初创公司尽调
尽调报告 Robotics / Industrial AI / Automation Private (late stage) 2026-06-20

Agile Robots AG

工业规模和战略伙伴都扎实,但当前定价和经营经济性仍过于不透明,不足以支撑高确信度入场。

Agile Robots 是一家可信且已具规模的工业机器人平台,制造深度真实、战略伙伴强,但当前入场判断仍被私营公司在定价、利润率、烧钱速度和并购后执行质量上的不透明卡住。

封面要素

成立时间 01
2018 [CO001]
总部 02
Munich Germany [CO004]
Series C(2021) 03
220 USD M [CO011]
最近公开估值基准 04
1000 USD M+ [CO012]
2024 年收入信号 05
200 EUR M [CO021]
已部署系统 06
20000 systems+ [CO020]
最新员工规模信号 07
2500 employees+ [CO018]

公司概况

Agile Robots AG 由 Zhaopeng Chen、Peter Meusel 等 DLR 机器人研究人员于 2018 年创立,已成长为一家总部位于慕尼黑的工业机器人平台,覆盖机械臂、灵巧操作、自动化工程、厂内物流软件和新的人形机器人项目。公开资料显示,公司业务靠 AI 驱动硬件和系统集成支撑,而不是纯软件经济性,运营横跨德国、中国、印度和美国。战略动能真实存在:SoftBank 投资了 2021 年 Series C,BMW 和 idealworks 拓宽了工业入口,Franka 与 thyssenkrupp Automation Engineering 扩大了资产底盘,Google DeepMind 以及与 NVIDIA 相关的 AI 基础设施强化了具身 AI 叙事。主要尽调约束在披露质量:收入、员工数、部署量和估值信号仍停留在新闻稿级别,而非申报级别;当前利润率、烧钱速度、现金跑道和控制权安排均未公开。

官网
www.agile-robots.com
成立时间
2018-01-01
创始人
Dr. Zhaopeng Chen, Peter Meusel
创立地点
Munich, Germany
总部
Munich, Germany
产品
Agile Robots 销售 AI 驱动的工业机器人系统,包括 Diana 7、Yu 5 Industrial、Agile Hand、AgileCore、AMR/AGV 与自动化产品、收购而来的工程能力,以及 Agile ONE 人形机器人路线图。
客户
面向汽车、消费电子、物流、医疗健康和工业制造客户,满足富接触自动化与系统集成需求。
商业模式
商业化由硬件、自动化系统集成、软件 / 控制栈、服务和制造牵引,而非纯经常性软件。
阶段
Private (late stage)
融资情况
公开融资历史从早期天使 / 种子轮和 Pre-A 轮延伸至 2020 年八位数 Series A,以及 2021 年 US$220 million Series C,后者让 Agile 跻身独角兽。公开资料还提到 2022 年后续融资,但当前估值和融资条款未披露。
[CO001, CO002, CO004, CO005, CO008, CO011, CO012, CO020]

执行摘要

主要优势

  • 真实工业规模清晰可见:部署系统 >20,000 套,披露 2024 年收入约 EUR200 million,并已形成多国运营足迹。
  • 公司把 DLR 衍生的机器人技术厚度,与制造、系统集成和合作伙伴杠杆结合起来,而不是只靠演示支撑叙事。
  • SoftBank、BMW/idealworks、Google DeepMind、Deutsche Telekom 和 NVIDIA 等战略关系,实质增强了市场进入能力和技术可信度。
  • 收购 Franka、BÄR、idealworks 和 thyssenkrupp Automation Engineering 后,Agile 的产品和客户基础比单一机械臂机器人初创公司更宽。

主要风险

  • 当前估值、利润率结构、烧钱速度、现金跑道和清算优先权均未披露,无法形成有把握的承销判断。
  • 多项收购叠加新的人形机器人项目,抬高了整合复杂度、资本强度和管理层带宽风险。
  • 公开收入、员工数和部署指标仍是新闻稿级别口径,不是审计或监管文件级披露。
  • 德国-中国跨境运营与先进 AI/机器人研发,带来地缘政治、出口管制和合规暴露。
  • 如果当前势头不能转化为耐久经济性,人形机器人和物理 AI 的上行空间可能已经先于可验证的工业 ROI 被计入价格。

未决问题

  • 2021 年独角兽基准之后的当前定价估值或二级市场证据
  • 经审计的收入质量、毛利率、运营支出、资本开支、烧钱速度和现金跑道披露
  • 主要融资轮次的股权结构表、董事会控制权和优先股堆叠细节
  • 客户留存、集中度、续约质量,以及软件 / 服务变现证据
  • Franka、idealworks、BÄR 和 thyssenkrupp Automation Engineering 资产的并购后整合 KPI

目录

Chapter 01

01公司概览

1.1 身份、版图与运营画像

Agile Robots 对外呈现的是一家总部位于慕尼黑、提供 AI 驱动自动化的公司,而不是单一产品机器人厂商。官方公司页面反复描述一套结合机器人硬件、传感器、计算机视觉、力控和软件的技术栈,并把公司放在汽车、消费电子、医疗健康、物流和服务流程之中。历史主线来自创始人的 DLR 背景:Zhaopeng Chen 和 Peter Meusel 2018 年从力控机器人和智能操作研究出发创办公司,随后扩张为多地制造和研发网络。公开运营版图证据在地理分布上较强,在经审计规模上较弱。慕尼黑是治理和研发中心,Kaufbeuren 是公开的生产锚点,官方页面还列出中国、印度和美国的运营。这个版图很关键,因为后续关于具身 AI、人形机器人和工厂级部署的战略主张,都需要公司在多地拥有内部制造、集成能力和工业客户入口。[CO001, CO002, CO003, CO004, CO005, CO008]

快照 KPI 表
指标数值 / 状态日期置信度缺口
成立20182018None
创始人Zhaopeng Chen;Peter Meusel2026 年视角None
当前法律形式Agile Robots SE(由 AG 转换)2024-03-26None
总部德国慕尼黑2026 年视角None
生产锚点巴伐利亚 Kaufbeuren2026 年视角None
公开员工数>1,900 至 >2,5002025-2026未披露经审计基线
公开部署数量>20,000 套机器人解决方案2026公司说法,非申报级披露
公开收入信号2024 年约 EUR200m2025-2026 年新闻稿引用私营公司说法,无财务报表

汇总官方与第三方公开快照;员工数和收入是方向性公开说法,不是经审计披露。

[CO001, CO004, CO006, CO016, CO017, CO018]
FO002: 公司快照逻辑

Agile Robots 把研究驱动的硬件、软件、资本和收购来的工业渠道接成一套运营栈。

[CO002, CO003, CO013, CO025, CO027, CO031]
FO003: 快照 KPI

公开规模标记显示,这是一家增长很快、但仍选择性披露的私营机器人公司。

[CO006, CO011, CO012, CO016, CO018, CO019]

1.2 领导层与治理信号

领导层证据最集中在创始人和 2024 年法律形式转换,公开董事会披露则薄得多。Chen 和 Meusel 不是泛泛的商业创始人;公司和第三方资料反复把他们与 DLR 的机器人工作相连,这也解释了 Agile Robots 的公开叙事为什么强调灵巧、力控和先进感知,而不是单纯替代人力。治理记录显示,2024 年 3 月 Agile Robots 从 AG 转为 SE,这是一个有意义的正式动作,更符合一家预期开展泛欧洲增长、跨境运营并扩大机构形象的公司。质量和运营体系信号也有支撑:ISO 9001 认证和反复提到的内部制造,都强化了 Agile Robots 希望被视为生产级公司的定位。未解的问题是所有权和控制权细节。可用资料中没有申报级公开股权表,没有清晰披露的董事会图谱,也没有详细债务或控制权披露,因此领导层质量比治理透明度更容易评估。[CO001, CO006, CO007, CO017, CO037]

领导层与创始人表
人物角色背景创始人-市场匹配 / 覆盖关键人物依赖
Zhaopeng Chen创始人兼 CEO前 DLR 机器人研究员、副实验室负责人技术创始人,锚定力控机器人和物理 AI 叙事很高
Peter Meusel联合创始人前 DLR 高级科学家;机器人与扭矩传感器专家硬件与操作技术可信度深
Rory Sexton执行董事 / 管理负责人官方解决方案与总部材料中的对外运营负责人衔接产品到工业交付的叙事
Yuekai Zhao董事总经理(Kaufbeuren 列名)Kaufbeuren 公开资料领导联系人中被点名显示生产基地有运营纵深
更广泛的 DLR 衍生联合创始人群体创始技术团队关于页面称公司由更多 DLR 专家共同创立强化深厚团队印象,但当前公开页面未完整披露姓名

行项目聚焦公开记录中的领导人物;董事会构成及与所有权挂钩的治理权利未公开详述。

[CO001, CO004, CO006, CO015, CO037]

1.3 融资历史与利益相关方图谱

公开融资证据呈现出熟悉的深科技路径:早期由中国相关创投投资,随后 SoftBank Vision Fund 2 领投的高能见度 Series C 奠定了 Agile Robots 的独角兽定位。公司自己的融资公告可以让我们还原从天使和种子支持,到 Pre-A,再到八位数 Series A,最后到 US$220 million Series C 的序列。这足以证明资本可得性、投资人质量,以及投资方愿意为资本密集型机器人扩张继续输血。但这些信息不足以还原控制权或当前持股比例。利益相关方图景不止股权资本。BMW 通过 idealworks 仍具战略相关性,Google DeepMind 则在 2026 年成为智能层伙伴。这些关系重要,因为 Agile Robots 的执行模式既依赖生态杠杆,也依赖融资本身。缺点是披露不透明:估值、员工数和收入信号出现在新闻稿和访谈里,但公司仍像一家私营企业,只在战略上有用时选择性披露动能指标。[CO009, CO010, CO011, CO012, CO013, CO014]

利益相关方或投资者图谱
利益相关方角色控制 / 经济重要性尽调要求
SoftBank Vision Fund 2后期投资者独角兽时代融资故事中最显眼的品牌背书方确认当前持股比例及任何董事会权利
Hillhouse Capital早期连续投资者出现在早期已披露轮次和 Series C 叙事中确认 Hillhouse 是否保留有意义的按比例跟投仓位
Sequoia Capital China早期投资者出现在天使 / pre-A 披露中;传递早期信心信号厘清当前持股和治理权利
BMW Group / idealworks 合作战略伙伴及前共同所有者背景通向内部物流和汽车自动化的重要渠道测试 2025 年全资收购后的商业依赖
Google DeepMind技术伙伴增加前沿 AI 可信度,并可能影响产品路线图厘清排他性、数据权利和商业化边界
thyssenkrupp 客户网络并入的工业关系基础增加 OEM 触达和工程足迹,而不是股权资本衡量转移后的客户留存

这张表混合了投资者与战略利益相关方,因为公开控制证据不完整,但战略影响仍与尽调相关。

[CO011, CO013, CO014, CO025, CO026, CO027]

1.4 里程碑、收购与整合风险

2023-2026 年的里程碑模式是当前公司故事的核心。Agile Robots 并不只是放大原有协作机器人产品线;它通过收购和相邻合作拼出一个更宽的工业平台。Franka 带来陷入困境但仍有价值的机器人 IP 和研究平台,BÄR 带来运输与自动化经验,idealworks 带来厂内物流软件和 AMR 暴露,thyssenkrupp Automation Engineering 则显著加深了工业纵深和 OEM 关系。到 2026 年初,公司在这一被收购的工业底座之上叠加 Google DeepMind 合作,明确推进自主、重推理的具身 AI。这一序列支持一个判断:Agile Robots 正在成为整合者和系统集成商,而不只是设备制造商。但执行风险也真实存在。Franka 来自破产程序,收购后的运营 KPI 仍稀薄,围绕员工、部署量和收入的公开规模主张仍是新闻稿级别而非经审计。整体里程碑记录因此令人印象深刻,但仍取决于公司能否把差异很大的资产、客户群和技术栈整合成功。[CO015, CO016, CO018, CO019, CO020, CO021]

里程碑表
日期事件类型金额 / 估值 / 状态参与方含义
2018公司由 DLR 研究人员创立创立N/AZhaopeng Chen、Peter Meusel、联合创始团队奠定深科技研究身份
2019-07-31完成 pre-A 融资融资金额未披露Hillhouse、Sequoia、Tinavi、Linear Venture 等投资方验证早期投资者支持
2020-02宣布 8 位数 Series A融资8 位数区间C-Ventures 领投轮次传递早期外部扩张资本信号
2021-09完成 Series C融资US$220m;估值 >US$1bnSoftBank Vision Fund 2 及现有投资者锁定独角兽地位叙事
2021-09宣布 ISO 9001 认证监管已认证Agile Robots、All-Cert增强制造可信度
2023-09多数股权收购 BÄR Automation合作已完成Agile Robots、BÄR Automation 交易方增加运输与专用自动化能力
2023-09多数股权投资 idealworks合作已完成Agile Robots、BMW Group、idealworks 三方增加 AMR 和内部物流足迹
2023-11从破产程序中收购 Franka反向债权人批准后完成Agile Robots、Franka Emika 债权人增加研究平台,但提高集成风险
2024-03-26从 AG 转换为 SE治理已完成Agile Robots为跨境规模化更新法律结构
2025-06全球总部在慕尼黑启用规模已完成Agile Robots、巴伐利亚官员、DLR 联系人集中研发与公众身份
2025-09全资收购 idealworks合作100% 所有权Agile Robots、BMW Group将战略投资转为完全控制
2025-11签署 thyssenkrupp Automation Engineering 资产协议合作已宣布协议Agile Robots、thyssenkrupp大幅扩展进入 OEM 项目
2026-03-24宣布 Google DeepMind 研究合作合作战略研究协议Agile Robots、Google DeepMind将装机基础连接到基础模型层
2026-04-01thyssenkrupp Automation Engineering 交易完成规模已交割Agile Robots、thyssenkrupp提高工业规模和集成压力

2021 年前日期基于公司公告,可能遗漏中间轮次;金额仅在公开来源提供时列出。

[CO001, CO006, CO007, CO010, CO011, CO013]
FO001: 公司里程碑时间线

Agile Robots 在 2018 到 2026 年间,从 DLR spinout 走向通过收购扩张的实体 AI 平台建设者。

Series A 采用公告月份作为里程碑日期,因为保留来源集中没有公开确切交割日期。

[CO001, CO010, CO011, CO022, CO027, CO031]

1.5 图表

Chapter 02

02市场分析

2.1 市场边界与纳入细分

本章最大的分析错误,是引用一个巨大的「机器人」TAM 后就停下。Agile Robots 的公开版图既不是广义消费机器人,也不只是某一类工业机械臂。保留下来的证据支持一个更窄、更可投资的边界:面向工厂及相邻物流的集成工业自动化,包括固定机械臂、协作机器人、工业软件、移动自动化,以及能解决真实生产任务的新兴人形机器人流程。这个边界与公司自己的应用页面一致,后者强调装配、机床上下料、搬运、检测和厂内物流,而不是消费或家庭场景。它也解释了相邻软件和移动能力为何重要。AgileCore 面向系统集成商和运营方,BÄR 与 idealworks 相关的移动机器人层则把产品组合延伸到 AMR、AGV 和移动操作项目。结果是一个基于证据、运营上自洽的市场定义,比无差别的机器人支出标题有用得多。[CM001, CM002, CM003, CM012, CM013, CM018]

市场定义表
细分 / 类别纳入支出排除支出买方 / 付款方相关性
固定式工业机械臂用于装配、搬运、检测的硬件、控制器、部署软件消费机器人、纯研究原型工厂工程 / 运营当前核心市场
协作机器人轻负载协作自动化单元和视觉驱动的 cobot 工作流没有柔性溢价的纯重工业焊接单元运营 / 自动化负责人与 Yu 5 及类似产品线高度相关
工业软件与编排机器人操作系统、车队控制、AI 助手、集成软件通用 ERP 或非机器人企业软件系统集成商 / 操作方 / IT-OT 负责人通过 AgileCore 体现重要性
移动机器人 / 内部物流AMR、AGV、移动操作臂、车队编排消费级配送机器人和最后一公里消费设备物流与仓储运营商通过 idealworks 和 BÄR 关联体现重要性
工业人形机器人面向生产车间和物流、数据密集型工业任务的人形机器人消费级人形娱乐或居家护理先进制造 / R&D / 创新预算新兴期权价值,而非当前核心销量
排除的广义机器人 TAM消费机器人、家用设备、通用服务机器人N/AN/A会高估 Agile 的相关市场

边界由保留的官方产品页和市场报告推断;广义机器人类别被有意排除,以保持本章可投资。

[CM001, CM002, CM003, CM014, CM017, CM018]
FM003: 买家 / 细分市场地图

Agile 的公开适配度最强处,落在柔性自动化、软件编排与工业数据深度的交汇点。

[CM012, CM013, CM014, CM015, CM016, CM017]

2.2 规模测算视角与相互矛盾的估计

公开工业机器人市场规模在方向上偏多,但数字并不一致。这种不一致本身就是发现,而不是噪音。Mordor、MarketsandMarkets、Verified Market Research 和 Future Market Insights 都显示多年增长可观,但它们对 2026 年绝对基数和隐含斜率分歧很大。保留的最低 2026 年估计约为 USD15.5 billion;最高为 USD65.1 billion。这个跨度太大,不能把任何一家发布方当成标准答案。方法论似乎驱动了差异:有些发布方更接近核心机器人硬件,另一些则纳入软件、集成、协作系统或更广的自动化层。尽调中更稳的结论是需求结构性真实存在,但 Agile Robots 的自下而上市场模型必须受产品匹配度和买方证据约束,而不能只靠自上而下 TAM 幻灯片。VDMA 背景有助于理解周期性:德国仍是自动化领导者,但本地行业机构仍把 2026 年环境描述为困难,提醒我们采用节奏会随制造业周期波动,而不是一条平滑 CAGR 曲线。[CM004, CM005, CM006, CM007, CM008, CM009]

TAM / SAM / SOM 或规模测算视角表
发布方年份 / 时间范围地理范围数值CAGR / 增长方法 / 局限置信度
Mordor Intelligence2026-2031全球USD54.28bn–USD94.38bn 区间11.7% CAGR市场定义更宽;包含主要工业供应商
MarketsandMarkets2026-2032全球USD15.5bn–USD20.8bn 区间5.0% CAGR范围更偏硬件;细分详细但基数更窄
Verified Market Research 机构2024-2031全球USD19.17bn–USD39.56bn 区间10.46% CAGR从 2024 年基数出发,采用不同细分框架
Future Market Insights 机构2026-2036全球USD65.1bn–USD343.8bn 区间18.1% CAGR可能更广泛纳入服务 / 集成;保留下来的最激进视角
VDMA2025-2026德国EUR13.8bn–EUR14.1bn,且预期 2026 年下滑2026 年收入变化 -5%行业协会收入口径,与全球市场报告不是同一指标
Agile 约束后的 SAM2026 年视角与 Agile 相关的全球工业自动化细分无法从公开来源拆出N/A需要自下而上的公司定价和结构数据;自上而下报告过宽

由于范围、层级纳入和时间跨度不同,各项数值不能直接比较;本表旨在保留矛盾,而不是强行制造虚假精度。

[CM004, CM005, CM006, CM007, CM008, CM010]
FM001: 市场规模视角

一个可用的 Agile 市场视角,需要从宽泛工业机器人收窄到软件驱动、灵活且与移动场景相关的工业自动化细分。

该金字塔是概念图,不是可相加模型;现有公开来源不足以支撑 Agile Robots 的单一清晰 SAM 数字。

[CM001, CM004, CM005, CM006, CM007, CM008]
FM002: 市场估算区间

保留的分析师页面显示,2026 年工业机器人市场区间很宽,但方向明确为正。

图中混用了不同年份、口径和方法;保留这些数值是为了呈现公开估算分歧,不代表它们可严格对比。

[CM004, CM005, CM006, CM007, CM008, CM010]

2.3 买方图谱与采用路径

保留资料指向一个多角色采购过程。汽车、电子、物流和相邻工业运营商是主要终端市场,但买方不总是终端用户。工厂工程和运营团队掌握流程痛点,采购和供应链团队感受可得性和交期压力,物流负责人关心厂内物流效率;涉及 AgileCore 或异构车队时,系统集成商也很关键。产品级证据进一步细化了买方图谱。Diana 7 和 Yu 5 服务对灵巧性和协作敏感的任务。移动机器人服务运输和厂内物流。Agile ONE 把故事延伸到跨工位劳动力替代或增强,但仍在工厂和物流内部,而不是消费场景。DeepMind 合作报道通过点名电子、汽车、数据中心和物流作为初始目标,强化了工业优先路径。ARENA2036 和工业 AI 云又提供了额外证据:采用可以通过研究、试点和数据合作来中介,而不只是直接一次性设备销售。[CM012, CM013, CM014, CM015, CM016, CM017]

