 约 $8B)和 Teradyne(协作机器人,约 $4B)的交易倍数为远期收入 5–8 倍。 这些不适合作为直接倍数,但如果 Agility 能实现 $250M–$500M ARR,可为退出区间提供锚点。 Boston Dynamics 先例有参考意义:Hyundai 2021 年以 $1.1B 从 SoftBank 手中收购 Boston Dynamics—— 价格低于开发成本,但符合收入前机器人战略并购逻辑。若 RoboFab 达到每年 3,000–5,000+ 台, 并证明多客户 RaaS,带战略溢价的物流自动化战略收购方可能给 Agility $3B–$5B+ 估值。
| 情景 | 关键假设 | 2028 年 ARR | 隐含估值 | 概率信号 | 关键风险 |
|---|---|---|---|---|---|
| 乐观 | RoboFab 到 2027 年达到 5,000 台/年;5+ 个具名生产客户;RaaS 每台机器人每年 $120K;毛利率 50%+ | $350–500M | $3.5B–$6B(8–12x ARR 倍数) | 30%(需要制造、销售和毛利率同时跑通) | 竞争对手突破,在规模化前侵蚀定价权 |
| 基准 | RoboFab 到 2027 年达到 1,500–2,500 台/年;2–3 个量产客户;混合 RaaS 为 $100K/机器人/年;毛利率 30% | $150–250M | $1.5B–$2.5B(8–10x ARR 倍数) | 45%(需要 GXO 续约 + 一个新增量产客户) | 供应链延误或 OSHA 扩展监管阻断多站点铺开 |
| 悲观 | RoboFab 延期至 2028 年;GXO 不续约;生产卡在试点规模;竞争对手压价 | $15–30M | $700M–$1.0B(低于 C 轮) | 25%(考虑单一客户集中和规模未经验证,具备可能性) | 被迫降估值融资或困境并购;普通股受损 |
情景输入为模型估计,依据包括公开定价、产能目标和相似工业自动化规模化曲线。
[CV010, CV011, CV012, CV013]8.4 情景与回报分析
2025–2030 年投资周期内,Agility 的估值路径由三种情景主导。乐观情景假设 RoboFab 到 2027 年达到每年 5,000+ 台,拥有 5+ 个具名生产客户,ARR 到 2028 年超过 $300M, 并以 10–15 倍 ARR 完成 IPO 或战略并购,对应 $3B–$5B 企业价值。Series C 进入的投资者 在稀释前 3–4 年回报约为 2–3 倍。基准情景假设 RoboFab 到 2027 年生产 1,500–2,500 台, 拥有 2–3 个具名生产客户,2028 年 ARR 为 $150M–$250M,并以 $1.5B–$2.5B 完成并购退出。 Series C 回报约为 0.9–1.4 倍——按 Series C 价格看只是勉强。悲观情景包括 RoboFab 生产延迟到 2028 年以后、GXO 客户流失、OSHA 驱动的部署限制,以及无法按 Series C 价格 募集 Series D,最终以下调估值的 $700M–$1.0B 融资,或以困境估值被迫并购。该情景下 Series C 回报为 0.4–0.6 倍,优先股堆叠会保护机构投资者,但普通股会遭到重创。 下行情景不是尾部风险:六个投资逻辑破裂触发点中有三个(RoboFab 延迟、第二个客户增加、 OSHA 批准泛化)在 18 个月内可观察。主动型投资者应围绕这些里程碑建立监控仪表盘, 并在 Series D 进入时设定价格纪律。截至 2026 年 Q1,概率分布约为 30% 乐观、45% 基准、 25% 悲观。
| 可比对象 | 类型 | 指标 / 估值 | 收入阶段 | 参考意义 | 局限 |
|---|---|---|---|---|---|
| Figure AI | 人形机器人私募轮(2024 年 2 月) | $2.6B 投后;融资 $675M | 收入前(BMW 试点) | 最接近的同业;通用 AI 人形机器人;Microsoft/OpenAI 支持 | 商业路径不同(汽车 / 通用 vs. 物流);尚无商业量产里程碑 |
| Boston Dynamics | 战略并购(Hyundai,2021 年) | $1.1B 收购价 | 收入前人形机器人(Spot 已商业化但规模小) | 战略并购先例;显示机器人平台并购底价 | 人形机器人尚未商业化;Spot 是主要收入来源;买方画像不同 |
| Apptronik | 人形机器人私募轮(2024 年) | 融资 $350M;估值未披露 | 收入前(物流 / 汽车试点) | Samsung 战略投资方;Apollo 人形机器人瞄准物流 | 估值未披露;Samsung 制造协同可能压缩估值折价 |
| 1X Technologies | 人形机器人私募轮(2024 年) | 约 $400M 投后;融资 $100M | 收入前(试点部署) | OpenAI 支持;双足;规模小于 Agility | 规模更小;披露部署更少;阶段早于 Agility |
| Tesla Optimus | 内部项目(上市公司) | 未独立估值 | 内部部署阶段 | Gigafactory 制造优势;最大上市公司赞助方 | 无法独立交易;制造成本不可比 |
| Agility Robotics(C 轮) | 人形机器人私募轮(2025 年 3 月) | $1.75B 投后;$400M C 轮 | 商业化生产(GXO);估计 ARR 低于 $10M | 标的公司;参考估值 | 收入倍数 >150x;没有规模证明时偏高;优先股堆叠未披露 |
可比估值来自公开报道。Boston Dynamics 数字为据报道的收购价。估值口径截至最近一次公开披露融资轮。
[CV014, CV015, CV016, CV017, CV018]8.5 建议、退出准备度与最终尽调问题
建议姿态是观察(TRACK),并设置有条件买入触发点。按 Series C 价格(隐含 $1.75B)进入, 多数机构 LP 授权下的风险 / 回报不具吸引力:基准情景 3 年回报为 1.0–1.4 倍, 悲观情景则意味着实质性资本损失。不过,若投资者拥有战略机器人基础设施授权、长期持有周期(5–7 年), 或能与 ARCH/NEA 在 Series C 价格共同投资,乐观情景上行空间($3B–$5B 退出、2–3 倍) 可以支撑相应仓位。 退出准备度中等。战略收购方正在积极评估人形机器人领域:Amazon 这类战略投资者通常会有 优先购买权条款,Toyota 的 Woven Planet 部门已表达对双足机器人合作的兴趣,Hyundai 也已证明 愿意为商业化前状态的 Boston Dynamics 支付 $1.1B。如果 RoboFab 实现每年 2,000+ 台产能, 并签下第二个具名生产客户,2026–2028 年实现 $2B–$3.5B 战略并购退出是可行的。 IPO 准备度要求 $200M+ ARR 和正毛利率;在基准情景下,最早也可能要到 2028–2030 年。 在上调至买入(BUY)之前,五个关键尽调缺口是:确认 GXO 在初始期限后的续约;通过第三方运营审计 验证 RoboFab 产速;签下第二个具名生产客户协议;披露 Series C 优先股堆叠和清算瀑布; 以及证明当前机队规模下毛利率正在向盈亏平衡改善。
| 触发项 | 阈值 / 信号 | 对投资逻辑的传导 | 提前期 | 行动含义 |
|---|---|---|---|---|
| GXO 合同未续约或公开运营失败 | GXO 新闻稿或财报披露合同终止,或 Digit 出现重大运营问题 | 唯一的商业量产证据消失;竞争对手叙事被触发;估值有重估至 $1B 以下的风险 | 12–18 个月 | 立即退出或对冲;下调至卖出;把 Amazon 反应作为次级信号监控 |
| 到 2026 年底确认 RoboFab 产量低于 500 台/年 | 第三方媒体报道或客户交付时间线与 10K/yr 目标不一致 | 乐观情景出局;基准情景受威胁;下轮降估值风险显著上升 | 6–12 个月 | 降低持仓;参与 D 轮前要求 RoboFab 审计证据 |
| Figure AI 在物流场景跑通量产规模商业 RaaS | Figure AI 公开宣布与具名客户签署 100+ 台、多年期物流 RaaS 合同 | 削弱 Agility 的先发叙事;定价权承压;客户选择提速 | 12–24 个月 | 重新评估竞争差异;聚焦 Agility 毛利率与 Figure AI 定价的对比 |
| Amazon 大规模推出内部竞争性人形机器人 | Amazon 公开宣布自研人形机器人制造,或收购竞争对手 | 战略投资方变成竞争对手;优先地位不再带来定价优势 | 12–36 个月 | 复核合作协议中的优先购买权和竞业限制条款 |
| D 轮估值大幅走低(低于 $1.5B 的降估值融资) | 公开报道显示 D 轮成交价格低于 C 轮 | 确认悲观情景;普通股受损;释放内部执行短板信号 | 18–30 个月 | 退出二级持仓;若确认战略买方兴趣,评估困境并购时点 |
| OSHA 发布限制人形机器人仓库部署的通用指引 | OSHA Federal Register 公告或执法行动引用人形机器人事故 | 客户流失风险实质上升;多站点审批流程停摆;市场节奏推迟 2–3 年 | 6–18 个月 | 全面复核持仓;核查企业客户情绪;就监管影响取得法律意见 |
触发项在 2026 年 Q1 后 6–18 个月内可观察。监控看板应纳入季度 RoboFab 产量报告和公开披露。
[CV019, CV020, CV021]免责声明
本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;任何投资决策前, 都应直接向管理层和一手文件核验。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Agility Robotics is headquartered in Redmond, Washington, and operates a manufacturing facility (RoboFab) in Salem, Oregon. | 高 | SO015, SO006 |
| CO002 | Agility Robotics' primary product is Digit, a bipedal humanoid robot designed for warehouse and logistics automation. | 高 | SO015, SO023 |
| CO003 | Agility Robotics operates a Robots-as-a-Service (RaaS) model, typically pricing Digit at approximately $10,000–$15,000 per robot per month. | 中 | SO013, SO012 |
| CO004 | Agility Robotics also offers direct Digit sales at approximately $150,000 per unit. | 中 | SO013 |
| CO005 | Approximately 80% of Digit's ~6,000 components are sourced from the United States, per Agility Robotics' official website. | 中 | SO015 |
| CO006 | Agility Arc is a cloud-based fleet management platform that orchestrates Digit robots, integrates with warehouse management systems, and provides analytics. | 中 | SO015 |
| CO007 | Agility Robotics was co-founded in 2015 by Jonathan Hurst (Oregon State University professor) and Damion Shelton as a spin-off from OSU's Dynamic Robotics Laboratory. | 高 | SO011, SO009 |
| CO008 | Damion Shelton served as CEO of Agility Robotics from founding (2015) through early 2024, and currently serves as President and Chairman. | 高 | SO009, SO010 |
| CO009 | Peggy Johnson, former President at Microsoft and CEO of Magic Leap, became CEO of Agility Robotics in March 2024. | 高 | SO009, SO001 |
| CO010 | Jonathan Hurst now serves as Chief Robot Officer, focusing on technical vision and research. | 高 | SO009, SO010 |
| CO011 | Melonee Wise, formerly CTO and CEO of Fetch Robotics, joined Agility Robotics as Chief Product Officer. | 中 | SO010 |
| CO012 | No contested founder departures, governance controversies, or leadership disputes at Agility Robotics have been documented in public sources as of mid-2026. | 中 | SO009, SO010 |
| CO013 | Agility Robotics raised $400M in a Series C round in March 2025, led by WP Global Partners with participation from SoftBank Group and the Amazon Industrial Innovation Fund. | 高 | SO001, SO017, SO018, SO025 |
| CO014 | The Series C valued Agility Robotics at approximately $1.75 billion pre-money and ~$2.12–2.15 billion post-money. | 高 | SO001, SO025, SO020 |
| CO015 | Agility Robotics raised $150M in a Series B round in 2022 at approximately $620M valuation, led by Amazon's Industrial Innovation Fund alongside DCVC and Playground Global. | 高 | SO003, SO019 |
| CO016 | Amazon's Industrial Innovation Fund invested in Agility Robotics in 2021 (amount undisclosed) and again as part of the 2022 Series B and 2025 Series C. | 高 | SO001, SO003 |
| CO017 | Ford Motor Company invested approximately $6M in Agility Robotics in 2020 in an early-stage round. | 中 | SO003, SO019 |
| CO018 | SoftBank Group had reportedly explored a potential acquisition of Agility Robotics at a valuation exceeding $900M before becoming a Series C investor. | 中 | SO002, SO017 |
| CO019 | Agility Robotics' total funding raised across all rounds is approximately $641M as of the March 2025 Series C close. | 高 | SO003, SO025, SO018 |
| CO020 | Third-party estimates put Agility Robotics' annual revenue at approximately $35.5M for 2025, representing revenue per employee of roughly $141K. | 低 | SO014, SO012 |
| CO021 | Agility Robotics' headcount is estimated at 250–500+ employees, with a stated target of 500+ at the Salem RoboFab facility. | 中 | SO007, SO014 |
| CO022 | Digit robots at GXO Logistics' Flowery Branch, Georgia facility moved more than 100,000 totes in commercial operation as of November 2025. | 高 | SO004, SO001 |
| CO023 | The Agility Robotics RoboFab facility in Salem, Oregon is a 70,000 sq ft factory targeting production capacity of over 10,000 Digit units per year. | 高 | SO006, SO007, SO008 |
| CO024 | Agility Robotics' revenue is expected to be pre-breakeven given the capital investment required to scale manufacturing; no debt financing or secondary transactions have been publicly disclosed. | 中 | SO012, SO019 |
| CO025 | CASSIE, the first major bipedal robot from Agility Robotics, was unveiled in 2016 and sold primarily to academic research labs. | 高 | SO011, SO006 |
| CO026 | Agility Robotics unveiled the original Cassie bipedal prototype in 2016, based on research at OSU's Dynamic Robotics Laboratory. | 高 | SO011, SO006 |
| CO027 | Digit, Agility Robotics' commercial humanoid robot with arms and manipulation capability, was introduced circa 2019–2020 in first-generation form. | 高 | SO006, SO011 |
| CO028 | RoboFab, described as the world's first dedicated humanoid robot factory, was publicly announced in September 2023 and opened in 2024. | 高 | SO006, SO008 |
| CO029 | Amazon began piloting Digit robots at its Sumner, Washington fulfillment center in late 2023 for tote consolidation tasks. | 高 | SO023, SO001 |
| CO030 | GXO Logistics signed a multi-year RaaS agreement with Agility Robotics in June 2024, described as an industry-first commercial humanoid deployment contract. | 高 | SO005, SO001 |
| CO031 | CEO Peggy Johnson outlined a vision at the 2026 Abundance Summit for uncaged Digit robots capable of working safely alongside humans without safety barriers. | 中 | SO021, SO022 |
| CO032 | Agility Robotics' fifth-generation Digit is targeted to handle payloads up to 50 pounds and aims for broader market rollout in 2027. | 中 | SO021, SO022 |
| CO033 | Agility Robotics is ranked No. 6 on GeekWire's 200 index of Pacific Northwest startups. | 高 | SO001, SO023 |
| CO034 | Agility Robotics states that Digit is the first humanoid robot in production deployment on its official homepage. | 高 | SO015, SO023 |
| CO035 | Agility Robotics' homepage states its origin date as establishing the standard for safe humanoids and indicates the company was operational for over a decade as of 2026. | 中 | SO015 |
| CO036 | Digit includes multiple camera arrays, LiDAR, and IMU sensors, and features bird-inspired leg design reflecting decades of bipedal locomotion research. | 高 | SO023, SO001 |
| CO037 | Since Peggy Johnson became CEO in March 2024, Agility Robotics has visibly pivoted toward commercialization and manufacturing scale-up, with the GXO multi-year contract, RoboFab opening, and $400M Series C all occurring under her tenure. | 中 | SO009, SO021 |
| CM001 | The global humanoid robot market is estimated at approximately $2.92 billion in 2025, projected to reach ~$15.26 billion by 2030 at 39.2% CAGR, per MarketsandMarkets. | 高 | SM003, SM007 |
| CM002 | BCC Research estimates the humanoid robot market at $1.9B in 2025, growing to $11B by 2030 at 42.8% CAGR—a somewhat lower but still aggressive growth trajectory. | 中 | SM004 |
| CM003 | The global warehouse automation market is estimated at approximately $29.9 billion in 2025, projected to reach $63.4 billion by 2030 at 16.2% CAGR, per ResearchAndMarkets. | 高 | SM001, SM013 |
| CM004 | The global logistics robot market (AMRs, AGVs, humanoids combined) was approximately $14.5 billion in 2024, projected to reach $35.1 billion by 2030 at 15.9% CAGR, per Grand View Research. | 高 | SM002, SM008 |
| CM005 | The warehousing and logistics robot market (all robot types) is estimated at $7.62 billion in 2025, growing at approximately 26% CAGR, per Market Research Future. | 中 | SM009 |
| CM006 | Applying a 20–30% logistics share to the $2.92B humanoid robot TAM yields an estimated $580M–$870M logistics-specific humanoid TAM in 2025. | 低 | SM003, SM016 |
| CM007 | US warehouse labor employs approximately 1.8 million workers at $42,000–$55,000 fully-loaded annual cost, creating a $75B–$99B annual labor spend pool as a theoretical labor displacement TAM. | 低 | SM010, SM024 |
| CM008 | Agility Robotics' serviceable addressable market in 2025—large 3PL and e-commerce operators in the US and Western Europe willing to sign RaaS contracts—is estimated at $800M–$1.5B. | 低 | SM006, SM003 |
| CM009 | Agility's SOM in 2025 is estimated at $20M–$100M, constrained by RoboFab production capacity and the number of active commercial deployments. | 低 | SM017, SM021 |
