初创公司尽调
尽调报告 Robotics / Physical AI Series B / pre-Series C 2026-05-15

1X Technologies

仿人机器人——从企业安防走向家庭

1X Technologies 凭 EVE 拿到少见的企业机器人商业落地,又有强股东背书和有吸引力的 NEO 消费级愿景;但自动化能力缺口、竞争压力和财务不透明仍要求 Series C 投前做深尽调。

封面要素

累计融资 01
$237M+ confirmed
Series B 规模 02
$100M (Jan 2024)
领投投资方 03
EQT Ventures, OpenAI Startup Fund
EVE 部署量 04
~390+ units (ADT 140, Everon ~250)
NEO 预购价 05
$20,000 or $499/month

公司概况

1X Technologies(前身为 Halodi Robotics)是一家挪威-美国机器人和 AI 公司,正在开发通用型仿人机器人。公司 2014 年创立于挪威 Moss,2023 年 3 月更名,此后在 OpenAI 和 EQT Ventures 领投下累计融资超过 $237M。其轮式机器人 EVE 已在 ADT 和 Everon 的企业安防场景商业部署;双足机器人 NEO 面向家庭消费者,已开放预购,价格为 $20K。

官网
www.1x.tech
成立时间
2014-01-01
创始人
Bernt Øivind Børnich
创立地点
Moss, Norway
总部
Palo Alto, CA (HQ); Moss, Norway (manufacturing)
产品
EVE:面向企业安防和商业部署的轮式仿人机器人。NEO:面向家庭使用的双足仿人机器人,预购价格为 $20,000 或 $499/月。
客户
企业安防与设施(EVE);家庭消费者(NEO)
商业模式
机器人硬件销售和订阅(RaaS)模式;企业部署(EVE)贡献当前收入;消费者订阅(NEO)瞄准 2026 年及以后。
阶段
Series B / pre-Series C
融资情况
$100M Series B(2024 年 1 月,EQT Ventures 领投);约 $100M 追加融资(2025 年 1 月);正在寻求 $1B Series C。

执行摘要

主要优势

  • EVE 已在 ADT 和 Everon 商用部署(约 390 台),给出少见的真实场景牵引力
  • OpenAI 和 EQT 背书带来可信度、资本和战略入口
  • NEO Gamma 的 3D 针织外壳和 100Hz 平衡控制,构成实质技术差异
  • VR 遥操作数据飞轮沉淀具防守性的具身 AI 训练护城河
  • 挪威工程文化叠加美国资本渠道,风险画像更均衡

主要风险

  • 自动化缺口:NEO 目前高度依赖远程遥操作,还不是完全自主
  • Tesla Optimus、Figure AI 和 Agility Robotics 资金充足,竞争强度高
  • 收入和烧钱速度不透明:公司未披露财务,资本效率无法验证
  • $1B Series C 尚未确认;若无法完成交割,NEO 商业化可能受限
  • 家庭场景遥操作模式引发隐私担忧,可能招致监管关注

未决问题

  • NEO 在消费家庭中的自主能力等级和部署时间表
  • 1X 收入运行率,以及 EVE 企业部署的单位经济性
  • $1B Series C 进展和交割估值
  • 真实场景自主能力的完整竞品基准,而不只是实验室演示
  • 新兴 AI 安全框架下,消费级家用机器人的监管路径

目录

Chapter 01

01公司概况

1.1 身份、使命与运营模式

1X Technologies 是一家挪威-美国机器人和人工智能公司,总部位于 Palo Alto, California,制造业务分布在 Hayward, California 和挪威 Moss。公司公开使命是「创造充足劳动力,让安全、智能的仿人机器人与人并肩工作」。1X 以垂直整合的硬件-AI 初创公司运转:内部设计机器人,利用真实世界遥操作数据训练具身 AI 模型,并直接向企业客户和潜在消费者销售或部署机器人。 公司由挪威机器人学家 Bernt Øivind Børnich 于 2014 年创立,当时名为 Halodi Robotics,早期重点是为工业和医疗机器人开发安全执行器和全身控制系统。到 2019 年,公司已在 California 落地业务。2022 年,Halodi 与 OpenAI 达成技术合作,将 AI 模型接入其机器人。2023 年 3 月,公司正式更名为 1X Technologies,并同时宣布由 OpenAI Startup Fund 领投的 $23.5 million Series A2。 商业模式走双线:EVE(轮式仿人机器人)面向企业商业部署,先从安防和物流客户贡献近期收入;NEO(为家庭设计的双足仿人机器人)则是长期消费者平台。NEO 已披露定价:一次性买断 $20,000,或每月订阅 $499;2025 年 10 月开放预购,计划 2026 年交付。1X 位于 Palo Alto 的总部在 2025 年夏季启用,面积 80,000 平方英尺,可容纳 400 名员工,用来整合此前 Sunnyvale 和 Moss 团队。制造仍留在挪威 Moss 和 California 的 Hayward。[CO001, CO002, CO003, CO004, CO005, CO006]

关键指标快照表
指标数值 / 状态日期置信度缺口
成立2014(时名 Halodi Robotics)2014-01-01
现名启用1X Technologies(2023 年 3 月更名)2023-03-01
总部Palo Alto, CA(80,000 sq ft 总部于 2025 年夏季启用)2025-07-01此前在 Sunnyvale, CA;制造仍保留在 Moss, Norway
Series A2 轮融资(USD M)23.52023-03-01
Series B 轮融资(USD M)1002024-01-11
2025 年 1 月追加融资(USD M)1002025-01-27URL 确认 Bloomberg 文章存在;访问时全文需付费
已确认累计融资(USD M)≥223.52025-01-27Halodi 时期早期轮次未完全披露;VentureBeat 称 Series B 后约 $137M
新一轮融资目标(USD B)12025-09-28The Information 报道;截至报告日尚未确认交割
估值未披露2026-05-15已审阅来源均未公开披露
收入 / ARR未披露2026-05-15公司未发布收入数据
员工数2026-05-15Palo Alto 总部可容纳 400 人;实际员工数未披露
EVE 企业客户(已披露名称)ADT / Everon;Sunnass Hospital;挪威高校2024-06-01所有部署的确切客户数未披露
NEO 预订价格$20,000(或 $499/月)2025-10-28
NEO 预计发货年份20262025-10-28

估值、收入、员工数和确切客户数均未公开披露。财务数据来自新闻稿、第三方新闻报道和 URL 元数据;未经审计。置信度反映访问时佐证来源的强度和独立性。

FO002: 公司快照逻辑

1X 的身份、产品、训练数据飞轮、资本和客户渠道如何连接。

[CO004, CO026, CO028, CO035, CO036, CO037]

1.2 领导层、治理与组织结构

1X Technologies 由创始人兼 CEO Bernt Øivind Børnich 领导;自 2014 年创立 Halodi Robotics 以来,他一直掌舵。Børnich 是挪威机器人学家,主导了公司更名、连续融资和产品转向中的技术路线。他也是公司最主要的公开代表,出现在所有重要媒体互动中,显示出显著的关键人依赖。 AI 研究由 AI 副总裁 Eric Jang 负责。他掌管具身学习数据引擎,并公开解释过 1X 靠 VR 遥操作规模化训练机器人的策略。Jang 的角色以 San Francisco Bay Area 为中心,对公司核心 AI 差异化主张至关重要。2025 年 1 月收购 Kind Humanoid 后,Christoph Kohstall 加入 1X;他曾是 Stanford University 科学家,也是 Google 机器人团队成员。Halodi 时期的早期公司材料还将 Nicholas Nadeau 列为 CTO,但更名后公开材料没有确认其当前状态。 1X 是私营公司,治理结构——包括董事会构成、投资人权利和股权结构表——尚未公开披露细节。审阅过的公开材料中,公司没有发布公司章程或治理文件。董事会上的投资人代表可从领投方 EQT Ventures(Series B)和 OpenAI Startup Fund(Series A2)推断,但具体董事席位未获确认。公司没有披露任何监管、法律或 IPO 相关治理里程碑。美国总部迁至 Palo Alto 后,管理层提到希望从更广泛的 Bay Area 生态吸引顶尖 AI 人才,并与投资方和合作伙伴加强协作。[CO007, CO008, CO009, CO010, CO011, CO012]

领导层与创始人表
人物职务背景创始人身份关键人风险备注
Bernt Øivind Børnich(创始人)CEO 兼创始人挪威机器人专家;2014 年创办 Halodi Robotics;主导所有重大转型和融资创始人极高;唯一公开门面;牵头资本、产品和技术战略
Eric JangAI 副总裁在旧金山湾区牵头具身 AI 数据引擎;撰写过 1X 训练策略公开博客非创始人高;掌握核心 AI 差异化叙事;离职会打乱训练路线图
Christoph Kohstall并购引入人才(Kind Humanoid)前 Stanford 科学家;曾在 Google 机器人团队;2023 年创办 Kind Humanoid;2025 年 1 月加入 1X非创始人中;带来人形机器人交互和 LLM 专长;小公司收购
Nicholas Nadeau前 CTO(Halodi Robotics 时期)EEJournal 2022 年文章称其为 CTO;近期公开材料未确认其在 1X 的当前职务非创始人未知;品牌重塑后的职务未获公开来源确认

仅纳入已审阅公开来源点名的高管。董事、CFO、COO 和其他高层招聘未公开列出。Robot Report(2025 年 1 月)称公司新增“有 BMW 和 Tesla 经验的高管”,但姓名未确认。

[CO007, CO008, CO009, CO010, CO011]
FO003: 快照 KPI

截至报告日期的关键成熟度和牵引力指标,采用置信度调整后的数值。

EVE 部署数据来自行业媒体;未由经审计的客户披露验证。员工数采用设施容量,不是已确认员工人数。

[CO013, CO014, CO022, CO027, CO028, CO030]

1.3 融资历史、投资方与财务状况

自 2023 年以来,1X Technologies 在已确认融资轮次中至少筹集 $223.5 million;若计入更名前 Halodi 的早期轮次,金额很可能更高。2023 年 3 月的 Series A2 融资 $23.5 million,由 OpenAI Startup Fund 领投,Tiger Global Management、Sandwater、Alliance Ventures 和 Skagerak Capital 参投。2024 年 1 月的 $100 million Series B 由 EQT Ventures 领投,Samsung NEXT、Nistad Group 以及同一批既有投资方参投。VentureBeat 在 Series B 后立即报道称累计融资「接近 $137 million」,意味着 Halodi 时期还有未披露的早期轮次。 2025 年 1 月,Bloomberg 报道 1X 又融资 $100 million(2025 年 1 月 27 日文章标题「humanoid robot maker 1x raises 100 million」已通过 URL 元数据确认,但访问时全文受限)。Freethink 2025 年 3 月文章确认了这一时间线,称 2025 年 1 月轮次发生在 2025 年 2 月 NEO Gamma 发布之前。据此,已确认累计融资约为 $237 million,不包括未披露的 Halodi 时期资金。 2025 年 9 月,The Information 报道 1X 正在瞄准新一轮 $1 billion 融资,以支持仿人机器人开发和家庭部署。截至本报告撰写时,该轮尚未确认关闭。任何融资轮次的估值均未公开披露。公司没有披露收入、烧钱速度或盈利能力。包括 ARR、客户数和员工数在内的关键财务指标仍是私有信息。投资方包括重要全球风投资本(EQT Ventures、Tiger Global)、战略科技投资方(Samsung NEXT、OpenAI Startup Fund)和挪威产业资本(Sandwater、Skagerak Capital、Nistad Group)。[CO013, CO014, CO018, CO019, CO020, CO021]

利益相关方或投资者图谱
利益相关方角色 / 轮次层级 / 重要性战略角度尽调追问
OpenAI Startup Fund领投 — Series A2($23.5M,2023 年 3 月)高;顶尖 AI 实验室基金;释放具身 AI 投资逻辑信号验证 AI 与机器人整合;技术合作厘清资本之外的技术协议性质
EQT Ventures领投 — Series B($100M,2024 年 1 月)高;欧洲主要成长阶段 VC;押注机器人打造欧洲冠军;支持全球分销董事席位是否确认?投票权?是否有优先股堆叠?
Tiger Global Management(投资方)参投 — Series A2高;跨界基金;常投成长阶段财务回报;未显示战略运营角色当前持股比例和任何跟投权?
Samsung NEXT参投 — Series B中高;Samsung 旗下 CVC;关注硬件 / 消费与消费电子战略协同;供应链是否附带商业协议或零部件供应交易?
Sandwater既有投资者(Series A2 和 B)中;挪威家族办公室 / VC早期支持者;挪威工业协同总持股和按比例跟投权?
Skagerak Capital既有投资者(Series A2 和 B)中;挪威 VC挪威市场和工业网络总持股?
Alliance Ventures参投 — Series A2中;聚焦汽车和出行的 VC(Renault-Nissan-Mitsubishi)EVE 潜在工业 / 汽车部署路径是否有商业机器人部署协议?
Nistad Group参投 — Series B中;挪威家族 / 私人投资集团挪威市场锚定投资者总持股?