细分 / 买方图谱
细分买方用户付款方工作流预算负责人采用触发因素
汽车装配工厂工程 / 制造运营产线工人与自动化团队工厂 capex / 自动化预算装配、搬运、检测运营 / 工程产能、质量、劳动力替代
电子制造工艺与自动化工程技术员与单元操作员自动化资本开支灵巧装配、搬运、洁净精密作业工厂运营精密且灵活的小负载自动化
厂内物流 / 仓库物流和仓库运营叉车 / 仓库团队物流自动化预算AMR/AGV 移动、车队协同供应链 / 仓库负责人交付周期、用工可得性、通行效率
系统集成项目系统集成商 / OEM 工程团队集成商部署团队项目业主和最终客户异构机器人编排集成商 P&L / 项目预算需要软件协同和更快部署
工业人形机器人试点创新 / 先进制造团队与机器人共用工位的工厂团队创新或战略自动化预算跨工位物料搬运和灵活任务R&D / 创新 / 制造战略固定单元表现不足时,需要可适应的自动化
采购和供应链间接采购团队内部买家和计划人员运营预算组件寻源和供应可得性采购负责人成本、供应可得性和交付周期改善

买方、用户和付款方角色来自产品页面、合作伙伴 / 客户案例和工业工作流描述的推断,而非明确的定价披露。

[CM012, CM013, CM014, CM015, CM017, CM018]
FM004: 采用漏斗或价值链地图

采用路径从流程痛点出发,进入集成,再走向数据积累和更深的自动化锁定。

[CM017, CM018, CM026, CM028, CM032, CM033]

2.4 增长驱动、约束与估值相关性

需求逻辑很容易看到:市场报告持续把智能制造、AI、IIoT、更高劳动力成本和工厂效率压力列为机器人行业的持久顺风。ResearchAndMarkets 更进一步,把市场描述为从自动化转向自主化,这与 Agile Robots 的具身 AI 叙事高度吻合。但约束对估值同样重要。MarketsandMarkets 强调,协作机器人高成本、集成复杂和互操作性有限仍是持续障碍。这些障碍对 Agile Robots 尤其相关,因为公司的系统价值越来越取决于整合硬件、软件、数据和被收购业务线,而不是销售一台孤立的机械臂。因此,切换成本很可能只有在客户把完整技术栈运营起来之后才会出现。这可以成为护城河,但也意味着实施风险高,商业化扩张需要耐心的证据点。公开定价仍缺席,所以回收期和单位经济性仍靠推断。投资人应把这个品类按战略重要且可能很大的市场来承销,但还不能把它视作一条测量清晰、无摩擦的采用曲线。[CM019, CM020, CM021, CM022, CM023, CM024]

增长驱动因素与约束表
驱动因素 / 约束方向时点含义尽调问题
智能制造、AI 和 IIoT 采用顺风当前且结构性支撑市场对自适应机器人系统的更广泛兴趣询问哪些部署需要软件带动,哪些只是独立硬件
劳动力成本压力和制造回流经济性顺风当前在生产成本更高的地区改善机器人 ROI按地区和垂直行业测试 Agile 经济性最强的位置
协作机器人作为高增长子赛道顺风当前和中期利好灵活、轻负载自动化用例衡量 Yu 5 和 Diana 从试点转向重复项目的转化率
人形系统从试点走向工业使用顺风但早期2026 年起如果 Agile ONE 证明可靠性,将创造期权价值要求提供试点管线、故障率和节拍基准
协作机器人和系统集成前期成本高逆风当前可能拖慢采用,并压缩 SME 需求要求提供定价、实施成本和回本案例
集成复杂度和互操作缺口约束当前抬高部署摩擦,拉长销售周期询问平均投产时间和集成人员需求
德国自动化需求周期性疲软逆风2026即便赛道结构性有吸引力,也可能压制本地订单时点按地区和行业拆分在手订单
公开定价透明度不足约束当前仅靠公开证据,SAM 和 ROI 模型可信度低要求提供产品级 ASP、毛利率区间和合同结构

驱动因素和约束的时点为定性判断;该表混合了自上而下的市场信号和公司层面的采用含义,因为估值同时取决于需求和落地摩擦。

[CM011, CM019, CM020, CM021, CM022, CM025]

2.5 图表

Chapter 03

03竞争对手

3.1 格局与买方替代方案

Agile Robots 并不在一个整齐的单一品类盒子里竞争。它的公开定位落在协作自动化、力控操作、AI 驱动机器人和高混线生产流程的交叉点。这意味着买方至少可以用四条替代路径解决同一任务。第一类是 Universal Robots、ABB、KUKA、FANUC 等广义协作自动化领导者,它们把机械臂与软件、培训、服务和集成支持打包。第二类是 Flexiv 和 Franka 等灵巧与力控专家,主要卖点是敏感操作,而不是目录广度。第三类是 Dobot 等低成本出口商,靠价格、伙伴触达和广泛入门级产品覆盖拉宽采用。第四类是买方侧替代:继续延长人工流程、混合使用集成商选择的组件,或选择更宽的工业自动化供应商,让机器人套件嵌入更大的工厂标准化项目。行业数据也重要。IFR、VDMA 和多家市场研究公司都指向庞大且仍在增长的机器人市场,但也显示买方注意力越来越被全球规模厂商、亚洲供应和 AI 重叙事牵引,而不是一个干净独立的「协作机器人」细分。[CP001, CP002, CP019, CP020, CP022, CP023]

竞争对手画像表
竞争对手 / 类别分类规模 / 覆盖信号目标细分市场差异化局限
Agile Robots具备 AI 的灵巧自动化新进入者公司称全球已安装 20,000+ 套解决方案汽车、电子、医疗健康和服务流程中的高混合工业自动化以 Diana 7 为锚点,讲述力控、灵巧性和 AI 驱动操控故事公开安装基数和服务证据少于既有厂商
Universal Robots协作机器人专家第三方指南称 50+ 个国家有 75,000+ 台 cobot;官网有 500+ 个 UR+ 项目SME 到企业级协作自动化易部署、培训界面、大型配件生态在力控研究和高负载重工业范围上的切入点不够明显
ABB大型工业自动化既有厂商大型全球自动化集团,集团层面 2025 年 R&D 为 $1.318bn跨多类机器人做标准化的企业客户产品组合和服务网络非常宽相比单点 cobot 新进入者,可能更重、更依赖集成商
KUKA汽车和工厂自动化既有厂商全球既有厂商,覆盖机器人、控制器、外设和移动解决方案需要系统集成的汽车、物流和 HRC 部署覆盖 cobot、移动解决方案和工业机器人编程和集成复杂度可能高于更简单的 cobot 优先平台
FANUC规模化工业机器人既有厂商第三方指南称已安装 1M+ 台机器人;官网列出 21 个系列和 100+ 个型号高稼动制造、汽车、电子、重载物料搬运硬件覆盖最广,并配套软件、仿真和服务不为低摩擦自助采用优化
Flexiv自适应力控专家更强调深科技定位,而非大众市场目录广度需要力敏自适应的复杂任务工业级力控与 AI 融合类似既有厂商的全球服务规模证据较少
Franka Robotics研究优先的操控平台学术和 AI 社区身份强重视直接控制和触觉行为的研究实验室与开发者参考平台地位和社区可信度工业覆盖窄于 ABB、FANUC 或 KUKA
Dobot低成本协作与教育出口商已售 100,000+ 台机器人;350+ 个合作伙伴;服务 100+ 个国家 / 地区预算敏感型工厂、教育和入门级协作自动化激进价格 / 合作伙伴触达和宽 starter 目录质量、支持和高级能力覆盖可能因应用而异

画像行汇总截至 2026-06-20 可从官方产品界面、母公司 / 投资者材料和可信市场概览看到的信息;在缺少经审计的竞争对手特定数据时,规模信号混合了上市公司事实、公司说法和第三方摘要。

[CP001, CP003, CP005, CP008, CP009, CP011]
FP001: 竞争定位图

以部署简易度(x 轴)和灵巧力控深度(y 轴)为序数维度,映射最相关的竞争者类别。

轴值是基于已审阅公开证据给出的 1-10 序数估计,不是已发布的基准分数;该图用于展示竞争格局形状,而非科学测量排名。

[CP003, CP005, CP009, CP014, CP019, CP020]

3.2 既有厂商画像与能力对比

最重要的竞争事实模式,是广度不对称。Universal Robots 强调简单、可重复和庞大的 UR+ 生态,即便载荷范围不能直接匹配每个工业任务,它仍是协作部署的强默认选择。ABB 和 KUKA 更像工厂自动化平台来切入市场:它们把协作机器人打包进更广的产品组合,后者包括多关节机械臂、专用机器人、软件、控制器、外设和服务触达。FANUC 在纯硬件广度上更宽,覆盖重载、协作 CRX、大型软件库和仿真工具。在这个背景下,Flexiv 和 Franka 重要,因为它们与 Agile 的灵巧叙事重叠更近。Flexiv 明确营销把工业级力控与 AI 融合的自适应机器人。Franka 仍是研究优先的参考平台,社区牵引和精细运动可信度较强。因此,当采购标准是敏感力交互、直接示教或受限环境中的更高精度操作时,Agile 最有竞争力。当买方优先考虑装机基础、全球支持密度或跨多类机器人实现广泛平台标准化时,它的处境就不那么有利。[CP003, CP004, CP005, CP006, CP007, CP008]

能力与战略方向对比
厂商核心能力信号软件 / 生态信号战略方向相比 Agile 的意义
Universal Robots3 kg 到 35 kg 协作机械臂,重复精度强UR+、培训、市场、应用套件简化采用,并借伙伴生态扩张当简单性比力控深度更重要时,是强默认选项
ABB覆盖关节、协作、delta、SCARA、喷涂和码垛的广泛产品组合更广服务网络和更宽自动化技术栈把机器人留在多品类自动化平台内客户想要一个规模化供应商时,很难被替换
KUKA传统机器人、cobot、移动解决方案、软件、控制器、外设偏系统级集成在一个品牌下融合 HRC、移动和工业自动化争夺复杂的工厂标准化预算
FANUC最高 2.3 吨负载范围,加上协作 CRX 选项250+ 个软件功能和 ROBOGUIDE把 AI、IoT 和供应韧性推入工业自动化抬高安装基数信任和生命周期支持门槛
Flexiv具备工业级力控的自适应机器人以 AI 为中心,但目录更窄赢下需要响应式力交互的任务与 Agile 的操控叙事重合度最高
Franka研究驱动的直接控制平台开发者和学术社区有吸引力继续做机器人和 AI 专业人士的参考平台买方优先看开放性和研究血统时,是强替代品
Dobot覆盖低端到中端市场的宽协作目录合作伙伴网络和云 / 控制器接口靠可负担价格和广度扩大出口把价格预期压到欧洲 / 日本溢价水平以下

该表关注战略而非穷尽:每行强调与 Agile Robots 最相关的厂商角度,而不是列出每个产品家族。

[CP005, CP008, CP009, CP010, CP012, CP014]
FP002: 功能广度 / 能力地图

围绕 Agile Robots 竞争组最关键的买家标准,对相对强项作比较。

[CP008, CP010, CP016, CP019, CP020, CP022]

3.3 定价、分销与切换成本

公开定价仍是这个市场最不透明的部分之一,但现有证据在方向上有用。第三方 2026 年指南把标准协作机器人的基础区间大致放在 $25,000 至 $60,000,尚未计入工具、视觉、合规工作和集成;高级配置则超过 $90,000。同一批指南把 Universal Robots、ABB 和 FANUC 放在重叠但差异明显的价格带,而 Dobot 等中国品牌则被描述为低成本挑战者。即便这些数字不等于真实标价,它们仍显示买方已经习惯预期的竞争压力。分销和切换成本比标价本身更重要。Universal Robots 用 UR+、培训和应用套件加速部署。ABB 强调服务网络广度。FANUC 卖的不只是机械臂,还包括控制器、软件、仿真、维护和长生命周期支持。一旦工厂在其中一个环境上实现标准化,替换就不只是换机械臂。通常还意味着重新培训员工、重新验证安全、调整集成,并可能重写工具流程。因此,Agile 需要足够强的操作和 AI 边缘,让买方愿意承担改变技术栈假设的成本,而不是购买最省事的既有厂商扩展。[CP008, CP010, CP016, CP032, CP033, CP034]

定价 / 包装与切换成本对比
厂商 / 类别公开定价信号包装 / 渠道模式切换成本驱动因素含义
Agile Robots已审阅界面没有广泛公开标价销售驱动的工业解决方案路径需要专有工作流证明来支撑换供应商竞争胜利必须来自能力,而非透明低价
Universal Robots第三方 2026 年指南给出 $30k-$60k 区间市场套件、培训、配件、认证伙伴UR+ 生态和操作员熟悉度工厂一旦在 UR 配件和培训上标准化,就具备粘性
ABB第三方 2026 年指南给出 $40k-$75k 区间集成商和企业驱动的自动化销售服务网络和更广自动化技术栈定价嵌在更大的工厂标准化决策里
FANUC第三方 2026 年指南给出 $35k 到 $90k+ 区间带控制器、软件和服务的工业平台销售软件、仿真、安装基数、服务流程在稼动率关键的工厂很难被替换
Dobot / 中国出口商第三方指南给出 $15k-$40k 预算框架合作伙伴占比较高的出口路径低前期价格可抵消较低的锁定深度向下压低协作机械臂价格预期
市场平均标准 cobot 在工具 / 集成前的基础价格为 $25k-$60k工具、合规、视觉和培训通常单独计费集成成本可能超过裸机械臂对比仅看标价会低估总竞争摩擦

定价数字是低可信度的第三方 2026 年可比值,而非官方厂商标价;这里只作为买方预期和预算带宽的方向性信号,不作为经审计 ASP 或报价合同。

[CP032, CP033, CP035, CP037, CP039]

3.4 护城河耐久性与替代风险

Agile 的护城河真实存在,但有条件。证据支持它围绕灵巧力控、研究级控制保真度和 AI 优先操作形成差异化位置。但这个护城河不等于免疫对手回应。ABB、FANUC、KUKA 和 UR 拥有更大的装机基础、更强的分销肌肉和更成熟的软件服务循环。Flexiv 直接在自适应力控话语上竞争。Franka 仍把研究与开发者社区表面做得有吸引力。Dobot 和其他中国出口商压低协作机器人价格栈底部。同时,第三方行业报道越来越把这个行业描述为从基础自动化走向自主化、AI 驱动机器人,并最终走向人形或通用系统,这会扩大未来进入者和相邻替代品清单。结果是,Agile 不能只靠品类新鲜感。它在精度要求高、触觉强、高混线制造任务中应最强,因为力控和学习能力会立刻发挥作用。当买方想要低前期成本、全球现场服务,或希望一家供应商已经掌握工厂更广自动化标准时,它应最弱。[CP011, CP018, CP019, CP020, CP021, CP023]

护城河耐久性 / 竞争风险登记表
护城河主张威胁严重性重要性缓解措施 / 尽调问题
力控和灵巧性领先Flexiv 和 Franka 瞄准类似的高敏感度操控需求与 Agile 支撑最充分的切入点功能重合最近询问触觉装配中的第三方节拍和任务成功率证明
AI 驱动差异化既有厂商越来越多把 AI、仿真和软件打包进更宽技术栈如果不绑定可量化工作流改善,AI 叙事可能商品化要求提供与 Agile 特定 AI 功能绑定的生产 KPI 变化
欧洲工业定位Dobot 和其他中国出口商压缩协作机械臂定价价格压力可能压窄客户为高端机械臂付费的意愿要求按地区和任务提供实际 ASP 以及赢 / 输数据
研究血统可信度Franka 已经占据很大一部分公开研究和开发者心智开发者偏好会影响试点选择和生态工具测试生产账户中买方要求 FCI/ROS 式开放性的频率
精密操控实力ABB、KUKA、FANUC 和 UR 拥有更强的分销、服务和工厂标准能力胜出仍可能输给既有采购和支持习惯量化集成时间节省和服务模式竞争力
品类新颖性市场正向自主性、AI 优先机器人和人形机器人拓宽新进入者或相邻平台可能快速重定义比较集合在押注更宽人形叙事前,监控 Agile 是否先靠现有机械臂获胜

严重性反映战略承销风险,而不是已观察到的故障率;每一行都是基于已审阅来源形成的尽调视角,并非确定的市场结果。

[CP018, CP019, CP020, CP021, CP030, CP031]
FP003: 护城河 / 就绪度 KPI

截至运行日,若干规模、生态和价格压力信号塑造了 Agile 的竞争环境。

[CP008, CP023, CP029, CP030, CP034, CP035]

3.5 图表

Chapter 04

04财务

4.1 收入模式与公开牵引

公开证据支持的是多元自动化技术栈,而不是单一产品业务。Agile Robots 营销机器人硬件、控制软件和应用 AI;2023-2026 年的收购增加了互补界面:Franka 扩展研究和敏感机械臂系统,idealworks 增加 AMR 硬件以及车队软件和仿真,thyssenkrupp Automation Engineering / Krause 增加交钥匙产线工程和安装。这构成了硬件收入、软件或支持收入、项目型集成收入的合理组合。难点在于,公开申报没有披露每个界面的收入占比、收入如何确认,软件是单独货币化还是被打包。最强公开收入信号是 Manufacturing Dive 引用公司口径称,2024 年收入约 €200 million,且同比翻倍。装机基础信号也很重要:BMW 提到生产中的 600 台、随后 850+ 台 iw.hubs,物流行业报道则称 Agile Robots 已交付超过 10,000 个自动化解决方案。这些都是可观的商业线索,但仍弱于经审计分部报告。因此,收入问题不是 Agile Robots 是否有商业活动——它显然有——而是收入组合是否足够经常性、高毛利且持久,足以支撑溢价机器人估值。[CI011, CI013, CI017, CI018, CI020, CI029]

收入流表
收入流机制单位 / 买方当前价值 / 状态质量尽调问题
机器人硬件(Diana、Yu、Franka、Thor、Agile Hand)一次性硬件销售,加集成 / 支持工业制造商;Franka 面向研究实验室活跃;产品家族公开营销,但未披露硬件收入拆分中 — 产品存在清楚,变现细节不透明要求提供 2024-2026 年按家族划分的硬件收入和毛利率
自动化软件(AgileCore / AI stack)软件随部署打包;可能有许可 / 支持收入系统集成商和运营方平台活跃,但未披露独立定价或软件 ARR低到中 — 平台角色清楚,合同模式不清楚要求提供软件许可结构和附加率
AMR / 厂内物流生态(idealworks)AMR 硬件、车队软件、仿真和服务BMW 以及外部物流 / 仓储客户安装基数已规模化,先有 600+、后有 850+ 个 iw.hub;外部客户群在扩大部署证明高,收入转化低要求提供 idealworks 收入、经常性软件占比和客户结构
交钥匙工厂集成(Krause / thyssenkrupp AE)项目工程、系统集成、交钥匙自动化产线汽车、电子、医疗科技、物流客户2026 年完成收购;带来长期客户关系中 — 客户相关性清楚,合同经济性未披露要求提供在手订单、利润率特征和里程碑付款结构
采购 / 供应链效率间接成本节省,而非收入流内部运营,尤其是电子品类Amazon Business 案例显示成本更低、交付周期更好中,作为成本代理指标,而非收入证据量化年度节省和营运资本影响

各行把有证据的业务线与内部效率杠杆分开。公开来源确认产品、部署和收购,但没有确认收入结构或确认政策。

[CI011, CI013, CI017, CI023, CI029, CI034]
定价 / 变现表
产品价格 / 单位 / 合同标价 vs. 实际成交折扣 / 未知项来源
Diana / Yu / Thor 硬件未公开披露Unknown标价、服务包和折扣梯度均未披露官方产品 / 公司页面
Franka 机器人未公开披露Unknown公开来源能证明需求和出货,但不能证明定价Franka 重启生产公告
idealworks AnyFleet / iw.hub / iw.sim 产品线未公开披露Unknown硬件销售、SaaS、部署费或支持服务各占多少仍不清楚idealworks 收购与 BMW 合作伙伴公告
Krause / 工厂集成项目项目制商业条款未披露可能按里程碑付款,但未公开积压订单、验收里程碑和留存机制未披露thyssenkrupp / 法律公告
Industrial AI Cloud / 仿真栈内部能力底座,不是公开对外销售产品定价N/A计算开支、预留容量和单模型成本未披露AI Cloud / NVIDIA 公告