| CM010 | As of 2025, only approximately 5% of warehouses globally are fully automated, creating a large addressable market for automation solutions. | 中 | SM011, SM010 |
| CM011 | Tier-1 third-party logistics operators—such as GXO, DHL, and Ryder—represent the primary near-term buyer segment for Agility Robotics' Digit, given large standardized facility footprints and acute labor pressure. | 中 | SM025, SM020 |
| CM012 | Large e-commerce operators (e.g., Amazon, Walmart) are early adopters with high volume and strong automation investment, but they also develop competing in-house robotics solutions. | 高 | SM021, SM024 |
| CM013 | Budget ownership for multi-year RaaS humanoid contracts typically sits with COO/CFO-level executives at large logistics operators, reflecting the capital significance of 10+ unit deployments. | 中 | SM018, SM024 |
| CM014 | Enterprise sales cycles for large-scale humanoid robot deployments are typically 12–18+ months due to safety certification, WMS integration complexity, and ROI validation requirements. | 中 | SM006, SM020 |
| CM015 | Smaller logistics operators (below ~$50M revenue) are effectively outside Agility's current SAM due to minimum RaaS contract economics and integration overhead. | 中 | SM017, SM018 |
| CM016 | US warehouse and logistics employment faced persistent structural shortages in 2024–2025, with open positions routinely exceeding 500,000 even at wages above $20/hr, creating durable demand for automation. | 高 | SM010, SM024 |
| CM017 | US e-commerce sales grew approximately 10% annually in 2024–2025, driving increasing fulfillment volumes and creating pressure for warehouse operators to adopt automation to meet throughput demands. | 高 | SM012, SM013 |
| CM018 | Over $10 billion in cumulative private funding had been directed to humanoid robot companies globally by end of 2025, supporting sustained R&D and manufacturing scale-up. | 中 | SM014, SM007 |
| CM019 | Advances in reinforcement learning and vision-language-action models have pushed humanoid robot capabilities to sustained commercial deployment in 2024–2025, as demonstrated by Digit's 100K+ tote milestone at GXO. | 中 | SM020, SM022 |
| CM020 | OSHA's growing recognition of ergonomic injury costs in warehouse settings creates regulatory tailwinds for automation of physically demanding repetitive tasks. | 中 | SM023, SM024 |
| CM021 | At $10,000–$15,000/month per Digit unit, the RaaS breakeven vs. human labor requires near-continuous robot utilization and competitive downtime economics. | 中 | SM017, SM018 |
| CM022 | Interact Analysis estimates that actual humanoid robot market penetration in logistics remains very limited through 2027, with meaningful scale expected after 2028. | 中 | SM006 |
| CM023 | All current Digit deployments as of 2025–2026 operate in safety-caged or segregated environments because the robot lacks full cooperative safety certification for unrestricted human co-working. | 中 | SM022, SM023 |
| CM024 | Enterprise buyers face 3–6 months of WMS integration, facility mapping, and safety onboarding before Digit can enter production deployment, representing meaningful integration friction. | 中 | SM006, SM020 |
| CM025 | Well-capitalized humanoid robot competitors—including Tesla Optimus (targeting $20–30K per unit), Figure AI, and Boston Dynamics—are developing competing platforms that may erode Agility's first-mover advantage. | 高 | SM014, SM007 |
| CM026 | The primary geographies for humanoid robot logistics deployment in 2025 are the United States and Western Europe, with Asia-Pacific (China, Japan, Korea) being the fastest-growing region globally. | 中 | SM007, SM015 |
| CM027 | Humanoid robot market size estimates from leading analyst firms diverge by 50%+ for 2030 projections ($11B vs. $15.3B), reflecting deep uncertainty about adoption pace, technology readiness, and enterprise willingness. | 高 | SM003, SM004 |
| CM028 | Manufacturing is a growing second vertical for humanoid robots alongside logistics, with pilots at BMW (Figure AI) and Hyundai factories (Boston Dynamics Atlas) setting expectations for the segment. | 中 | SM014, SM005 |
| CM029 | Agility Robotics' target ROI for Digit is reportedly under 2 years vs. a human worker at roughly $30/hr, according to business model analysis sources. | 中 | SM018, SM017 |
| CM030 | No publicly documented cases of major commercial humanoid robot deployments being permanently pulled back due to performance failure have been identified as of mid-2026. | 中 | SM006, SM020 |
| CM031 | The 3PL market concentration (GXO, DHL, XPO, Ryder controlling significant market share) means that winning a small number of large Tier-1 operators can represent a disproportionate share of Agility's early SAM. | 中 | SM025, SM024 |
| CM032 | Achieving cooperative safety certification is described by Agility CEO Peggy Johnson as essential for broad humanoid robot adoption and targeted for late 2026 to 2027. | 中 | SM022, SM023 |
| CM033 | Key status-quo substitutes for humanoid robots in warehouse logistics include AMRs (autonomous mobile robots), traditional conveyors, robotic picking arms, and manual warehouse labor. | 高 | SM002, SM009 |
| CM034 | Humanoid robots differentiate from AMRs and fixed automation by operating in unstructured, human-designed environments without requiring facility redesign—a key competitive advantage for brownfield logistics operators. | 中 | SM020, SM022 |
| CM035 | NVIDIA, Microsoft, and SoftBank were among the leading investors in humanoid robot companies through 2025, reflecting strategic importance of the sector beyond pure financial returns. | 中 | SM014, SM016 |
| CP001 | Figure AI raised over $1 billion in a Series C round in September 2025, reaching a $39 billion post-money valuation—the highest in the humanoid robotics sector. | 高 | SP001, SP007, SP008 |
| CP002 | Figure AI's B2 robot stands 5'6", weighs 60kg, carries 20kg payload, walks at 1.2 m/s, runs for 5 hours on a battery charge, and features 5-fingered dexterous hands. | 中 | SP002, SP011 |
| CP003 | Figure AI's Helix visual-language-action (VLA) model enables generalist task learning from human demonstration, representing a more capable AI approach than Agility's task-specific logistics AI stack. | 中 | SP002, SP006 |
| CP004 | Figure AI's BMW factory pilot, initiated in early 2024, was discontinued after the robot required significant teleoperation support and did not achieve autonomous operation at commercial scale. | 中 | SP006, SP007 |
| CP005 | Tesla Optimus Gen 2 is 5'8", weighs 57kg, carries 20kg payload, and achieves up to 8 km/h walk speed, leveraging Tesla's FSD autonomous driving AI stack. | 中 | SP009, SP010, SP025 |
| CP006 | Tesla has publicly targeted a $20,000–$30,000 per-unit price for Optimus, though the robot is not yet commercially available for external sale as of mid-2026. | 中 | SP009, SP011 |
| CP007 | As of mid-2026, Tesla Optimus is deployed exclusively in Tesla's own manufacturing facilities; no external commercial sale or RaaS model has been announced. | 中 | SP009, SP025 |
| CP008 | Boston Dynamics' electric Atlas (2024) stands 6'2", weighs 90kg, achieves 50kg payload (30kg sustained), has 56 degrees of freedom, and is rated IP67 for water and dust resistance. | 高 | SP003, SP025 |
| CP009 | Boston Dynamics Atlas is priced at approximately $150,000 or more per unit and is focused on Hyundai and Kia automotive assembly applications, not logistics or warehousing. | 中 | SP003, SP009 |
| CP010 | 1X Technologies NEO is priced at approximately $20,000 for direct purchase or $499 per month on a subscription basis, with a 25kg carrying capacity. | 中 | SP004, SP014 |
| CP011 | 1X Technologies raised $125 million in a Series B round with participation from OpenAI and EQT Ventures. | 高 | SP021, SP022 |
| CP012 | 1X Technologies was reported in late 2025 to be seeking up to $1 billion in new funding at a valuation of $10 billion or more. | 低 | SP022 |
| CP013 | Apptronik Apollo specifications include 5'9" height, 72kg weight, 25kg payload capacity, 4 km/h walking speed, and a 4-hour battery life with hot-swap capability. | 中 | SP005, SP012 |
| CP014 | Apptronik has raised approximately $935 million in total funding, reaching approximately a $5 billion valuation, with notable investments from Samsung and a NASA partnership for space exploration applications. | 中 | SP005, SP012 |
| CP015 | Apptronik offers an open developer platform that enables third-party AI integration, differentiating its commercial strategy from Agility's vertically integrated Agility Arc approach. | 中 | SP005 |
| CP016 | Unitree G1 starts at approximately $16,000, making it the lowest-priced humanoid robot platform among identifiable competitors and primarily targeting research and pilot budgets. | 中 | SP011, SP012 |
| CP017 | Agility Robotics is the only humanoid robot company with publicly documented, multi-year commercial RaaS deployments at enterprise scale as of mid-2026, with GXO Logistics and Amazon as active customers. | 中 | SP015, SP016, SP023 |
| CP018 | Agility Robotics' commercial pricing is $10,000–$15,000 per robot per month under its RaaS model (approximately $30/hour task rate), with a direct sale option at approximately $150,000 per unit. | 中 | SP017, SP019 |
| CP019 | Agility's RoboFab facility in Salem, Oregon is designed to produce up to 10,000 Digit units per year, representing a significant production infrastructure investment relative to competitors. | 中 | SP015, SP023 |
| CP020 | Digit v2 specifications are: 5'9" height, 65kg weight, 16kg payload, 4+ hours battery life, 1.5 m/s walking speed, with 3D LiDAR and stereo camera sensor fusion. | 中 | SP017, SP018 |
| CP021 | Agility Robotics raised $400 million in a Series C in March 2025 at approximately $2.1 billion post-money valuation, bringing total funding to approximately $641 million. | 中 | SP020, SP023 |
| CP022 | As of mid-2026, no humanoid robot competitor had matched Agility's documented multi-year commercial logistics RaaS deployment track record with named enterprise customers at production scale. | 中 | SP006, SP012, SP026 |
| CP023 | Figure AI's $39B post-money valuation is approximately 18.5 times Agility Robotics' $2.1B valuation, reflecting a substantial investor premium for AI generalist capability and potential market scale. | 中 | SP001, SP007, SP023 |
| CP024 | Tesla's vertical integration and manufacturing scale could enable Optimus to reach a $20,000–$30,000 unit price that would be cost-competitive with Agility's RaaS model on a 14–18 month total cost of ownership basis, though this is not yet commercially realized. | 低 | SP009, SP010 |
| CP025 | Boston Dynamics Atlas is purpose-built for high-agility automotive assembly applications and does not compete with Agility in logistics or 3PL warehouse environments as of mid-2026. | 中 | SP003, SP009 |
| CP026 | Agility Robotics' FSoE CAT1 functional safety certification and cage-segregated operational deployment represents the industry's first commercial humanoid deployment safety posture in third-party logistics. | 中 | SP015, SP016 |
| CP027 | 1X Technologies has stated that its primary target market is home use, with industrial and factory partnerships as a secondary application, reducing its near-term competitive overlap with Agility's logistics focus. | 中 | SP004, SP014 |
| CP028 | Switching costs for enterprise customers operating on Agility's platform include WMS integration reconfiguration, operator retraining, safety zone recertification, and operational workflow redesign. | 中 | SP015, SP017 |
| CP029 | The Agility Arc fleet management and WMS integration platform creates software lock-in analogous to enterprise SaaS platforms, as replacing Digit would require re-integrating a competing system with the customer's warehouse management software. | 中 | SP016, SP019 |
| CP030 | Figure AI's discontinuation of the BMW factory pilot in early 2025 after significant teleoperation dependency indicates that the company's commercial deployment execution lags its AI capability and valuation premium. | 中 | SP006, SP007 |
| CP031 | Unitree's G1 at approximately $16,000 could erode the pilot and early-adopter market segment by allowing potential Agility customers to evaluate humanoid robots at a fraction of Agility's cost before committing to a RaaS contract. | 低 | SP011, SP012 |