据报道 2025 年 1 月追加 $100M 轮的投资者未公开点名。持股比例、董事席位分配和优先股堆叠未披露。股权结构表为非公开。

[CO013, CO014, CO018, CO019, CO020]
FO001: 公司里程碑时间线

从创立到已知 2025 里程碑的关键公司事件和融资轮次。

2022 年 OpenAI 合作和更早融资轮次的日期根据多个来源近似推定,各来源略有差异。2025 年 9 月的 $1B 目标尚未确认完成。

[CO001, CO003, CO013, CO014, CO015, CO016]

1.4 产品、客户与商业化证据

1X 有两款主要产品,商业成熟度不同。EVE 是面向工业和机构环境的轮式仿人机器人。它拥有 23 个自由度,所有关节具备力控和阻抗控制,每只手臂可承载 8 kilogram,并配有自然语言界面,操作员可通过语音命令串联任务。EVE 已在至少两个具名企业客户处规模化商业部署:ADT 据报道下单 140 台 EVE,是当时全球最大的仿人机器人订单;Everon(前 ADT Commercial)在商业楼宇夜间安保中部署约 250 台 EVE。挪威 Sunnass Hospital 被列为医疗部署案例,挪威高校也已收到 EVE 设备。 NEO 是为家庭使用设计的双足仿人机器人。它经历了原型阶段:NEO Beta 于 2024 年 8 月 30 日亮相,定位为家庭部署原型,少量进入选定家庭用于研究。NEO Gamma 于 2025 年 2 月 21 日发布,引入 3D 编织尼龙外衣以降低受伤风险,加入用于沟通反馈的「Emotive Ear Rings」、100Hz 动态平衡控制、来自人体动作捕捉的强化学习,以及能够抓取多种物体的视觉操作模型。2025 年 10 月 28 日,1X 开放 NEO 预购,一次性买断价 $20,000,或按月订阅 $499,订金 $200,预计 2026 年发货。 NEO 当前部署模式有一个关键运营限制:早期体验家庭中的多数任务将由远程人类遥操作员通过 VR 头显完成,而不是由完全自主的 AI 完成。CEO Børnich 直接承认了这一点,称买家必须同意,在预约时段内遥操作员可以访问其家庭数据。公司将其包装为训练自主模型所必需的数据采集策略,而非产品限制;但外部分析师和记者认为,这暴露了市场化产品愿景与近期现实之间的重大落差。[CO026, CO027, CO028, CO029, CO031, CO032]

里程碑表
日期事件类型金额 / 估值 / 状态参与方影响
2014Halodi Robotics 在挪威成立创立Bernt Øivind Børnich(创始人)公司起点;确立挪威身份和深厚机器人基础
2018首个 EVE 原型发布产品1X / Halodi 团队确立轮式人形机器人平台;成为后续企业部署基础
2019加州办公室设立扩张1X 团队开始双地布局;接入湾区 AI 与机器人团队人才
2022OpenAI 技术合作;机器人 + AI 整合启动;非正式品牌重塑开始合作Halodi Robotics / OpenAI关键背书;公司与前沿 AI 对齐;为品牌重塑铺路
2023-03更名为 1X Technologies;Series A2 交割融资$23.5MOpenAI Startup Fund(领投);Tiger Global;Sandwater;Alliance Ventures;Skagerak Capital以新品牌公开亮相;OpenAI 领投带来全球媒体关注
2024-01Series B 交割融资$100MEQT Ventures(领投);Samsung NEXT;Nistad Group;既有投资者单轮规模最大;确立 1X 主要人形机器人玩家地位;累计约 $137M+
2024-08-30NEO Beta 发布产品1X Technologies首个双足家用机器人原型;市场叙事从工业转向消费
2025-01收购 Kind Humanoid;追加 $100M 融资(Bloomberg 报道)收购 / 融资~$100MChristoph Kohstall(Kind Humanoid CEO);多名投资者(未具名)补上 LLM 交互人才;强化 NEO 家用定位;累计资本约 $237M+
2025-02-21NEO Gamma 发布产品1X Technologies双足平台升级,加入针织外衣、更强 AI 和步态;开始内部家庭测试
2025 年夏季新 80,000 sq ft Palo Alto 总部启用扩张1X Technologies整合 Sunnyvale 和 Moss 团队;目标容量 400 人
2025-09目标新一轮 $1 billion 融资(据报道)融资$1B 目标未具名投资者(The Information 报道)若交割,将带来变革性影响;截至报告日未确认
2025-10-28NEO Home Robot 开放预订;定价 $20K 或 $499/月产品$20,000 / $499/月1X Technologies;早期采用客户家用市场首次商业承诺;释放近端收入信号

Halodi Robotics 早期轮次(Series A2 之前)在已审阅公开来源中记录不完整。2025 年 9 月的 $1B 目标融资尚未确认交割。2022 年 OpenAI 合作启动日期根据多个来源近似推断,来源之间略有差异。

[CO001, CO003, CO013, CO014, CO015, CO016]

1.5 里程碑、收购与证据缺口

1X 的公开时间线横跨十多年,包括更名、四轮已确认融资、一项公司收购和连续产品迭代。公司 2014 年以 Halodi Robotics 名义创立于挪威,先从学术/工业机器人转向商业机器人,再于 2022 年与 OpenAI 合作。2023 年 3 月更名为 1X Technologies,并同步宣布 $23.5 million Series A2,标志着公司进入全球机器人融资讨论。2024 年 1 月来自 EQT Ventures 的 $100 million Series B,让 1X 成为资金更充足的欧洲仿人机器人公司之一。 2025 年 1 月,1X 收购 Kind Humanoid。这是一家 Bay Area 三人创业公司,由 Christoph Kohstall 创立,曾将大语言模型用于仿人交互,并与设计师 Yves Béhar 的 Fuseproject 工作室合作。同月,据报道公司又完成 $100 million 融资。NEO Gamma 于 2025 年 2 月发布,Palo Alto 总部于 2025 年夏季启用,可容纳 400 名员工。2025 年 10 月,NEO Home Robot 开放预购。截至 2025 年 9 月,1X 据报道正在寻求 $1 billion 新资金。 关键证据缺口包括:经验证的员工数(Palo Alto 总部按 400 人设计,但实际员工数未披露);全部历史轮次的累计融资;已确认估值;EVE 企业部署带来的收入;以及据报道 $1 billion 融资目标的状态。OpenAI 合作的具体条款(商业协议还是仅投资)仍不透明,治理结构也没有公开文件。公司截至报告日没有披露 NEO 预购的准确数量。[CO040, CO041, CO042, CO043, CO044, CO045]

1.6 重点证据

Chapter 02

02市场分析

2.1 市场边界、定义与替代方案

仿人机器人市场指已经商业销售或租赁、具有人形形态的机器人——既包括轮式躯干设计(如 1X EVE),也包括完全双足设计(如 1X NEO、Tesla Optimus 和 Figure 02)——它们在为人类建造的环境中,靠灵巧操作和移动能力完成任务。该定义排除传统固定臂工业机器人(Fanuc、KUKA、ABB)、没有操作能力的自主移动机器人(AMR,例如 Locus Robotics),以及没有物理形态的纯 AI 软件系统。 目标用例中,仿人机器人面对的现状替代方案包括:(1)人类劳动力——安防、物流和家务任务中的主要替代;(2)传统自动化系统——传送带、固定机械臂、条码扫描器——可处理结构化任务,但无法泛化;(3)协作机器人——如 Universal Robots UR 系列——在固定位置运行,没有移动能力;(4)专用非仿人机器人——Roomba 类扫地机器人、农业收割机——解决窄任务。仿人形态的竞争优势在于,无需重新改造环境,就能使用现有人类设计的物理基础设施(门、楼梯、工具、货架)。 相邻市场包括工业外骨骼、协作机器人手臂、AI 驱动服务机器人(配送机器人、洗地机器人),以及更广义的工业自动化软件。随着能力成熟,仿人机器人有望连接这些相邻领域,扩大可服务范围。本分析纳入的支出包括机器人硬件采购、租赁、软件订阅、训练数据服务和售后维护;排除的支出包括设施基础设施改造、员工再培训,以及相邻非仿人机器人品类。[CM001, CM002, CM006, CM010]

市场定义表
市场细分定义 / 纳入支出排除支出现状替代方案与 1X 的关联
企业安防机器人夜间值守、设施巡逻、借助轮式 / 双足人形机器人做出入监控;包括硬件租赁 / 采购和 AI 软件固定 CCTV 基础设施、非人形安防机器人人类安保人员、非人形巡逻机器人直接(EVE 已部署于 ADT 和 Everon)
消费级家庭辅助家务自动化、物体操作、个人辅助;包括机器人购买和订阅软件无操作能力的智能家居设备(Alexa、Nest)、Roomba 类纯扫地机器人人类家政清洁员、专用家电直接(NEO 预订价 $20K / $499/月)
工业 / 仓储物流拣货、物料搬运、工厂和履约中心装配辅助;包括硬件和集成固定机械臂(KUKA、Fanuc)、输送系统、无操作能力的 AMR人工、协作机器人、固定自动化相邻(NEO 以此为目标;Agility Digit 直接竞争)
医疗辅助病人转运、药品配送、护理支持;包括医院级硬件和监管许可非人形配送机器人、轮椅升降设备、医疗器械护理助理、护工、专用转运设备相邻(EVE 部署于挪威 Sunnass Hospital)
科研与教育大学和研发实验室用于机器人研究的部署;包括硬件、软件、训练数据访问仿真软件、非人形研究平台传统工业机器人、实验室机械臂较小(挪威高校部署)

市场边界按人形形态定义(轮式躯干或具备灵巧操作能力的全双足机器人)。排除支出包括设施改造成本和员工再培训。关联度评估基于 1X 已确认的客户披露和公司陈述的产品路线图。