所有已查阅来源都没有给出对外定价。本表列出可变现的界面,并把披露缺口明确标出。

[CI029, CI033]
FI001: 收入模型桥接图

有公开证据支撑的变现面,从硬件和软件延伸到厂内物流和全厂自动化。

收入权重未披露;该图只映射已有证据的变现面和客户结果。

[CI011, CI017, CI029, CI034]

4.2 规模、收购与资本建设

公司的公开战略明确是规模优先且资本密集。员工数引用从 2024 年 3 月的超过 1,700 人,升至 2026 年 Amazon Business 案例研究中的超过 2,300 人,以及 2025 年末 / 2026 年报道中的超过 2,500 人;仅 thyssenkrupp 交易就增加约 650 名专家和约 10 个地点。这种扩张不只是有机增长:它由一轮有意的并购整合推动,收购对象包括困境资产、战略资产和能力增厚资产。Franka 在破产后保住了敏感机器人装机基础;BÄR 增加系统集成传承;idealworks 增加带有 BMW 可信度的规模化厂内物流生态;tkAE / Krause 增加工厂工程纵深和长期 OEM 关系。独立物流报道还称,Agile Robots 每年在德国投入超过 €80 million 研发,并已落地超过 10,000 个自动化解决方案。与此同时,Agile One 生产、Kaufbeuren 的 Franka 重启,以及横跨欧洲、中国和印度的生产运营,都意味着沉重的制造和集成负担。如果收入质量跟上,这一战略可以形成强护城河;但它也意味着资本正同时被招聘、收购、制造和算力消耗,而不是流向单一可控项目。[CI008, CI009, CI010, CI016, CI019, CI020]

单位经济模型表
指标数值 / 空值置信度重要性尽调要求
2024 年收入规模~€200M(公司经 Manufacturing Dive 引用)目前找到的唯一公开收入规模信号获取经审计的 2024 年收入和 2025 年运行率
公开员工数1,700+(2024 年 3 月)至 2,300+(2026 年)至 2,500+(2025 年末 / 2026 年引用)暗示运营费用快速扩张,整合压力也在加重要求按职能和地域拆分月度员工数桥表
已交付自动化解决方案10,000+指向较广的部署足迹,以及已装机支持负担拆分活跃部署与历史部署
德国年度研发投入>€80M反映持续固定成本基础和资本化创新负担确认总研发投入、资本化政策和地点组合
采购效率代理指标电子成本下降;可得性和交期改善公开来源中找到的唯一运营效率代理指标量化节省金额、库存天数和供应商集中度
毛利率未公开披露没有毛利率数据,就无法承销硬件 / 集成业务质量要求按硬件、软件和服务拆分毛利率
烧钱额 / 现金跑道未公开披露公开证据不足以建模资本充足性要求月度烧钱额和账面现金

本章只有一个公开收入数据点和若干规模代理指标,但没有经审计的单位经济模型栈。空值是真实披露缺口,不是撰写遗漏。

[CI018, CI019, CI020, CI021, CI023, CI031]
融资与工业规模时间线
事件日期规模信号财务相关性
Series C 融资2021-09US$220M 轮融资;>US$1B 估值信号界定最后一次明确披露的大额股权事件
Franka 收购2023-11保留约 100 名员工;巴伐利亚生产延续增加困境资产整合负担;价格未披露
转为 SE2024-03欧洲法律形式;保留慕尼黑总部支撑跨境扩张,但没有给出账目
慕尼黑总部启用2026-06 报道2,300+ 名员工;>10,000 套解决方案;德国研发投入 >€80M反映沉重固定成本野心和品牌投入
tkAE 公告 / 交割2025-11 / 2026-04+650 名专家;+10 个站点;新增行业和 OEM 关系可能同时增加收入规模和整合成本
Industrial AI Cloud / Cosmos2026在工业云 + 仿真栈上训练基础模型带来计算强度和数据基础设施依赖

时间线把资本事件和规模承诺放在一起,因为公司公开讲述工业扩建,多于披露常规财务报表。

[CI006, CI007, CI008, CI009, CI016, CI021]
FI002: 公开财务估算区间

公开记录给出少数规模锚点,但不足以拼出完整财务模型。

收入代理值采用 Manufacturing Dive 的 €200M 数字及其美元换算;员工数和 AMR 区间对应不同时间点,并非同时值。

[CI018, CI019, CI021]
FI003: 资本强度 / 现金流地图

虽然没有现金流量表,公开可见的资本消耗点已经很清楚。

由于没有公开资产负债表或现金流量表,矩阵采用定性现金需求分档。

[CI019, CI021, CI022, CI024, CI026, CI037]

4.3 算力依赖与运营效率代理指标

Agile Robots 的不同寻常之处还在于,公开资本故事如今包括 AI 基础设施,而不只是机器人和工厂。公司承诺从 2026 年开始在 Deutsche Telekom 和 NVIDIA 的 Industrial AI Cloud 上训练基础模型,并另行公开称已提前接入 NVIDIA Cosmos 3 用于仿真。公司自己对 Industrial AI Cloud 的解释把取舍说得很清楚:通用机器人智能需要大数据集、强云基础设施和可扩展训练循环。换句话说,模型雄心在成为收入线之前,先变成运营成本线。公开资料没有披露预留 GPU 容量、云支出或模型训练单位经济性,因此投资人无法判断算力是可管理的差异化因素,还是拖累现金跑道的隐性负担。公开资料中唯一具体的运营效率代理指标来自采购侧:Amazon Business 引用 Agile Robots 采购负责人称,一些电子品类已经实现成本降低,同时可得性和交期改善。这证明供应链纪律在组织内部重要,但仍是狭窄代理指标,不能证明强毛利率或 CAC 效率。[CI023, CI024, CI025, CI026, CI039]

资本充足性表
项目公开状态含义尽调要求
种子轮至 Series C 融资种子轮、pre-A 轮和 Series C 已披露;Series C 总额为 US$220M历史股权支持可见,但时间已久要求披露 2021 年后的融资和当前股权结构表
收购购买价格Franka、idealworks 和 tkAE 价格未披露无法还原 M&A 现金使用要求披露购买对价、业绩对赌和承接负债
账面现金未披露现金跑道无法判断要求披露季度末现金和受限现金
月度烧钱额未披露无法评估融资依赖或下一轮触发点要求提供 12 个月烧钱额桥表
债务 / 项目融资义务已查阅来源未发现公开义务可能不存在,也可能只是未披露要求债务明细、担保和表外承诺
AI 计算承诺Industrial AI Cloud 和仿真项目已公开;成本承诺未公开计算可能成为不断增长的现金消耗项要求披露已签约云支出和 GPU 预留条款

本表把真正披露的事项与仍不透明的事项分开。它应被视为融资依赖清单,而不是资产负债表。

[CI004, CI005, CI006, CI024, CI026, CI030]

4.4 财务可见度缺口与结论

决定性尽调问题不是 Agile Robots 是否在扩张,而是公开证据是否足以判断扩张质量。答案是否定的。已审阅资料披露了历史融资、法律形式变更、收购逻辑、部署规模以及部分招聘 / 研发代理指标。它们没有披露当前现金、月度烧钱、现金跑道、毛利率、库存强度、应收账款、订单积压转化、重大收购价格、债务义务或实际定价。即便找到的一条类似申报的线索,也只是前身 AG 的注册摘录供应商结果,不含账目。这种不透明比通常更重要,因为公司正同时整合多项业务,维持巴伐利亚和亚洲制造,并叠加算力密集型 AI 项目。因此,财务结论是混合的:Agile Robots 有足够公开证据显示市场牵引和工业雄心,值得继续尽调;但常规财务披露不足,无法承销收入质量、利润率路径或资本充足性。任何投资判断都必须依赖直接进入 data room,而不能只靠公开叙事。[CI030, CI031, CI032, CI033, CI035, CI036]

公开财务缺口表
缺失的私有指标影响精确尽调路径
按硬件 / 软件 / 集成拆分的收入结构无法判断收入质量或周期性要求按业务线和地域提供 2024–2026 年收入桥表
按业务线拆分的毛利率无法区分可扩展的软件经济性与低毛利集成工作要求按产品家族提供月度毛利率和服务交付成本
现金、烧钱额和现金跑道无法评估对下一轮股权融资或债务的依赖要求最新现金余额、烧钱额桥表和下行情景跑道
营运资本 / 库存 / 应收款无法测算制造现金占用或里程碑付款风险敞口要求库存周转、应收账龄、递延收入和应付条款
M&A 购买对价和负债整合式扩张经济性不透明;整合成本无法建模要求已签署 SPA 经济条款、整合预算和协同计划
客户定价 / 合同结构无法推断实际定价权或折扣力度要求前 20 大合同、续约条款和价格瀑布

每一行都是重大尽调阻塞点。公开证据对规模和战略最强,对收入质量和资产负债表可见性最弱。

[CI030, CI031, CI032, CI033, CI038, CI035]

4.5 图表

Chapter 05

05产品与技术

5.1 产品组合与客户流程

Agile Robots 已不再像单一机械臂厂商。公开产品组合覆盖精密协作机器人、力控研究友好型机械臂、移动机器人、人形机器人系统,以及一个用于协调这些系统的 AI 软件层。放到客户流程里,公司卖的是智能生产系统的想法,而不是彼此割裂的机器人单元:机械臂处理操作,移动平台处理工位之间的移动和编排,AgileCore 处理配置和优化,更新的基础模型界面承诺随着时间推移减少手工编程。最高确信度证据仍来自具体硬件页面。Diana 7 明确定位在力敏感、高精度操作;Yu 5 Industrial 强调视觉驱动协作自动化;Thor 产品线覆盖更广载荷和臂展区间;Agile Hand 把技术栈延伸到灵巧末端执行器;移动机器人页面展示机械臂如何与伙伴 AMR 或 AGV 平台组合。这种广度有战略价值,因为 Agile 可以向工厂推销一整套集成自动化项目,而不是一次卖一台机器人。它也提高了执行门槛,因为每一层都必须与技术栈其余部分可靠协同。[CE001, CE002, CE003, CE004, CE017, CE019]

产品模块 / 资产矩阵
模块 / 产品线主要用户状态 / 成熟度核心差异化尽调缺口
Diana 7工业机械臂买家;通过 FCI 使用的研究用户在售产品,规格明确,并有研究扩展更新7 轴扭矩传感、0.5 N 力控,FCI 接入 Franka 生态未公开按垂直行业拆分的现场可靠性或商业采用情况
Yu 5 Industrial协作单元操作方在售产品,带视觉能力和 TÜV 认证集成摄像头 / NPU,加上协作安全和 0.5 N 力控声明公开客户案例结果有限
Thor 系列通用工业自动化买家组合已发布;Thor 20 标记为即将推出载荷从 3 kg 到 20 kg 形成阶梯,Thor 7 Pro 可选力控按型号划分的公开部署证据稀少
Agile Hand灵巧操作 / 研究买家在售专用末端执行器21 个关节、触觉、1 kHz 通信,支持 ROS 和 API与更广 Agile 系统的商业附着率未公开
AgileCore系统集成商 / 操作方软件平台正在积极营销自然语言编程、基于 RAG 的设置辅助、自优化未公开设置时间缩短或模型准确率基准
移动机器人栈工厂物流 / 移动机械臂买家由合作伙伴驱动的现行方案idealworks + BÄR 生态把能力扩展到固定机械臂单元之外合作伙伴依赖和所有权边界让产品责任更复杂
Agile ONE更长期的工业人形机器人买家公开证据处于发布 / 演示阶段分层 AI、灵巧手、真实工业训练数据公开证明仍偏向发布和演示内容,而非运营指标

各行概括截至 2026-06-20 与客户工作流和技术尽调最相关的公开产品模块;成熟度标签反映可见证明深度,而不是管理层内部路线图阶段。

[CE004, CE005, CE006, CE013, CE015, CE017]
工作流 / 用例表
用户任务当前工作流Agile 方案可量化收益信号限制
精密力敏装配手工或半自动操作,设置高度依赖示教Diana 7 + AgileCore0.5 N 力控声明和易示教界面公开周期时间和良率变化未披露
协作抓取 / 放置与检测标准协作机器人搭配独立视觉栈Yu 5 Industrial集成摄像头 / NPU 和协作安全声明未公开与 UR / ABB / Dobot 替代方案的基准对比
中重型上下料传统工业机械臂或定制搬运单元Thor 系列载荷阶梯和臂展范围让一个家族覆盖多类任务实际装机型号组合的公开证明有限
灵巧末端执行器研究或精细搬运定制夹爪或实验室自制手Agile Hand21 关节手,支持 API 和 ROS商业生产用例不如研究用途可见
操作员编程 / 集成手工编码、接线和集成商设置AgileCore自然语言任务设置和引导式外设集成未公开量化的上手时间缩短
厂内物流与移动操作AMR 和机械臂供应商分离idealworks/BÄR + Agile 栈AnyFleet、iw.hub、iw.os 和定制移动机械臂归入同一伞下共同责任模式可能模糊合作伙伴之间的问责

收益信号只限于公开来源陈述或暗示的内容;缺少量化 ROI 本身就是尽调事项,并不证明性能弱。

[CE006, CE016, CE017, CE019, CE021, CE022]
FE001: 产品架构图

从硬件模块到 AI 训练和编排,分层展示 Agile 的公开产品栈。

[CE004, CE015, CE021, CE025, CE027, CE035]
FE002: 客户工作流 / 运营流

Agile 如何呈现工业客户从任务定义走向自主运行和模型改进。

[CE015, CE016, CE021, CE023, CE025, CE026]

5.2 架构、控制与开发者界面

Agile 的公开架构故事在继承或连接 Franka 研究技术栈的地方最强。Diana 7 有最清晰的技术规格界面:七轴扭矩传感、7 kg 载荷、923 mm 臂展、±0.05 mm 重复定位精度,以及 0.5 N 力控精度。关键产品开发动作是加入 Franka Control Interface 支持,暴露 1 kHz 实时控制,并把 Diana 7 接入 libfranka、ROS、ROS 2、MATLAB 和 Simulink 生态。这是一座有意义的技术桥梁,因为它让 Agile 从纯工业叙事进入研究和开发者流程,而该流程已经拥有工具和社区预期。GitHub 证据强化了这一点。libfranka 明确是低层级、实时的;franka_ros2 明确处于活跃开发中,并以 ROS 2 Humble 为中心。与此同时,Agile Hand 暴露自己的 1 kHz 通信路径和 C++ / Python / ROS 兼容性,而 AgileCore 走向相反方向,用自然语言任务创建、基于 RAG 的设置指导和自优化行为来抽象控制。合在一起,这些来源指向双层架构:面向高级用户的低层级直接控制可信度,以及面向更广部署的高层级 AI 抽象层。这很有吸引力,但也意味着公司必须同时维护硬核控制保真度和易用编排。[CE005, CE006, CE007, CE008, CE009, CE010]

技术 / 运营架构表
层级 / 组件角色关键依赖风险
Diana 7 硬件力控 7 轴机械臂扭矩传感器、AgileCore、可选 FCI 桥接性能差异化取决于控制栈能否持续强于大型在位者
FCI / libfranka / franka_ros2 开发栈直接底层控制和开发者集成Franka 生态、ROS 2、实时网络开发者信号强度部分落在 Agile 自有仓库之外
Agile Hand灵巧末端执行器1 kHz 协议、带传感器关节、API 兼容性专用硬件增加集成和维护复杂度
AgileCore / AgileAI高层编排和自然语言任务创建LLM/VLM 模型、RAG 数据池、内部遥测模型安全和故障处理的公开透明度有限
Industrial AI Cloud基础模型训练和扩展层Deutsche Telekom 与 NVIDIA 基础设施外部计算伙伴依赖,以及数据治理问题
仿真 / 遥操作栈合成数据生成和难例数据采集Cosmos 3、Isaac Sim/Lab、GTC 双臂设置工具链依赖 NVIDIA 迭代节奏和兼容性
移动机器人合作伙伴运输、编排和混合车队协调idealworks AnyFleet/iw.os/iw.hub 与 BÄR 工程能力合作伙伴执行风险和产品边界模糊
Agile ONE 人形机器人面向未来工业任务的通用具身自动化层分层 AI 模型、真实世界数据、内部制造相比组合中其他产品,公开成熟度仍更偏演示

架构由产品页、生态公告、GitHub 仓库和合作伙伴界面还原;Agile 没有发布单一技术架构文件,因此部分连接由推断支撑,而不是公司明确画出的图。

[CE008, CE009, CE010, CE011, CE012, CE013]
FE003: 关键依赖地图

外部技术和生态依赖实质影响产品交付与路线图执行。

[CE010, CE012, CE021, CE023, CE025, CE027]
FE004: 产品成熟度 / 能力图谱

按证据深度和部署标准,比较主要模块在公开信息中的相对成熟度。

[CE006, CE012, CE013, CE019, CE023, CE031]

5.3 工业 AI 技术栈与生态依赖

Agile 的产品叙事越来越像基础模型叙事。Google DeepMind 合作明确把 Gemini Robotics 模型与 Agile 硬件和真实部署数据相连。Industrial AI Cloud 界面更进一步,主张真实工业数据、合成仿真数据和人类示范数据会汇入一个专有机器人基础模型,并在 NVIDIA 和 Deutsche Telekom 基础设施上训练。NVIDIA 在多个层面都重要:Agile 称自己多年使用 NVIDIA 工具,正在测试 Cosmos 3 用于仿真和数据生成,并出现在 NVIDIA 自己列出、使用 Cosmos、Isaac Sim 和 Isaac Lab 的人形机器人建设者名单中。生态也通过 Franka 和 GTC 演示循环,Agile 在那里围绕双臂 Diana 7 设置展示遥操作和数据采集。这套架构可能成为真实优势,因为它把现有部署转成模型改进飞轮。风险同样清楚。产品智能并非完全自给自足。它依赖外部模型伙伴、外部算力平台和一条公众无法独立审计的大数据管线。这使伙伴可靠性、数据治理和训练经济性成为核心产品问题,而不是边缘议题。[CE023, CE024, CE025, CE026, CE027, CE028]

信任 / 质量 / 合规表
控制 / 认证截至 2026-06-20 的状态范围缺口 / 限制
ISO 9001 管理体系已认证开发和生产机器人系统质量体系认证不等于产品级安全或网络安全证据
Safety Core for Robot Applications 安全核心2023 年获得 TÜV SÜD 证书基于 IEC 61508 的软件库,具备功率 / 力限制和碰撞检测公开证据证明证书存在,但没有完整披露安全论证
Yu 5 Industrial 硬件安全2024 年获 TÜV SÜD 认证IEC 61508、ISO 13849-1、ISO 10218-1;控制器安全功能认证证明基线,但不能证明每个部署场景中的现场表现
协作应用安全Yu 5 页面有声明ISO/TS 15066 协作场景下的力监测公开范围细节比认证公告更薄
内部制造和巴伐利亚重启生产可从 Kaufbeuren 更新看到Franka 和 Yu 5 生产延续未公开良率、报废率或工厂产能指标
AI 云模型训练控制部分可见在 Industrial AI Cloud 上进行基础模型计算和数据训练未公开相当于 SOC 2 或 ISO 27001 的安全 / 隐私 / 控制框架

本表区分明确证书和更宽泛的控制声明;公开文件没有给出范围细节时,保留缺口,不用推断填平。

[CE020, CE029, CE031, CE032, CE033, CE039]
路线图 / 发布 / 开发阶段表
日期 / 阶段功能 / 里程碑状态含义来源
2021ISO 9001 质量管理认证已完成早期证据表明,正式质量流程在后续产品认证前已经就位Agile ISO 9001 新闻
2023Safety Core TÜV SÜD 证书已完成显示其在更广泛硬件认证之前,已投入基础安全库Agile Safety Core 新闻
2024Yu 5 Industrial TÜV SÜD 认证和 Kaufbeuren 重启生产已完成表明协作硬件和制造连续性已准备就绪Agile Yu 5 与 Kaufbeuren 新闻
2025Diana 7 FCI 支持与 GTC 双臂遥操作演示已完成把 Agile 延伸到研究 / 开发者和数据采集工作流Agile FCI 与 GTC 新闻
2026DeepMind 合作,以及 Industrial AI Cloud 锚定客户身份进行中产品栈进一步嵌入基础模型和云算力依赖Agile DeepMind 与 AI Cloud 新闻
2026Agile ONE 发布、Hannover 演示与 AI 工厂启用早期商业化 / 演示人形机器人叙事具备战略意义,但阶段仍早于机械臂产品Agile ONE 新闻线索

路线图阶段根据公开发布证据推断,不应等同于内部产品管理关口;主要尽调价值在于排序,而不是精确判断待办节奏。

[CE008, CE023, CE025, CE029, CE031, CE032]