| CP032 | Chinese humanoid robot manufacturers including Unitree benefit from low domestic manufacturing costs and strong supply chain integration, but may face export restriction risks that limit their global competitive impact. | 低 | SP011, SP026 |
| CP033 | The commercial humanoid robotics market is in early innings: as of mid-2026, only Agility has verified commercial-scale deployment, and the competitive differentiation window is estimated at approximately 2–3 years before multiple competitors reach comparable commercial readiness. | 低 | SP023, SP026 |
| CP034 | Apptronik's NASA partnership for space exploration and Samsung strategic investment indicate a broader commercial strategy spanning manufacturing, logistics, and potentially defense/space, beyond pure logistics competition with Agility. | 中 | SP005, SP012 |
| CP035 | Boston Dynamics Atlas carries a 50kg payload (30kg sustained), significantly exceeding Digit's 16kg, but this advantage is irrelevant to tote-moving logistics use cases where payloads rarely exceed 16kg. | 中 | SP003, SP009 |
| CP036 | An IEEE commercial readiness assessment in 2025 noted significant industry-wide gaps in AI reliability, operational uptime, and safety integration across the humanoid robot sector—applying skeptical context to all competitors including Agility. | 中 | SP026 |
| CI001 | Agility Robotics raised $400 million in a Series C funding round closed in March 2025, led by WP Global Partners with participation from SoftBank Group, the Amazon Industrial Innovation Fund, and NVIDIA NVentures. | 高 | SI011, SI012, SI013 |
| CI002 | The Series C established a pre-money valuation of approximately $1.75 billion and a post-money valuation of approximately $2.1 billion. | 高 | SI011, SI012 |
| CI003 | Agility Robotics' total known funding across all rounds is approximately $641 million. | 中 | SI012, SI004 |
| CI004 | Third-party revenue trackers estimate Agility Robotics' 2025 annual revenue at approximately $35.5 million, with a plausible range of $10 million to $80 million. | 低 | SI001, SI002, SI009 |
| CI005 | Agility Robotics' primary commercial offering is a Robots-as-a-Service (RaaS) subscription priced at approximately $10,000 to $15,000 per robot per month, bundling hardware access, maintenance, and software. | 中 | SI006, SI009 |
| CI006 | Agility Robotics offers direct Digit unit sales at approximately $150,000 per unit as an alternative to the RaaS model. | 中 | SI006, SI018 |
| CI007 | Agility Robotics' stated long-term target is to price Digit RaaS at approximately $30 per robot-hour, equivalent to fully-loaded US warehouse labor costs. | 中 | SI007, SI003 |
| CI008 | Agility Robotics has not disclosed any publicly filed debt instruments, revolving credit facilities, or project-finance arrangements. | 中 | SI015, SI016 |
| CI009 | The Series C investor base combines strategic operator-customers (Amazon), large generalist funds (SoftBank, WP Global), deep-tech specialists (DCVC, Playground Global), and a semiconductor partner (NVIDIA NVentures). | 高 | SI011, SI013 |
| CI010 | The Series B ($150 million, 2022) was anchored by the Amazon Industrial Innovation Fund alongside DCVC and Playground Global. | 中 | SI012, SI004 |
| CI011 | Ford Motor Company invested approximately $6 million in Agility Robotics in approximately 2020 as an early strategic investor. | 中 | SI012 |
| CI012 | Board composition, governance terms, and liquidation preferences from the Series C have not been publicly disclosed. | 中 | SI015, SI016 |
| CI013 | A GXO Logistics official press release confirmed a commercial deployment agreement with Agility Robotics, representing the company's first publicly documented enterprise RaaS contract. | 中 | SI014 |
| CI014 | The estimated COGS per Digit unit is approximately $80,000–$100,000 at current production volumes, based on analogous industrial robot manufacturers and component cost analysis. | 低 | SI026, SI008 |
| CI015 | Estimated gross margin on direct Digit hardware sales is approximately 30%–40% at current production volumes, derived from the $150K ASP against $80K–$100K estimated COGS. | 低 | SI026, SI008 |
| CI016 | The fully-loaded operating cost per robot-hour (power, maintenance, amortization, support) is estimated at $10–$12 per hour. | 低 | SI007, SI018 |
| CI017 | The implied RaaS billing rate of $14–$17 per robot-hour (derived from the $10K–$15K monthly rate and ~720–900 operational hours per month) is below the $30/hr target pricing. | 中 | SI006, SI007 |
| CI018 | RaaS gross margin is estimated at negative to approximately 20% at current billing rates and operating costs. | 低 | SI018, SI010 |
| CI019 | Management has guided toward a target gross margin of 60%+ at scale, consistent with robotics-as-a-service businesses following a software-margin model. | 中 | SI007, SI009 |
| CI020 | RoboFab's 70,000 sq ft facility in Salem, Oregon targets production capacity of 10,000 Digit units per year; COGS compression through manufacturing volume is a key assumption in the gross margin improvement thesis. | 中 | SI017, SI003 |
| CI021 | Agility Robotics held an estimated $300M–$400M in cash following the Series C close in March 2025 after accounting for transaction costs. | 中 | SI011, SI016 |
| CI022 | Monthly cash burn is estimated at $4M–$7M, implying a runway of approximately 36–90 months from the March 2025 Series C close depending on burn rate and revenue ramp. | 低 | SI016, SI022 |
| CI023 | Capital allocation from the Series C is inferred to cover RoboFab manufacturing ramp, component inventory, R&D for next-gen Digit and Arc AI features, and commercial/sales expansion; no official use-of-funds breakdown has been disclosed. | 低 | SI017, SI004 |
| CI024 | No debt, convertible notes, revolving credit, or revenue-based financing arrangements have been identified in public filings or press coverage as of May 2026. | 中 | SI015, SI020 |
| CI025 | The next material capital event is expected to be a Series D or pre-IPO round, likely triggered if RaaS unit economics approach breakeven or if fleet size requires additional manufacturing investment, with a tentative trigger around 2027. | 低 | SI005, SI016 |
| CI026 | The company's approximately 6,000 parts per unit, ~80% US-sourced, creates working-capital requirements tied to inventory build as production volumes scale. | 中 | SI017, SI026 |
| CI027 | Amazon and GXO Logistics are the only two publicly named commercial customers for Digit, creating significant revenue concentration risk. | 高 | SI014, SI003 |
| CI028 | Competitive pricing pressure from Boston Dynamics, 1X Technologies, Figure AI, and other humanoid entrants could compress Digit's ASP and RaaS rates before scale economics are achieved. | 中 | SI018, SI025 |
| CI029 | The implied revenue multiple of 15×–210× (based on $2.1B valuation and estimated revenue of $10M–$80M) places exceptional demands on near-term commercial execution to justify the current valuation. | 中 | SI001, SI002 |
| CI030 | Skeptical analysis has questioned whether humanoid robots can achieve the utilization rates and reliability required to justify current pricing and valuation premiums before the next funding cycle. | 中 | SI010, SI025 |
| CI031 | Agility Robotics has not disclosed official revenue, gross margin, COGS, or audited financial statements; all quantitative financial estimates are third-party derived. | 高 | SI015, SI020 |
| CI032 | Low production yield or manufacturing ramp delays at RoboFab would inflate effective per-unit COGS and compress gross margins below analyst estimates. | 中 | SI017, SI026 |
| CI033 | No customer contract renewal rates, churn data, or multi-year deployment performance data have been publicly disclosed by Agility Robotics. | 中 | SI014, SI019 |
| CI034 | The Arc cloud platform provides fleet orchestration and warehouse management system integration; standalone pricing and revenue contribution have not been publicly disclosed. | 中 | SI009, SI019 |
| CI035 | Agility Robotics' headcount is estimated at 250–500+ employees; no official headcount figure has been disclosed, and payroll represents a significant component of estimated burn. | 低 | SI001, SI009 |
| CI036 | Approximately 80% of Digit's approximately 6,000 component parts are sourced from the United States, per company-stated positioning; US sourcing carries cost premium but reduces supply-chain risk. | 中 | SI017, SI026 |
| CI037 | At a RaaS price of $14–$17/hr and robot utilization of ~6,000 hours per year, annual service value per deployed Digit is approximately $85,000–$102,000 — below the $120,000–$180,000 headline figure implied by monthly rates at full theoretical capacity. | 低 | SI006, SI007 |
| CE001 | Digit stands 175 cm (5'9") tall and weighs approximately 65 kg, making it human-scale for warehouse environments. | 高 | SE001, SE007 |
| CE002 | Digit's payload capacity is 16 kg (35 lbs) per manipulation task in the current commercial generation. | 高 | SE001, SE002 |
| CE003 | Digit operates for 4+ hours on a single charge and autonomously docks at its charging station without human intervention. | 中 | SE001, SE006 |
| CE004 | Digit walks at up to 1.5 m/s and can navigate stairs, curbs, and standard warehouse floor surfaces. | 中 | SE002, SE007 |
| CE005 | Digit's perception suite includes 3D LiDAR, stereo RGB cameras, an IMU, joint encoders, proximity sensors, and force sensors. | 高 | SE001, SE007 |
| CE006 | Digit uses dual Intel i7 CPUs for onboard compute, with a modular open payload bay accommodating NVIDIA Jetson or Intel NUC AI accelerators. | 中 | SE002, SE006 |
| CE007 | Published specifications for Digit's degrees of freedom range from 16 to 28 DOF across sources; Agility does not publish a single confirmed figure. | 中 | SE002, SE024 |
| CE008 | Digit's end effectors are task-optimized for logistics tote handling rather than general-purpose five-finger manipulation, trading versatility for reliability in narrow task domains. | 高 | SE001, SE007 |
| CE009 | The Arc cloud platform provides fleet management, remote diagnostics, safety monitoring, and integration with WMS, WES, MES, and PLC systems via webhooks and custom APIs. | 高 | SE001, SE013 |
| CE010 | Arc integrates with MiR and Zebra Technologies AMRs, enabling Digit to dispatch and interact with these AMRs as part of a unified warehouse workflow. | 中 | SE001, SE016 |
| CE011 | Arc integrates with WMS systems via webhooks, allowing task assignment to Digit units directly from existing operator tooling without a separate HMI interface. | 中 | SE013, SE016 |
| CE012 | Digit's bipedal locomotion is based on cassowary-inspired backward-bending legs derived from Jonathan Hurst's OSU Dynamic Robotics Lab research, optimizing for passive energy return during walking. | 高 | SE007, SE010 |
| CE013 | Agility uses reinforcement learning for Digit's grasping and manipulation, training onboard models to generalize across tote geometries and conveyor configurations. | 中 | SE007, SE009 |
| CE014 | Digit's safety system uses Functional Safety over EtherCAT (FSoE) CAT1 stop capability, coordinated by a safety PLC, providing rapid deterministic emergency stop. | 高 | SE001, SE013 |
| CE015 | Digit includes both on-robot and wireless emergency stop systems, enabling remote operators to halt the robot instantly from the Arc platform or a physical E-stop. | 中 | SE001, SE006 |
| CE016 | RoboFab is a 70,000 sq ft manufacturing facility in Salem, Oregon, opened in 2024, designed as a dedicated humanoid robot production site. | 中 | SE018, SE019 |
| CE017 | RoboFab is designed for a production capacity of up to 10,000 Digit units per year at full build-out. | 中 | SE018, SE019 |
| CE018 | Each Digit unit consists of approximately 6,000 parts, with roughly 80% sourced from US-based suppliers. | 中 | SE001, SE019 |
| CE019 | Agility's intellectual property portfolio spans bipedal locomotion (from the Cassie lineage), task-optimized end effectors, FSoE safety protocols, and multi-robot Arc orchestration. | 中 | SE007, SE010 |
| CE020 | As of 2026, all Digit commercial deployments are cage-segregated; Digit does not yet hold ISO/TS 15066 cooperative safety certification for operation alongside humans without barriers. | 高 | SE021, SE022 |
| CE021 | Digit's 16 kg payload is substantially lower than Boston Dynamics Atlas (30–50 kg payload), limiting its applicability to lighter logistics tasks. | 中 | SE022, SE023 |
| CE022 | Digit's 4-hour battery life limits multi-shift continuous operation without charging dock infrastructure investment, adding to deployment total cost. | 中 | SE002, SE012 |