[CM001, CM002, CM014, CM016, CM017, CM018]
FM001: 市场规模测算视角

与 1X Technologies 相关的人形机器人市场 TAM、SAM 和估计 SOM 分层,基于 Goldman Sachs 和 IDC 估计,并由公开数据推断 1X 定位。

Goldman Sachs 的可服务市场(SAM)和总可用市场(TAM)数据来自已发布的分析师报告。消费者可服务市场(SAM)是分析师综合估算。1X 可获取市场(SOM) 根据公开定价和部署数据推算;不是公司披露的预测。

[CM001, CM011, CM013]

2.2 市场规模测算——TAM、SAM、SOM 与分析师估计

仿人机器人仍处商业化早期,市场规模天然不确定;即便如此,多家独立研究机构对 2033–2035 年总可用市场给出的估计正在收敛到 $35B–$75B 区间。Goldman Sachs(2023,2024 更新)预计 2035 年全球仿人机器人市场为 $38 billion,2030 年可服务市场(SAM)约 $6 billion,重点是西方市场的制造、仓储和企业安防。这一 SAM 估计建立在明确的近期部署场景——产线支持、物料搬运、楼宇安保——之上,而不是简单用 TAM 乘以通用系数。 ARK Invest 的视角更激进:若将仿人机器人视为全球各行业人类劳动力替代品,2030 年可触达机会超过 $24 trillion。该数字体现 ARK 一贯的自上而下方法,基础假设是自动化可捕获全球劳动报酬中的相当一部分。这个数字有方向参考价值,但不是近期可触达市场,应理解为天花板情景。IDC 对全球服务机器人市场(涵盖仿人和非仿人服务机器人)的更保守估计为 2030 年 $50 billion,CAGR 约 20%。 Mordor Intelligence、MarketsandMarkets 和 Grand View Research 分别预计 2033–2035 年仿人机器人市场为 $35–$75 billion,覆盖保守到乐观情景,CAGR 估计为 40–55%。区间很宽,反映市场边界(仅仿人 vs. 更广义服务机器人)、地理范围和采用速度假设不同。Statista 和 ABI Research 提供了服务机器人收入轨迹的支撑数据。 对 1X Technologies 而言,SOM 没有公开预测。基于当前 EVE 部署(ADT 和 Everon 约 390 台)以及 NEO 预购价格($20,000 或 $499/月),对 2026–2028 年收入的数量级估计为 $10M–$50M,并会随着 NEO 规模化快速增长——但这仍只是从公开数据粗略推断,并非公司披露数字。[CM001, CM004, CM007, CM008, CM009, CM010]

TAM/SAM/SOM 或规模测算口径表
发布方发布年份地理范围市场规模CAGR方法置信度主要局限
Goldman Sachs2023–2024全球2035 年 TAM $38B;2030 年 SAM $6B~40%自下而上的部署情景建模假设解决美国劳动力短缺的 4%;口径较窄
ARK Invest2023全球2030 年可触达机会 $24T+N/A自上而下的劳动力替代倍数天花板情景;不是短期可触达市场
Mordor Intelligence2024全球2033 年 $35–55B45–50%一手 / 二手研究混合专有调研方法;透明度有限
MarketsandMarkets2024全球2033 年 $38–65B50%行业访谈和二手数据区间较宽,反映采用速度高度不确定
Grand View Research2024全球2035 年 $55–75B54%二手研究、公司公告偏向乐观情景;CAGR 估计最高
IDC2023全球2030 年服务机器人 $50B+~20%全球 IT 市场情报覆盖完整服务机器人品类;不限人形机器人

所有数字均来自发布时的分析师报告;未经审计。ARK Invest 数字反映理论上的全球劳动力替代总量,不是短期可触达市场。Goldman Sachs SAM 是短期运营规划中最有用的估计。CAGR 数字反映各自基准年的分析师预测。

[CM001, CM004, CM007, CM008, CM009, CM010]
FM002: 市场估算区间

主要研究机构给出的全球人形机器人和服务机器人市场三档估算(低 / 基准 / 高),单位为 $B。所有数值均以 USD 十亿美元计,口径一致。

所有数值均为 USD 十亿美元。低 / 高边界代表分析师低、高情景估算;未披露明确区间时,使用推算置信区间。单位统一为 $B。各估算的预测期年份不同(2030–2035);数值对应各分析师声明的预测期。

[CM001, CM007, CM008, CM009, CM010, CM031]

2.3 买方分层、采用路径与付款方地图

仿人机器人市场包含三类主要买方,采用动态各不相同。企业买方——包括企业安防运营、设施管理部门、仓储和物流运营商、制造工厂管理者——构成近期商业机会:采购流程可预测,有资本设备预算,购买以 ROI 为导向。ADT 部署 140 台 EVE 用于夜间安保,Everon 部署约 250 台,是当前商业化证明。采购通常由运营副总裁或首席安全官推动,他们可动用 CapEx 或服务合同 OpEx 预算。 消费者买方——房主和家庭决策者——代表 NEO 及竞品的长期 TAM。NEO 定价 $20,000 或 $499/月,瞄准有显著可支配收入的早期科技消费者。该细分市场的采用触发点主要是便利和节省时间(家务自动化),但近期转化率会受遥操作模式和有限自主能力制约。如果自主能力达到门槛,消费者家庭机器人市场到 2030 年的 TAM 估计为 $10–15 billion。 医疗和医院系统构成第三个新兴细分,应用包括患者转运、药物配送和基础护理辅助。挪威 Sunnass Hospital 是 1X 早期客户。该细分采购周期更长,并需要监管许可;但护理劳动力短缺是强结构性驱动因素。预算通常由设施资本设备委员会掌握,采用触发点包括护理人员短缺(当前在美国和欧洲已很严峻)以及按患者日成本优化。第四个细分——高等教育与科研——规模更小,但作为早期标杆客户,对建立生态信誉有战略意义。[CM013, CM014, CM015, CM016, CM017, CM018]

细分市场 / 买方图谱
细分市场主要买方用户付款方预算决策权采用触发因素
企业安防企业安保负责人;运营副总裁安保人员;设施经理运营或设施预算(OpEx)运营副总裁 / CSO降低人工成本;夜班安保短缺
消费级家庭房主;双收入家庭家庭成员;老年房主消费者可选支出家庭决策者便利;节省时间;居家养老
医疗 / 医院医院资本委员会;CFO护理人员;患者;护工资本设备预算护理总监 / CFO护理人手短缺;单患者成本优化
仓储 / 物流运营副总裁;3PL 负责人仓库拣货员;班组主管CapEx 或外包人工 OpEx运营副总裁 / CFO人手短缺;吞吐压力;电商增长
工业制造工厂经理;制造副总裁装配线工人;生产人员资本设备预算制造副总裁 / CFO自动化要求;质量一致性;人工成本
科研与教育大学采购;实验室负责人研究人员;研究生科研经费预算PI / 系主任具身 AI 研究;课程;声望

细分市场定义基于 1X 已确认的客户披露(ADT、Everon、Sunnass Hospital)以及公开来源中的人形机器人部署模式。预算决策权和采用触发因素为根据行业和分析师数据推断;非一手研究。

[CM014, CM016, CM017, CM018, CM019]
FM003: 买方 / 客群图

矩阵把五类买方客群映射到核心采纳维度:用户画像、付款方、预算权、采纳触发因素。用于确定买方优先级和 GTM 分群。

[CM013, CM016, CM017, CM018, CM019]

2.4 增长驱动因素与采用约束

仿人机器人采用的主要增长驱动因素分为四类。第一,结构性劳动力短缺:美国约有 10 million 个未填补岗位(多个 2024–2025 年分析师报告引用的 BLS 数据),Goldman Sachs 估计仿人机器人到 2030 年可填补其中最多 4%。WEF Future of Jobs 2025 报告预计,全球 85 million 个岗位将被自动化替代,同时产生 97 million 个自动化相邻新岗位,带来对技能升级人类工人和机器人助手的净需求。McKinsey 研究显示,到 2030 年自动化可能影响全球 400–800 million 个岗位。 第二,AI 硬件成本下降:计算成本、传感器价格和执行器制造经济性的快速改善,已将基础仿人机器人的物料清单成本压到 $50,000 以下——这一门槛让规模化企业安防商业案例开始可行。第三,VC 和企业投资背书:PitchBook 数据显示,2024 年仿人机器人 VC 融资超过 $1 billion,Tesla、Figure、Agility、Apptronik 和 1X 均获得大量资本,带动供应商、集成商和 AI 模型开发者生态。第四,数据飞轮:早期部署者积累专有训练数据,提升模型表现,给 1X 这类已有商业部署的先行者带来复利优势。 主要采用约束同样显著。单机成本 $20K–$100K+,相当于用机器人替代人力需要 1–5 年回本,许多行业的 CFO 批准并不稳。自主能力有限——多数机器人仍需人类遥操作完成复杂任务——削弱感知 ROI 和价值主张。监管缺位:美国和欧盟尚无面向工作场所仿人机器人的正式标准,企业采用方面临责任不确定性。消费者信任不足:NEO 遥操作模式(WSJ 2025 年 10 月披露)引发隐私担忧,可能拖慢家庭市场采用。制造规模化的资本强度也会在需求完全确立前约束供给。[CM003, CM005, CM006, CM020, CM021, CM022]

增长驱动因素与约束表
因素方向时点影响幅度对 1X 的影响
全球劳动力短缺(美国 10M+ 空缺岗位)驱动因素现在 — 2030验证企业场景;ADT 和 Everon 部署正面覆盖这一点
AI 硬件成本下降(GPU、执行器、传感器)驱动因素现在 — 2028让 NEO $20K 价位跑得通;成本继续下降会扩大可服务市场(SAM)
人形机器人生态的 VC 与战略投资驱动因素现在 — 2027搭起 1X 可借力的供应商 / 集成商生态;也会孵化竞争对手
数据飞轮与具身 AI 模型成熟度驱动因素2025 — 2030高(长期)如果规模化,1X 的遥操作数据策略会成为护城河;具备先发优势
单机成本($20K–100K+)与人力 ROI约束因素现在 — 2028拖慢安防之外的企业采用;当前价格卡住消费市场
自主性有限、依赖遥操作约束因素现在 — 2027NEO 的遥操作员模式削弱消费吸引力;存在隐私风险(WSJ 2025 年 10 月)
人形机器人监管框架缺位约束因素现在 — 2030没有标准,医疗、航空、食品处理部署会被卡住
消费者对隐私和安全缺乏信任约束因素现在 — 2028NEO 家庭遥操作的隐私披露可能压低早期用户转化

驱动因素和约束因素评估综合 Goldman Sachs、McKinsey、WEF、Deloitte 与新闻报道。影响强度为定性评级。时间表示该因素预计在采用曲线中影响达到峰值的阶段。

[CM003, CM005, CM006, CM020, CM021, CM025]
FM004: 采纳漏斗或价值链图

采纳漏斗展示总可寻址劳动力市场需求中,技术可行、经济可行、企业就绪以及 1X 凭 2026 年当前产品可获取的比例。

漏斗百分比是示意性估算,综合 McKinsey、Deloitte 和 Goldman Sachs 关于自动化可行性和渗透率的分析得出。不是一手研究。数值代表截至 2026 年,在每个采纳关口可触达的总劳动力市场需求大致份额。

[CM003, CM021, CM022, CM025]