5.4 信任、质量与成熟度风险

信任故事可信但不均衡。Agile 可以指向 ISO 9001 质量管理认证、经 TÜV SÜD 认证且按 IEC 61508 评估的 Safety Core,以及 Yu 5 Industrial 通过 TÜV SÜD 对 IEC 61508、ISO 13849-1 和 ISO 10218-1 的认证。这些都是有意义的信号,说明公司投入了正式质量和安全流程,而不是把 AI 与机器人当作纯软件实验。生产控制也变得更清楚:Kaufbeuren 已从巴伐利亚重启 Franka 发货,Agile 称关键产品由内部制造。同一批证据也显示成熟度仍未完全证明。Agile ONE 在战略上令人兴奋,但公开记录仍由发布信息、Hannover 演示和基础设施公告主导,而不是广泛现场数据。整个技术栈的公开正常运行时间、MTBF、现场事故和安全治理披露仍稀少。DLR Agile Justin 血脉有助于解释为什么力控、触觉感知和全身 AI 反复作为优势出现,但血脉不等于商业可靠性。因此,投资人应把当前产品故事视为架构强、生态整合有前景、规模化工业运营证据只被部分证明。[CE029, CE030, CE031, CE032, CE033, CE034]

5.5 图表

Chapter 06

06客户

6.1 客户分层与采用界面

对一家仍为私营的机器人公司来说,Agile Robots 面向的客户界面异常宽。产品组合覆盖协作机械臂、敏感研究机器人、AMR、编排软件、仿真,如今还包括人形机器人系统。这种广度至少形成五个公开客户簇:汽车制造与厂内物流、非汽车制造与仓储、研究 / 学术、电子与精密装配,以及新兴、可承接人形机器人的工厂流程。最强分层证据来自 idealworks 和 Franka,而不是大量公开的 Agile 核心品牌终端客户名单。BMW 是规模化汽车使用的锚定参考账户;MoldTecs 和其他 idealworks 故事证明了非汽车仓储和制造相关性;Franka 和 Diana 7 则通过点名机构建立了单独的研究与学术细分。产品组合还反复提到电子和精密装配,并由 Amazon Business 的采购证据和公司材料中的行业引用间接支撑。这种广度是优势,因为它降低了对单一产品形态的依赖;但它也让披露挑战更难:公开记录显示 Agile Robots 能卖到哪里,却没有显示每个细分贡献多少收入。[CU001, CU002, CU003, CU011, CU018, CU030]

客户分群表
分群买方 / 用户 / 付款方用例规模收入 / 战略价值缺口
汽车制造 / 厂内物流BMW 生产和物流负责人 / 工厂操作人员 / BMW 预算负责人AMR 车队编排、零部件搬运、混行物流2023 年 600+ 台 iw.hub;到 2025 年 BMW 各站点 850+ 台最强的具名规模化生产证明;基石参考客户未披露收入占比或合同期限
汽车以外的制造 / 仓储运营和供应链团队 / 工厂用户 / 企业运营预算AMR 工作流、仓储、生产供给BMW 之外的 idealworks 具名客户,包括 MoldTecs;外部客户基础被描述为在增长证明其可走出内部孵化历史、具备市场化能力客户名单不完整,收入贡献未披露
研究 / 学术实验室负责人 / 研究人员 / 资助或机构预算力敏机器人研究、AI、HRI、控制实验MIT、Stanford、ETH Zurich、Max Planck、NVIDIA 被列为用户Franka 和研究级控制形成较强装机基础信号没有单机构支出或续约数据
电子与精密装配制造工程 / 操作人员 / 资本开支负责人精密装配和智能视觉应用公司材料反复提及该行业;电子领域有采购效率证据若放量,可支撑工业相关性和利润率上行公开具名终端客户证明很少
物流 / 仓储 / 服务机器人仓库运营 / 工厂物流 / 企业运营预算AMR 车队和新兴人形机器人工作流idealworks 和 Agile ONE 材料直接瞄准该分群交叉销售和未来软件附加的主要邻近市场人形机器人部署仍未规模化,客户证明偏薄

分群同时采用直接具名客户证明和行业层面证据。规模单元格优先看部署证明,而不是未披露收入估算。

[CU001, CU002, CU005, CU008, CU011, CU018]
客户增长 / 采用轨迹表
指标数值日期来源置信度含义缺失分母
BMW 部署的 iw.hub600+2023-09BMW / Agile idealworks 融资发布证明锚定客户内部已规模化生产使用缺少 BMW 工厂覆盖比例或单站点支出
BMW 部署的 iw.hub850+2025-09Agile / 物流报道显示全资收购后仍在扩张缺少外部客户拆分
BMW 每日任务量~30,0002026-06 访问idealworks BMW 案例高频运营使用,不只是试点存在缺少单次任务节省成本
BMW AMR 可用性98%2026-06 访问idealworks BMW 案例指向运营可靠性缺少相对替代方案基线
MoldTecs 每日任务量1,7002026-06 访问idealworks 客户案例证实非汽车场景已有运营使用只有一个公开案例
Agile Robots 全球部署20,000+ 套解决方案2026-03Agile / DeepMind 发布显示历史装机足迹较广不是客户数量
Franka 恢复发货客户已重新从 Kaufbeuren 收货2024-03Agile / Franka 发布破产后连续性已恢复未披露发货量

采用可见度最强的是部署数量和任务量,不是客户数、ARR 或转化率。重复出现的 BMW 数据点对应两个时间戳,不是写作重复错误。

[CU004, CU005, CU006, CU008, CU010, CU015]
FU001: 客户旅程图

公开可见路径从行业锁定和设计导入开始,延伸到工厂部署、任务强度和扩张。

旅程阶段来自公开证据,并非公司内部披露的销售流程。

[CU022, CU036, CU037]

6.2 具名客户证据与结果证据

本章最强的直接客户证据是 BMW / idealworks。BMW 先在 2023 年披露厂内物流流程中有超过 600 台 iw.hubs;2025-2026 年 idealworks 证据集把这一数字推进到全球 BMW 生产基地的 850+ 台机器人,每天约 30,000 次任务,98% 可用性。这些是强运营指标,尤其因为它们绑定具名工厂和持续生产环境,而不是会议 demo。非汽车证据不那么丰富,但仍真实存在:idealworks 的客户故事页面点名 MoldTecs,并称其 Sonneberg 工厂运行八台 iw.hubs,每天约 1,700 次任务。Amazon Business 提供另一类具名证据——不是产品收入,而是一个实时供应商关系,Agile 报告电子件成本降低、交期改善。Franka 又增加一层证据:2023 年破产后,Agile 从 Kaufbeuren 重启发货,并引用 MIT、Stanford、ETH Zurich、Max Planck 和 NVIDIA 的持续使用。结果是,部署和连续性证据可信,但 Agile 品牌工业机器人仍缺少广泛公开的具名生产买方名单。[CU004, CU005, CU006, CU007, CU008, CU009]

具名客户证明表
客户分群部署 / 用例生产 / 试点结果局限
BMW Group汽车制造 / 厂内物流idealworks AMR,加上跨多个工厂的编排与仿真生产 / 规模化600+ 到 850+ 台 iw.hub;约 30,000 次任务 / 天;98% 可用性;全球工厂推出未披露收入占比、合同期限或节省金额
MoldTecs制造idealworks AMR 在 Sonneberg 自动化生产供给中使用生产 / 规模化8 台 iw.hub;约 1,700 次任务 / 天;有直接供应链引述客户案例页上只看到一个公开的非汽车案例细节
Agile Robots(作为 Amazon Business 客户)采购 / 电子供应链企业采购和寻源优化生产 / 活跃供应商关系电子领域成本下降,同时可得性和交期改善证明的是采购支持,不是 Agile Robots 产品带来的收入
MIT / Stanford / ETH Zurich / NVIDIA 研究生态研究 / 学术Franka 敏感机器人和研究工作流生产 / 活跃装机基础具名机构继续依赖 Franka 触觉能力;2023 年困境后已恢复发货没有单位数量或续约数据

公司不发布完整具名客户名单,因此本表有意保持不完整。最强直接证明来自 BMW;Franka 研究用户有具名,但没有量化。

[CU006, CU008, CU009, CU010, CU011, CU032]
FU003: 客户证明矩阵

BMW 在规模和具体性上都最高;其他证明点真实存在,但变现透明度较低。

定性矩阵反映证据质量,而非客户价值。它用于比较不同关系类型下的证明强度。

[CU006, CU008, CU009, CU011, CU024, CU039]

6.3 留存、耐久性与扩张动态

传统留存指标没有出现在公开记录中。已审阅资料均未披露客户数、NRR、GRR、流失率、续约率或合同期限。这意味着客户耐久性必须从运营行为而不是财务指标推断。最好的耐久性代理指标是 BMW:Agile 收购 idealworks 全部股权后,BMW Group 仍是长期伙伴,装机基础还在按地域扩张。第二个耐久性代理指标是 Franka:2023 年破产相关中断后,2024 年恢复生产,研究生态看起来仍完整,具名机构仍被引用为活跃用户。MoldTecs 等 idealworks 外部案例显示,受制于母公司之外也可能具备一定可复制性,尽管公开证据仍浅。扩张动态也高度依赖渠道。Agile 越来越通过被收购生态触达客户——厂内物流中的 idealworks、研究中的 Franka、系统集成中的 BÄR,以及偏 OEM 自动化中的 Krause——而不是通过公开披露的直销漏斗。这让扩张显得合理,但也意味着交叉销售论点如果没有 CRM 和 cohort 数据,很难验证。[CU010, CU012, CU014, CU019, CU020, CU021]

留存 / 重复使用 / 满意度表
指标数值 / 空值分群置信度尽调要求
净收入留存(NRR)未公开披露全部要求按产品家族和收购单元提供 NRR
总收入留存(GRR)未公开披露全部要求按分群提供 GRR 和客户流失
合同期限 / 续约条款未公开披露BMW / idealworks / Agile 直销要求提供头部客户合同条款和续约节奏
耐久性代理指标:BMW 合作关系全资收购后,BMW 仍是长期合作伙伴汽车确认合同期限和排他性
耐久性代理指标:Franka 连续性破产后已恢复发货,并提到研究需求研究 / 学术要求提供积压订单、复购率和支持 SLA
耐久性代理指标:任务强度BMW 30,000 次任务 / 天,MoldTecs 1,700 次任务 / 天汽车 / 制造要求提供月度任务趋势和停机历史
公开满意度评价已审阅来源中未找到全部要求提供 NPS、客户推荐和第三方评价语料

公开记录提供了耐久性代理指标,但没有标准 SaaS 式留存披露。空值是真实数据缺口,不是漏做。

[CU022, CU026, CU027, CU037]
扩张与集中度风险表
扩张驱动因素集中度风险影响尽调路径
BMW / idealworks 参考客户锚定客户验证产品,但也可能主导叙事和管线可信度要求提供前 10 大客户集中度和 BMW 收入占比
外部 idealworks 客户中等显示制造 / 仓储场景已有非内部客户采用要求提供完整具名客户名单和年度经常性服务收入
研究生态(Franka / Diana 7)中等强生态可孕育未来开发者和买方要求提供实验室到商业转化案例和复购率
收购带来的渠道扩张(BÄR、Krause、Franka)中等偏高增长取决于能否整合不同渠道和支持团队审查收购后客户留存和交叉销售计分卡
人形机器人 / Physical AI 扩张高执行风险可能打开新工作流,但公开客户证明仍未规模化要求按用例提供付费试点、积压订单和转化率
采购效率合作直接集中度低,直接收入证明少支撑运营,但不能分散客户收入量化节省金额,并对照任何直接收入影响

集中度按证据质量评估,而不只看账户数量。即使外部扩张正在推进,BMW 仍主导最强公开证明。

[CU021, CU024, CU028, CU029, CU030, CU032]
FU002: 采用 / 部署漏斗

公开证明从宽泛行业野心收窄到少数高度具体的具名客户。

该漏斗是披露漏斗,不是已签收入漏斗:计数代表公开证据覆盖的层数,而非内部管线规模。

[CU024, CU025, CU036]

6.4 集中度风险、盲点与结论

客户结论令人鼓舞但不完整。Agile Robots 显然不只是幻灯片:BMW / idealworks 已有规模,MoldTecs 已在运营,Amazon 是正在发生的企业关系,Franka 展示了一个经历困境事件后仍存续的研究装机基础。但 BMW 以巨大优势主导公开证据质量。最强证据点、最密集指标和最清晰具名账户引用都来自同一圈层,即便实际收入集中度未来证明更低,这仍会抬高集中度风险。同时,公开记录很少说明留存、定价权、合同结构或头部账户经济性。公司也反复把电子和精密装配称为战略终端市场,但那里具名终端客户证据仍薄。正确尽调结论因此是平衡的:客户牵引真实且多细分,但收入耐久性和集中度仍需要资料室证据。在卖方流程中,缺失的客户指标包可能几乎和产品叙事同样重要。[CU024, CU025, CU026, CU027, CU029, CU036]

公开客户盲点表
缺失指标重要性精确尽调路径
客户数量没有客户数,20,000 次部署无法转化为账户广度要求按产品和地域提供活跃客户数
头部客户收入集中度BMW 可见度可能掩盖依赖要求提供前 10 大账户收入、毛利率和续约计划
试点到生产转化当前漏斗缺少分母要求按季度提供机会、试点和生产批次
留存 / 流失直接商业客户没有公开续约证据要求按批次提供 NRR、GRR、流失和扩张
合同期限 / 定价部署数量推不出耐久性和定价权要求提供样本合同和价格瀑布
被收购单元的客户重叠交叉销售假设取决于重叠度和附加率审查 CRM 去重、重叠地图和交叉销售管线

这些盲点重要,是因为 Agile Robots 有可信的部署证明,但客户层面的公开收入质量披露偏薄。

[CU024, CU025, CU026, CU027, CU036, CU039]

6.5 图表

Chapter 07

07风险

7.1 按严重度排序的风险图景

Agile Robots 进入 2026 年时,同时背负的执行要求异常密集。公司不只是卖出更多设备;它还要整合破产后的 Franka,把 thyssenkrupp Automation Engineering 转为 Krause Automation,整合 idealworks,并把 DeepMind、NVIDIA 和 Deutsche Telekom 依赖叠进下一轮产品周期。公开证据显示规模真实存在——2024 年收入约 €200 million、超过 2,500 名员工、已部署超过 20,000 套机器人系统——但规模也会放大任何运营失误的冲击半径。因此,最高严重度风险不是单一产品缺陷,而是系统性风险:整合滑坡、伙伴依赖和合规工作量可能传导为交付失误、利润率压缩和估值叙事变弱。下方风险热力图和传导图说明,投资论点应按整合与控制故事来承销,而不只是机器人增长故事。[CR013, CR014, CR015, CR016, CR017, CR026]

运营 / 质量 / 安全风险登记表
失效模式公开信号可能性严重性缓解成熟度剩余暴露未解决缺口
Franka 整合和复产未达预期破产和发货中断后,生产直到 2024 年 3 月才恢复中等复产后的良率、正常运行时间或积压清理没有公开 KPI
人形机器人和基础模型路线图跑得快于可靠性控制Agile ONE 发布更依赖 AI、云和模型迭代,而不是长期运行历史建设中没有公开现场可靠性或安全事故统计
多站点制造质量漂移生产足迹横跨欧洲、中国和印度,全球超过 15 个站点中等没有公开站点级质量看板或供应商计分卡
云 / 数据管线中断拖慢模型训练Industrial AI Cloud 是 2026 年基础模型开发的既定依赖发展中未公开备用算力或冗余披露

运营项聚焦公开可见的规模化和整合线索;未找到独立事故数据库来验证车队可靠性。

[CR007, CR008, CR009, CR028, CR029, CR030]
FR001: 风险热力图

按发生可能性和影响,定性梳理 Agile Robots 的主要公开风险向量。

发生可能性和影响是基于公开披露与监管的分析判断,并非量化概率。

[CR004, CR005, CR014, CR026, CR029, CR033]
FR002: 风险传导图

上游整合、合规与合作伙伴风险如何传导到收入、利润率和估值结果。

边表示从运营模型和公开合规依赖中推断出的方向性分析关系。

[CR013, CR017, CR026, CR029, CR030, CR039]

7.2 监管与法律暴露

监管 / 法律负担很重,因为 Agile 结合了实体机器人、工业软件、跨境生产和模型训练基础设施。欧盟 dual-use 和美国 EAR 框架并不能证明 Agile 目前每一笔出货都需要许可,但它们确立了一个事实:先进机器人硬件、软件和技术援助可能因目的地和最终用途而受控。另行看,AI Act 已在欧洲施加基于风险的制度,禁止性做法已经生效,高风险义务也接近实施里程碑。公开证据足以说明合规成本和文档负担正在上升,但不足以显示逐产品准备情况。遗留收购资产问题也重要。Franka 经历破产,Munich Startup 报道了围绕收购的补贴欺诈调查,Franka 出售还引发德国 - 中国审查公开关注。这种组合提高了客户、贷款方和监管机构要求比标准自动化供应商更扎实尽调包的概率。[CR001, CR002, CR004, CR005, CR006, CR021]

监管 / 法律风险登记表
风险司法辖区 / 触发因素当前证据可能性严重性缓解成熟度剩余暴露尽调路径
出口管制或外贸审批摩擦欧盟 / 德国 / 美国针对机器人、软件和技术协助的两用规则Franka 流程中已出现中国关联审查;欧盟两用和 EAR 框架仍相关建设中获取产品级出口分类和目的地筛查历史
EU AI Act 高风险义务欧盟工作场所和工业 AI 部署AI Act 现已按风险适用,禁止性实践已经生效;高风险映射未披露建设中要求 AI Act 合规路线图和法律负责人
被收购资产的历史法律尾部风险德国 / 被收购子公司Munich Startup 报道,Franka 收购时存在补贴欺诈调查;公开结果未披露Unknown要求律师备忘录说明继承诉讼、调查和赔偿安排
安全认证范围漂移欧盟 / 全球工业客户Agile 已有 ISO 和 TÜV 里程碑,但公开证据没有显示全车队认证覆盖中等审查逐产品认证矩阵和续期日历

各行是基于公开监管文本、媒体报道和公司披露作出的严重性排序定性判断;缓解成熟度是分析估计。

[CR004, CR005, CR006, CR010, CR011, CR012]

7.3 运营与伙伴依赖风险

运营上,Agile 已公开拿出质量里程碑,也重启了生产,但披露材料仍有大块盲区。ISO 9001 和 TÜV 证书能支撑流程可信度,Kaufbeuren 复产也说明 Agile 能较快稳住 Franka 并恢复发货。问题是,公开材料没有给出全车队的正常运行时间、质保或事故数据。2026 年路线图又高度押注伙伴基础设施:DeepMind 提供基础模型,Deutsche Telekom 与 NVIDIA 提供工业 AI 云算力,Krause Automation、idealworks 等并入实体负责拓宽市场路径和装机基础。每项依赖单看都合理,但叠在一起就是一个耦合系统:一份合同出问题、一个算力瓶颈、一次集成失手,都可能同时打断多条收入线。因此,下方依赖图是一件实用的承保工具:它标出哪些看似商业合作的关系,可能变成硬执行约束。[CR007, CR008, CR009, CR010, CR011, CR012]

合作伙伴 / 依赖风险登记表
依赖项交易对手 / 资产作用集中度失效情景严重性缓释信号剩余暴露
基础模型伙伴Google DeepMind模型智能和训练闭环若 Gemini 访问、定价或优先级变化,路线图会放慢已公开宣布长期研究合作
工业 AI 算力栈Deutsche Telekom / NVIDIA云端训练基础设施算力瓶颈或主权要求拖慢模型迭代欧洲基础设施叙事降低部分政策风险
新收购的工业平台Krause Automation(原 thyssenkrupp 部门)打入 OEM 关系并获得数亿级收入底盘整合失手会削弱客户留存和协同兑现保留 Krause Automation 传统品牌
AMR / 物流生态idealworks / BMW 网络装机基数扩张和软件生态联合路线图复杂,或大客户集中限制利润率杠杆850+ 台 iw.hubs 装机车队显示真实运营底盘

若合作关系走弱,交易对手按其对收入交付、模型训练能力或客户扩张的直接影响排序。

[CR013, CR014, CR015, CR018, CR023, CR024]
FR003: 依赖图谱

梳理 Agile 在 2025–2026 扩张周期中披露的最重要外部依赖。

该图聚焦外部披露的单点依赖,而非完整供应商基础。

[CR023, CR025, CR026, CR029, CR031, CR032]