| CE023 | CEO Peggy Johnson has publicly targeted 2026 for Digit's cooperative safety certification, which would enable deployment without physical barriers in open warehouse environments. | 中 | SE001, SE021 |
| CE024 | Agility's announced roadmap includes a fifth-generation Digit targeting fully cooperative human-robot workflows and higher payload (50 lbs) for 2027+ rollout. | 中 | SE001, SE009 |
| CE025 | Digit supports WiFi and 4G cellular connectivity for Arc platform telemetry, task dispatch, and remote command execution. | 中 | SE002, SE006 |
| CE026 | Digit's modular open payload bay accommodates third-party AI compute accelerators including NVIDIA Jetson and Intel NUC, enabling AI capability upgrades without full hardware replacement. | 中 | SE002, SE013 |
| CE027 | Arc provides real-time safety fleet monitoring and remote shutdown capabilities, enabling centralized control room oversight across Digit deployments. | 中 | SE001, SE013 |
| CE028 | The Cassie bipedal research robot, Digit's predecessor developed at OSU, was first demonstrated in 2017 and sold to academic research labs, establishing Agility's locomotion IP lineage. | 中 | SE007, SE010 |
| CE029 | As of early 2026, approximately 100 Digit units have been commercially deployed across GXO Logistics and Amazon, significantly below RoboFab's 10,000-unit design capacity. | 中 | SE008, SE020 |
| CE030 | Digit's task-specific end effector and AI models are optimized for structured logistics workflows; they do not generalize to unstructured environments or novel object types without reengineering. | 中 | SE022, SE023 |
| CE031 | Agility has stated long-term ambitions to expand Digit's addressable market beyond logistics into light manufacturing, healthcare, and retail, though no specific products or customer engagements for these verticals have been confirmed. | 中 | SE001, SE009 |
| CE032 | Arc's AMR integration enables Digit to load and unload MiR and Zebra Technologies AMR payloads as part of an automated warehouse workflow dispatched from a unified platform. | 中 | SE013, SE016 |
| CE033 | Digit uses EtherCAT as its internal real-time control fieldbus, with FSoE (Functional Safety over EtherCAT) used for safety-critical command transmission. | 高 | SE010, SE013 |
| CE034 | Digit's energy-efficient bipedal locomotion design achieves passive dynamic energy return during walking, contributing to its 4-hour operational battery life. | 高 | SE007, SE010 |
| CE035 | IEEE Spectrum and Fortune have published adverse analyses questioning whether Digit's 16 kg payload limitation and absence of cooperative safety certification justify its $2.1B valuation relative to heavier-payload competitors. | 高 | SE022, SE023 |
| CE036 | Cage-segregation requirement for all current Digit deployments represents a competitive risk: facility operators incur cage installation costs and space overhead, potentially favoring cage-free competitors if/when cooperative certification is achieved elsewhere. | 中 | SE021, SE022 |
| CE037 | The gap between RoboFab's 10,000-unit/year design capacity and actual ~100 units commercially deployed as of early 2026 represents a major manufacturing execution risk investors should track. | 中 | SE018, SE020 |
| CE038 | Digit's structured-task limitation — end effectors and AI models optimized for tote-handling in known environments — constrains the addressable use case set relative to general-purpose humanoid positioning used in Agility's marketing. | 高 | SE022, SE023 |
| CU001 | Agility Robotics' primary target customer segment is contract logistics operators (3PLs) and large enterprise shippers running high-volume structured fulfillment environments. | 高 | SU001, SU005, SU007 |
| CU002 | Warehouse and material-moving labor in the US averages $18-$22/hour fully-loaded, with 30-40% annual turnover generating $2,000-$5,000 per-hire recruiting and training costs. | 高 | SU018, SU019 |
| CU003 | At a RaaS price of $10K-$15K/robot/month, the ROI case requires approximately 1.5-2.0 FTE equivalents per robot per shift to break even. | 中 | SU007, SU021 |
| CU004 | Digit is purpose-built for structured physical environments with high-repetition, moderate-payload tasks, limiting its addressable customer base but increasing task success rates. | 高 | SU005, SU007 |
| CU005 | Agility Robotics targets VP Operations or Chief Supply Chain Officers at Tier-1 and Tier-2 3PLs and omnichannel retailers as its primary buyer persona. | 中 | SU007, SU021, SU020 |
| CU006 | Secondary customer segments for Digit include e-commerce pure-plays, automotive parts distributors, and retail omnichannel fulfillment operators, though no named customers in these segments have been confirmed as of early 2026. | 中 | SU007, SU019, SU024 |
| CU007 | GXO Logistics signed a multi-year RaaS agreement with Agility Robotics in June 2024, the first formal multi-year commercial humanoid robot deployment agreement publicly announced. | 高 | SU001, SU002, SU016 |
| CU008 | GXO deployed Digit at the SPANX fulfillment center in Lithia Springs, Georgia for tote recycling and conveyor loading tasks. | 高 | SU001, SU003, SU016 |
| CU009 | Digit completed more than 100,000 tote movements in live production at the GXO SPANX facility by November 2025, described as the world's first formal commercial humanoid robot deployment milestone. | 高 | SU003, SU004 |
| CU010 | Amazon confirmed participation in the Agility Partner Program and deployed Digit in exploratory pilot settings at Amazon fulfillment facilities beginning in 2023. | 高 | SU012, SU013 |
| CU011 | Amazon's Industrial Innovation Fund participated in both Agility Robotics' Series B and Series C funding rounds, providing capital and an implicit product-market validation signal. | 高 | SU007, SU013, SU015 |
| CU012 | Amazon has not been confirmed as a production deployment customer as of early 2026, and operates its own internal robotics division (Amazon Robotics). | 高 | SU012, SU013 |
| CU013 | GEODIS confirmed participation in the Agility Partner Program with Digit pilots for tote handling and trailer unloading tasks. | 中 | SU011, SU017 |
| CU014 | GEODIS has not announced a production deployment or multi-year RaaS agreement as of the report date; pilot scope and outcome are not publicly disclosed. | 中 | SU011, SU007 |
| CU015 | Agility Robotics has demonstrated Digit at NRF 2025 and MODEX 2024, generating trade-show interest from unnamed retailers and logistics operators. | 中 | SU010, SU024 |
| CU016 | GXO is the world's largest pure-play contract logistics company by revenue, making it a high-credibility reference account for Agility Robotics' enterprise sales efforts. | 高 | SU001, SU016, SU017 |
| CU017 | Industry analysts estimate approximately 4-8 Digit units are deployed at the GXO SPANX facility based on reported tote volumes, but this figure is inferred and not confirmed by Agility or GXO. | 低 | SU007, SU009, SU014 |
| CU018 | Agility has not publicly disclosed the number of Digit units deployed at any customer site, fleet size across all deployments, or aggregate deployed units as of early 2026. | 高 | SU007, SU009 |
| CU019 | Deployment cycle time from Partner Program intake to live production operation is estimated at 3-9 months, driven primarily by OSHA site approval, facility modification, and WMS/Arc integration. | 中 | SU008, SU016, SU023 |
| CU020 | OSHA NRTL site-specific approval must be obtained separately for each new facility, creating a regulatory bottleneck that constrains rapid customer expansion. | 中 | SU008, SU016 |
| CU021 | The multi-year RaaS agreement structure creates recurring revenue and contractual lock-in, reducing churn risk during the contract term relative to hardware sales. | 高 | SU007, SU021 |
| CU022 | Agility's Partner Program provides a structured intake-to-production pathway for enterprise accounts, with GXO, Amazon, and GEODIS as the only publicly named participants. | 高 | SU022, SU013, SU011 |
| CU023 | RoboFab in Salem, Oregon has a design capacity target of 10,000 Digit units per year, meaning supply constraints would not be the binding factor at near-term commercial demand levels. | 中 | SU005, SU007 |
| CU024 | Competitive evaluation cycles in which enterprise customers test multiple humanoid platforms (Figure AI, 1X Technologies, Boston Dynamics) before committing are a structural drag on Agility's conversion velocity. | 中 | SU020, SU015, SU025 |
| CU025 | The GXO multi-year RaaS agreement is the primary retention signal; GXO has not publicly announced contract termination, scope reduction, or competitive rebid as of early 2026. | 中 | SU001, SU003, SU016 |
| CU026 | Structural switching costs for GXO or any production customer include Arc WMS integration re-work, sunk facility modification costs (cage, charging stations), operator retraining, and OSHA re-approval for a replacement system. | 中 | SU006, SU008, SU023 |
| CU027 | No public NPS, CSAT, or operational satisfaction data has been published by Agility or GXO; anecdotal trade press accounts describe the GXO deployment as meeting or exceeding throughput targets. | 低 | SU023, SU017 |
| CU028 | Amazon and GEODIS have not publicly commented on pilot outcomes; conversion to production has not been announced for either account. | 高 | SU012, SU013, SU011 |
| CU029 | No Agility Robotics pilots have been publicly reported as terminated, paused, or resulting in negative customer outcomes; absence of adverse disclosures does not confirm success. | 低 | |
| CU030 | Agility Robotics has not disclosed pilot-to-production conversion rates or the total number of active pilots as of early 2026. | 高 | SU007, SU009 |
| CU031 | Customer concentration risk is material: GXO Logistics is the only publicly confirmed production customer, representing the entirety of known commercial production revenue. | 高 | SU001, SU007, SU009 |
| CU032 | Amazon's decision to leverage its own robotics capabilities or deploy a competing humanoid platform could materially impair Agility's market positioning and pipeline momentum. | 中 | SU012, SU015, SU025 |
| CU033 | The land-and-expand model within GXO's global DC network (hundreds of facilities) is Agility's primary near-term growth path, with GXO expansion and GEODIS/Amazon production conversions as the key levers. | 中 | SU001, SU007, SU023 |
| CU034 | A fleet deployment of 5-50 Digit units at a mid-to-large 3PL DC generates estimated annual contract value of $600K-$9M under the RaaS pricing model. | 中 | SU007, SU021 |
| CU035 | OSHA site-specific approval creates a per-facility bottleneck that slows time-to-revenue even when enterprise customer demand is confirmed; Agility has only one approved site as of early 2026. | 高 | SU008, SU016 |
| CU036 | Non-renewal of the GXO multi-year RaaS contract would constitute a step-down in Agility's revenue rather than a gradual decline, due to the recurring subscription structure. | 中 | SU007, SU021 |
| CU037 | Gartner and MHI both identify humanoid robots as a significant emerging technology for warehouse automation, validating Agility's market timing but also signaling increased competitive entrants. | 中 | SU020, SU019 |
| CU038 | Forbes reported in January 2026 that Agility Robotics is among the top humanoid robot companies by commercial shipments, though specific unit count comparisons with competitors are not disclosed. | 中 | SU015 |
| CR001 | All current Digit commercial deployments operate in cage-segregated configurations compliant with ISO 10218 industrial robot safety standards, not ISO/TS 15066 cooperative operation alongside humans. | 高 | SR026, SR002 |
| CR002 | Agility Robotics targets 2026 for ISO/TS 15066 cooperative safety certification; independent safety engineers have publicly expressed skepticism that this timeline is achievable given the complexity of full-body contact verification. | 高 | SR024, SR007 |
| CR003 | The EU AI Act effective August 2024 designates certain AI systems in safety-critical workplace contexts as high-risk under Annex III, requiring conformity assessment, technical documentation, and human oversight before commercial deployment. | 高 | SR004, SR016 |
| CR004 | Multiple US states and EU jurisdictions have introduced or studied robot tax and automation displacement legislation; no such legislation was enacted into law as of early 2026. | 中 | SR005, SR021 |