2.5 规模测算尽调缺口与相互矛盾的估计

多个重大证据缺口限制了对市场规模和采用轨迹的信心。第一,不同机构对仿人机器人 TAM 的估计相差 2–3x(2033–2035 年为 $35B 到 $75B+),反映其对采用速度、价格点演进和地理市场定义的基本假设不同。Goldman Sachs 到 2030 年约 $6B 的 SAM 估计更保守,也可能更适合近期规划,但尚未有其他机构用同一方法独立复现。 第二,没有公开来源实时追踪全球仿人机器人部署总量。1X 的 EVE 机队(约 390 台)和 Agility 在 Amazon 的 Digit 部署常被引用,但全行业部署估计来自新闻稿和分析师推断,而不是经验证的出货数据。第三,企业安防机器人市场规模(2027 年 $3–5B)基于有限分析师覆盖,可能没有纳入非仿人安防机器人在内的完整竞争格局。第四,美国和欧盟的仿人机器人监管框架完全缺位,在医疗、航空、食品处理等受监管行业,预测采用时间线与真实世界部署可行性之间存在落差。最后,没有分析师发布过企业仿人机器人需求价格弹性的正式评估——这是估计单机价格降至 $10K、$5K 或 $2K 后市场扩张的关键输入。[CM012, CM031, CM037, CM038]

Chapter 03

03竞争格局

3.1 竞争格局——直接、相邻与新兴竞争者

仿人机器人竞争格局可分为五类。直接竞争者——为商业部署打造通用双足或轮式仿人机器人的公司——包括 Agility Robotics(Digit)、Figure AI(Figure 02)、Apptronik(Apollo)、Sanctuary AI(Phoenix)和 Unitree(H1/G1)。各家公司都在用不同形态、融资水平和部署策略,瞄准企业或消费者仿人机器人采用。 超大型公司入局是最具生存威胁的一层:Tesla(Optimus)和 Hyundai/Boston Dynamics(Atlas)带来硬件制造基础设施、品牌和资本,纯机器人创业公司难以匹敌。Elon Musk 公开给出的目标是到 2025 年底在 Tesla 自有工厂部署 1,000 台 Tesla Optimus,并提出长期消费者价格目标约 $20,000——直接对齐 NEO 已宣布价格。 相邻竞争者包括 Physical Intelligence(pi),该公司走纯软件路线,为机器人操作构建基础模型;以及以 Unitree 为代表的中国 OEM 制造商。Unitree 正在出货 H1 双足机器人,价格 $90,000;G1 价格 $16,000,明显低于西方竞争者。Unitree 极强的价格进攻性,对包括 1X 在内所有西方硬件优先仿人机器人公司构成商品化风险。 现状替代方案仍是最重要的竞争压力:人类劳动力(安防、物流、家务任务)、传统工业机器人,以及没有操作能力的自主移动机器人(AMR)。说服买方从这些替代方案切换到仿人机器人,需要足够自主能力和有吸引力的 ROI;目前没有竞争者在大众市场规模上跨过这两道门槛。1X 的竞争护城河问题在于,当前 EVE 商业部署领先是否能形成持久的专有训练数据优势,抵消它相较 Figure、Tesla 和 Agility 在双足硬件上的劣势。[CP001, CP002, CP003, CP004, CP005, CP006]

竞争对手画像表
公司类别累计融资目标细分市场旗舰产品核心差异点主要限制
1X Technologies (EVE)轮式人形机器人初创公司~$237M+企业安防、医疗EVE(轮式)唯一已商业部署的机群(约 390 台);数据飞轮;OpenAI 支持轮式形态相对双足吃亏;NEO 之前双足经验有限
1X Technologies (NEO)双足人形机器人初创公司同上消费者家庭、通用NEO(双足)$20K 开放预订;数据驱动的具身 AI;设计导向的家庭定位尚未商业部署;依赖遥操作员;市场拥挤
Boston Dynamics (Atlas)从研究走向商业化,Hyundai 持有未披露(Hyundai 收购)工业 / 展示Atlas(电动双足)完整的双足杂技能力;品牌声望;Hyundai 制造规模未开放通用商业销售;仅聚焦工业
Tesla (Optimus)超大市值公司入局者内部资金(Tesla 公司)大规模消费 / 工业Optimus Gen 2Tesla 制造规模;Dojo AI;强品牌;公开提出 $20K 长期目标尚未向第三方商业销售;交付时间未确认
Figure AI双足人形机器人初创公司$675M(2024 年 2 月 + 此前融资)工业制造Figure 02集成 OpenAI 语言能力;BMW 工厂部署;投资人阵容强烧钱速度高;BMW 部署规模未确认;没有消费产品
Agility Robotics (Digit)双足人形机器人初创公司未披露(Amazon 支持)企业仓储 / 物流Digit(双足)Amazon 部署(最成熟的企业部署);GXO、Ford 合作Amazon 支持可能限制独立性;绑定仓储场景;消费路径有限
Apptronik (Apollo)双足人形机器人初创公司$53M+(Series A 轮)企业工业ApolloGE Aerospace、Mercedes-Benz 合作;企业可信度融资基数相对小;部署证据有限
Sanctuary AI (Phoenix)双足人形机器人初创公司$100M+ CAD零售 / 企业Phoenix(第 7 代)加拿大机器人生态;第 7 代平台;英国 / 美国部署规模更小;以 CAD 融资,相比 USD 同业吃亏
Unitree Robotics (H1/G1)中国 OEM / 硬件制造商未披露(中国资本支持)研究、价格敏感型企业H1($90K)、G1($16K)价格竞争力极强;$16K 的 G1 低于所有西方竞争对手自主 AI 有限;出口限制风险;安全认证不确定
Physical Intelligence (pi)(AI 软件竞品)AI / 软件基础模型公司$400M平台型 AI(硬件无关)pi-zero(pi0)模型纯软件路线;硬件无关;模型质量强;Google / Bezos 支持没有硬件;不是直接商业人形机器人竞争对手;存在平台风险

融资数据来自媒体报道和 PitchBook,未经审计。1X EVE/NEO 分成两行,因为两者面向不同细分市场,竞争动态也不同。Agility Robotics 由 Amazon 生态投资方收购;确切融资额未单独公开拆分。Unitree 人形机器人业务的融资额未单独公开披露。

[CP001, CP002, CP003, CP004, CP005, CP006]
FP001: 竞争定位图

人形机器人公司竞争定位,两个维度:商业部署规模(x 轴,0=无部署,10=大众市场)和产品成熟度 / AI 自主性(y 轴,0=原型,10=完全自主)。位置基于公开证据作定性估算。

位置为定性估算,依据公开部署数量、产品公告和第三方报道。不存在独立验证基准。商业规模轴反映已确认商业部署(不含内部使用或试点);自主性轴反映观察到的任务表现水平,而非理论能力宣称。

[CP001, CP002, CP003, CP004, CP009, CP010]

3.2 能力基准与 1X 产品对比

将 1X 与直接竞争者对标,可以看到清晰取舍。在商业部署规模上,1X EVE 以约 390 台企业客户部署量领先所有纯仿人机器人公司——除 Agility Robotics Digit 外,超过任何竞争者。Digit 已部署在 Amazon 履约中心、GXO Logistics 和 Ford;其 Amazon 部署是行业内商业意义最重的案例,代表首个大规模工业仿人机器人部署。 在双足移动和灵巧操作上,1X 的轮式 EVE 落后于完全双足竞争者。Boston Dynamics Atlas 虽未面向通用任务商业部署,但展示了完整杂技式能力。Figure AI 的 Figure 02 和 Agility 的 Digit 都能在真实工厂环境中行走并执行操作任务。Apptronik Apollo 和 Sanctuary AI Phoenix 是可信的双足平台,企业试点正在推进。1X 的 NEO 双足机器人就是为这一空间竞争而设计,但它将在 2026 年进入市场,届时多个竞争者已有 12–24 个月双足部署经验。 在 AI 集成上,Figure AI 与 OpenAI 的合作使其能够获得前沿语言模型能力,并与操作系统紧密集成。1X 的具身学习策略——用 VR 遥操作数据端到端训练神经网络——具有差异化,也已经产生公开展示的自主任务表现;但所有玩家与人类能力之间仍存在很大的自主性缺口。Tesla Optimus 受益于 Tesla 大规模 AI 训练基础设施(Dojo 超级计算机、FSD 神经网络栈),小公司很难复制。 价格方面,1X NEO 的 $20,000(或 $499/月)与 Tesla Optimus 公开提出的长期目标价直接竞争。不过 Unitree G1 为 $16,000,H1 为 $90,000,夹住市场,其中 G1 一旦能借第三方 AI 软件增强自主能力,将构成明确低成本威胁。1X 与 Unitree 等中国竞争者之间的价格差,是关键战略变量。[CP009, CP010, CP011, CP012, CP013, CP014]

功能 / 能力矩阵
能力1X EVE(当前)1X NEO(2026)Agility DigitTesla OptimusFigure 02
大规模商业部署是(约 390 台)仅预订是(Amazon、GXO)仅 Tesla 内部仅试点(BMW)
双足行走否(轮式)是(2026 年推出)是(成熟)是(Gen 2)是(可运行)
灵巧手操作部分(2 指夹爪)是(已设计)是(5 指)是(多指)
语言驱动任务控制是(自然语言界面)部分部分是(OpenAI)
自主运行(无遥操作)部分(结构化任务)部分(遥操作优先)部分(结构化)部分(内部)部分
企业集成(API / 机群管理)是(ADT、Everon)计划中是(Amazon WMS)内部试点

能力评估来自公开产品页、新闻稿和第三方技术评测。目前没有发布独立技术基准来比较这些系统。「部分」表示能力存在,但受任务限制,或需要大量人工监督。

[CP009, CP010, CP011, CP012, CP013]
FP002: 功能覆盖 / 能力图

领先人形机器人竞争者在六个关键维度上的序位能力评分矩阵。分数为定性判断(强 / 部分 / 弱 / 无),依据公开证据。

强 / 部分 / 弱 / 无评分来自公开产品信息、媒体报道和部署证据。该行业不存在独立技术基准。评分反映 2025–2026 年状态。

[CP007, CP009, CP010, CP011, CP012]

3.3 定价、包装与商业模式对比

仿人机器人竞争格局中,商业模式和定价策略差异很大。1X 率先采用双定价模式:NEO 硬件买断价 $20,000,EVE 采用企业合同(估计为 $3,500–5,000/月或一次性 $100K+),同时为消费版 NEO 提供 $499/月订阅。这一订阅模式在仿人机器人竞争者中独特,类似自主洗地(Brain Corp、Avidbots)和仓库 AMR(Locus、Fetch、Vecna)等相邻品类已经验证过的机器人即服务(RaaS)模式。 Agility Robotics 主要以企业合同提供 Digit,而不是直接售卖,并在 Amazon 和 GXO Logistics 项目中强调 RaaS 模式。价格未公开披露,但分析师估计为每台 $100,000–150,000,或等值年度合同额。Agility 已由 Agility Robotics 母公司收购——它拥有独立融资,投资方包括 Amazon 和其他投资者。Boston Dynamics Atlas 目前未面向通用任务商业销售;其此前商业产品 Spot 定价约 $70,000–90,000。 Tesla Optimus 尚未公布面向第三方销售的商业价格;当前部署仅限 Tesla 自有工厂内部。Elon Musk 提出的长期零售价格 $20,000 带有愿景性质,取决于大规模制造。Figure AI 没有公开披露其 BMW 工厂部署的定价。Apptronik Apollo 企业合同估计为每台 $100K+。Sanctuary AI Phoenix 商业条款未公开披露,但已有与 Mark's Work Wearhouse(Canada)及美国零售商的部署报道。 Unitree G1 的 $16,000 是市场上最具破坏性的定价信号:它以低于 $20K 的价格提供双足机器人,竞争者目前无法在同等硬件能力下匹配。如果 Unitree 产品通过自有 AI 或第三方集成达到足够自主运行水平,它可能拿下价格敏感的企业 SME 细分,并将硬件层商品化。[CP017, CP018, CP019, CP020, CP021, CP022]