7.4 缓释措施、监控项与论点失效标准

投资含义不是 Agile 不能投,而是尽调必须从口号式 AI 愿景转向运营证据。公开缓释措施确实存在:Franka 保留本地生产,大额交易配备法律顾问,安全认证里程碑已经拿到,欧洲云定位明确,慕尼黑研发基地每年投入超过 €80 million。这些动作足以支持继续跟进。但在管理层证明模型与算力伙伴的合同韧性、产品级出口分类、AI Act 工作流和现场可靠性看板之前,剩余风险仍然很高。投资者应把下方五项否决标准视为强制监控闸门。如果出口管制摩擦出现,AI Act 责任仍含糊,Krause / Franka / idealworks 集成滑坡,或重大安全事件浮现,正确反应不只是减仓,而是重新承保整个 AI 溢价倍数。这就是本章的核心教训:Agile 的上行空间真实存在,低估运营控制风险的代价也同样真实。[CR017, CR020, CR026, CR033, CR037, CR041]

人员 / 执行风险登记表
角色 / 职能依赖或缺口可能性严重性缓释信号尽调路径
整合领导力吸收 Franka、idealworks 和 Krause Automation,同时不丢速度Agile 让被收购运营单元保持活跃,并保留本地品牌 / 团队审阅整合 PMO、100 天计划和领导层留任指标
安全与合规人才AI Act、出口管制和产品认证都需要稀缺专家公司公开披露专门安全里程碑,并在 M&A 中聘用监管顾问要求提供安全、出口和 AI 治理负责人的组织架构图
跨境制造管理15 个以上站点,以及横跨欧洲、中国和印度的生产,推高协同负担慕尼黑总部和德国 R&D 投入显示集中管控意图审阅站点 KPI、供应商审计和升级节奏
R&D 烧钱纪律德国每年 €80M+ R&D 投入和人形机器人开发,增加固定成本承诺收入规模已达约 €200M,但利润率未披露按项目要求提供烧钱、资本开支和招聘计划

执行项强调组织压力:M&A、合规搭建以及内部出资的硬件 / AI 项目同时推进。

[CR003, CR014, CR016, CR017, CR034, CR035]
缓释与终止标准表
风险可监测触发点阈值 / 事件行动含义
出口管制升级新增许可要求、出货受阻,或客户最终用途受限一个季度内两个或更多战略出货延迟或被拒在目的地管制被梳理清楚前,暂停承销跨境上行空间
AI Act 合规滞后2027 年预算周期前没有具名负责人、风险映射或落地计划管理层无法展示产品级合规工作流将 AI 赋能扩张带来的估值上行视为未经验证
收购整合失手Krause / Franka / idealworks 未兑现客户承诺,或出现领导层流失12 个月内出现重大 OEM 延误、品牌反转或披露重组下调工业协同论点,并假设更低收入倍数
云 / 模型伙伴依赖兑现成风险伙伴合同变化、算力瓶颈,或没有备用技术栈出现重大训练放缓或排他性约束将 Agile 重新定价为系统集成商,AI 可选性更弱
安全 / 质量事故召回、严重事故,或保修趋势上行浮出水面旗舰产品出现一次重大安全事件,或反复现场故障在审阅根因证据前,停止承销人形机器人快速爬坡

触发点是实操尽调关口,不是法律确定性测试;每一项都用于判断投资论点应暂停还是重新定价。

[CR007, CR010, CR011, CR012, CR017, CR026]

7.5 展示材料

Chapter 08

08估值

8.1 建议与承保立场

公开证据足以让投资者认真看待 Agile Robots:它已经是一家有规模的工业机器人平台。但这些证据还不足以支持在未说明的高溢价估值上把公司承保为明确买入。公司确实有硬资产:超过 20,000 套已部署系统、公开约 €200 million 收入数据点、庞大员工规模、多国制造能力,以及围绕 DeepMind、NVIDIA、Deutsche Telekom 和慕尼黑 AI 生态的势能。这些不是表面信号。问题在价格敏感性。已抓取来源披露了历史融资轮次和 2021 年独角兽里程碑,却没有披露 2026 年当前估值、清算优先栈或运营经济性。因此,近期正确立场是跟踪 / 继续研究:保持贴近,拿到资料室,在毛利、烧钱和商业韧性得到证明前,不把叙事势能折算成溢价入场价。下方建议表和逻辑图说明,单靠规模还不够。这个披露缺口正是入场需要耐心的原因。[CV004, CV014, CV016, CV017, CV035, CV039]

建议摘要表
维度当前判断公开支撑决策含义
建议跟踪 / 继续研究战略动能强,但定价和经济性透明度有限没有数据室前,不要按假设溢价承诺投资
置信度中低很多事实公开,但当前估值和利润率数据没有公开在把上行空间视为可投前,先通过尽调关口
风险评级资本强度、整合负担和合规压力仍然偏高要求下行保护或更低价格
估值立场合理到偏高最近披露的独角兽标记和情景估算,尚未显示明确折价承销进入纪律,而不是叙事动能

本表把本章证据转成可提交 IC 的判断;估值立场是分析结论,不是已披露的市场标记。

[CV035, CV039, CV040, CV041, CV043]
FV001: 建议逻辑

决策流把规模、市场动能、估值不透明度和下行触发因素,与本章建议连接起来。

该流程是定性的,且刻意不量化;它说明为什么建议取决于缺失的尽调事项。

[CV014, CV016, CV023, CV035, CV039, CV041]

8.2 融资背景与可比公司集合

Agile 的融资历史亮眼,而且跨度够长,值得重视。公司从早期 8 位数融资推进到 2019 年 pre-A,再到 SoftBank 领投的 2021 年 Series C,第三方档案还指向 2022 年后续融资。这段历史支撑了一个判断:风险投资人反复为这个平台出资。但它并没有告诉我们公司今天值多少钱。公开可比公司覆盖也不干净。ABB、KUKA、FANUC、Teradyne / Universal Robots 可以作为披露质量、规模和竞争广度的参照,但没有一家能与 Agile 这种仍为私营、AI 权重高、收购活跃的平台做严格同类对比。这里最安全的可比用法是定性:它们说明在位者业务宽、资本充足、还要接受公开市场问责,这会抬高任何私营公司估值溢价的门槛。因此,下方可比表聚焦可比性边界,而不是假装上市同业倍数可以直接平移。[CV001, CV002, CV003, CV004, CV006, CV007]

论点 / 反论点表
论点公开支撑什么会改变判断
具备真实规模的 AI 赋能工业平台20,000+ 套已部署系统、€200M 收入数据点、DeepMind 和 NVIDIA 相关动能需要披露利润率、客户留存和当前价格,才可上调判断
植根德国、全球足迹扩张的机器人冠军慕尼黑总部、€80M+ 德国 R&D 投入、制造横跨欧洲 / 中国 / 印度需要证据证明跨境复杂性带来利润率杠杆,而不是成本外溢
反论点:估值可能比证据更快计入热度当前私募标记未披露,人形机器人经济性仍有争议当前定价轮加单位经济性披露,可以缩小这个缺口
反论点:收购和平台宽度稀释焦点Franka、idealworks、Krause、人形机器人和 AI 云都在争夺资本和执行带宽若能证明整合成功、现金转化稳定,反论点会减弱

每个论点都刻意配上反转条件,让建议随证据变化,而不是被叙事牵着走。

[CV014, CV016, CV017, CV035, CV039, CV041]
可比估值表
可比对象公开状态 / 申报信号与 Agile 的相关性局限
ABB Robotics产品组合覆盖广泛关节机器人和协作机器人的上市龙头展示多元化工业龙头如何讲述机器人广度和客户覆盖比 Agile 多元化得多,不是干净的纯机器人倍数
KUKA欧洲上市机器人和自动化运营商可作为德国 / 工厂自动化参照点业务组合包含更重的系统集成和成熟汽车敞口
FANUC拥有综合报告历史的日本上市龙头可用于规模、报告纪律和工业装机基准利润率结构和区域组合与 Agile 明显不同
Teradyne / Universal Robots上市母公司申报历史,加上通过 Universal Robots 获得的协作机器人敞口可作为机器人相邻公开市场治理和披露基准母公司除协作机器人外,还包括半导体测试和更广业务
Universal Robots 产品节奏官方新闻流显示协作机器人持续创新凸显协作 / 工业自动化竞争压力不是独立上市估值参照

可比组刻意保持定性:抓取到的公开申报 / 页面提供了覆盖和披露锚点,但没有给 Agile 提供当前干净的同业倍数数据集。

[CV028, CV029, CV030, CV031, CV032, CV042]

8.3 牛市、基准与熊市框架

当前私有估值未披露,场景分析是把证据转成价值的唯一可辩护方法。最后一个公开收入数据点约 €200 million,可用来勾勒粗略区间,而不是精确目标。基准情形下,Agile 更像一个规模较大、仍在增长、但披露不透明的工业机器人平台,值得尊重,却不该获得不受约束的 AI 稀缺性定价。对应粗略区间为 €0.8 billion 至 €1.2 billion。牛市情形不能只靠标题,需要证明 DeepMind 加持的智能、自研人形机器人生产和并入平台能转化为持久收入质量与利润率扩张。熊市情形同样合理:人形机器人经济性若仍停留在宣传层面,集成负荷若压缩利润率,或新资本以偏向内部人的条款进入,都可能触发下行。换句话说,场景区间宽,是因为披露薄,不是因为市场机会小。[CV014, CV023, CV024, CV033, CV035, CV036]

牛 / 基准 / 熊情景表
情景核心假设指示性 EV 区间概率信号关键下行 / 上行驱动
AI 合作转化为持久软件 / 服务变现,人形机器人爬坡落地,整合成功€1.2B-€1.6B需要数据室给出强于公开来源的证据AI 可选性转成可复制的工业 ROI
基准工业增长延续,收购逐步消化,但经济性仍好坏参半,投资人要求纪律€0.8B-€1.2B最符合当前公开证据执行可信,但还不足以支撑溢价定价
人形机器人经济性滞后,整合拖累,市场情绪向传统自动化压缩€0.6B-€0.8B若当前披露缺口掩盖弱利润率或重度稀释,则情景成立倍数压缩和更高融资需求

企业价值区间是低置信情景估算,锚定最后公开收入数据点和行业增长背景,而不是已披露的市场报价。

[CV014, CV023, CV024, CV036, CV037, CV038]
FV002: 估值敏感性

以最后一个公开收入数据点为锚,在不同倍数假设下给出示意性企业价值情景。

数值是以最后公开的约 €200M 收入说法估算的 EV,单位为百万欧元;它们是情景工具,不是观察到的市场价格。

[CV014, CV036, CV037, CV038]
FV003: 估值 / 回报区间

区间图比较公开证据情形下的熊市、基准和牛市企业价值带。

区间是低置信度估算,因为抓取来源未披露当前估值、稀释结构和利润率。

[CV036, CV037, CV038, CV040]

8.4 最终尽调要求与论点失效触发器

Agile 的尽调负担很明确,也可操作,但任何委员会上调建议前都必须把这些问题清掉。第一,投资者需要当前定价证据:最新融资轮、董事会估值或真实二级交易价格。第二,需要经济性,包括毛利率、运营开支、资本开支以及主要项目的现金消耗。第三,需要清算优先栈和分配瀑布,因为企业价值看起来健康时,普通股结果仍可能很差。第四,需要证明部署规模正在转化为粘性客户,以及软件或服务变现。在这些要求得到回答前,正确动作是监控否决触发器,而不是追逐上行标题。如果量产滑坡、集成破裂、AI 伙伴变成集中风险,或公司以高度结构化条款融资,估值就应向传统自动化供应商压缩,而不是按 AI 溢价平台定价。[CV012, CV021, CV035, CV039, CV040, CV041]

论点破裂与终止触发点表
触发点阈值 / 事件对论点的传导行动含义
无披露却要求溢价估值管理层寻求明显高于公开独角兽基准的价格,但没有当前经济性材料包动能变成未定价的不透明拒绝,或要求进入下一步前开放数据室透明度
人形机器人商业化停滞量产推迟,或到 2026 年 ROI 证据仍停留在个案牛情景可选性塌缩成成本中心只使用基准 / 熊区间,并削减战略溢价
被收购单元整合失手出现重大客户流失、重组或交付延迟平台论点变成复杂性论点重新定价到更接近传统自动化,并降低兴趣
AI 伙伴依赖收紧没有备用算力 / 模型路径,或出现不利合同条款可选性变成集中风险要求合同尽调并压低进入价格
资本需求加速新融资带有惩罚性优先权或重度稀释即便企业增长,普通股上行也可能收缩任何投资决定前重建 waterfall

这些终止触发点设计用于第一次会面到最终 IC 批准之间的监测,而不只是投后监测。

[CV012, CV016, CV035, CV039, CV040, CV041]
最终尽调问题表
主题缺失证据为什么重要负责人 / 尽调路径
当前估值最新定价轮、二级交易或董事会标记进入纪律取决于今天的价格,而不是 2021 年独角兽标签CFO / 董事会材料
利润率和烧钱按产品族拆分的毛利率、opex、capex 和现金 runway决定 AI 和人形机器人可选性是增值还是稀释财务数据室
优先权栈清算优先权栈和反稀释条款普通股结果可能与企业价值明显背离法务 / 融资文件
商业耐久性留存、集中度和软件 attach验证部署规模能否转化为经常性经济性收入运营 / 客户分析

每个尽调问题都对应一个证据缺口;仅凭公开证据,建议还不能从跟踪 / 继续研究上调。

[CV035, CV039, CV040]
FV004: 投资 KPI

IC 风格评分卡,覆盖市场、证明、护城河、经济性、风险和披露质量。

评分带有分析性和相对性;它们总结本章证据,而不是套用标准化第三方评分框架。

[CV017, CV023, CV024, CV026, CV033, CV039]

8.5 展示材料

附录 A: 最终尽调问题

  • 当前价格锚:最新轮次、董事会估值或真实二级市场成交
  • 单位经济性:毛利率、服务收入占比、支持负担和资本开支强度
  • 资本结构:优先权堆栈、反稀释条款和控制权图谱
  • 商业质量:按主要客户群组拆分的续约、集中度和扩张证据
  • 被收购业务的整合评分卡,以及人形机器人投入保持纪律的证明
[CV035, CV039, CV040, CV041, CV043]

免责声明

本报告仅用于研究和尽调目的,不构成投资建议。报告依赖截至 2026-06-20 可获得的公开材料。Agile Robots 是私营公司,关键运营和估值事实仍未披露;任何投资决策前,都应在一手尽调中验证这些事实。