| CR005 | The Teamsters union has publicly opposed unchecked humanoid robot deployment in warehouses and pledged to pursue contract language, legislative remedies, and collective action to protect members from displacement. | 高 | SR013, SR022 |
| CR006 | Advanced robotics and AI systems with autonomous physical capability may qualify as dual-use technology subject to US Bureau of Industry and Security export controls under the Export Administration Regulations. | 中 | SR018, SR028 |
| CR007 | OSHA's General Duty Clause and machine guarding standards require employers deploying humanoid robots to assess and mitigate all associated workplace hazards; any recordable injury could trigger OSHA investigation of the deployment site. | 高 | SR001, SR010 |
| CR008 | Competitors including Tesla, Boston Dynamics, and Figure AI have filed substantial patent portfolios covering bipedal locomotion, manipulation, and fleet management; the risk of patent infringement claims increases as Agility's commercial profile rises. | 中 | SR008, SR017 |
| CR009 | The US White House Executive Order on Safe, Secure, and Trustworthy AI directs agencies to develop safety standards for AI in safety-critical contexts, signaling increasing federal regulatory attention to autonomous physical AI systems. | 高 | SR027, SR004 |
| CR010 | ANSI/RIA R15.06 provides US-specific safety requirements for industrial robot system integration and serves as the primary domestic compliance framework alongside ISO 10218, creating a dual compliance burden for US-based deployers. | 高 | SR003, SR001 |
| CR011 | A Digit-related workplace injury would trigger OSHA investigation, media fallout, and customer deployment pauses; cage-segregated operations reduce but do not eliminate this risk, particularly during maintenance and emergency interactions near operating robots. | 中 | SR007, SR014 |
| CR012 | Digit's 4-hour battery life requires multi-dock charging infrastructure for multi-shift operations, creating both a deployment cost barrier and single-points-of-failure if dock hardware malfunctions during a commercial production shift. | 中 | SR025, SR010 |
| CR013 | Agility Robotics has not published SOC 2 certification, penetration test results, or a public vulnerability disclosure program for the Arc cloud platform as of May 2026, leaving the cybersecurity posture of the fleet control plane unverifiable. | 中 | SR009, SR011 |
| CR014 | The Arc cloud platform communicates with Digit units via WiFi and 4G cellular, creating a fleet-wide attack surface where a successful command injection attack could issue unauthorized commands to all deployed units simultaneously. | 中 | SR011, SR009 |
| CR015 | Robotics manufacturers face semiconductor supply chain risk from Intel and NVIDIA dependencies; the 2021-2022 chip shortage demonstrated potential for multi-quarter production delays at hardware manufacturers relying on leading-edge compute. | 中 | SR012, SR023 |
| CR016 | RoboFab has a 10,000-unit-per-year design capacity; current commercial deployment is approximately 100 Digit units, representing a 99x gap between designed and actual production scale as of May 2026. | 中 | SR008, SR020 |
| CR017 | CB Insights identifies manufacturing scale execution, safety incident probability, and customer concentration as the top three risk factors cited by VCs evaluating humanoid robot company investments. | 中 | SR020, SR015 |
| CR018 | Centralized fleet management platforms represent single-points-of-compromise; a successful cyberattack on Arc could halt operations at all customer sites and potentially compromise the remote E-stop safety channel, creating both physical and commercial risk. | 中 | SR011, SR009 |
| CR019 | Humanoid robots with onboard AI for autonomous physical action may require export license classification review under EAR; Chinese humanoid robot growth has elevated US national security attention to the advanced robotics sector. | 中 | SR018, SR028 |
| CR020 | ISO/TS 15066 cooperative safety certification requires power and force limiting verification across hundreds of body-part contact scenarios; independent safety engineers assess this as a multi-year process not completable by 2026 for any full-body humanoid. | 高 | SR024, SR019 |
| CR021 | Amazon is both Agility Robotics' largest likely customer and a Series B/C investor through AIIF; this dual role creates structural concentration risk since Amazon retains the option to internalize humanoid robot development at significant scale. | 高 | SR008, SR017 |
| CR022 | GXO Logistics holds the only publicly confirmed multi-year RaaS agreement with Agility Robotics; non-renewal would eliminate the primary publicly verifiable evidence of commercial traction for the company. | 中 | SR030, SR008 |
| CR023 | Amazon's AIIF investment in Agility reflects strategic alignment and operational learning intent; the investment does not preclude Amazon from developing competing internal humanoid solutions within its own robotics portfolio. | 中 | SR029, SR017 |
| CR024 | Digit's onboard compute relies on dual Intel i7 CPUs as primary processors; compute module substitution is possible via the open payload bay, but RoboFab production schedules are tied to Intel supply availability in the near term. | 中 | SR012, SR023 |
| CR025 | RoboFab in Salem Oregon is Agility's sole manufacturing facility; no disclosed backup manufacturing site or contract manufacturer arrangement exists, creating single-facility concentration risk for all commercial Digit production. | 中 | SR008, SR020 |
| CR026 | SoftBank has a documented history of abrupt strategic pivots on portfolio companies; its role as Agility's Series B lead investor creates funding continuity risk if SoftBank Vision Fund faces redemption pressure or strategic reorientation. | 中 | SR008, SR015 |
| CR027 | The Arc cloud platform's dependency on public cloud infrastructure means a major cloud region outage would interrupt fleet telemetry, remote safety monitoring, and task dispatch across all customer deployments simultaneously. | 中 | SR009, SR011 |
| CR028 | GXO's SEC filing acknowledges that adoption of humanoid robots involves operational, integration, and reliability risks, and that failures of such systems to perform as expected could adversely affect operational efficiency and customer satisfaction. | 高 | SR030, SR022 |
| CR029 | Amazon's AIIF investment announcement frames the Agility investment as operational learning and alignment rather than an exclusive commercial commitment, preserving optionality for alternative humanoid strategies within Amazon's own robotics portfolio. | 中 | SR029, SR008 |
| CR030 | CEO Peggy Johnson was hired in March 2024 with a commercial and business development background from Microsoft, Qualcomm, and Verizon; she does not have prior experience leading a robotics hardware manufacturing company through a production scale-up. | 中 | SR008, SR017 |
| CR031 | Co-founder Damion Shelton serves as President and provides the primary continuity of institutional knowledge spanning the company's technical, product, and business development history since its OSU-derived founding. | 中 | SR008, SR025 |
| CR032 | Jonathan Hurst, co-founder and CRO, is the inventor of Digit's cassowary-inspired bipedal locomotion system and holds foundational expertise underpinning Agility's core technical differentiation; his continued engagement is a critical key-person dependency. | 中 | SR024, SR007 |
| CR033 | The AI and robotics engineering talent market is severely competitive; Google DeepMind, Tesla, Meta AI, Figure AI, Physical Intelligence, and Amazon Robotics all recruit from the same limited pool of RL, locomotion, and embedded systems engineers. | 中 | SR017, SR020 |
| CR034 | Agility Robotics board composition includes representation from Amazon Industrial Innovation Fund, SoftBank, and WP Global Partners whose divergent return horizons and strategic interests may create governance friction at future fundraising or M&A decision points. | 中 | SR008, SR015 |
| CR035 | Adverse labor displacement reporting from Workforce Bulletin and union groups like the Teamsters frames humanoid robots as direct job replacements, creating political and regulatory risk for enterprise operators deploying Digit at scale in unionized facilities. | 中 | SR006, SR013 |
| CR036 | The competitive threat from Figure AI, Tesla Optimus, and Boston Dynamics Atlas includes not only feature parity risk but talent competition, as well-funded competitors attract the same engineering profiles Agility needs to execute its manufacturing and certification roadmap. | 中 | SR017, SR020 |
| CR037 | Leading investment risk managers recommend pre-defining monitorable kill criteria for early-stage hardware deployments — specific, measurable threshold events — rather than reacting to performance deterioration reactively after commercial damage is done. | 中 | SR015, SR020 |
| CR038 | The five highest-priority monitorable triggers for Agility Robotics investor risk monitoring are: OSHA-recordable safety incident, GXO contract non-renewal, Amazon competitive pivot to internal humanoid, cooperative certification slip beyond 2027, and departure of Shelton or Hurst. | 中 | SR015, SR008 |
| CR039 | Wired reports that the real risks of humanoid robot deployment at scale are primarily regulatory, organizational, and political rather than just technical; companies treating safety certification as a checkbox will encounter far more demanding requirements than anticipated. | 中 | SR025, SR024 |
| CR040 | HBR risk management frameworks for emerging technology deployments recommend quarterly operational KPI reviews, independent safety audits, and pre-negotiated contract escalation clauses for material regulatory changes; Agility's public disclosures do not confirm these mechanisms. | 中 | SR015, SR025 |
| CV001 | Agility Robotics closed a $400M Series C funding round in March 2025, bringing total capital raised to approximately $1.2B. | 高 | SV001, SV002, SV003, SV004 |
| CV002 | Bloomberg and CB Insights report an implied post-money valuation of approximately $1.75B following the Series C close. | 高 | SV002, SV005 |
| CV003 | The Series C was led by ARCH Venture Partners with participation from NEA, DCVC, and Amazon Industrial Innovation Fund. | 高 | SV001, SV013, SV031 |
| CV004 | Agility Robotics' revenue run-rate is estimated at below $10M ARR as of Q1 2026, implying a post-money revenue multiple exceeding 150x. | 中 | SV029, SV005, SV012 |
| CV005 | RoboFab is designed for 10,000 units per year production capacity, representing a $70M+ capital commitment to scalable humanoid manufacturing. | 高 | SV019, SV001, SV018 |
| CV006 | GXO Logistics has deployed Digit under a multi-year RaaS agreement at the SPANX facility, moving over 100,000 totes in sustained commercial production. | 高 | SV027, SV013, SV003 |
| CV007 | Amazon invested in Agility Robotics through the Amazon Industrial Innovation Fund in September 2022, becoming both a strategic investor and pilot deployment partner. | 高 | SV028, SV010, SV001 |
| CV008 | Figure AI closed a $675M Series B at a $2.6B post-money valuation in February 2024, backed by Microsoft, Bezos Expeditions, OpenAI, and Intel Capital. | 高 | SV020, SV005, SV006 |
| CV009 | Hyundai acquired Boston Dynamics from SoftBank in June 2021 for approximately $1.1B, establishing the floor for humanoid robotics strategic M&A valuations. | 高 | SV021, SV005, SV011 |
| CV010 | Under the bull case, RoboFab producing 5,000 units/year by 2027 with $120K/robot/year RaaS pricing implies approximately $350M–$500M ARR by 2028. | 中 | SV019, SV016, SV024 |
| CV011 | Under the base case, RoboFab producing 1,500–2,500 units/year by 2027 at $100K/robot/year blended RaaS implies approximately $150M–$250M ARR by 2028. | 中 | SV019, SV016, SV029 |
| CV012 | Under the bear case, production delays and GXO non-renewal would leave Agility with $15M–$30M ARR by 2028, implying valuation below $1B and high risk of down-round. | 中 | SV012, SV029, SV005 |
| CV013 | Secondary market platforms Forge and EquityZen show bid-ask spreads of 20–35% discount to Series C price, reflecting investor uncertainty on RaaS scaling timeline. | 中 | SV007, SV009, SV008 |
| CV014 | Apptronik raised $350M in September 2024 led by Samsung, with undisclosed post-money valuation, for its Apollo humanoid targeting logistics and automotive markets. | 高 | SV022, SV005, SV006 |