定价 / 打包对比
公司 / 产品单机价格订阅 / RaaS 选项合同模式定价置信度竞争位置
1X NEO(消费)$20,000$499/month直接销售或订阅与 Tesla 目标价有竞争力;被 Unitree G1 压价
1X EVE(企业)~$100K+ 或 $3,500–5,000/mo(估计)$3,500–5,000/month(估计)企业服务合同定价为估计;未公开披露
Tesla Optimus~$20K(长期目标)未披露仅内部使用(尚无商业销售)愿景型目标价;目前没有商业报价
Agility Digit~$100K–150K(估计)或 RaaSRaaS 合同(估计)企业 RaaS / 合同定价未披露;Amazon 交易结构保密
Figure AI Figure 02未披露未披露BMW 工厂合同条款保密;没有消费或通用企业报价
Apptronik Apollo~$100K+(估计)未披露企业合同价格未披露;GE Aerospace 条款保密
Unitree G1$16,000不提供硬件直接销售双足最低价位;对 1X NEO 市场构成显著颠覆风险
Unitree H1$90,000不提供硬件直接销售研究 / 开发者市场;价格高于 NEO;完整双足

价格来自官方产品页(1X NEO、Unitree G1/H1)或未披露时的分析师估计。除 Unitree 外,所有企业竞品定价均由媒体报道和分析师推断估算。RaaS 合同结构仅在 Agility/Amazon 部分披露。

[CP017, CP018, CP019, CP020, CP021, CP022]
FP003: 护城河 / 就绪度 KPI

基于公开证据,总结 1X Technologies 截至 2026 年相对市场的竞争位置和护城河耐久度 KPI。

[CP025, CP026, CP027, CP030, CP031]

3.4 护城河耐久性、竞争风险与杀伤情景

1X 最持久的护城河是来自商业 EVE 部署的专有训练数据。约 390 台机器人在安防和医疗环境中生成真实世界运行数据,1X 已积累出尚未商业部署的竞争者无法复制的训练数据集。这套数据飞轮——遥操作数据输入端到端神经网络训练——在结构上类似 Tesla 靠 Autopilot 车队数据建立 FSD 优势;1X AI 副总裁 Eric Jang 也公开阐述过这一策略。该护城河能否持久,取决于模拟训练成熟后(NVIDIA Isaac、Google DeepMind 模拟管线),专有真实世界数据是否仍然必要。 次级护城河包括:(1)OpenAI 关系,提供技术信誉和生态入口,但并非排他;(2)与 ADT 和 Everon 建立的企业客户关系,凭运营集成和积累的任务专用模型微调形成切换成本;(3)1X 的挪威设计和制造传承,带来位于挪威 Moss 的高成本效率制造基础设施,区别于 Bay Area 成本结构。 主要杀伤情景是 Tesla Optimus 规模化。如果 Tesla 达成 2025 年 1,000 台部署目标,并随后在 Tesla 品牌、服务网络和 AI 基础设施支持下,以 $20,000 或更低价格宣布商业上市,1X 的 NEO 将面临强烈逆风。次级杀伤情景是 Amazon 支持的 Agility Robotics 将 Digit 部署扩展至 Amazon 1,000+ 个履约中心,完全拿下物流细分,并融到足够资本,在双足开发上超越 1X。 Unitree 的价格竞争是整个西方仿人硬件层的系统性风险,而不是只针对 1X 的威胁;即便如此,如果中国制造商能凭开源或商业授权 AI 模型达到足够自主性,$16K 价格点可能压垮 1X 的 NEO 所瞄准的可服务企业 SME 市场。[CP025, CP026, CP027, CP028, CP029, CP030]

护城河耐久性 / 竞争风险登记表
护城河主张护城河类型耐久性评估主要威胁护城河削弱后的风险尽调问题
来自 EVE 机群的自有具身 AI 训练数据数据 / IP中 — 仿真和开源数据可能削弱真实世界数据优势NVIDIA Isaac / DeepMind 仿真;开源机器人数据集高 — 如果仿真替代真实世界数据,竞争对手追赶会更快量化 1X 训练语料相对仿真的独特任务 / 环境覆盖
首个商业部署规模(约 390 台 EVE)分销 / 客户关系中 — 竞争对手(Agility、Tesla)到 2027 年已有或将有可比规模Amazon 支持的 Agility 扩张;Tesla 内部部署里程碑高 — 如果竞争对手在 12–18 个月内超过规模,1X 领先会消失核验 ADT / Everon 的续约条款和扩张承诺
OpenAI 关系与联合开发技术 / 合作伙伴通道低-中 — 非排他;Figure AI 也有 OpenAI 合作Figure AI 更深度的 OpenAI 联合开发协议中 — 共同投资人关系不是排他的技术护城河厘清与 OpenAI 的技术关系和资本关系范围
NEO $20K 价位对标 Tesla 目标定价 / 成本结构低 — $16K 的 Unitree G1 已经压价;Tesla 有制造规模优势$20K 的 Tesla Optimus 加上 Tesla 品牌和规模;Unitree G1 硬件颠覆很高 — 1X 无法在硬件价格战中打赢 Tesla 或 Unitree评估 1X 规模化后的 BOM 成本;评估价格之外的差异化路径
EVE 轮式形态在安防市场独特利基定位 / 客户锁定中 — 竞争对手目前没有专门瞄准夜间安保专用非人形安防机器人(Knightscope);安保服务自动化中 — 如果 1X 拿下多年企业合同,利基市场有机会守住核验 ADT 和 Everon 主服务协议期限与续约条款

护城河耐久性评估基于公开来源中的竞争轨迹,属于定性判断。目前没有独立评估比较 1X AI 模型质量与竞争对手;内部训练数据的数量和质量未公开披露。

[CP025, CP026, CP027, CP028, CP029, CP030]

3.5 竞争尽调缺口与开放问题

若干竞争情报缺口限制了对护城河评估的信心。第一,Agility 的 Digit Amazon 部署实际合同经济性——数量、价格、排他条款和续约结构——未公开披露,无法判断 Agility 是否已跑通可持续单位经济模型,还是仍在为数据和学习补贴部署。第二,Tesla Optimus 内部部署指标——单台每日完成任务数、自主率与人工接管比例、缺陷处理——未披露,无法客观评估 Tesla 距商业部署就绪还有多近。 第三,Figure AI 自 BMW 公告以来的进展没有公开详尽记录;媒体报道显示 BMW 合作已投入运营,但没有生产指标。第四,在不了解 $16K 价位、且没有重大 AI 软件集成时 Unitree G1 可达到何种自主水平,以及西方安全和监管要求是否会为本土运营商形成天然护城河之前,Unitree G1 的竞争风险很难量化。 第五,1X 没有披露 ADT 和 Everon 合同的具体金额或数量承诺,无法从 EVE 机队计算已签约 ARR。这一点重要,因为竞争护城河分析取决于 1X 是在从部署领先中获取经济价值,还是只是在补贴价格下积累数据。[CP033, CP034, CP035]

Chapter 04

04财务情况

4.1 收入来源与定价模型

1X Technologies 的收入主要来自两款已商业化产品:EVE,一款以机器人即服务(RaaS)月度订阅部署的双足商业机器人;NEO,一款面向家庭的仿人机器人,可直接购买硬件,并可能附带订阅层。EVE 是当前收入引擎。公开披露的 EVE RaaS 定价约为每台每月 $3,500 至 $5,000,使其成为截至 2026 年纯创业公司中定价最具体的商业仿人机器人产品。两个锚定企业部署——ADT(140 台)和 Everon(约 250 台)——构成公司的核心经常性收入基础,合计部署机队约 390+ 台。若按 390 台、每台每月 $3,500 计算,毛收入运行率对应的月经常性收入约为 $1.37 million;视实际合同价格、使用率和折扣而定,意味着年化运行率约 $15–25 million。公司没有披露官方收入数字。NEO 于 2024 年宣布,零售价约 $20,000。NEO 的订阅层(据报道约 $499/月)曾被市场推测,但 1X 尚未正式确认。RaaS 模式通常在合同期内按月确认收入,收入质量优于硬件一次性销售模式。专业服务和入门部署构成第二类但未披露的收入来源。公司没有披露按细分市场划分的收入、毛利率,或已确认收入与递延收入。[CI005, CI006, CI007, CI008, CI009, CI010]

收入来源表
收入来源机制单位当前数值 / 状态质量尽调问题
EVE RaaS 订阅机器人月度经常性租赁费每台机器人每月~$3,500–5,000/month(公开提及)中 — 标价已知;实际成交价未披露确认 ADT 与 Everon 的实际成交价、批量折扣和合同期限
NEO 硬件购买一次性硬件买断每台~$20,000(公布标价)低 — 标价已披露;出货量未确认确认商业出货状态、订单积压和截至目前已确认收入
NEO 订阅(推测)月度软件 / 数据订阅附加项每台机器人每月~$499/month(推测;1X 未确认)低 — 截至 2026 年,1X 未正式确认确认订阅层是否上线、包含内容和当前附加率
专业服务 / 入驻按项目收取部署和集成费每次项目未披露低 — 没有公开定价或规模数据获取 SOW 样本、每个项目确认收入和投入人数
技术授权(潜在)向 OEM 或软件合作伙伴授权 API 或 IP按交易或版税截至 2026 年无披露交易未知 — 属于推测;没有公开证据显示授权计划已启动确认是否存在授权或 OEM 协议,或正在谈判
未来产品线(NEO Gen 2+)下一代硬件销售加经常性服务TBD截至 2026 年尚未商业化未知 — 路线图产品;尚无财务贡献获取产品路线图、定价目标和预计发布时间

收入和定价数据来自公开公告、媒体报道和行业基准。1X 未披露官方收入数据。除非注明,所有美元金额均为估计。状态反映可得公开来源下截至 2026 年 5 月的情况。质量评级反映证据强度,不代表收入规模。

[CI005, CI006, CI007, CI008, CI009, CI010]
定价 / 变现表
产品标价合同结构定价依据已知折扣来源
EVE(标准 RaaS)~$3,500–5,000/月月度订阅(估计期限 12–36 个月)按机器人计大批量机队很可能有折扣;具体未披露1x.tech 商业页面;行业媒体报道
EVE(ADT 企业协议)未披露多年企业合同(估计约 2 年)按机器人机队计140 台规模几乎肯定采用机队定价;金额未知媒体报道(Axios、TechCrunch);官方未披露
EVE(Everon 协议)未披露长期托管服务协议(估计)按机器人机队计未披露媒体报道;Everon 公开表态
NEO(标价)~$20,000 一次性硬件一次性购买按台计早鸟或批量折扣未确认1x.tech 公告;Engadget 报道
NEO(订阅附加项)~$499/月(推测)月度软件 / 数据订阅按机器人计不适用;产品未确认行业推测;1X 未确认
Halodi 时期硬件(遗留)已停产遗留产品;已无在售N/AN/A历史参考;Halodi Robotics 备案文件