证据索引

结论
编号陈述可信度来源
CO001 Agile Robots was founded in 2018 by DLR robotics researchers Zhaopeng Chen and Peter Meusel. SO002, SO005, SO018, SO024
CO002 Agile Robots positions itself as an AI-powered automation company combining robotics hardware, sensors, computer vision, and software. SO001, SO002, SO003
CO003 The company says its solutions are used across automotive, consumer electronics, healthcare, logistics, and service workflows rather than a single vertical. SO001, SO002, SO028
CO004 Agile Robots states that Munich is its headquarters and research hub while Kaufbeuren serves as a production site. SO002, SO008, SO024
CO005 The company publicly lists sites in Germany, China, India, and the United States, indicating a globally distributed operating footprint. SO002, SO008
CO006 Agile Robots converted from a German AG into the European legal form SE in March 2024. SO007
CO007 The company’s ISO 9001 certification covers development and production of robotic systems. SO017
CO008 Agile Robots describes Diana 7, Yu 5 Industrial, Agile Hand, and AgileCore as part of its current core product portfolio. SO001, SO002, SO024
CO009 The earliest disclosed institutional funding sequence publicly referenced by Agile Robots runs from angel and seed financing into a pre-A round before later larger rounds. SO005
CO010 Agile Robots announced an eight-figure Series A round led by C-Ventures in early 2020. SO004
CO011 Agile Robots said it completed a Series C financing of US$220 million in 2021. SO006
CO012 The same Series C announcement described Agile Robots as having raised over US$130 million in 2020 and reaching a valuation above US$1 billion. SO006, SO018
CO013 SoftBank Vision Fund 2 participated in the Series C round and became the company’s most visible brand-name financial backer. SO006
CO014 Earlier disclosed backers include Hillhouse Capital, Sequoia Capital China, Tinavi, Linear Venture, and C-Ventures. SO004, SO005, SO006
CO015 Agile Robots opened a new global headquarters in Munich in 2025 with DLR affiliation and Bavarian political support highlighted in the launch narrative. SO008
CO016 The Munich headquarters release said Agile Robots employed more than 2,300 people worldwide in mid-2025. SO008
CO017 A 2025 Kaufbeuren profile described Agile Robots as employing more than 1,900 people worldwide. SO024
CO018 CNBC reported in March 2026 that Agile Robots employed more than 2,500 people globally after its recent expansion. SO028
CO019 Public headcount references therefore show a fast-moving range from more than 1,900 to more than 2,500 employees rather than a single audited figure. SO008, SO024, SO028
CO020 Agile Robots says it has installed more than 20,000 robotic solutions worldwide. SO014, SO015
CO021 The company’s 2025 acquisition announcement for thyssenkrupp Automation Engineering said Agile Robots had doubled revenue annually since founding and reached about EUR200 million in 2024. SO013, SO020, SO028
CO022 The Franka transaction followed Franka Emika’s insolvency and creditor-committee approval, making it both a capability acquisition and a distressed-asset integration. SO009, SO025, SO026
CO023 Independent reporting said the Franka deal preserved roughly 100 jobs and secured continuity for the research platform. SO026
CO024 Agile Robots acquired a majority stake in BÄR Automation in 2023 to add driverless transport and special-purpose automation capabilities. SO010, SO024
CO025 Agile Robots became idealworks’ majority shareholder in 2023, with BMW describing idealworks as a robotics ecosystem for industrial automation and intralogistics. SO011, SO021
CO026 Agile Robots acquired the remaining idealworks shares in 2025, leaving BMW as a long-term commercial partner rather than a co-owner. SO012
CO027 Agile Robots agreed to buy assets of thyssenkrupp Automation Engineering in late 2025 and closed the transaction in April 2026. SO013, SO014, SO022, SO023, SO027
CO028 Both Agile Robots and thyssenkrupp framed the acquisition as a way to combine automation engineering with Agile’s robotics software and to open new end markets and OEM relationships. SO013, SO022, SO023
CO029 Manufacturing Dive reported that the thyssenkrupp Automation Engineering deal broadens Agile Robots beyond its earlier sector concentration and gives it more access to factory automation programs. SO020
CO030 Agile Robots joined ARENA2036 in 2026 to deepen industrial research links around wiring harnesses and smart automation experiments. SO016
CO031 The company’s Google DeepMind partnership is structured as a research collaboration to integrate Gemini Robotics foundation models with Agile Robots hardware. SO015, SO028
CO032 Agile Robots and Google DeepMind said the first high-value industrial use cases include electronics, automotive, data centers, and logistics. SO015, SO028
CO033 The combination of AI foundation-model partnerships, in-house manufacturing, and acquisitions suggests Agile Robots is pursuing a vertically integrated industrial robotics stack. SO012, SO013, SO015, SO024
CO034 The Franka acquisition also creates integration risk because the target entered insolvency before closing and public post-merger performance metrics remain limited. SO009, SO025, SO026
CO035 The headcount, revenue, and deployment figures used in press materials are unaudited public claims rather than filing-grade disclosures. SO008, SO013, SO015, SO024, SO028
CO036 Despite European headquarters and DLR roots, Agile Robots’ funding and expansion profile is international, spanning Chinese investors, SoftBank, BMW, Google DeepMind, and global manufacturing sites. SO005, SO006, SO008, SO011, SO015
CO037 Agile Robots remains privately held with no public debt, cap-table, or board-composition disclosures sufficient to verify control rights. SO006, SO007, SO018
CO038 By 2026, the company narrative had shifted from standalone cobot supplier to broader physical-AI platform consolidator. SO001, SO013, SO015, SO019
CM001 The relevant market for Agile Robots is narrower than all robotics and broader than a single arm category: it spans integrated industrial automation systems, software, and adjacent mobility for factory and logistics workflows. SM004, SM008, SM012, SM015
CM002 Agile Robots’ own solutions pages anchor demand in manufacturing tasks such as machine tending, assembly, material handling, quality inspection, dispensing, and finishing. SM008
CM003 Official product pages show that Agile Robots sells both fixed robot hardware and software, while its mobile robotics offering extends into AGV and AMR-linked workflows via BÄR Automation and idealworks. SM008, SM012, SM015
CM004 Mordor Intelligence estimates the industrial robotics market at USD54.28 billion in 2026, reaching USD94.38 billion by 2031 at an 11.7% CAGR. SM003
CM005 MarketsandMarkets estimates the industrial robotics market at USD15.5 billion in 2026 and USD20.8 billion by 2032, implying a 5.0% CAGR. SM004
CM006 Verified Market Research estimates the industrial robotics market at USD19.17 billion in 2024 and USD39.56 billion by 2031, implying a 10.46% CAGR. SM005
CM007 Future Market Insights estimates the industrial robotics market at USD65.1 billion in 2026 and USD343.8 billion by 2036, implying an 18.1% CAGR. SM006
CM008 The wide spread across public TAM estimates indicates that publisher methodologies are not directly comparable and often mix hardware, software, integration, and adjacent automation layers differently. SM003, SM004, SM005, SM006
CM009 MarketsandMarkets says Asia Pacific accounted for 67.3% of industrial-robotics revenue in 2025, highlighting a demand center outside Agile Robots’ Bavarian home base. SM004
CM010 VDMA describes German robotics and automation as a >400-member industry body with forecast 2025 revenue of EUR13.8 billion. SM002
CM011 The same VDMA page says Germany’s robotics and automation industry remained in difficult waters in 2026 and expected a 5% revenue decline to EUR14.1 billion. SM002
CM012 MarketsandMarkets identifies automotive, electrical and electronics, food and beverage, precision engineering, pharmaceuticals, and other industrial categories as core end-use segments. SM004
CM013 Agile Robots’ official solutions page specifically highlights automotive, consumer electronics, healthcare, and smart manufacturing as current target sectors. SM008
CM014 Agile ONE is positioned for manufacturing and logistics environments where it can move between workstations and collaborate with humans. SM010
CM015 Yu 5 Industrial is positioned as a fast-to-deploy collaborative robot with integrated vision for picking, placing, and varied industrial tasks. SM014
CM016 Diana 7 is marketed around torque sensing, seven-axis dexterity, and intuitive setup, supporting precise assembly and sensitive manipulation use cases. SM009, SM022
CM017 AgileCore targets system integrators and operators, showing that software buyers and deployment partners matter alongside robot-unit buyers. SM012
CM018 The mobile robotics page shows that buyer workflows extend into intralogistics, AGVs, AMRs, and mobile manipulators rather than only fixed arms. SM015
CM019 ResearchAndMarkets’ 2025 summary says industrial robots are shifting from automation toward autonomy through AI, computer vision, digital twins, and collaborative systems. SM007
CM020 MarketsandMarkets says smart manufacturing, AI, IIoT, and demand for operational efficiency are major market drivers. SM004
CM021 Mordor Intelligence attributes market growth partly to higher factory wages, tighter reshoring economics, and government subsidies. SM003
CM022 MarketsandMarkets says collaborative robots are the fastest-growing segment because of flexible human-robot collaboration and lower deployment barriers. SM004
CM023 The same report says handling applications currently hold the largest share of industrial robotics demand. SM004
CM024 MarketsandMarkets says robots with up to 16kg payload hold the largest share, which aligns better with Agile’s lighter-arm portfolio than with very heavy industrial cells. SM004, SM009, SM014
CM025 The ResearchAndMarkets summary highlights humanoid, collaborative, and AI-driven robotics as reshaping manufacturing, but still frames the trend inside broader industrial automation rather than consumer robotics. SM007
CM026 Agile Robots says its humanoid foundation models are trained on real production-floor data, simulation data, and teleoperation data, which implies an adoption path tied to data-rich industrial customers first. SM010, SM016, SM020
CM027 The Google DeepMind partnership coverage says first use cases will center on electronics, automotive, data centers, and logistics. SM023, SM025, SM026
CM028 The industrial AI Cloud announcements show Agile Robots wants European cloud and NVIDIA-linked infrastructure to train industrial foundation models on production data from 2026 onward. SM018, SM019, SM020
CM029 Amazon Business’s customer case says procurement cost, availability, and lead time improvements matter economically to Agile Robots, reinforcing that supply-chain execution is part of buyer ROI. SM021
CM030 Agile ONE’s official page says the humanoid can walk at 2 meters per second and is intended as a co-worker on production floors. SM010
CM031 The Agile ONE launch announcement said full production would start in early 2026 in Bavaria and that Agile Robots would manufacture the platform in-house. SM016
CM032 MarketsandMarkets lists high cobot costs, integration complexity, and lack of standardization or interoperability as key restraints and challenges. SM004
CM033 Because Agile ONE, mobile robotics, and AgileCore all depend on workflow integration, switching costs are likely created at the system level rather than at the robot-arm level alone. SM012, SM015, SM019
CM034 The official solutions pages imply multiple buyer roles: plant engineering, operations, procurement, system integrators, and logistics owners. SM008, SM012, SM015, SM021
CM035 ARENA2036 membership gives Agile Robots a research-campus route into customer-adjacent experimentation and validation rather than only direct sales channels. SM024
CM036 The market evidence supports strong top-down demand, but it does not isolate a clean SAM for Agile Robots because public reports aggregate very different robot categories and services. SM003, SM004, SM005, SM006, SM007
CM037 Agile’s portfolio is better aligned with flexible production, electronics, assembly, and intralogistics than with the heaviest fixed-cell industrial robot categories dominated by incumbents. SM008, SM009, SM014, SM015, SM022
CM038 The available public evidence does not disclose product-level pricing or contract structures, so ROI must be inferred from operational language and partner case studies rather than measured directly. SM008, SM021
CP001 Agile Robots says it builds AI-driven automation solutions for automotive, consumer electronics, healthcare, and service industries. SP001
CP002 Agile Robots says its solutions combine robotic arms, mobile platforms, and software for demanding industrial customers. SP001
CP003 Agile Robots markets Diana 7 as a force-controlled 7-axis arm with torque sensors on all seven axes. SP002
CP004 Agile Robots says Diana 7 offers 7 kg payload, 923 mm reach, ±0.05 mm repeatability, and 0.5 N force-control accuracy. SP002
CP005 Universal Robots presents its cobots as a scalable industrial automation platform rather than a single-arm product. SP003
CP006 Universal Robots publicly lists collaborative arms from UR3e through UR30, covering payloads from 3 kg to 35 kg and reaches up to 1750 mm. SP003
CP007 Universal Robots advertises repeatability down to ±0.03 mm on its public cobot range. SP003
CP008 Universal Robots says the UR+ marketplace includes more than 500 certified kits, components, software, and accessories. SP003
CP009 ABB presents a broad robotics portfolio covering articulated, collaborative, delta, SCARA, paint, and palletizing robots. SP005
CP010 ABB says its collaborative robots are easy to set up, program, operate, scale, and support through a broad service network. SP005
CP011 ABB reported $1.318 billion of 2025 R&D investment at the group level in its annual reporting suite. SP006
CP012 KUKA says its portfolio spans traditional industrial robots, cobots, mobile solutions, software, controllers, and peripherals. SP007
CP013 KUKA explicitly frames direct human-robot collaboration and Industrie 4.0 mobile solutions as part of the same product spectrum. SP007
CP014 FANUC says it has over 50 years of robotics experience and supports industrial automation through robots, software, simulation, and service. SP009
CP015 FANUC advertises payloads up to 2.3 tons, collaborative payloads up to 50 kg, 21 series, and more than 100 specialized models. SP009
CP016 FANUC says its software catalog includes more than 250 advanced functions plus ROBOGUIDE simulation. SP009
CP017 FANUC’s 2025 integrated report says automation is expanding beyond factories into logistics, construction, food, pharmaceuticals, cosmetics, and agriculture. SP010
CP018 FANUC’s 2025 strategy emphasizes quality, customer-oriented advanced technologies, supply resilience, and more explicit AI and IoT application. SP010
CP019 Flexiv defines its main wedge as adaptive robots that fuse industrial-grade force control with advanced artificial intelligence. SP011
CP020 Franka describes itself as a research-driven robotics company building a reference platform for robotics and AI professionals. SP012
CP021 Franka’s Diana 7 page says the arm is easy to set up and program and is positioned for sensitive, versatile operation. SP013
CP022 Dobot markets collaborative robots from 3 kg to 20 kg payload as well as desktop robots and a humanoid line. SP014
CP023 Dobot claims it serves more than 100 countries and regions, has sold more than 100,000 robots, and works with more than 350 global partners. SP014
CP024 IFR’s World Robotics product includes industrial-robot installation and operational-stock data through 2024 plus forecasts for 2025 to 2028. SP015
CP025 VDMA says the German robotics and automation sector had a 2025 revenue forecast of €13.8 billion and more than 400 members. SP016
CP026 VDMA also says the German robotics and automation sector is expected to see a 5 percent revenue decline in 2026 to €14.1 billion. SP016
CP027 Mordor Intelligence estimates the industrial robotics market at $54.28 billion in 2026 and $94.38 billion by 2031. SP017
CP028 Grand View Research, MarketsandMarkets, Verified Market Research, and Future Market Insights all publish large-growth industrial robotics forecasts, but their methodologies and ranges differ materially. SP018, SP019, SP020, SP021
CP029 Robotics & Automation News says 4.28 million industrial robots were operating globally in 2025 and annual installations topped half a million for the third straight year. SP023
CP030 Robotics & Automation News says Asia accounted for 70 percent of new deployments and China alone represented 51 percent of global installations. SP023
CP031 Business Wire summarized 2025 industrial-robot research around a shift from automation toward autonomy and AI-driven robotics. SP022
CP032 Standard Bots’ 2026 pricing guide says standard cobots commonly price from $25,000 to $60,000 before integration add-ons, with advanced setups above $90,000. SP025
CP033 The same pricing guide places Universal Robots around $30,000 to $60,000, ABB around $40,000 to $75,000, and FANUC from $35,000 to more than $90,000 depending on configuration. SP025
CP034 Standard Bots’ manufacturer guide says FANUC has more than 1 million robots installed, ABB more than 400,000, and Universal Robots more than 75,000 cobots deployed in more than 50 countries. SP024
CP035 The same guide says Chinese brands such as Dobot compete with budget automation in roughly the $15,000 to $40,000 range. SP024
CP036 Teradyne’s 2025 reporting groups Universal Robots within a Robotics Group and says that group achieved three consecutive quarters of growth in 2025. SP026, SP027
CP037 Incumbents such as ABB, FANUC, KUKA, and UR pair hardware with software, service, simulation, and training surfaces that increase switching cost after deployment. SP003, SP005, SP007, SP009
CP038 Agile’s most credible competitive wedge appears to be dexterous force-control and AI-led manipulation rather than sheer install base or public price leadership. SP001, SP002, SP011, SP012, SP013
CP039 Realized pricing, reseller density, and discounting remain only partially visible in public sources even though they heavily influence buying decisions. SP004, SP008, SP025
CP040 The competitive field is crowded enough that Agile cannot assume a clean market category; buyers can choose broad incumbents, adaptive-force specialists, research-first platforms, or low-cost Chinese exporters. SP003, SP005, SP009, SP011, SP012, SP014, SP023
CI001 Agile Robots was founded in 2018 as a spin-off of the German Aerospace Center (DLR). SI001, SI017
CI002 Agile Robots announced an 8-figure financing round before its pre-A financing. SI002
CI003 Agile Robots said its pre-A round was jointly backed by Hillhouse Capital, Sequoia Capital, Tinavi, and Linear Venture. SI003
CI004 Agile Robots said it had raised more than US$130 million in 2020 before completing Series C. SI004
CI005 Agile Robots said the Series C financing totaled US$220 million. SI004
CI006 Agile Robots said Series C made it the only intelligent-robotics unicorn globally with valuation above US$1 billion at that time. SI004, SI017
CI007 Agile Robots converted from a German AG to a European SE in March 2024 while remaining headquartered in Munich. SI005, SI011
CI008 Agile Robots acquired Franka Emika in November 2023 and did not disclose financial terms. SI006, SI029
CI009 The Franka acquisition was a distressed transaction that preserved roughly 100 employees and ongoing operations after insolvency proceedings. SI029
CI010 Agile Robots acquired a majority share in BÄR Automation in September 2023. SI007
CI011 The 2023 idealworks investment added AnyFleet, iw.hub, and iw.os to Agile Robots' orbit of automation products. SI008, SI021
CI012 BMW said more than 600 iw.hubs were already running in BMW intralogistics workflows at the time of the 2023 idealworks financing. SI021, SI008
CI013 After acquiring the remaining idealworks shares in 2025, Agile Robots said more than 850 iw.hubs were operating across BMW Group production sites. SI021, SI009
CI014 Agile Robots agreed in November 2025 to acquire thyssenkrupp Automation Engineering assets in Europe and North America. SI009, SI022
CI015 The thyssenkrupp Automation Engineering transaction closed in April 2026 and the business now operates as Krause Automation within Agile Robots Group. SI010, SI023
CI016 Public 2025-2026 sources say the thyssenkrupp deal added roughly 650 experts and around ten new locations to Agile Robots. SI018, SI019
CI017 Public coverage says the thyssenkrupp deal broadened Agile Robots into consumer electronics, medical technology, and logistics. SI018, SI019, SI022
CI018 Manufacturing Dive reported that Agile Robots had doubled revenue year over year to about €200 million in 2024, citing company information. SI019
CI019 Agile Robots was publicly described with more than 1,700 employees in March 2024, more than 2,300 employees in 2026, and more than 2,500 people in late-2025/2026 growth coverage. SI015, SI016, SI018, SI019, SI027
CI020 Independent 2026 logistics coverage says Agile Robots has implemented more than 10,000 automation solutions for global industrial customers. SI027
CI021 Independent 2026 coverage says Agile Robots invests more than €80 million annually in research and development in Germany. SI027
CI022 Public sources describe production operations in Europe, China, and India alongside Munich headquarters and Bavarian manufacturing. SI019, SI027, SI028
CI023 Amazon Business quotes Agile Robots procurement leadership saying electronics categories have seen lower costs and better availability and lead times. SI016
CI024 Agile Robots said it will start using Deutsche Telekom and NVIDIA's Industrial AI Cloud from 2026 to train foundation models on real production data. SI012, SI013
CI025 Agile Robots said its foundation-model program relies on storing, processing, and scaling very large datasets on purpose-built cloud infrastructure. SI013
CI026 Agile Robots said early access to NVIDIA Cosmos 3 expands simulation infrastructure for generalized robot intelligence. SI014
CI027 Agile Robots restarted Franka robot production in Kaufbeuren in March 2024 and said systems were shipping to customers again from the Bavarian site. SI015
CI028 Munich Startup reported that Agile ONE series production was due to begin at Agile Robots' own Bavarian plant in early 2026. SI018
CI029 Agile Robots monetizes through robot hardware, automation software, AMR ecosystems, and turnkey plant-integration capability gathered through acquisitions. SI001, SI008, SI009, SI010, SI021, SI022
CI030 No purchase price was publicly disclosed for the Franka, idealworks, or thyssenkrupp Automation Engineering transactions reviewed for this chapter. SI006, SI021, SI025, SI029
CI031 No public cash-on-hand, burn-rate, or runway figures were found in the reviewed sources. SI001, SI011, SI017, SI019, SI026
CI032 No public gross-margin, service-delivery-cost, or working-capital disclosures were found in the reviewed sources. SI001, SI011, SI017, SI019, SI026