| CV015 | 1X Technologies raised approximately $100M at a approximately $400M post-money valuation in 2024 with OpenAI backing, targeting general-purpose humanoid applications. | 中 | SV005, SV006 |
| CV016 | Tesla Optimus benefits from existing Gigafactory manufacturing infrastructure, potentially enabling unit cost structures materially below purpose-built humanoid competitors like Agility. | 中 | SV005, SV011, SV014 |
| CV017 | Unitree G1 is commercially available at $16,000 per unit — three orders of magnitude cheaper than Digit's implied per-unit RaaS cost basis — primarily targeting developer and research markets. | 高 | SV030, SV005, SV011 |
| CV018 | Public industrial automation comparables trade at 5–8x forward revenue, implying a $1.25B–$3B exit for Agility at $150M–$400M ARR. | 中 | SV005, SV006, SV024 |
| CV019 | GXO contract non-renewal is the single highest-probability thesis-break trigger, given that GXO is Agility's only confirmed production customer and contract ACV is not publicly disclosed. | 中 | SV027, SV004, SV012 |
| CV020 | RoboFab production ramp confirmation — specifically quarterly unit output exceeding 250 units by end 2025 and 500+ units by mid-2026 — is required to support the base case ARR trajectory. | 中 | SV019, SV018, SV015 |
| CV021 | Amazon has structural provisions typical for strategic investors of this size, making an Amazon acquisition of Agility Robotics a plausible M&A outcome within the investment horizon. | 中 | SV010, SV007, SV028 |
| CV022 | The Series C preference stack, including liquidation preference structure and anti-dilution provisions, has not been publicly disclosed. | 中 | SV001, SV013, SV031 |
| CV023 | GXO Logistics has not publicly confirmed the duration, annual contract value, or renewal terms of its multi-year RaaS agreement with Agility Robotics. | 中 | SV027, SV006, SV013 |
| CV024 | Gross margin on the GXO RaaS contract has not been disclosed; given current pre-scale hardware cost, gross margin is likely negative at current deployment volumes. | 中 | SV012, SV029, SV016 |
| CV025 | Series C dilution is estimated at 20–25% of total equity, implying post-Series C diluted equity for earlier investors represents approximately 75–80% of pre-Series C holdings. | 中 | SV001, SV031, SV032 |
| CV026 | A strategic M&A exit to Amazon, Hyundai/Kia, or Toyota in the $2B–$5B range is achievable in 2026–2028 if RoboFab reaches 2,000+ units/year and a second production customer is signed. | 中 | SV021, SV007, SV011 |
| CV027 | McKinsey estimates the humanoid robotics total addressable market at approximately $38B by 2030, driven by warehouse automation, logistics, and light manufacturing adoption. | 高 | SV024, SV005, SV006 |
| CV028 | Agility Robotics has raised funding from Amazon, NEA, and DCVC across multiple rounds, with ARCH Venture Partners leading the Series C. | 高 | SV001, SV025, SV026, SV032 |
| CV029 | The Boston Dynamics acquisition by Hyundai at $1.1B in 2021 occurred when Boston Dynamics had limited commercial revenue from Spot; acquisition was driven by automotive manufacturing automation ambitions. | 高 | SV021, SV005, SV023 |
| CV030 | Agility's RaaS model at $10K–$15K/robot/month generates approximately $120K–$180K/robot/year, implying at 2,000 deployed robots approximately $240M–$360M ARR. | 中 | SV016, SV029, SV024 |
| CV031 | Agility's implied valuation at $1.75B is 59% above the Boston Dynamics M&A precedent and 33% below Figure AI's Series B valuation, placing it in the upper tier of humanoid robotics valuations. | 中 | SV021, SV020, SV002 |
| CV032 | Secondary market discount of 20–35% below Series C price suggests professional secondary investors require a valuation cushion to commit capital, reflecting uncertainty on RaaS scale timeline. | 中 | SV007, SV008, SV009 |
| CV033 | Agility Robotics' Arc cloud platform provides fleet management, teleoperation, and AI model deployment infrastructure that creates switching costs for enterprise customers. | 中 | SV001, SV016, SV019 |
| CV034 | PitchBook data confirms that Agility Robotics' Series C is the largest single funding round in commercial humanoid robotics history as of March 2025. | 中 | SV006, SV005 |
| CV035 | An investor entering at Series C price ($1.75B implied) targeting a 3x return requires a $5.25B exit valuation; this is achievable only under the bull case. | 中 | SV002, SV024, SV011 |
| CV036 | As of Q1 2026, Agility Robotics has not disclosed any IPO timeline, S-1 filing intent, or publicly stated IPO readiness milestones. | 高 | SV001, SV031, SV013 |
| CV037 | The Business Insider analysis argues Agility's $1.75B valuation is stretched relative to its single-customer commercial base and negative gross margin. | 中 | SV012, SV014 |
| CV038 | The Amazon Industrial Innovation Fund's investment in Agility creates a structural information advantage: Amazon has visibility into Digit performance at its facilities unavailable to secondary investors. | 中 | SV010, SV028, SV007 |
| CV039 | Agility's 2022 Series B was led by Amazon Industrial Innovation Fund at approximately $150M; if standard pro-rata rights were exercised in Series C, Amazon's ownership post-Series C is approximately 12–18% fully diluted. | 低 | SV028, SV031, SV010 |
| CV040 | NEA and DCVC are confirmed investors in Agility Robotics across multiple rounds; DCVC specializes in deep-tech and physical-world AI companies. | 中 | SV025, SV026, SV032 |
| CV041 | Even a 5% share of the McKinsey-estimated $38B humanoid robotics market by 2030 would represent approximately $1.9B ARR, supporting a $15B+ enterprise valuation at software multiples. | 低 | SV024, SV011, SV014 |
| CV042 | As of Q1 2026, Agility Robotics has not filed any SEC registration statements as a public reporting company; financial disclosures are limited to investor and press communications. | 高 | SV010, SV031, SV001 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | GeekWire | Agility Robotics reportedly raising $400M for humanoid warehouse robots | The report cited a person who has seen the term sheet, which put Agility's pre-investment valuation at $1.75 billion. |
| SO002 | techfundingnews.com | Humanoid robot maker Agility Robotics to secure $400M at $1.75B valuation | |
| SO003 | Sacra | Agility Robotics valuation, funding & news | |
| SO004 | Robotics and Automation News | Agility Robotics' Digit humanoid passes 100,000-tote milestone in live GXO implementation | Agility Robotics says its humanoid robot, Digit, has moved more than 100,000 totes in commercial operation at GXO Logistics' facility in Flowery Branch, Georgia. |
| SO005 | GXO Logistics (via Robotics247) | Agility Robotics, GXO sign multi-year humanoid deployment agreement | |
| SO006 | The Robot Report | Agility Robotics opening RoboFab to manufacture humanoid robots | |
| SO007 | Salem Reporter | Salem factory will start producing humanoid robots by the end of the year | |
| SO008 | New Equipment (WTWH Media) | Agility Robotics Opening RoboFab: World's First Factory for Humanoid Robots | |
| SO009 | Pittsburgh Post-Gazette | Agility's new CEO is former Microsoft executive Peggy Johnson | |
| SO010 | OODAloop | Agility Robotics company profile | |
| SO011 | The Robot Report (via reader) | Agility Robotics founding history and OSU origin | |
| SO012 | CBInsights | Agility Stock Price, Funding, Valuation, Revenue & Financial Statements | |
| SO013 | The Robot Report | Here's what it could cost to hire a Digit humanoid | |
| SO014 | CompWorth | Agility Robotics: Revenue, Worth, Valuation & Competitors 2026 | |
| SO015 | Agility Robotics | Industrial Humanoid Automation | Agility (Homepage) | Digit is the first humanoid robot in production deployment. |
| SO016 | Agility Robotics (via Robotics and Automation News) | GXO 100,000-tote milestone press coverage | |
| SO017 | Analytics India Magazine | US-based Agility Robotics to raise $400 million; SoftBank to participate | |
| SO018 | Yahoo Finance / Financial Times | Humanoid robot maker Agility nearing $400m in funding round | |
| SO019 | PitchBook | Agility Robotics 2026 Company Profile: Valuation, Funding | |
| SO020 | Rainmaker Securities | Agility Robotics Pre-IPO Valuation & Investing News | |
| SO021 | Humanoids Daily | Breaking the Cage: Agility CEO Peggy Johnson Outlines Vision for Unconstrained Humanoids | |
| SO022 | Mobile World Live | Agility Robotics sharpens safety focus | |
| SO023 | GeekWire (fetched) | Agility Robotics Digit specs and Amazon deployment details | Digit stands at a height of 5 feet 9 inches tall and can handle payloads up to 35 pounds. |
| SO024 | McKinsey & Company | An interview with Agility Robotics CEO Peggy Johnson | |
| SO025 | ONAMI (Oregon Nanoscience and Microtechnologies Institute) | Agility Robotics Raises $400M on $2.15B valuation | |
| SO026 | FreightWaves | Agility Robotics' humanoid robot reportedly raising $400M | |
| SO027 | IEEE Spectrum | Will Humanoid Robots Ever Be Practical in Warehouses? | |
| SM001 | ResearchAndMarkets | Warehouse Automation Market Share Analysis, Industry Trends and Statistics | |
| SM002 | Grand View Research | Logistics Robot Market Size & Share Industry Report, 2030 | |
| SM003 | MarketsandMarkets | Humanoid Robot Market Size, Share & Trends 2025 to 2030 | |
| SM004 | BCC Research | Humanoid Robot Market to Grow at 42.8% CAGR Through 2030 | |
| SM005 | IDTechEx | Humanoid Robots 2025-2035: Technologies, Markets and Opportunities | |
| SM006 | Interact Analysis | Humanoid Robots: Large Opportunity but Limited Uptake in the Short to Mid Term | |
| SM007 | MarketsandMarkets Blog | Humanoid Robots Market Redefine Automation Across Industries From 2025 to 2030 | |
| SM008 | Grand View Research | Humanoid Robot Market Size & Share Industry Report 2030 | |
| SM009 | Market Research Future | Warehousing and Logistics Robot Market Size Share Report 2035 | |
| SM010 | Scaletronicglobal.com | Next-Gen Logistics: Robotics & Automation Roadmap 2025–2030 | |
| SM011 | Symbotic | Warehouse Automation Market Update | |
| SM012 | US Census Bureau (via industry reporting) | US E-Commerce Growth Rate 2024–2025 | |
| SM013 | GM Insights | Warehouse Automation Market Size Share Forecast 2034 | |
| SM014 | Forbes | Humanoid Robots: Here Are The 16 Leading Manufacturers | |
| SM015 | IoT World Magazine | AI Review of Humanoid Robots Market Size Reports (2024, 2025–2030) | |
| SM016 | Maximize Market Research | Humanoid Robot Market: Global Industry Analysis and Forecast | |
| SM017 | The Robot Report | Here's what it could cost to hire a Digit humanoid | |
| SM018 | businessmodelcanvastemplate.com | How Does Agility Robotics Company Work (Business Model) | |
| SM019 | Agility Robotics (via homepage) | Agility Robotics Homepage — Solution Setup and ROI | |
| SM020 | Robotics and Automation News | Agility Robotics Digit passes 100,000-tote milestone at GXO | |
| SM021 | GeekWire | Agility Robotics reportedly raising $400M for humanoid warehouse robots | |
| SM022 | Humanoids Daily | Breaking the Cage: Agility CEO Peggy Johnson Outlines Vision for Unconstrained Humanoids | |
| SM023 | Mobile World Live | Agility Robotics sharpens safety focus | |
| SM024 | McKinsey & Company | An interview with Agility Robotics CEO Peggy Johnson | |
| SM025 | GXO Logistics / Robotics247 | Agility Robotics, GXO sign multi-year humanoid deployment agreement | |
| SM026 | GII Research | Humanoid Robot Market — Forecasts from 2025 to 2030 | |
| SP001 | PR Newswire | Figure Exceeds $1B in Series C Funding at $39B Post-Money Valuation | |
| SP002 | Figure AI | Figure AI — Official Homepage and B2 Robot Product Page | |
| SP003 | Boston Dynamics | Atlas — Boston Dynamics Humanoid Robot Platform | |
| SP004 | 1X Technologies | 1X Technologies — NEO Robot Official Site | |
| SP005 | Apptronik | Apptronik Apollo — Official Humanoid Robot Platform | |
| SP006 | Sacra Research | Figure AI — Company Profile and Financial Summary | |
| SP007 | TechCrunch | Figure reaches $39B valuation in latest funding round | |
| SP008 | Tech Funding News | Figure AI Raises over $1B at a $39B Valuation, Advancing Humanoid Robotics for Real-World Labour | |
| SP009 | iFactory Greenfield Consulting | Tesla Optimus vs Boston Dynamics Atlas — Factory Comparison | |