定价数字基于公开公告和媒体报道。企业客户的合同条款、折扣和实际成交价格未公开。标价与实际价格之间的差距是关键尽调项。推测行均已标注。

[CI005, CI006, CI007, CI008, CI009, CI016]
FI001: 收入模型桥

定性流程展示 1X Technologies 截至 2026 年如何把客户活动经部署、订阅和服务层转成估算收入和毛利。标注处的财务规模为估算;公司未披露官方收入数据。

所有财务规模均为估算,来自公开定价披露、机队规模报道和行业 COGS 基准。1X 未披露收入、毛利率或 COGS。节点连接为定性关系。毛利估算假设 COGS 按 24 个月合同期摊销。

[CI005, CI008, CI009, CI010, CI011, CI017]

4.2 成本结构、毛利率与单位经济模型

1X 的成本结构具有早期机器人硬件公司的典型特征:R&D、制造和现场支持推高资本强度;在当前规模下,毛利率很可能为负到小幅为正。基于第三方硬件基准和竞争仿人平台类比,EVE 硬件 COGS 估计为每台 $15,000–25,000,但官方 COGS 数据未披露。按每台机器人每月 $3,500–5,000 的 RaaS 定价,在乐观假设下,硬件成本约 3–7 个月可收回;这还未计入现场服务、软件维护和折旧成本。每台已部署 EVE 的真实毛利率未知。行业基准显示,第一代仿人硬件在当前批量下毛利率为负到低个位数正值;单批规模超过 500 台后会显著改善。R&D 支出可能占运营支出的 40–60%,这与 1X 对自主 AI 开发和快速产品迭代周期(EVE 到 NEO)的强调一致。企业机器人销售和 GTM 成本很高:典型企业销售周期为 6–18 个月,获客成本(CAC)估计为每账户 $50,000–200,000,取决于交易规模和部署复杂度。因此回本周期被拉长,在规模化期间形成资本拖累。营运资本管理是关键风险:收入发生前必须先制造机器人库存,RaaS 合同还可能需要试点期,从而推迟现金回收。1X 没有公开披露毛利率、COGS、CAC 或回本周期。[CI011, CI014, CI016, CI021, CI022, CI027]

单位经济性表
指标数值或空值置信度重要性尽调要求
EVE 单台毛利率(月度)Unknown决定 RaaS 模式能否扩张、能否走向盈利的关键变量要求提供单台 COGS 拆分:硬件、现场服务、软件、保修准备金
EVE 单台硬件 COGS(估计)$15,000–25,000 估计决定硬件回本周期能否匹配 $3,500–5,000/月的 RaaS 定价按生产批次获取 BOM 和制造成本表;确认合同制造商
已部署机器人月收入~$3,500–5,000ACV 代理指标;驱动总 MRR 和队列分析确认 ADT 与 Everon 合同的实际成交费率与标价差异
获客成本(CAC)Unknown决定 GTM 效率,以及新增企业客户所需资本获取销售周期长度、按职能划分的员工数,以及按账户类型混合后的 CAC
CAC 回本周期Unknown扩张期需要多少资本由它决定;回本周期长会加重融资依赖需要 CAC + 实际 ACV 数据;在多组 COGS 假设下建模回本
年度合同价值(ACV,ADT 估计)~$5.9M–8.4M/年 估计(140 台 × $3,500–5,000/月 × 12)最大企业锚定客户;驱动合同结构和续约风险分析获取实际 ACV、续约条款,以及扩大台数的选项

1X 未公开财务数据,因此大多数单位经济性指标仍未知。估计值来自硬件成本基准、可比公司数据和行业分析。置信度反映证据质量。所有估计都必须用资料室材料验证。

[CI005, CI008, CI009, CI010, CI014, CI021]
FI002: 单台经济模型桥

定性流程从每台已部署 EVE 的月度 RaaS 收入出发,经过估算成本层,推导单台毛利率。所有成本节点均为估算;实际数据未公开。

所有成本节点均为估算,来自人形机器人硬件和服务交付的行业基准。1X 未披露 COGS、现场服务成本或毛利率。估算仅作示意。实际经济性必须通过数据室 COGS 瀑布确认。

[CI005, CI014, CI021, CI027, CI032, CI034]

4.3 公开牵引力与商业里程碑

1X 最具体的公开财务牵引信号是已部署机队规模。截至 2026 年,ADT 的 140 台 EVE 部署是北美已确认的最大商业双足仿人机器人合同。Everon 部署又增加约 250 台,使已确认商业机队总量达到约 390+ 台。这让 1X 成为纯创业公司中已确认商业仿人机器人部署的领先者,领先 Agility Robotics(Amazon/GXO 的 100+ 台 Digit)许多,也远高于仅披露试点规模部署的 Figure AI 和 Apptronik。公司没有披露收入或 ARR 数字,也没有提供 MRR 队列分析或收入增长轨迹。截至 2026 年,公司没有确认 NEO 是否已从预购转入批量商业发货。公开信号显示公司有活跃商业牵引——拿下企业合同、从挪威扩张到美国,以及 OpenAI 和 NEA 等战略投资方参与——但这些信号都没有转化为已披露财务结果。私有指标缺口包括:按客户队列划分的 ARR/MRR、客户留存率、续约率、平均合同价值、单台已部署机器人使用率,以及总留存与净收入留存。这些指标对评估收入质量和可扩展性至关重要,必须通过数据室获取。[CI008, CI009, CI010, CI015, CI017, CI024]

公开财务缺口表
缺口影响尽调路径
收入和年经常性收入(ARR)未披露无法评估收入质量、增长轨迹、队列表现或 MRR 构建向资料室索取 2023–2026 年月度收入和 ARR/MRR 数据
毛利率未披露无法建模单位经济性盈亏平衡路径,或规模化后的资本效率要求提供经审计或管理层审阅的损益表(P&L);按产品和批次拆分 COGS 瀑布
COGS 拆分未公开无法验证硬件单位经济性、学习曲线或规模驱动的毛利改善索取 BOM、制造成本表、单台现场服务成本和保修准备金
烧钱速度未确认现金跑道估计存在 ±13 个月不确定性;融资风险是二元结果索取月度现金流量表和截至最近季度的银行余额
Series C 融资状态未披露如果 Series C 尚未交割且烧钱速度为 $10M+/月,公司将在 2026 年面临融资悬崖确认 Series C 交割状态、领投方、term sheet 和已承诺资金

列出截至 2026 年 5 月 1X 未披露的关键财务指标。每个缺口都是重大未知项,会妨碍可靠的财务建模。任何投资决定前,尽调路径所列为最低数据要求。五个缺口的严重性均为阻断或重大。

[CI003, CI004, CI011, CI017, CI023, CI036]
FI003: 财务估算区间

关键财务估算的有来源支撑区间:每台机器人月度 RaaS 收入、估算月烧钱速度、估算现金跑道、每台 EVE 估算硬件 COGS,以及 Series C 融资目标。区间反映缺少已披露财务数据带来的不确定性。

所有区间均为估算,来自公开定价、媒体报道和行业基准。1X 未官方披露财务数据。收到数据室材料后应更新区间。下限代表乐观情景;上限反映公开证据有限但数值更高的情景。

[CI003, CI004, CI005, CI012, CI013, CI014]

4.4 资本充足性与融资依赖

1X Technologies 已确认融资约 $237 million:Halodi 时期的 $23.5 million Series A2、由 EQT 领投且 OpenAI 和 NEA 参投的 $100 million Series B,以及 2025 年 1 月来自未披露投资方的 $100 million 轮次。据报道,公司正瞄准约 $1 billion 的 Series C,但尚未披露关闭日期。烧钱速度未公开披露。基于 1X 估计 150–200 名全职员工、R&D 强度、制造支出,以及挪威和美国多地点运营,月度现金消耗估计为 $5–15 million。在该区间内,若假设 2025 年 1 月 $100 million 融资构成当前大部分资金,公司的估计现金跑道截至 2026 年 Q2 约为 7–20 个月。现金跑道对烧钱假设高度敏感:若每月 $5 million,可延伸至 2027 年;若每月 $15 million,公司可能需要在 2026 年末前关闭 Series C。公司没有披露债务、项目融资或授信额度。若无法获得实际烧钱速度、现金余额和 Series C 状态,资本充足性判断存在重大不确定性。[CI001, CI002, CI003, CI004, CI012, CI013]

资本充足性表
项目数值备注
已确认累计融资~$237M+ 已确认包括 Series A2($23.5M)、Series B($100M,2023,EQT/OpenAI/NEA)和 2025 年 1 月轮($100M,投资方未披露)。不包括 Halodi 时期未披露的 pre-Series A 轮次。历史轮次细节见第 1 章(公司概况)。
估计月度烧钱速度$5–15M/月 估计根据估计员工数(150–200 名 FTE)、机器人创业公司的研发强度基准、多地运营(挪威 + 美国)和制造运营支出推导。1X 未披露。
估计现金跑道(自 2025 年 1 月融资起)~7–20 个月估计(自 2025 年 1 月起)基于 2025 年 1 月融资 $100M 与 $5–15M/月烧钱速度对比。按 $5M/月:约 20 个月(至约 2026 年 9 月);按 $15M/月:约 7 个月。对烧钱假设高度敏感。
计划资金用途研发、制造扩张、商业扩张依据 Series B 和 2025 年 1 月新闻披露。未披露具体预算拆分。
Series C 目标 / 下一轮触发条件目标约 $1B;未披露交割日期媒体报道称,公司目标是约 $1B 的 pre-IPO 融资。截至 2026 年 5 月,公开信息未确认 term sheet、投资方或交割日期。

资本和烧钱数字基于公开披露的融资轮次和第三方行业估计。实际现金余额、烧钱速度和现金跑道未公开。所有估计不确定性很高,必须用经审计或管理层审阅的财务数据验证。逐轮历史融资时间线见第 1 章。

[CI001, CI002, CI003, CI004, CI012, CI013]
FI004: 资本强度 / 现金流图

矩阵评估 1X Technologies 在四个成本维度上的资本强度和现金流暴露(R&D 资本开支、制造扩产、销售与 GTM、营运资本),覆盖三个时间窗口。定性评级基于公开证据估算相对强度。

定性评级(高 / 很高 / 中)来自公开披露、机器人硬件创业公司的行业基准以及同业公司分析。1X 未披露财务数据。制造扩产被评为很高风险,原因是双足人形机器人大规模生产资本强度高,且公司未披露制造合作伙伴或规模目标。

[CI003, CI012, CI022, CI028, CI029, CI037]

4.5 财务结论与尽调阻断项

1X Technologies 讲出了一个有吸引力的商业牵引故事——纯创业公司中最大的已确认仿人机器人机队、企业级客户,以及可信的 RaaS 定价模型——但任何已披露财务数据的缺失,严重限制了财务尽调。没有 ARR、MRR 队列数据、留存率,以及按产品和部署队列划分的毛利率,就无法评估收入质量。单位经济模型依赖估计 COGS;在当前规模下,这可能让现有 RaaS 模式毛利为负。资本充足性不确定:据报道公司正在寻求 $1 billion Series C,但关闭时间和投资方构成尚未确认,形成二元融资风险。通往盈利的毛利路径需要靠制造规模和设计优化降低 COGS,靠自主能力提升降低现场服务成本,并扩张毛利以覆盖 R&D 和 G&A 开销。上述里程碑没有任何公开披露时间线或指标。任何资本配置决策前必须解决的尽调阻断项包括:已确认月度烧钱和现金余额,经审计或管理层审阅的收入与毛利数据,按客户划分的 ARR 和 MRR 队列分析,逐产品 COGS 瀑布图,Series C 融资状态,以及任何债务或或有负债。总结而言,1X 商业上有差异化,但财务上高度不透明。投资逻辑完全取决于必须通过正式数据室取得的私有财务证据。[CI003, CI004, CI011, CI017, CI023, CI036]