CI033 No public list pricing or realized-pricing disclosures were found for Diana, Franka, idealworks, AgileCore, or turnkey integration projects. SI001, SI008, SI009, SI010, SI015
CI034 thyssenkrupp, ARQIS, and Taylor Wessing all frame Automation Engineering as a business with long-standing customer relationships and turnkey plant expertise, supporting a project-based revenue component after the acquisition. SI022, SI024, SI025
CI035 Register-excerpt vendors show the SE entity as active in Munich and the predecessor AG trace as inactive, which helps legal-entity continuity checks but still does not provide operating accounts. SI005, SI026, SI031
CI036 thyssenkrupp publicly described Agile Robots as having the financial strength needed to sustainably develop Automation Engineering under new ownership. SI022
CI037 The Franka transaction shows Agile Robots is willing to absorb distressed assets and their continuity obligations as part of its expansion strategy. SI006, SI029
CI038 The chapter's financial verdict is that public evidence supports meaningful top-line scale and industrial investment, but revenue quality and capital adequacy remain underwritten mostly by company statements rather than audited disclosures. SI018, SI019, SI022, SI027
CI039 Agile Robots said Agile ONE ceremonially switched on Europe's first industrial AI cloud in Munich, underscoring that its humanoid roadmap depends on dedicated compute infrastructure. SI030, SI012
CE001 Agile Robots describes itself as a Physical AI company delivering automation solutions for automotive, consumer electronics, healthcare, and service industries. SE001, SE002
CE002 Agile says it designs and develops robotic arms, hands, mobile platforms, and AgileCore software in-house across seven production sites and two R&D centers. SE003
CE003 Agile says customers already use more than 20,000 of its robotic solutions worldwide. SE003, SE012, SE024
CE004 Agile’s public portfolio includes Diana 7, Agile ONE, Agile Hand, AgileCore, the Thor series, Yu 5 Industrial, and mobile-robotics solutions. SE004, SE005, SE006, SE007, SE008, SE009, SE010, SE011
CE005 Agile says Diana 7 is a force-controlled robot arm with torque sensors in all seven axes and AI-driven software. SE005
CE006 Agile says Diana 7 offers 7 kg payload, 923 mm reach, ±0.05 mm repeatability, and 0.5 N force-control accuracy. SE005
CE007 Franka says Diana 7 is backed by AgileCore and is easy to set up, integrate, program, maintain, and repair. SE025
CE008 Agile says Diana 7 now supports the Franka Control Interface, enabling direct access to robot control algorithms. SE016, SE035
CE009 Agile says FCI allows 1 kHz real-time commands for position, velocity, and torque in joint space and position, velocity, and force in Cartesian space. SE016, SE035
CE010 Agile says FCI connects Diana 7 to libfranka, ROS, ROS 2, MATLAB, and Simulink ecosystems. SE016, SE035
CE011 The libfranka repository describes itself as a low-level C++ client library for real-time control of Franka research robots and states that the robot must have the FCI feature installed. SE031
CE012 The franka_ros2 repository says it provides ROS 2 Humble integration for Franka research robots and is in rapid development with expected breaking changes. SE032
CE013 Agile Hand is described as a five-finger anthropomorphic hand with 21 joints, 15 or 16 degrees of freedom, 10 N active fingertip force, and 1 kHz communication. SE007
CE014 Agile Hand exposes C++ and Python APIs and is compatible with ROS according to the official product page. SE007
CE015 AgileCore is described as a software platform for operators and system integrators built around an AI assistant that uses robotic learning methods plus LLM and VLM models. SE008
CE016 Agile says AgileCore supports natural-language programming, RAG-based integration guidance, and self-optimizing robot behaviors. SE008
CE017 The Thor series officially spans 3 kg, 7 kg, 12 kg, and 20 kg payload options with reaches up to 1700 mm. SE009
CE018 Agile says Thor 7 Pro adds joint torque sensors, relative force detection accuracy of 0.5 N, and hand-guiding capabilities. SE009
CE019 Agile says Yu 5 Industrial combines an integrated camera and NPU with teaching by demonstration, AgileTags localization, and force sensing. SE010
CE020 Agile says Yu 5 Industrial aligns with ISO 10218-1 and ISO 13849, while force monitoring supports collaborative applications under ISO/TS 15066. SE010
CE021 Agile says its mobile-robotics offer is built with idealworks and BÄR Automation and covers AGV and AMR hardware, software, and custom mobile manipulators. SE011
CE022 BMW says idealworks contributes AnyFleet, iw.hub, and iw.os while Agile contributes robotics and AI expertise for joint industrial-automation expansion. SE033
CE023 Agile and Google DeepMind say the partnership integrates Gemini Robotics foundation models with Agile hardware for adaptable industrial robots. SE012, SE028, SE029, SE030
CE024 TechCrunch and CNBC both say the DeepMind collaboration targets industrial use cases such as electronics manufacturing, automotive, data centers, and logistics. SE028, SE029
CE025 Agile says it is an anchor customer of Deutsche Telekom and NVIDIA’s Industrial AI Cloud and will train foundation models on European infrastructure. SE013, SE024
CE026 Agile says its foundation models use real industrial data, synthetic simulation data, and human data from data farms. SE013, SE014, SE022
CE027 Agile says it has collaborated with NVIDIA for years and is testing Cosmos 3 for simulation, data curation, generation, and evaluation. SE013, SE015
CE028 NVIDIA’s 2026 physical AI release lists Agile Robots among humanoid builders using Cosmos world models, Isaac Sim, and Isaac Lab, while ABB, FANUC, KUKA, and Universal Robots use NVIDIA tools for industrial deployment and digital twins. SE027
CE029 Agile says Kaufbeuren production restarted shipping for Franka robots and Yu 5 Industrial in Bavaria after the Franka acquisition. SE017
CE030 Agile says the torque sensors in Franka and Yu 5 axes originated at DLR and that the systems continue to be developed by Agile and Franka R&D teams in Munich. SE017
CE031 Agile’s ISO 9001 certification covers the company’s management system for developing and producing robotic systems. SE019
CE032 Agile says its Safety Core for Robot Applications was assessed to IEC 61508 and includes power and force limiting plus safe collision detection. SE021
CE033 Agile says Yu 5 Industrial received TÜV SÜD certification covering IEC 61508, ISO 13849-1, and ISO 10218-1. SE020
CE034 DLR says Agile Justin has 53 degrees of freedom, torque sensors in all joints, 1 kHz whole-body control, tactile skin, and access to external GPU and cloud compute. SE026
CE035 Agile ONE is presented as a humanoid for manufacturing and logistics with cameras, LiDAR, speech recognition, proximity sensing, 21-joint hands, and 2 m/s walking speed. SE006
CE036 Agile says Agile ONE uses a layered AI system and trains on real-world industrial, simulated, and teleoperation data. SE022, SE023, SE024
CE037 Agile says Agile ONE is not meant to stand alone but to work inside a broader production ecosystem linked through AgileCore and other robotic solutions. SE006, SE023
CE038 Agile’s public product proof is strongest on technical specifications, certifications, and ecosystem announcements rather than public field reliability or uptime metrics. SE005, SE010, SE019, SE020, SE021, SE023
CE039 Agile’s AI-led stack increases dependence on proprietary data, partner compute, and external ecosystems such as NVIDIA and Google DeepMind. SE012, SE013, SE014, SE015, SE027
CE040 Public developer signal is concentrated in the Franka research stack through libfranka and franka_ros2 rather than in a large Agile-owned public repository surface. SE016, SE031, SE032
CU001 Agile Robots publicly targets automation customers across manufacturing, logistics, warehousing, service, and research environments rather than a single vertical. SU001, SU002
CU002 Diana 7, Yu 5 Industrial, Thor, AgileCore, and Agile ONE collectively position the company for precision assembly, collaborative robotics, intralogistics, and humanoid use cases. SU003, SU004, SU005, SU006, SU007
CU003 idealworks contributes AnyFleet, iw.sim, and iw.hub to Agile Robots' customer-facing logistics portfolio. SU008, SU009
CU004 BMW said more than 600 iw.hubs were operating in BMW intralogistics workflows when Agile first joined idealworks in 2023. SU008, SU018
CU005 After the 2025 full acquisition, Agile Robots said more than 850 iw.hubs were operating across BMW Group production sites worldwide. SU009, SU022, SU023
CU006 idealworks' BMW customer story says close to 600 iw.hubs run roughly 30,000 missions per day at BMW plants with 98% availability. SU030
CU007 idealworks says BMW deployments already span Regensburg, Dingolfing, Munich, and Spartanburg, with further deployments planned for Steyr, Hams Hall, and Leipzig. SU030
CU008 idealworks' customer stories page names non-automotive users including MoldTecs and says the Sonneberg site runs eight iw.hubs for roughly 1,700 missions every day. SU029
CU009 Amazon Business presents Agile Robots as a live customer reference and quotes lower costs plus better availability and lead times in electronics procurement. SU017
CU010 Agile Robots restarted Franka production in Kaufbeuren in March 2024 and said customer shipments resumed from the Bavarian site. SU013, SU020
CU011 Agile Robots says Franka robots remain widely used by MIT, Stanford, ETH Zurich, Max Planck, and NVIDIA. SU013, SU020
CU012 Diana 7 gained Franka Control Interface support in 2025, enabling 1 kHz real-time control and compatibility with ROS, ROS 2, MATLAB, and Simulink for research users. SU012
CU013 Franka Research 3 is marketed as a force-sensitive reference platform for AI and robotics with 3 kg payload, 855 mm reach, 94.5% workspace efficiency, 7 degrees of freedom, and integrated torque sensors. SU031
CU014 Franka's community page shows an active ecosystem of community integrations and control tools, which is a developer-signal for installed-base durability in research. SU032
CU015 Agile Robots says it has already installed more than 20,000 robotics solutions worldwide. SU010
CU016 The Google DeepMind partnership says the first target use cases are high-value industrial and manufacturing tasks where reliability and scale matter. SU010, SU026
CU017 Manufacturing-side use cases discussed publicly for the DeepMind collaboration include complex assembly and cable routing. SU026
CU018 Agile ONE is marketed as an autonomous co-worker for manufacturing, logistics, and service environments that moves between workstations and collaborates with humans. SU004, SU015, SU016
CU019 ARENA2036 membership gives Agile Robots access to automotive research networks and wiring-harness transformation projects that can seed future customer relationships. SU011, SU021
CU020 BÄR Automation broadens customer access in system integration and industrial automation beyond Agile's core robot hardware. SU027
CU021 The thyssenkrupp / Krause acquisition adds close partnerships with leading OEMs and new access to logistics and industrial customer bases. SU028
CU022 BMW remains a long-term partner of idealworks even after Agile Robots acquired the remaining shares in 2025. SU009, SU022, SU023
CU023 BMW founded idealworks in 2020 specifically to commercialize robotics and management software for logistics beyond internal use. SU018, SU033
CU024 The clearest named scaled production customer proof in the public record is BMW / idealworks; most Agile-branded customer references are either sectoral or research-oriented rather than fully named end customers. SU009, SU013, SU017, SU029, SU030
CU025 No public customer-count disclosure for Agile Robots itself was found in the reviewed sources. SU001, SU002, SU010
CU026 No public NRR, GRR, logo churn, or cohort retention metrics were found in the reviewed sources. SU001, SU009, SU017, SU029
CU027 No public contract-length or renewal-rate disclosure was found for BMW, idealworks external customers, or Agile-branded deployments. SU009, SU018, SU029, SU030
CU028 Customer expansion is currently evidenced more through acquired channels and ecosystems — idealworks, Franka, BÄR, and Krause — than through disclosed direct-sales metrics. SU009, SU013, SU027, SU028
CU029 BMW concentration risk is elevated because BMW is both the most specific scaled production user and the foundational reference behind idealworks' external credibility. SU018, SU022, SU030, SU033
CU030 Non-automotive expansion is nevertheless visible: idealworks cites customers and markets in manufacturing, logistics, and warehousing outside BMW. SU009, SU022, SU023, SU029
CU031 Research and academia are a distinct customer segment for Franka and now Diana 7, not merely a marketing adjacency. SU012, SU013, SU019, SU031
CU032 The Amazon Business reference is customer proof of procurement efficiency rather than proof of revenue from Agile Robots products. SU017
CU033 Franka's 2023 insolvency created a continuity risk for customers that Agile had to address by restarting production and preserving the team. SU013, SU024, SU025
CU034 Automationspraxis reported a rumored Franka purchase price above €30 million, while official sources kept financial terms undisclosed. SU024, SU025
CU035 The idealworks 2023 release said its first successfully completed customer projects in the USA were already in place and Asia-Pacific expansion was due the following year. SU018
CU036 Public adoption evidence is strongest at the deployment-milestone level — missions, plant rollouts, restarted shipments, and named research users — but weak on denominator metrics such as pilot-to-production conversion. SU009, SU013, SU017, SU029, SU030
CU037 Durability proxies come from BMW's continued partnership, recurring missions at BMW plants, restarted Franka shipments, and an active research community rather than disclosed renewal data. SU009, SU013, SU030, SU032
CU038 Named customer proof is fresh enough for 2026 because the reviewed record includes 2025-2026 idealworks, DeepMind, and Agile ONE disclosures plus 2024 Franka shipping evidence still relevant to current continuity. SU009, SU010, SU013, SU022, SU023, SU026
CU039 The customer-chapter verdict is that Agile Robots has credible deployment proof, especially through idealworks and Franka, but public retention and concentration data remain too thin to underwrite durable revenue quality. SU017, SU022, SU029, SU030
CR001 Franka Emika's creditors' committee approved Agile Robots' purchase agreement before the transaction closed. SR001, SR019
CR002 The Munich insolvency court opened Franka Emika's insolvency proceedings on 2023-11-01. SR019
CR003 Agile Robots said it would continue Franka Emika's operations with roughly 100 employees in Bavaria. SR001, SR019, SR033
CR004 Munich Startup reported that Munich I prosecutors were investigating Franka Emika over alleged subsidy fraud when Agile announced the takeover. SR031
CR005 Automationspraxis reported that rival parties raised Germany-China control concerns around the Franka sale process. SR020
CR006 A complaint cited by Automationspraxis argued that Franka technology could effectively be transferred to China through Agile's ownership links. SR020
CR007 Agile restarted Franka robot production in Kaufbeuren in March 2024 after the acquisition. SR011
CR008 Agile said Franka systems were shipping again once local production resumed in Kaufbeuren. SR011
CR009 Agile described Franka's robotic arms as designed and developed in Germany. SR011
CR010 Agile's management system is ISO 9001 certified for developing and producing robotic systems. SR012
CR011 Agile's Yu 5 Industrial robot received TÜV SÜD certification against IEC 61508, ISO 13849-1, and ISO 10218-1 related norms. SR013
CR012 Agile's Safety Core for Robot Applications received a TÜV SÜD certificate tied to IEC 61508 functional-safety development. SR014
CR013 Agile announced the thyssenkrupp Automation Engineering asset acquisition in November 2025. SR002, SR015
CR014 Agile said the thyssenkrupp transaction would add about 650 experts and 10 new locations. SR002, SR026
CR015 Agile said thyssenkrupp Automation Engineering generated 2024 revenue in the hundreds of millions of euros. SR002, SR022
CR016 Agile said its own revenue reached around €200 million in 2024 after doubling annually since founding. SR002, SR021
CR017 Agile closed the thyssenkrupp asset acquisition on 2026-04-01 and took over assets in Europe and North America. SR003, SR016
CR018 thyssenkrupp Automation Engineering now operates as Krause Automation within the Agile Robots Group. SR003, SR017
CR019 Agile said the thyssenkrupp deal expands the company beyond automotive into logistics, electronics, and medical technology. SR002, SR015
CR020 thyssenkrupp described the divestiture as part of sharpening its portfolio toward capital-market-ready businesses. SR015, SR016
CR021 ARQIS said it advised Agile on all legal aspects of the thyssenkrupp Automation Engineering acquisition. SR017
CR022 Taylor Wessing staffed corporate and regulatory lawyers on the thyssenkrupp sale mandate. SR018
CR023 Agile became idealworks' majority shareholder through a Series A financing round in July 2025. SR004, SR023
CR024 Agile acquired the remaining idealworks shares in September 2025. SR005
CR025 idealworks said more than 850 iw.hubs were operating across BMW Group production sites when Agile bought the remaining shares. SR005
CR026 Agile and Google DeepMind announced a strategic research partnership in March 2026. SR006, SR024
CR027 The DeepMind partnership combines Gemini Robotics foundation models with Agile's industrial robotics hardware stack. SR006, SR025
CR028 CNBC reported that Agile already had more than 20,000 deployed robotic systems globally when the DeepMind partnership was announced. SR024
CR029 Agile became an anchor customer of Deutsche Telekom and NVIDIA's Industrial AI Cloud starting in 2026. SR007
CR030 Agile said it will train foundation models on European cloud infrastructure using real production data. SR007, SR008
CR031 Agile described the Industrial AI Cloud as infrastructure built with Deutsche Telekom and NVIDIA. SR007, SR008
CR032 Agile publicized early access to NVIDIA Cosmos 3 as part of its simulation and data-generation workflow. SR009
CR033 Agile ONE was presented as a factory-floor humanoid whose learning loop depends on the company's broader AI, cloud, and partner ecosystem. SR010, SR006, SR007
CR034 Agile said it had over 2,500 employees from around 60 countries when the thyssenkrupp deal was announced. SR002, SR026
CR035 Agile's about-us materials describe production sites in Europe, China, and India and more than 15 sites globally. SR027
CR036 Manufacturing Dive also described Agile as operating production in China, India, and Germany. SR021
CR037 Logistik Heute reported that Agile invests more than €80 million per year in research and development in Germany. SR027
CR038 BMW said idealworks collaborates with NVIDIA, Microsoft, and ADLINK, adding partner-management complexity to the investment case. SR023
CR039 The EU dual-use regime controls export, transit, brokering, and technical assistance for dual-use items and can require authorizations. SR028, SR035
CR040 BIS maintains the EAR as the U.S. framework for classifying and licensing dual-use exports. SR030
CR041 The EU AI Act imposes a risk-based regime with special obligations for high-risk AI systems and bans certain practices. SR029, SR032, SR034
CR042 The European Commission says the AI Act's prohibited practices have applied since February 2025. SR032, SR029
CR043 A robotics company that combines workplace automation, industrial data collection, and foundation-model training faces a rising compliance burden under export-control and AI-governance rules. SR028, SR029, SR030, SR032
CR044 Agile's 2026 risk profile is dominated by simultaneous integration, compliance, partner-dependency, and capital-intensity demands rather than by a single isolated failure mode. SR002, SR006, SR007, SR021, SR027
CV001 Agile announced an initial 8-figure funding round in 2018. SV001
CV002 Agile completed a pre-A round in 2019 backed by Hillhouse, Sequoia, Tinavi, and Linear Venture. SV002
CV003 Agile completed a Series C financing round in 2021 led by SoftBank Vision Fund 2. SV003
CV004 Agile said its 2021 Series C made it the first German robotics unicorn with valuation above US$1 billion. SV003, SV011
CV005 Agile said it raised more than US$130 million in total during 2020. SV003
CV006 TechCrunch reported that Agile had raised more than US$270 million in venture funding by March 2026. SV016
CV007 Munich Startup listed a last investment in April 2022 and a US$30 million Series C follow-on round for Agile. SV011
CV008 Agile converted from an AG into an SE in March 2024 to simplify cross-border European operations. SV004
CV009 Agile said the SE conversion followed growth to more than 1,700 employees worldwide, including about 600 in Germany. SV004
CV010 Agile said it invests more than €80 million a year in research and development in Germany. SV005, SV013
CV011 Agile operates manufacturing facilities in Europe, China, and India. SV005, SV010
CV012 Agile said Agile ONE will enter full production in Bavaria in early 2026. SV006
CV013 Agile said it manufactures robotics hardware in-house rather than outsourcing final production. SV006
CV014 Agile said revenue had doubled year on year since founding and reached roughly €200 million by the time it launched Agile ONE. SV006, SV015
CV015 Agile said the Industrial AI Cloud will be used to train its foundation models on infrastructure built by Deutsche Telekom and NVIDIA in Munich. SV007
CV016 TechCrunch and CNBC described the DeepMind partnership as a long-term effort to deploy Gemini Robotics models on Agile's industrial robots at scale. SV016, SV017, SV031
CV017 CNBC reported that Agile had already deployed more than 20,000 robotic systems globally by March 2026. SV017
CV018 ARENA2036 said Agile had more than 2,300 employees worldwide and had already acquired a majority stake in BÄR Automation. SV018
CV019 Amazon Business also described Agile as employing more than 2,300 experts in Munich-focused operations. SV019
CV020 Munich Startup said Agile remained the first robotics unicorn worldwide after the 2021 Series C. SV011
CV021 Munich Startup's 2025 thyssenkrupp article said the deal added roughly 650 experts and 10 new locations. SV012
CV022 VDMA forecast 2025 revenue of about €13.8 billion for Germany's robotics and automation association members. SV021
CV023 Mordor Intelligence estimated the industrial robotics market at US$54.28 billion in 2026 with US$94.38 billion by 2031 at an 11.7% CAGR. SV022
CV024 MarketsandMarkets described the industrial robots market as a US$20.8 billion opportunity by 2032 growing 5.0% CAGR from 2026. SV023
CV025 IFR's World Robotics report remains one of the sector's main benchmark datasets for installations, stock, applications, and industries. SV020, SV033
CV026 BusinessWire's ResearchAndMarkets summary said robotics deal value reached US$7.3 billion in H1 2025. SV030
CV027 The same ResearchAndMarkets summary said SMEs still face cost and skills barriers despite strong robotics investment momentum. SV030