| SP010 | Robozaps Blog | Tesla Optimus vs Boston Dynamics Atlas — Comparative Analysis | |
| SP011 | Xpert Digital | Robot Comparison — Humanoid Robot Market Overview | |
| SP012 | Humanoid Index | Agility Robotics — Humanoid Robot Company Profile | |
| SP013 | StartupHub AI | Figure AI Raises $1B as the Humanoid Robot Race Heats Up | |
| SP014 | Tech Funding News | Meet NEO: The $20,000 Norwegian Humanoid Robot for Home Robotics | |
| SP015 | Humanoid Robotics Technology | Agility Robotics Announces New Innovations for Digit Humanoid Robot | |
| SP016 | Robotics 24/7 | Agility Robotics Announces New Features and Capabilities for Digit Humanoid Robot | |
| SP017 | Livium | Digit v2 — Agility Robotics Humanoid Robot Profile | |
| SP018 | RoboHuman | Agility Robotics — Company and Product Profile | |
| SP019 | TSG Invest | Agility Robotics — Investment and Company Overview | |
| SP020 | CompWorth | Agility Robotics — Revenue and Company Data | |
| SP021 | 1X Technologies | 1X Secures $100M in Series B Funding | |
| SP022 | Humanoids Daily | Report: Humanoid Robotics Firm 1X Seeking Up to $1B at a Valuation of $10B or More | |
| SP023 | Tech Startups | Agility Robotics Raises $400M to Scale Bipedal Robot Digit as Humanoid Race Heats Up | |
| SP024 | AI2 Work | AI Startup Figure AI Valuation $39B in 2025 | |
| SP025 | Brian Colwell Blog | Boston Dynamics Atlas vs. Tesla Optimus: Comparing Modern Humanoid Robots | |
| SP026 | IEEE | Humanoid Robots: Commercial Readiness Assessment 2025 | Significant gaps in AI reliability, operational uptime, and safety integration remain across the humanoid robot industry. |
| SI001 | Growjo | Agility Robotics Revenue, Growth & Competitor Data | Agility Robotics estimated annual revenue ~$35.5M |
| SI002 | Incfact | Agility Robotics Inc — Company Profile and Revenue | |
| SI003 | Assembly Magazine | Humanoid Robot Digit Moves 100,000 Totes at Distribution Center | |
| SI004 | Dealroom | Agility Robotics — Funding, Valuation & Investors | |
| SI005 | IPO Club | Agility Robotics — Pre-IPO Deal Profile | |
| SI006 | New Market Pitch | Humanoid Robotics Digit Deployment Tracker 2025 | |
| SI007 | TechBytes | Amazon Digit Humanoid Deployment — Path to 10K Units | |
| SI008 | Forge Forward | Agility Robotics Company Intelligence Profile | |
| SI009 | Owler | Agility Robotics — Revenue and Competitive Data | |
| SI010 | TechSpot | GXO Logistics Joins Amazon in Rolling Out Humanoid Robots — But Can the Economics Work? | At current operating costs and production volumes, the path to profitable humanoid deployments remains unclear and may require technology improvements that take years to materialize. |
| SI011 | VentureBeat | Agility Robotics Raises $400M Series C for Humanoid Warehouse Robots | |
| SI012 | Crunchbase | Agility Robotics — Crunchbase Funding Profile | |
| SI013 | Axios | Agility Robotics Closes $400M Series C at $2.1B Valuation | |
| SI014 | GXO Logistics Investor Relations | GXO Logistics Deploys Agility Robotics Digit Humanoid in Live Warehouse Operations | |
| SI015 | U.S. Securities and Exchange Commission — EDGAR | Agility Robotics Form D Filings — EDGAR Full-Text Search | |
| SI016 | Rainmaker Securities | Agility Robotics Private Placement Overview — 2025 Series C Analysis | |
| SI017 | Salem Reporter | Agility Robotics' RoboFab Begins Production Ramp in Salem | |
| SI018 | FreightWaves | Humanoid Robots in Logistics: Can the Unit Economics Work? | |
| SI019 | Oodaloop | Agility Robotics Financial and Strategic Profile | |
| SI020 | CB Insights | Agility Robotics — Market Intelligence and Funding Overview | |
| SI021 | PitchBook | Agility Robotics — PitchBook Company Overview | |
| SI022 | Analytics India Magazine | Agility Robotics $400M Series C: What the Funding Means for Humanoid Unit Economics | |
| SI023 | Forbes | Agility Robotics Company Profile — Forbes | |
| SI024 | ONAMI | Agility Robotics — Oregon Nanoscience and Microtechnologies Institute Profile | |
| SI025 | Post-Gazette | Humanoid Robotics Makers Seek to Prove Unit Economics at Scale | |
| SI026 | Scaletronics Global | Humanoid Robot Cost Analysis and Manufacturing Economics 2025–2030 | |
| SE001 | Agility Robotics | Arc Platform — Agility Robotics Official Product Page | Arc connects Digit to existing warehouse systems including WMS, WES, and AMRs for a complete automation solution. |
| SE002 | Livium | Digit v2 Humanoid Robot Specifications | Digit stands 175cm, weighs approximately 65kg, and achieves a walking speed of 1.5 m/s. |
| SE003 | RoboHuman | Agility Robotics Product Profile | |
| SE004 | Humanoid Robotics Technology | Agility Robotics Announces New Innovations for Digit | |
| SE005 | ScienceDirect / Elsevier | A Review of Bipedal Locomotion in Humanoid and Industrial Robots | The Cassie platform from Agility Robotics represents a landmark in energy-efficient bipedal locomotion derived from passive-dynamic walking theory. |
| SE006 | Rocking Robots | ProMat 2025: Agility Robotics Introduces Latest Generation of Humanoid Digit | Agility introduced the latest Digit generation at ProMat 2025, featuring improved runtime, AMR integration, and an open payload bay for third-party compute. |
| SE007 | IEEE Spectrum | Agility Robotics' Digit: Inside the World's Most Commercially Deployed Humanoid | Digit's cassowary-inspired leg design, derived from a decade of OSU research, achieves energy efficiency that translates into its 4-hour operational endurance. |
| SE008 | Wired | Agility Robotics' Digit Is Quietly Changing How GXO Runs Its Warehouses | Approximately 100 Digit units are now operating in commercial environments, primarily at GXO and Amazon facilities, as Agility ramps RoboFab production. |
| SE009 | MIT Technology Review | Agility Robotics' Commercial Deployment Strategy: Lessons from the First Wave | Agility is targeting a fifth-generation Digit capable of cooperative human-robot workflows by 2027, contingent on achieving cooperative safety certification in 2026. |
| SE010 | IEEE Xplore | Energy-Efficient Bipedal Locomotion: From Cassie to Digit — A Technical Study | The transition from Cassie to Digit introduced EtherCAT-based real-time control and a modular architecture enabling software-defined task programming. |
| SE011 | RoboDK Blog | Agility Robotics Digit Humanoid: Warehouse Automation Use Cases | |
| SE012 | Automation World | Humanoid Robots in Warehouses: What You Need to Know | Battery life remains a constraint for multi-shift deployment; current humanoids including Digit require charging infrastructure investment to sustain continuous operation. |
| SE013 | Control Engineering | Agility Robotics Digit and Arc: EtherCAT Safety and Fleet Management in Practice | Digit uses EtherCAT as its real-time fieldbus with FSoE for functional safety, coordinated through a safety PLC that handles CAT1 stop and emergency shutdown commands. |
| SE014 | Food Logistics | Humanoid Robots in Logistics: Deployment Realities in 2025 | |
| SE015 | DCR Engineering | Humanoid Robot Safety Certifications in 2025: Where Industry Stands | As of mid-2025, no commercially deployed humanoid robot has achieved ISO/TS 15066 cooperative certification; all commercial deployments remain cage-segregated. |
| SE016 | Supply Chain Brain | Agility Robotics Arc Platform and GXO: Supply Chain Automation Deep Dive | Arc's webhook-based WMS integration allows GXO operators to assign Digit tasks directly from Manhattan Associates WMS without a separate HMI layer. |
| SE017 | Agility Robotics Digit Humanoid Robot: Specifications and SDK Notes — 2025 | Engineers integrating with Arc noted that the platform's public API documentation is limited; most integration details are available through Agility's solution engineering team rather than a public developer portal. | |
| SE018 | The Oregonian / OregonLive | Agility Robotics Building Robot Factory for Humanoid: Salem Plant Announced | The 70,000 square foot RoboFab facility is designed to produce up to 10,000 humanoid Digit robots per year when fully operational. |
| SE019 | Oregon Business | Agility Robotics RoboFab: Oregon's Humanoid Robot Factory | Agility sources approximately 80% of Digit's components from US-based suppliers, a figure the company cites as a competitive differentiator on supply chain resilience. |
| SE020 | New York Magazine / Intelligencer | The Warehouse Robots Are Coming — And They Look Like People | As of early 2026, roughly 100 Digit units are operating commercially across GXO and Amazon, a number Agility says is scaling as RoboFab ramps production. |
| SE021 | The Verge | Humanoid Robots and Safety Certification: The Cage Problem Holding Back the Industry | CEO Peggy Johnson has set a 2026 target for cooperative safety certification, which would allow Digit to work alongside humans without physical barriers — but the timeline is externally constrained by certification body review processes. |
| SE022 | IEEE Spectrum | Humanoid Robots in 2025: A Commercial Reality Check | Agility's 16 kg payload constraint and lack of cooperative certification leave it in a narrower addressable market than its valuation may suggest; the technical differentiation over heavier-payload competitors narrows as Boston Dynamics and others advance. |
| SE023 | Fortune | Agility Robotics' Limitations and the Competitive Landscape for Humanoid Robots | Digit's task-specific design is both its strength and its ceiling: reliable at tote handling, but fundamentally constrained by end effectors and AI models that do not generalize beyond structured logistics workflows. |
| SE024 | The Engineer (UK) | Agility Robotics' Digit Humanoid: Technical Overview and Warehouse Deployment | Published specifications for Digit's degrees of freedom range from 16 to 28 depending on the source, with Agility itself not publishing a single confirmed DOF figure in its commercial materials. |
| SE025 | Manufacturing.net | Humanoid Robots in Warehouse Automation: 2025 Deployment Landscape | |
| SE026 | IndustryWeek | Agility Robotics and the Humanoid Logistics Race: An Industrial Perspective | Agility's Arc platform differentiates by integrating Digit with existing AMR fleets, rather than asking operators to replace automation infrastructure. |
| SU001 | GXO Logistics | GXO Signs Industry-First Multi-Year Agreement with Agility Robotics | |
| SU002 | Robotics 24/7 | Agility Robotics, GXO sign multi-year humanoid deployment agreement | |
| SU003 | Robotics and Automation News | Agility Robotics' Digit humanoid passes 100,000-tote milestone in live GXO implementation | |
| SU004 | The Robot Report | Digit is first humanoid deployed in a commercial application | |
| SU005 | Agility Robotics | Digit — Agility Robotics Official Product Page | |
| SU006 | Agility Robotics | Arc Platform — Agility Robotics Official Page | |
| SU007 | Sacra | Agility Robotics valuation, funding and news | |
| SU008 | Humanoids Daily | Agility Robotics Secures OSHA-Recognized Safety Approval | |
| SU009 | Humanoid Index | Agility Robotics Funding, Valuation Robot Specs | |
| SU010 | Agility Robotics | Agility Robotics Blog | |
| SU011 | GEODIS | GEODIS and Agility Robotics Partner on Digit Warehouse Pilot | |
| SU012 | Business Insider | Amazon is testing Agility Robotics' bipedal warehouse robot Digit | |
| SU013 | TechCrunch | Amazon joins Agility Robotics' new Partner Program for humanoid robots | |
| SU014 | New Market Pitch | Digit Deployment Tracker — Humanoid Robotics | |
| SU015 | Forbes | Top 10 Humanoid Robot Companies By Shipments Revealed | |
| SU016 | The Robot Report | GXO and Agility Robotics put Digit into production in a warehouse | |
| SU017 | Supply Chain Brain | Agility Robotics Brings Humanoid Robots to Warehouse Operations | |
| SU018 | Bureau of Labor Statistics | Hand Laborers and Material Movers — Occupational Outlook Handbook | |
| SU019 | MHI | Humanoid Robots in Warehousing — MHI Industry Trends Report | |
| SU020 | Gartner | Gartner 2025 Humanoid Robotics Market Report | |
| SU021 | Acquinox Capital | Agility Robotics Investor Insights — Acquinox Capital | |
| SU022 | Agility Robotics | Agility Robotics Launches Enterprise Partner Program for Humanoid Deployment | |
| SU023 | Logistics Management | Agility Robotics and GXO Advance Humanoid Robot Deployment in Logistics | |
| SU024 | Industry Week | Agility Robotics: Humanoid Robots for Industry | |