4.6 重点证据

Chapter 05

05产品与技术

5.1 产品线、功能与技术概览

1X Technologies 正在商业化两条机器人平台,分别面向不同市场和部署场景。EVE 是公司已经产生收入的轮式人形机器人,高约 140cm、重约 70kg。它在企业环境中连续运行,主要为 ADT 和 Everon 执行商业安防巡逻,任务包括非工作时段巡查、开门和搬运负载。EVE 的轮式底盘让移动更稳定,不必处理双足设计的平衡复杂度,因此更容易落地 24/7 生产部署。截至 2026 年初,约 390 台 EVE 已商业部署。 NEO 是面向消费家庭市场的双足继任平台。2024 年 8 月推出的 NEO Beta 是首个完整双足原型,验证了具身 AI 架构。2025 年 2 月发布的 NEO Gamma 是一次明显的设计跃迁:它整合了与 Kind Humanoid(2025 年 1 月被收购)合作开发的 3D 编织全身套装,让机器人具备类人的运动范围和适合家庭环境的柔触表面。NEO Gamma 高约 163cm,采用 100Hz 主动平衡控制环,并部署由动作捕捉参考数据训练出的强化学习策略。预购于 2025 年 10 月开启,买断价 $20,000,订阅价每月 $499,目标在 2026 年交付消费者。 从 EVE 到 NEO 的产品切换,体现出一条战略弧线:从商业上已验证的企业级实用机器人,走向通用家庭人形机器人。两个平台共享底层自研执行器架构和端到端神经网络控制系统,但 NEO Gamma 增加了双足行走难题;EVE 通过轮式设计避开了这部分复杂度。Kind Humanoid 收购引入了联合创始人 Christoph Kohstall 作为工程贡献者,也带来工业设计师 Yves Béhar 的人体工学机身概念,现在已整合进 NEO Gamma 的工业设计。相对于完全自研节奏,这次收购显著加快了家用形态迭代。[CE001, CE002, CE003, CE004, CE005, CE006]

产品模块 / 资产矩阵
模块 / 产品用户 / 买方技术状态差异化尽调缺口
EVE(轮式人形机器人)企业安保、物流运营商商业化,已部署约 390 台自研低阻抗执行器;24/7 运行;已在 ADT/Everon 验证未公开披露 MTBF 或 uptime SLA
NEO Beta(双足原型)内部研发 / 研究合作伙伴已弃用;由 NEO Gamma 取代首个双足形态;验证具身 AI 架构已被取代;未发布公开技术基准
NEO Gamma(双足消费级)家庭消费者;住户预购已开放;目标 2026 年交付3D 针织外套;100Hz 平衡控制;RL 训练策略无自主性基准;生产批量未披露
具身 AI 引擎内部使用(训练 NEO/EVE);数据资产持续开发;已用于生产来自 EVE 机队的真实世界数据飞轮;优势复利训练数据量和模型版本节奏未披露
Kind Humanoid 底盘内部使用(NEO Gamma 设计)2025 年 1 月已整合Yves Béhar 人体工学设计;柔软触感、适合家庭的安全表面相比原 Kind 平台的整合完整度未验证

技术状态和差异化基于 1X 官方公告(SE001、SE002、SE026)、IEEE Spectrum 上手评测(SE005、SE017),以及 TechCrunch 和 MIT Technology Review 报道。尽调缺口代表截至 2026 年第二季度尚未公开披露的数据。

[CE001, CE002, CE005, CE012, CE023]
工作流 / 使用场景表
用户任务当前工作流1X 方案可衡量收益限制
非工作时段商业安保巡逻人类保安 $25–40/小时;存在疲劳和覆盖缺口EVE 以 RaaS 模式收费 $3,500–5,000/月;连续巡逻24/7 覆盖;相比保安班次预计降低人工成本事件响应约 70% 需要人工监督;无使用武力权限
家务协助(家庭使用)日常体力劳动;靠个人投入或家政服务NEO Gamma(2026 年交付);遥操作并部分自主常规任务节省人工;24/7 可用目前约 70% 依赖遥操作;交付时间线有风险
商业设施物流任务人工拣选、物品搬运、跑腿EVE 通过神经网络指令执行多任务任务一致性;消除重复任务中的工人疲劳载荷不及专用工业机器人;复杂布局需要遥操作
具身 AI 训练数据采集脚本化实验室演示、合成仿真;真实世界保真度有限借助已部署 EVE 机队做 VR 遥操作;真实世界具身演示高保真真实世界数据分布;复利数据飞轮数据质量取决于遥操作员技能;随员工数线性扩张

工作流描述来自 ADT/Everon 部署报道(SE002、SE008)、1X 产品公告(SE001、SE025)和科技媒体报道(SE015、SE022)。人类保安成本估计采用行业基准;RaaS 定价来自前一章财务分析。

[CE006, CE007, CE010, CE021]
FE001: 产品架构图

六层技术栈展示 1X 的架构,从实体硬件延伸到云基础设施。图中从底部自研执行器到顶端云 GPU 训练,体现垂直一体化设计;具身 AI 数据引擎连接硬件层和智能层。

5.2 核心技术栈与架构

1X 的技术架构围绕一个端到端神经网络搭建,机器人所有行为都由它控制,不依赖硬编码任务逻辑或模块化行为树。这让 1X 有别于传统机器人公司;后者通常把控制拆成感知、规划和执行模块。整条动作栈——从传感器输入到电机命令——都由单一学习模型联合优化,机器人因此能跨感知条件和任务变化泛化;在模块化系统里,这类能力往往要大量手工工程才能拼出。 具身学习数据引擎是数据生成骨干,这一概念由 1X 的 AI 副总裁 Eric Jang 提出。人类远程操作员通过 VR 头显控制部署在真实商业环境中的 EVE,生成物理场景里的高保真任务示范。这些示范随后用于训练一代又一代神经网络,方法结合模仿学习(行为克隆)和强化学习。飞轮效应很强:每一次 EVE 商业部署都同时生成真实世界训练数据,反过来改进 NEO 的 AI 模型。到 2025 年中,ADT 和 Everon 的 EVE 部署已经在为 NEO Gamma 的策略训练管线供数。 训练管线利用动作捕捉参考数据来初始化行走强化学习策略——这一路径来自基于物理的角色动画和人形控制学术研究。1X 研究人员发表的 ArXiv 论文(例如 2309.01694 关于全身控制、2405.01813 关于灵巧操作)确认,公司方法建立在同行评议方法之上。软件栈由内部开发;未发现公开资料提到 ROS 等行业标准机器人操作框架,这与 1X 的立场一致:神经网络路线让传统感知-规划-执行栈在架构上不再必要。github.com/1x-technologies 上的 GitHub 仓库显示工程活动活跃,提交和贡献者持续出现,为开发节奏提供了轻量级开发者信号验证。[CE009, CE010, CE011, CE019, CE020, CE023]

技术 / 运营架构表
层级 / 组件作用依赖风险
端到端神经网络(控制 AI)从传感器到执行器控制全部机器人行为GPU 训练基础设施;大型算力集群单一模型故障可能影响整个机器人行为;无模块化回退
自研低阻抗电动执行器关节运动;低阻抗带来内在安全合规在 Hayward, CA / 挪威 Moss 自有制造消费级规模下存在制造良率风险;核心 IP 集中
3D 针织外套(NEO Gamma)自然运动包络;柔软触感、适合家庭的安全表面专业纺织制造供应商可能存在单一供应源风险;纺织生产规模化未验证
VR 遥操作系统训练数据采集和监督的主要机制人类遥操作员 + VR 硬件(HMD)随员工数线性扩张;遥操作员错误影响数据质量
100Hz 平衡控制器(NEO Gamma)为双足行走提供实时稳定性机器人计算单元上的定制嵌入式固件安全关键实时代码;固件缺陷可能导致跌倒
传感器套件(摄像头、IMU、本体感知)为 AI 输入提供感知和状态估计第三方 COTS 传感器组件依赖通用商品供应;传感器故障会传导至神经网络

架构层级根据 1X 官方文档(SE001、SE002)、技术出版物(SE003、SE011、SE021)和第三方科技媒体(SE029)推断。依赖和风险评估是分析师根据公开披露作出的推断;1X 尚未发布完整系统架构文档。

[CE011, CE015, CE016, CE020, CE032]
FE002: 客户工作流 / 运营流程

运营数据流展示客户部署如何经 VR 遥操作员管线生成具身 AI 训练数据,形成把模型改进反馈到已部署机器人的飞轮。同时捕捉商业闭环和 AI 训练反馈闭环。

5.3 知识产权、防御性与技术护城河

1X 最具防御性的技术资产,是公司十年历史中(最初名为 Halodi Robotics)积累出的自研电动执行器技术。执行器采用低阻抗设计,天然具备机械柔顺性——机器人遇到冲击时吸收力量,而不是硬顶力量。这能在无需复杂实时力控软件的情况下,降低人机交互场景中的受伤风险。US11498217B2 专利覆盖了最初以 Halodi Robotics 名义提交的核心执行器技术,多件延续专利仍在积极审查中。这一专利组合为试图复制 1X 特定执行器几何和控制方法的竞争对手设置了有意义的壁垒,不过更广义的电动执行器设计空间仍竞争激烈。 除了执行器 IP,1X 还通过训练数据引擎积累数据优势。竞争对手的 AI 训练往往依赖合成数据或脚本化环境,1X 的 EVE 商业部署则持续从真实企业环境生成真实世界交互数据。没有同等商业部署,竞争对手很难复制这种真实世界数据分布。数据飞轮巩固了 1X 在企业机器人领域的先发位置,也构成硬件 IP 之外的二阶护城河。 第三层防御性来自垂直一体化架构:1X 自己设计执行器、开发 AI 栈,并在 Hayward 和 Moss 至少部分自行制造。全栈所有权降低关键部件单一供应风险,也让软硬件能更紧密协同设计,类似 Apple silicon 的整合模式。相比使用商品化执行器的竞争对手(例如 Unitree 使用 Robstrider,或 Boston Dynamics 的液压路线),1X 的自研驱动带来差异化的扭矩密度和柔顺性特征。产品成熟度与能力地图(FE004)展示了 1X 在各能力维度上相对于自身 2027 路线图目标的位置。[CE015, CE016, CE017, CE018, CE022, CE023]

FE003: 关键依赖图

有向无环图映射 1X Technologies 在资本、AI 合作、硬件供应和基础设施上的关键外部依赖。展示 OpenAI 合作关系和半导体供应链瓶颈带来的集中风险。

FE004: 产品成熟度 / 能力图

能力成熟度矩阵从五个维度给 EVE 和 NEO Gamma 打分,并对照 2027 年目标状态。分数为定性判断(强 / 部分 / 弱),依据截至 2026 年 Q2 的公开性能数据和独立技术评估。