CV028 ABB markets both articulated and collaborative robots across a broad application set, highlighting the breadth of incumbent industrial competition. SV024
CV029 KUKA markets robot systems ranging from fenced high-speed cells to direct human-robot collaboration and mobile solutions. SV025
CV030 Universal Robots launched the UR18 in October 2025, showing that cobot competitors continue to push payload and speed improvements. SV026
CV031 FANUC maintains integrated report archives for investors, signalling deep public reporting history among incumbent robotics peers. SV027, SV034
CV032 Teradyne's annual reports page shows a 2026 10-K and annual report to shareholders, providing public filing history for Universal Robots' parent. SV028, SV029, SV035
CV033 Forbes argued that purpose-built automation may deliver better manufacturing ROI than headline-driven humanoid deployments. SV032
CV034 Forbes framed 2026 factory budgeting as a choice between humanoid hype and automation systems that can already prove ROI. SV032
CV035 Agile's public evidence supports a scenario-based valuation approach because no current post-2021 private-market mark is disclosed in fetched sources. SV003, SV011, SV016
CV036 A 4x-6x revenue range on the last public €200 million revenue datapoint implies an indicative enterprise value range of roughly €800 million to €1.2 billion. SV014, SV015, SV022, SV023
CV037 A bull 6x-8x scenario implies roughly €1.2 billion to €1.6 billion if AI optionality converts into durable industrial demand and margins. SV015, SV016, SV022, SV023
CV038 A bear 3x-4x scenario implies roughly €600 million to €800 million if integration, margins, or humanoid economics disappoint. SV015, SV022, SV030, SV032
CV039 The public evidence supports a track or research-more posture rather than a buy call because current valuation, margins, burn, and preference stack are undisclosed. SV011, SV016, SV029
CV040 At any entry price that assumes a material premium to the last disclosed unicorn benchmark, the evidence set looks fair-to-stretched rather than clearly attractive. SV003, SV011, SV014, SV032
CV041 DeepMind, NVIDIA, and Munich AI-cloud momentum expand upside optionality, but they do not yet replace the need for proof on margins and repeatable ROI. SV007, SV016, SV017, SV032
CV042 Public comparables are imperfect because incumbents such as ABB, KUKA, FANUC, Teradyne, and Universal Robots differ in diversification, public float, and product mix. SV024, SV025, SV026, SV027, SV029
CV043 Agile's capital intensity is likely to remain elevated because it is simultaneously funding in-house hardware, humanoid production, acquisitions, and foundation-model training. SV005, SV006, SV007, SV012
CV044 The most credible upside case is a large industrial-platform outcome, but the most credible downside case is multiple compression into a conventional automation supplier profile. SV022, SV023, SV030, SV032
来源
编号出版方标题引文
SO001 Agile Robots SE We are Agile Robots | Driving industries forward
SO002 Agile Robots SE About Agile Robots | How we transform industries
SO003 Agile Robots SE Create the future of robotics
SO004 Agile Robots SE Agile Robots announces 8-figure funding
SO005 Agile Robots SE Agile Robots AG successfully completes pre-A round financing
SO006 Agile Robots SE Agile Robots AG completes Series C financing
SO007 Agile Robots SE Agile Robots becomes an SE
SO008 Agile Robots SE Smart Robotics “Made in Bavaria”: Agile Robots opens global headquarters in Munich
SO009 Agile Robots SE Agile Robots AG acquires robotics specialist Franka Emika
SO010 Agile Robots SE Agile Robots acquires majority share in BÄR Automation
SO011 Agile Robots SE Agile Robots AG becomes the majority shareholder of idealworks, a BMW Group subsidiary
SO012 Agile Robots SE Agile Robots completes acquisition of BMW Group spin-off idealworks
SO013 Agile Robots SE Agile Robots acquires thyssenkrupp Auto­mation Engineering: Physical AI as a catalyst for Industrial Transformation
SO014 Agile Robots SE Agile Robots closes acquisition of thyssenkrupp Automation Engineering
SO015 Agile Robots SE Agile Robots and Google DeepMind partner to bring intelligence to robotics
SO016 Agile Robots SE Agile Robots becomes a member of ARENA2036
SO017 Agile Robots SE Agile Robots AG receives ISO 9001 certification
SO018 Munich Startup Agile Robots AG - Munich Startup
SO019 Munich Startup Agile Robots: Acquisition and Humanoid Drive Next Growth Spurt - Munich Startup
SO020 Manufacturing Dive Agile Robots to tap into new sectors with latest acquisition
SO021 BMW Group idealworks: BMW Group subsidiary gains Agile Robots AG as majority shareholder
SO022 thyssenkrupp AG thyssenkrupp to sell Automation Engineering to Agile Robots
SO023 thyssenkrupp AG thyssenkrupp successfully completes sale of Automation Engineering to Agile Robots
SO024 Wir sind Kaufbeuren Agile Robots SE
SO025 Automationspraxis Starkes Signal für den Robotik-Standort Deutschland: Agile Robots kauft Franka Emika
SO026 Verlag INDat Robotik Startup Franka Emika: Zukunft des Unternehmens durch Verkauf an Agile Robots gesichert - Verlag INDat
SO027 ARQIS ARQIS advises Agile Robots SE on the acquisition of thyssenkrupp Automation Engineering - ARQIS
SO028 CNBC Google partners with Agile Robots, growing its AI robotics footprint
SM001 International Federation of Robotics International Federation of Robotics
SM002 VDMA Robotik + Automation - vdma.org - VDMA
SM003 Mordor Intelligence Industrial Robotics Market Size, Analysis, Share & Growth Trends 2031
SM004 MarketsandMarkets Industrial Robotics Market Size, Share and Growth
SM005 Verified Market Research Industrial Robotics Market Report: Size, Growth, Trends & Forecast (2025–2033)
SM006 Future Market Insights Industrial Robotics Market Size, Trends & Forecast 2026-2036
SM007 Business Wire / ResearchAndMarkets.com Industrial Robots Research Report 2025: Moving from Automation to Autonomy - Humanoid, Collaborative & AI-Driven Robotics Reshape Manufacturing as BMW, Mercedes-Benz & Tesla Pilot Factory Deployments - ResearchAndMarkets.com
SM008 Agile Robots SE Solutions at Agile Robots | Cutting-edge automation
SM009 Agile Robots SE Diana 7 – the dexterous robot | Agile Robots
SM010 Agile Robots SE The next evolution in physical AI: Agile ONE
SM011 Agile Robots SE Agile Hand – the human-like robotic hand | Agile Robots
SM012 Agile Robots SE AgileCore – the AI-powered software platform | Agile Robots
SM013 Agile Robots SE Thor series
SM014 Agile Robots SE Yu 5 Industrial – the cobot with smart vision | Agile Robots
SM015 Agile Robots SE Mobile robotics – shared expertise | Agile Robots
SM016 Agile Robots SE Agile Robots launches humanoid robot for industry: Agile ONE
SM017 Agile Robots SE Humanoid Agile ONE embodies Physical AI at Hannover Messe 2026
SM018 Agile Robots SE Agile ONE powers up Europe’s largest AI factory
SM019 Agile Robots SE Agile Robots is an anchor-customer for Europe's first Industrial AI Cloud
SM020 Agile Robots SE Industrial AI Cloud – The cloud infrastructure behind our Foundation Model
SM021 Amazon Business Agile Robots - Customer Stories | Amazon Business
SM022 Franka Robotics Diana 7
SM023 The Robot Report Agile Robots to deploy Google DeepMind foundation models on its humanoid
SM024 ARENA2036 Agile Robots becomes a member of ARENA2036
SM025 CNBC Google partners with Agile Robots, growing its AI robotics footprint
SM026 TechCrunch Agile Robots becomes the latest robotics company to partner with Google DeepMind | TechCrunch
SP001 Agile Robots SE We are Agile Robots | Driving industries forward
SP002 Agile Robots SE Diana 7 – the dexterous robot | Agile Robots
SP003 Universal Robots Collaborative Robots & Cobots | Universal Robots
SP004 Universal Robots News Center - Universal Robots
SP005 ABB Robotics | ABB
SP006 ABB Annual Reporting Suite 2025 | ABB
SP007 KUKA Robot systems | KUKA Germany
SP008 KUKA Reports | KUKA Germany
SP009 FANUC FANUC Industrial Robots
SP010 FANUC Integrated Report 2025
SP011 Flexiv Home | Flexiv
SP012 Franka Robotics Homepage
SP013 Franka Robotics Diana 7
SP014 Dobot Robotics Dobot Robotics | Official website
SP015 International Federation of Robotics International Federation of Robotics
SP016 VDMA Robotik + Automation - vdma.org - VDMA
SP017 Mordor Intelligence Industrial Robotics Market Size, Analysis, Share & Growth Trends 2031
SP018 Grand View Research Industrial Robotics Market Size, Share | Industry Report, 2030
SP019 MarketsandMarkets Industrial Robotics Market Size, Share and Growth
SP020 Verified Market Research Industrial Robotics Market Report: Size, Growth, Trends & Forecast (2025–2033)
SP021 Future Market Insights Industrial Robotics Market Size, Trends & Forecast 2026-2036
SP022 Business Wire / ResearchAndMarkets Industrial Robots Research Report 2025: Moving from Automation to Autonomy - Humanoid, Collaborative & AI-Driven Robotics Reshape Manufacturing as BMW, Mercedes-Benz & Tesla Pilot Factory Deployments - ResearchAndMarkets.com
SP023 Robotics & Automation News Top 30 industrial robotics companies
SP024 Standard Bots Robot manufacturers’ guide 2026: Who builds the world’s best robots? - Standard Bots
SP025 Standard Bots Cobot price explained: 2026 guide to collaborative robot costs - Standard Bots
SP026 Teradyne Teradyne Reports Fourth Quarter and Full Year 2025 Results
SP027 Teradyne Untitled source
SI001 Agile Robots SE About Agile Robots | How we transform industries
SI002 Agile Robots SE Agile Robots announces 8-figure funding
SI003 Agile Robots SE Agile Robots AG successfully completes pre-A round financing
SI004 Agile Robots SE Agile Robots AG completes Series C financing
SI005 Agile Robots SE Agile Robots becomes an SE
SI006 Agile Robots SE Agile Robots AG acquires robotics specialist Franka Emika
SI007 Agile Robots SE Agile Robots acquires majority share in BÄR Automation
SI008 Agile Robots SE Agile Robots AG becomes the majority shareholder of idealworks, a BMW Group subsidiary
SI009 Agile Robots SE Agile Robots acquires thyssenkrupp Automation Engineering: Physical AI as a catalyst for Industrial Transformation
SI010 Agile Robots SE Agile Robots closes acquisition of thyssenkrupp Automation Engineering
SI011 Agile Robots SE Smart Robotics “Made in Bavaria”: Agile Robots opens global headquarters in Munich
SI012 Agile Robots SE Agile Robots is an anchor-customer for Europe's first Industrial AI Cloud
SI013 Agile Robots SE Industrial AI Cloud – The cloud infrastructure behind our Foundation Model
SI014 Agile Robots SE Simulating worlds: Agile Robots early access to NVIDIA Cosmos 3
SI015 Agile Robots SE Ready for shipment: Agile Robots ramps up production of Franka robots in Kaufbeuren
SI016 Amazon Business Agile Robots - Customer Stories | Amazon Business
SI017 Munich Startup Agile Robots AG - Munich Startup
SI018 Munich Startup Agile Robots: Acquisition and Humanoid Drive Next Growth Spurt
SI019 Manufacturing Dive Agile Robots to tap into new sectors with latest acquisition
SI020 The Robot Report Agile Robots acquires thyssenkrupp Automation Engineering
SI021 BMW Group PressClub idealworks: BMW Group subsidiary gains Agile Robots AG as majority shareholder
SI022 thyssenkrupp AG thyssenkrupp to sell Automation Engineering to Agile Robots
SI023 thyssenkrupp AG thyssenkrupp successfully completes sale of Automation Engineering to Agile Robots
SI024 ARQIS ARQIS advises Agile Robots SE on the acquisition of thyssenkrupp Automation Engineering
SI025 Taylor Wessing Taylor Wessing advises thyssenkrupp on the sale of Automation Engineering to Agile Robots
SI026 CompanyHouse Agile Robots AG, Gilching
SI027 Logistik Heute KI-gestützte Robotik: Agile Robots eröffnet Hauptsitz in München
SI028 MM Logistik Agile Robots eröffnet globales Hauptquartier in München
SI029 INDat Robotik Startup Franka Emika: Zukunft des Unternehmens durch Verkauf an Agile Robots gesichert
SI030 Agile Robots SE Agile ONE powers up Europe's largest AI factory
SI031 online-handelsregister.de Handelsregisterauszug von Agile Robots SE aus München (HRB 292066)
SE001 Agile Robots SE We are Agile Robots | Driving industries forward
SE002 Agile Robots SE About Agile Robots | How we transform industries
SE003 Agile Robots SE Innovation at Agile Robots | AI-driven automation
SE004 Agile Robots SE Solutions at Agile Robots | Cutting-edge automation
SE005 Agile Robots SE Diana 7 – the dexterous robot | Agile Robots
SE006 Agile Robots SE The next evolution in physical AI: Agile ONE
SE007 Agile Robots SE Agile Hand – the human-like robotic hand | Agile Robots
SE008 Agile Robots SE AgileCore – the AI-powered software platform | Agile Robots
SE009 Agile Robots SE Thor series
SE010 Agile Robots SE Yu 5 Industrial – the cobot with smart vision | Agile Robots
SE011 Agile Robots SE Mobile robotics – shared expertise | Agile Robots
SE012 Agile Robots SE Agile Robots and Google DeepMind partner to bring intelligence to robotics
SE013 Agile Robots SE Agile Robots is an anchor-customer for Europe's first Industrial AI Cloud
SE014 Agile Robots SE Industrial AI Cloud – The cloud infrastructure behind our Foundation Model
SE015 Agile Robots SE Simulating worlds: Agile Robots early access to NVIDIA Cosmos 3
SE016 Agile Robots SE Agile Robots released new features on Diana 7 to fully support Franka Control Interface
SE017 Agile Robots SE Ready for shipment: Agile Robots ramps up production of Franka robots in Kaufbeuren
SE018 Agile Robots SE Agile Robots and Franka Robotics at NVIDIA GTC 2025
SE019 Agile Robots SE Agile Robots AG receives ISO 9001 certification
SE020 Agile Robots SE Yu 5 Industrial from Agile Robots AG receives TÜV SÜD certification
SE021 Agile Robots SE Agile Robots AG receives TÜV-SÜD certificate
SE022 Agile Robots SE Agile Robots launches humanoid robot for industry: Agile ONE
SE023 Agile Robots SE Humanoid Agile ONE embodies Physical AI at Hannover Messe 2026
SE024 Agile Robots SE Agile ONE powers up Europe’s largest AI factory
SE025 Franka Robotics Diana 7
SE026 German Aerospace Center (DLR) Agile Justin
SE027 NVIDIA Newsroom NVIDIA and Global Robotics Leaders Take Physical AI to the Real World
SE028 TechCrunch Agile Robots becomes the latest robotics company to partner with Google DeepMind | TechCrunch
SE029 CNBC Google partners with Agile Robots, growing its AI robotics footprint
SE030 The Robot Report Agile Robots to deploy Google DeepMind foundation models on its humanoid
SE031 GitHub / Franka Robotics GitHub - frankarobotics/libfranka: C++ client library to control Franka robots in real-time
SE032 GitHub / Franka Robotics GitHub - frankarobotics/franka_ros2: ROS 2 integration for Franka research robots
SE033 BMW Group idealworks: BMW Group subsidiary gains Agile Robots AG as majority shareholder
SE034 ARENA2036 Agile Robots becomes a member of ARENA2036
SE035 PR Newswire Now compatible with FCI: Franka Robotics integrates Diana 7 from Agile Robots into its ecosystem
SU001 Agile Robots SE About Agile Robots | How we transform industries
SU002 Agile Robots SE Solutions at Agile Robots | Cutting-edge automation
SU003 Agile Robots SE Diana 7 – the dexterous robot
SU004 Agile Robots SE The next evolution in physical AI: Agile ONE
SU005 Agile Robots SE AgileCore – the AI-powered software platform
SU006 Agile Robots SE Thor series
SU007 Agile Robots SE Yu 5 Industrial – the cobot with smart vision
SU008 Agile Robots SE Agile Robots AG becomes the majority shareholder of idealworks, a BMW Group subsidiary
SU009 Agile Robots SE Agile Robots completes acquisition of BMW Group spin-off idealworks
SU010 Agile Robots SE Agile Robots and Google DeepMind partner to bring intelligence to robotics
SU011 Agile Robots SE Agile Robots becomes a member of ARENA2036
SU012 Agile Robots SE Agile Robots released new features on Diana 7 to fully support Franka Control Interface
SU013 Agile Robots SE Ready for shipment: Agile Robots ramps up production of Franka robots in Kaufbeuren
SU014 Agile Robots SE Agile Robots and Franka Robotics at NVIDIA GTC 2025
SU015 Agile Robots SE Agile Robots launches humanoid robot for industry: Agile ONE
SU016 Agile Robots SE Humanoid Agile ONE embodies Physical AI at Hannover Messe 2026
SU017 Amazon Business Agile Robots - Customer Stories | Amazon Business
SU018 BMW Group PressClub idealworks: BMW Group subsidiary gains Agile Robots AG as majority shareholder
SU019 Franka Robotics Diana 7
SU020 Franka Robotics Ready for shipment: Agile Robots ramps up production of Franka robots in Kaufbeuren
SU021 ARENA2036 Agile Robots becomes a member of ARENA2036
SU022 LOGISTRA AMR takeover: Agile Robots fully acquires BMW spin-off Idealworks.
SU023 Munich Startup Agile Robots acquires Idealworks
SU024 Automationspraxis Starkes Signal für den Robotik-Standort Deutschland: Agile Robots kauft Franka Emika
SU025 INDat Robotik Startup Franka Emika: Zukunft des Unternehmens durch Verkauf an Agile Robots gesichert
SU026 The Robot Report Agile Robots to deploy Google DeepMind foundation models on its humanoid
SU027 Agile Robots SE Agile Robots acquires majority share in BÄR Automation
SU028 Agile Robots SE Agile Robots closes acquisition of thyssenkrupp Automation Engineering
SU029 idealworks Customer Stories | Idealworks
SU030 idealworks Driving success at BMW Group: Idealworks AMRs revolutionize automotive manufacturing
SU031 Franka Robotics Franka Research 3
SU032 Franka Robotics Community Contributions
SU033 BMW Group PressClub BMW Group founds company: IDEALworks GmbH to develop and distribute innovative robots and management software for logistics solutions
SR001 Agile Robots Agile Robots AG acquires robotics specialist Franka Emika The creditors‘ committee of insolvent FRANKA EMIKA had previously approved the corresponding agreement.
SR002 Agile Robots Agile Robots acquires thyssenkrupp Automation Engineering: Physical AI as a catalyst for Industrial Transformation
SR003 Agile Robots Agile Robots closes acquisition of thyssenkrupp Automation Engineering
SR004 Agile Robots Agile Robots AG becomes the majority shareholder of idealworks, a BMW Group subsidiary
SR005 Agile Robots Agile Robots completes acquisition of BMW Group spin-off idealworks
SR006 Agile Robots Agile Robots and Google DeepMind partner to bring intelligence to robotics
SR007 Agile Robots Agile Robots is an anchor-customer for Europe's first Industrial AI Cloud
SR008 Agile Robots Industrial AI Cloud – The cloud infrastructure behind our Foundation Model
SR009 Agile Robots Simulating worlds: Agile Robots early access to NVIDIA Cosmos 3
SR010 Agile Robots Humanoid Agile ONE embodies Physical AI at Hannover Messe 2026
SR011 Agile Robots Ready for shipment: Agile Robots ramps up production of Franka robots in Kaufbeuren
SR012 Agile Robots Agile Robots AG receives ISO 9001 certification
SR013 Agile Robots Yu 5 Industrial from Agile Robots AG receives TÜV SÜD certification
SR014 Agile Robots Agile Robots AG receives TÜV-SÜD certificate
SR015 thyssenkrupp thyssenkrupp to sell Automation Engineering to Agile Robots
SR016 thyssenkrupp thyssenkrupp successfully completes sale of Automation Engineering to Agile Robots
SR017 ARQIS ARQIS advises Agile Robots SE on the acquisition of thyssenkrupp Automation Engineering
SR018 Taylor Wessing Taylor Wessing advises thyssenkrupp on the sale of Automation Engineering to Agile Robots
SR019 Verlag INDat Robotik Startup Franka Emika: Zukunft des Unternehmens durch Verkauf an Agile Robots gesichert
SR020 Automationspraxis Starkes Signal für den Robotik-Standort Deutschland: Agile Robots kauft Franka Emika Wir müssen davon ausgehen, dass es sich bei Agile faktisch um ein durch in China ansässige Gesellschaften und Institutionen kontrolliertes Unternehmen handelt.
SR021 Manufacturing Dive Agile Robots to tap into new sectors with latest acquisition
SR022 The Robot Report Agile Robots acquires thyssenkrupp Automation Engineering
SR023 BMW Group idealworks: BMW Group subsidiary gains Agile Robots AG as majority shareholder
SR024 CNBC Google partners with Agile Robots, growing its AI robotics footprint
SR025 TechCrunch Agile Robots becomes the latest robotics company to partner with Google DeepMind
SR026 Munich Startup Agile Robots: Acquisition and Humanoid Drive Next Growth Spurt
SR027 Logistik Heute KI-gestützte Robotik: Agile Robots eröffnet Hauptsitz in München
SR028 European Commission Exporting dual-use items
SR029 EUR-Lex Regulation - EU - 2024/1689 - EN
SR030 Bureau of Industry and Security EAR
SR031 Munich Startup Agile Robots acquires Franka Emika
SR032 European Commission AI Act
SR033 Falkensteg Falkensteg berät beim Verkauf des Robotik-Start-Ups Franka Emika an Agile Robots
SR034 CEN-CENELEC Artificial Intelligence
SR035 EUR-Lex EUR-Lex - 02021R0821-20241108 - EN
SV001 Agile Robots Agile Robots announces 8-figure funding
SV002 Agile Robots Agile Robots AG successfully completes pre-A round financing
SV003 Agile Robots Agile Robots AG completes Series C financing
SV004 Agile Robots Agile Robots becomes an SE
SV005 Agile Robots Smart Robotics “Made in Bavaria”: Agile Robots opens global headquarters in Munich
SV006 Agile Robots Agile Robots launches humanoid robot for industry: Agile ONE
SV007 Agile Robots Agile ONE powers up Europe’s largest AI factory
SV008 Agile Robots Agile Robots becomes a member of ARENA2036
SV009 Agile Robots Agile Robots and Franka Robotics at NVIDIA GTC 2025
SV010 Agile Robots About Agile Robots | How we transform industries
SV011 Munich Startup Agile Robots AG - Munich Startup
SV012 Munich Startup Agile Robots: Acquisition and Humanoid Drive Next Growth Spurt
SV013 Logistik Heute KI-gestützte Robotik: Agile Robots eröffnet Hauptsitz in München
SV014 MM Logistik Agile Robots eröffnet globales Hauptquartier in München
SV015 Manufacturing Dive Agile Robots to tap into new sectors with latest acquisition
SV016 TechCrunch Agile Robots becomes the latest robotics company to partner with Google DeepMind
SV017 CNBC Google partners with Agile Robots, growing its AI robotics footprint
SV018 ARENA2036 Agile Robots becomes a member of ARENA2036
SV019 Amazon Business Agile Robots - Customer Stories
SV020 International Federation of Robotics International Federation of Robotics
SV021 VDMA Robotik + Automation - vdma.org - VDMA
SV022 Mordor Intelligence Industrial Robotics Market Size, Analysis, Share & Growth Trends 2031
SV023 MarketsandMarkets Industrial Robotics Market Size, Share and Growth
SV024 ABB Robotics | ABB
SV025 KUKA Robot systems | KUKA Germany
SV026 Universal Robots News Center - Universal Robots
SV027 FANUC Integrated Reports - Library - Investors
SV028 Securities and Exchange Commission Teradyne 2025 Form 10-K XBRL Viewer
SV029 Teradyne Annual Reports
SV030 Business Wire / ResearchAndMarkets Industrial Robots Research Report 2025: Moving from Automation to Autonomy - Humanoid, Collaborative & AI-Driven Robotics Reshape Manufacturing as BMW, Mercedes-Benz & Tesla Pilot Factory Deployments - ResearchAndMarkets.com
SV031 The Robot Report Agile Robots to deploy Google DeepMind foundation models on its humanoid
SV032 Forbes Humanoid Robots Won't Save Manufacturing. Here's What Will. As factories finalize 2026 budgets, a costly choice looms: invest in human-shaped machines capturing headlines, or purpose-built systems capturing ROI.
SV033 Research and Markets Industrial Robotics Market Report 2026 - Research and Markets
SV034 FANUC Investors - FANUC CORPORATION
SV035 Teradyne Robotics | Teradyne