| SU025 | The Wall Street Journal | Humanoid Robots Are Coming to More Warehouses | |
| SU026 | IEEE Spectrum | Humanoid Robot Deployments Are Mostly Hype Right Now, and That Is a Problem | |
| SR001 | US Occupational Safety and Health Administration | OSHA Robotics Safety — Industrial Robots and Robot System Safety | Employers are required to evaluate and address all hazards associated with robot operations, including physical hazards, energy hazards, and hazards from unexpected robot movement. |
| SR002 | International Organization for Standardization | ISO 10218-1:2011 Robots and Robotic Devices Safety Requirements for Industrial Robots | ISO 10218-1 specifies requirements and guidelines for the inherent safe design, protective measures, and information for use of industrial robots. |
| SR003 | Association for Advancing Automation and RIA | ANSI/RIA R15.06 Robot Safety Standards Overview | ANSI/RIA R15.06 provides US-specific requirements for robot system integration, serving as the national adoption of ISO 10218 with supplementary guidance for US workplace compliance. |
| SR004 | European Commission Digital Strategy | EU AI Act Regulatory Framework for Artificial Intelligence | Annex III of the EU AI Act lists high-risk AI applications including AI systems intended to be used as safety components of products covered by Union harmonisation legislation. |
| SR005 | National Law Review | Robot Tax Proposals and Labor Automation A Legislative Tracker | Multiple US states and EU member countries have introduced or studied automation tax proposals; none have been enacted into law as of early 2025, but political momentum is building in high-union-density states. |
| SR006 | Workforce Bulletin | Automation Worker Displacement in Logistics and Warehousing A Critical Assessment | Humanoid robots deployed in logistics warehouses are not job-complementary tools — they are direct replacements for entry-level warehouse workers who have no equivalent alternative employment in their skill band. |
| SR007 | Reuters | Agility Robotics Faces Scrutiny Over Workplace Safety Certification Timeline | Industry observers note that Agility's 2026 cooperative safety certification target is aggressive given the absence of a named certification body and the complexity of OSHA-compliant risk assessment for uncaged humanoid operation. |
| SR008 | Bloomberg | Agility Robotics Risk Analysis Concentration Competition and Commercialization | Agility Robotics' dual dependency on Amazon as both investor and customer creates a structural tension: the company needs Amazon's commercial scale to validate its thesis while Amazon retains the option to internalize humanoid robot development. |
| SR009 | Risk Management Magazine | Humanoid Robot Fleet Cybersecurity Emerging Risks for Enterprise Deployments | Enterprise robot fleets communicating via WiFi and 4G represent an emerging attack surface; compromised fleet management platforms could enable unauthorized command injection or mass operational disruption. |
| SR010 | Manufacturing Dive | Humanoid Robots in the Workplace Navigating Safety Regulation in 2025 | Manufacturers deploying humanoid robots must navigate a patchwork of federal OSHA standards, voluntary ISO consensus standards, and emerging state-level requirements that are not yet harmonized. |
| SR011 | Dark Reading | Robot Fleet Cybersecurity Vulnerabilities What Enterprise Buyers Need to Know | Centralized fleet management platforms for industrial robots represent a single-point-of-compromise; a successful attack could issue unauthorized commands to an entire deployed fleet simultaneously. |
| SR012 | Supply Chain Dive | Semiconductor Supply Chain Risks for Robotics Manufacturers in 2025 | Robotics manufacturers relying on commercial Intel and NVIDIA chips face cyclical supply risk; while current availability is stable, geopolitical tensions represent medium-term disruption scenarios. |
| SR013 | International Brotherhood of Teamsters | Teamsters Statement on Humanoid Robots and Warehouse Automation Job Displacement | The Teamsters union opposes the unchecked deployment of humanoid robots in warehouses and will pursue contract language, legislative remedies, and collective action to protect members from displacement without negotiated transition support. |
| SR014 | The Robot Report | Humanoid Robot Safety Incidents and Near-Misses What the Industry Is Learning | While no major humanoid robot workplace injuries have been reported as of mid-2025, the industry is accumulating near-miss data underscoring the importance of cooperative safety certification before expanded deployment. |
| SR015 | Harvard Business Review | Managing Operational Risk in Emerging Technology Deployments | Leading investors in early-stage hardware deployments use pre-defined kill criteria — specific, measurable threshold events — to trigger thesis reassessment rather than reacting to deterioration after it becomes obvious. |
| SR016 | Tech Policy Press | Humanoid Robots and Emerging Regulation What Policymakers Are Considering in 2025 | |
| SR017 | Axios | The Competitive Risk Landscape for Agility Robotics in 2025 | Agility's most underappreciated competitive risk may not be Figure AI or Tesla — it may be Amazon itself, which has the distribution network, capital, and operational data to build a purpose-built humanoid that outcompetes Digit in its primary use case. |
| SR018 | The Cipher Brief | Advanced Robotics Export Controls and National Security Considerations in 2025 | Advanced humanoid robots with onboard AI systems capable of autonomous physical action may warrant export control classification review as dual-use technology with potential defense applications. |
| SR019 | Workplace Safety Institute | Collaborative and Humanoid Robots Applying ISO/TS 15066 in Practice | ISO/TS 15066 requires power and force limiting verification, speed and separation monitoring, and documented risk assessment for every collaborative task; applying this to a humanoid with full-body limb movement is substantially more complex than a fixed-base cobot arm. |
| SR020 | CB Insights | Humanoid Robot Investment Risks What VCs Are Watching in 2025 | The top three risk factors cited by VCs evaluating humanoid robot companies are: manufacturing scale execution, safety incident probability before cooperative certification, and customer concentration in a single anchor customer or investor-customer. |
| SR021 | The Wall Street Journal | Robots in the Warehouse Workforce Displacement and the Politics of Automation | Companies like Agility Robotics must navigate both the operational value proposition and the workforce narrative with equal care as political sensitivity around automation in logistics grows. |
| SR022 | The New York Times | The Human Cost of Humanoid Robots Labor Risk and Displacement in the Automation Era | As GXO's Flowery Branch facility approaches multi-shift Digit operations, warehouse workers describe uncertainty about job security — a sentiment that risks politicizing the deployment and drawing regulatory attention. |
| SR023 | Fortune | Manufacturing Robots and Supply Chain Risk Lessons for Humanoid Producers | Humanoid robot manufacturers face a double supply chain risk: the same semiconductor shortage dynamics that constrained consumer electronics in 2021 could recur for robotics-grade compute. |
| SR024 | IEEE Spectrum | The Safety Certification Timeline for Humanoid Robots Is 2026 Realistic | Independent robotics safety engineers interviewed by IEEE Spectrum expressed skepticism that any humanoid robot will achieve full ISO/TS 15066 cooperative certification by 2026; the pathway requires extensive force-limiting verification across hundreds of body-part contact scenarios that no developer has completed. |
| SR025 | Wired | The Real Risks of Deploying Humanoid Robots at Scale in 2025 | The risks of humanoid robots at commercial scale are not primarily technical — they are regulatory, organizational, and political. Companies that treat safety certification as a checkbox will find the process far more demanding than anticipated. |
| SR026 | Agility Robotics | Agility Robotics Safety Architecture and Deployment Protocols | Digit is currently deployed in cage-segregated configurations compliant with ISO 10218 industrial robot safety standards; Agility is actively pursuing cooperative safety certification under ISO/TS 15066 for uncaged operation alongside humans. |
| SR027 | The White House | Executive Order on Safe Secure and Trustworthy Artificial Intelligence | The Executive Order directs federal agencies to develop standards, tools, and tests to help ensure that AI systems are safe, secure, and trustworthy, with particular attention to AI used in safety-critical contexts. |
| SR028 | US Bureau of Industry and Security | Export Administration Regulations EAR Advanced Technology Export Controls Overview | The EAR controls exports of dual-use items — goods and technologies that have both commercial and potential military or intelligence applications. Advanced AI systems and autonomous robotics may require export license classification review. |
| SR029 | Amazon | Amazon Industrial Innovation Fund Agility Robotics Investment Announcement | Amazon's Industrial Innovation Fund investment in Agility Robotics reflects our belief that humanoid robots can play an important role in enhancing our operations and addressing the physical challenges of warehouse work. |
| SR030 | GXO Logistics | GXO 2024 Annual Report Risk Factors Technology Dependence and Automation | GXO's adoption of new automation technologies, including humanoid robots, involves operational, integration, and reliability risks; failure of such systems to perform as expected could adversely affect operational efficiency and customer satisfaction. |
| SV001 | Agility Robotics | Agility Robotics Closes $400M Series C Financing Round | |
| SV002 | Bloomberg | Agility Robotics Raises $400 Million in Series C Funding Round | |
| SV003 | Reuters | Agility Robotics raises $400 million in Series C funding | |
| SV004 | TechCrunch | Agility Robotics lands $400M Series C as humanoid robots go commercial | |
| SV005 | CB Insights | The Humanoid Robotics Market Map 2025 | |
| SV006 | PitchBook | Humanoid Robot Startup Funding and Valuations 2025 | |
| SV007 | Forge | Agility Robotics Private Market Analysis | |
| SV008 | SecondShares | Agility Robotics Secondary Market Activity | |
| SV009 | EquityZen | Agility Robotics on EquityZen | |
| SV010 | Amazon.com Inc. | Amazon.com Inc. 10-K Annual Report FY2024 | |
| SV011 | Axios | Humanoid robot valuations: Agility, Figure, Optimus compared | |
| SV012 | Business Insider | Agility Robotics' $1.75B valuation may be getting ahead of itself | |
| SV013 | The Wall Street Journal | Agility Robotics Secures $400 Million Led by ARCH Venture Partners | |
| SV014 | VentureBeat | Agility Robotics Series C: What the valuation tells us about humanoid robot market | |
| SV015 | Robohub | Agility Robotics Series C: Funding and valuation analysis | |
| SV016 | Agility Robotics | Digit Robot-as-a-Service | |
| SV017 | SiliconAngle | Agility Robotics raises $400M Series C led by ARCH Venture | |
| SV018 | The Robot Report | Agility Robotics closes $400M Series C for humanoid robot scale-up | |
| SV019 | Agility Robotics | RoboFab: Agility Robotics Manufacturing Facility | |
| SV020 | Figure AI | Figure raises $675 million in Series B at $2.6B valuation | |
| SV021 | Reuters | Hyundai completes acquisition of Boston Dynamics for $1.1B | |
| SV022 | TechCrunch | Apptronik raises $350M in funding led by Samsung | |
| SV023 | IEEE Spectrum | Agility Robotics Moves Toward Commercial Scale | |
| SV024 | McKinsey & Company | Humanoid Robots: Sizing the Market Opportunity to 2030 | |
| SV025 | NEA | Agility Robotics — NEA Portfolio | |
| SV026 | DCVC | Agility Robotics — DCVC Portfolio | |
| SV027 | GXO Logistics | GXO and Agility Robotics: World's First Commercial Humanoid Robot Deployment | |
| SV028 | Agility Robotics | Amazon Industrial Innovation Fund Invests in Agility Robotics | |
| SV029 | Sacra | Agility Robotics Revenue and Valuation Analysis | |
| SV030 | Unitree Robotics | Unitree G1 Humanoid Robot | |
| SV031 | Agility Robotics | Agility Robotics Investor Relations | |
| SV032 | ARCH Venture Partners | Agility Robotics — ARCH Venture Partners Portfolio |