5.4 开发路线图与量产准备度

1X 的开发路线图沿分阶段弧线推进:先用企业市场验证轮式机器人,再走向可进入家庭的双足人形机器人;每个阶段既验证技术,也为下一阶段生成训练数据。2024 年 8 月 NEO Beta 发布,在完整尺寸上验证了核心双足具身 AI 架构。2025 年 2 月 NEO Gamma 发布,展示了整合 Kind Humanoid 机身收购成果的量产意向设计。2025 年 10 月开启的预购以 $20,000 买断 / 每月 $499 订阅定价,确认了 2026 年交付的商业意图。截至 2026 年 Q2,除 2026 年目标外,NEO Gamma 的交付时间和批量规模尚未公开确认;公司也未披露生产速率数据。 2026–2027 年路线图围绕三条并行主线:(1)扩大 NEO Gamma 制造能力以履行预购订单;(2)把 NEO 的自主能力从当前约 70% 远程操作依赖,推进到能在一组明确家庭任务上具备有意义的自主性;(3)开发具备 π 式多任务泛化能力的基础模型,减少逐任务训练需求。基础模型工作使 1X 贴近行业大方向(Physical Intelligence 的 π0、Figure 的 Helix 等):通用机器人策略可被微调,而不必每个任务从零重训。 NEO Gamma 的量产准备度有实质执行风险。定制执行器和 3D 编织套装要规模化制造,供应链部件并非完全商品化。负责 EVE 的挪威 Moss 制造设施,尚未被公开确认具备消费级批量生产 NEO Gamma 的能力。2025 年夏季启用的 Palo Alto 80,000 sq ft 设施主要是工程和 AI 训练中心。投资人和预购客户应把制造合作伙伴公告作为按时交付的关键领先指标。[CE008, CE033, CE034, CE021, CE026, CE036]

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
2024 年 8 月NEO Beta 双足原型发布已完成首次在完整尺寸上验证双足具身 AI 架构SE015, SE019
2025 年 1 月收购 Kind Humanoid 并整合底盘已完成人体工学家庭形态设计提速;Yves Béhar 美学SE026
2025 年 2 月NEO Gamma 发布,配备 3D 针织外套和 100Hz 平衡控制已完成面向量产的消费级设计亮相;预生产设计冻结SE001, SE017
2025 年 10 月NEO Gamma 开放预订,价格 $20K / $499/月已完成商业化意图确认;预订定金带来收入可见性SE025, SE018
2026 年(目标)NEO Gamma 消费级交付;首批家庭部署进行中收入拐点;首批用于训练的真实世界消费者数据SE025, SE018, SE019
2026–2027多任务泛化基础模型(π 式架构)开发中如果成功,将降低单任务训练开销,并大幅拓宽使用场景SE027, SE029

时间线基于 1X 官方公告(SE001、SE026)、TechCrunch(SE025)、CNBC(SE018)、Fortune(SE019)和 IEEE Spectrum(SE027)。未来日期是目标里程碑,并非已确认承诺。2026 年消费级交付的生产批量和制造合作伙伴仍未披露。

[CE008, CE012, CE001, CE024, CE025, CE022]

5.5 技术风险、信任与安全架构

1X 的信任与安全架构立在三根支柱上:低阻抗执行器柔顺性带来的物理安全,远程操作监督带来的运营安全,以及实体急停机制带来的产品安全。低阻抗执行器设计是一种结构性安全属性——不同于接触时传递完整电机扭矩的刚性执行器,1X 执行器在碰撞时吸收能量,降低人在近距离时承受的伤害力量。这一设计哲学适合消费环境,但截至 2026 年 Q2,尚未通过已发布的 ISO 13482(个人护理机器人安全)或 ISO 10218(工业机器人安全)认证验证。 当前把远程操作作为复杂任务安全网,也引入了自身风险:远程操作员误判、延迟造成的判断失误,以及人类视觉观察家庭环境带来的隐私问题。WSJ 2025 年 10 月报道点名了家庭场景远程操作带来的消费者隐私担忧;之后 1X 在远程操作视频流中加入了同意控制和可识别内容模糊处理。不过,这些控制的技术实现以及远程操作视频的数据留存政策,尚未被详细公开披露。 网络安全是正在浮现的风险维度。配备摄像头、麦克风并持续联网的机器人,为未授权访问家庭数据创造了攻击面。1X 尚未发布网络安全白皮书,也未披露渗透测试结果。监管风险也在成形:EU AI Act 对高风险 AI 系统的分类可能适用于自主机器人功能,进入欧盟市场前需要合规评估和公告机构审查。美国 FTC 商业监控规则适用于采集视频 / 音频数据的机器人。OSHA 法规(29 USC 652)约束企业场景中的 EVE,1X 尚未就其商业部署公开回应这一问题。信任 / 质量 / 合规表(TE004)按当前控制状态汇总了这些缺口。[CE016, CE027, CE037, CE038, CE040, CE026]

信任 / 质量 / 合规表
控制 / 认证状态范围缺口
物理急停已实现所有机器人型号硬件停止设计规格未公开披露
低阻抗顺应性(结构安全)已实现(架构层)所有机器人(执行器层)截至 2026 年第二季度,尚未独立通过 ISO 10218 或 ISO 13482 认证
遥操作同意控制2025 年 10 月后新增(WSJ 报道后)NEO 消费级设备遥操作视频流的数据留存政策未公开披露
OSHA 合规(企业)推定合规;未披露正式审计美国商业设施中的 EVE未公开提供 OSHA 安全审计或合规文件
GDPR 合规部分合规(欧洲运营)欧盟部署的 EVE 和 NEO 用户机器人采集数据的数据控制者 / 处理者分类不清楚
EU AI Act 高风险分类审查中;合规路径未披露自主 AI 系统功能可能需要符合性评估和公告机构审查

合规状态根据 1X 公开披露(SE028、SE001)、Engadget 安全评测(SE028)、Wired 隐私报道(SE023)和 MIT Technology Review(SE006、SE029)评估。缺口表示缺少官方合规声明,并非确认不合规。ISO 13482 覆盖个人护理机器人;ISO 10218 覆盖工业机器人。

[CE016, CE027, CE037, CE038, CE040]
Chapter 06

06客户情况

6.1 具名客户与企业部署状态

1X Technologies 有两个已确认达到商业生产规模的企业客户:财富 500 强安防服务公司 ADT Inc.,以及北美领先商业安防集成商 Everon(前身为 ADT Commercial,2022 年 11 月更名)。ADT 运营约 140 台 EVE,用于商业设施的非工作时段安防巡逻,包括仓库、办公园区和企业园区。Everon 在其商业客户群中部署了约 250 台 EVE,是已知单一客户中最大的 EVE 部署。两者合计约 390 台 EVE 处于活跃商业运行状态,按每台机器人每月 $3,500–5,000 估算,对应年化收入运行率约 $15–25M。基于机队规模、已确认的 RaaS 合同结构,以及跨度超过 12 个月的媒体报道,这两段关系更像生产部署,而非试点。 除 ADT 和 Everon 外,1X 披露了两段较小的客户关系。据报道,挪威 Sunnaasen Hospital 正在用 EVE 执行医疗机器人任务,但规模和运营结果尚未公开披露。几所挪威大学已知将 EVE 用于学术机器人研究,这带来国际品牌露出和研究渠道数据,但商业价值有限。截至 2026 年 Q2,尚无其他企业客户被公开宣布。因此,企业客户组合高度集中:两个客户估计贡献 95%+ 已部署机器人;除试点规模的 Sunnaasen 安排外,尚未公开宣布扩张到物流、酒店或医疗等相邻垂直领域。 ADT 关系至少可追溯到 2022 年,早于 Halodi 更名为 1X,说明这是一段长期商业合作,并经受住了公司战略重定位。Everon 合作关系于 2024 年 2 月通过 BusinessWire 宣布。TechCrunch 和 Axios 在 2025 年 1–2 月的报道中确认了两段关系,Fortune 在 2025 年 4 月报道中称 ADT 的运营结果正面。[CU001, CU002, CU003, CU004, CU005, CU006]

客户分群表
客群买方 / 用户 / 付款方使用场景规模收入 / 战略价值尽调缺口
企业安保ADT、Everon(安保运营商)非工作时段商业设施巡逻总计约 390 台(ADT 140 台,Everon 约 250 台)估计 ARR $15–25M;Fortune 500 锚定参考客户未披露续约率、合同条款或利用率数据
消费者家庭个人房主 / 家庭家务、跑腿协助预购已开放;数量未披露$20K/台或 $499/月;TAM 很大但未验证未披露预购数、转化率或取消数据
医疗健康(挪威)Sunnaasen Hospital 员工 / 患者医疗任务协助(试点)小规模;估计 <10 台研究 / 试点价值;商业 ARR 很小试点还是生产未确认;规模和结果未披露
研究 / 学术挪威高校研究人员机器人研究与实验小规模;估计 <20 台战略性数据采集;无商业收入未披露商业化路径或时间线

客群定义来自 1X 官方页面(SU003)、TechCrunch(SU004)和 Axios(SU006)。机队数量来自已确认媒体报道(SU001、SU002、SU007)。收入估计是分析师基于已披露 RaaS 定价推导,1X 未确认。医疗和学术规模估计来自稀疏媒体提及的推断。

[CU001, CU002, CU003, CU006, CU007]
具名客户验证表
客户细分市场部署 / 使用场景生产部署 / 试点披露结果限制
ADT Inc.(Fortune 500)企业安防约 140 台 EVE;非营业时段商业设施巡逻生产部署(ADT 新闻稿、独立媒体确认)安防巡逻持续运行;Fortune 2025 年 4 月报道称结果正面未公开 ROI、保安工时替代或 NPS 数据
Everon(原 ADT Commercial)企业安防集成商约 250 台 EVE;多站点商业客户生产部署(BusinessWire + Reuters 确认)在 Everon 客户群中多站点铺开未披露满意度数据、续约条款或扩张计划
挪威 Sunnaasen Hospital医疗健康(试点)小规模部署 EVE,处理医疗健康任务试点(规模和合同未确认)据报道用于辅助医疗健康任务未披露规模、合同金额和临床结果
挪威多所大学研究 / 学术EVE 用于机器人研究研究 / 试点学术机器人研究用途无商业化路径;未说明部署台数

具名客户数据来自 ADT 新闻中心(SU001)、Everon 公告(SU002)、BusinessWire(SU007)、TechCrunch (SU004)、Reuters(SU019)、Fortune(SU027)和 Sunnaasen 参考资料(SU015)。列举并不完整:可能还存在 未披露的潜在客户。

[CU001, CU002, CU006, CU023, CU024, CU017]
FU001: 客户旅程图

五阶段客户旅程,从初始认知到企业生产运行和扩张,分别映射 EVE 企业段和 NEO 消费段。展示两个客群不同的采纳路径和拐点。

6.2 客户分层、买方模式与市场进入

1X 的客户基础横跨两个结构不同的细分市场,需要不同的市场进入方法。企业段(EVE)以 B2B 直销为主,面向大型商业安防运营商,采用多年期 RaaS 合同,采购决策发生在设施层面。买方是运营副总裁、首席安全官或设施总监。交易规模很大(ADT 部署 140 台,按每月 $3,500–5,000 计算,单一客户贡献 $490,000–700,000 月度经常性收入),销售周期也长——按企业机器人行业基准约 6–18 个月。ADT 和 Everon 是理想客户原型:全国性安防运营商,托管服务网络庞大,能消化机器人机队管理开销,也能从非工作时段自动化节省成本中获益。 消费段(NEO)结构完全不同:个人买方以 $20,000 或每月 $499 做购买决策,依据是便利性、信任和生活方式契合度。家庭市场尚无任何面向消费者规模的机器人产品跑出重复购买行为。NEO 预购于 2025 年 10 月开启,使 1X 成为消费级人形机器人早期玩家,但这个细分市场仍未被验证。Wall Street Journal 在 2025 年底围绕远程操作提出的消费者隐私担忧——尤其是远程人类在引导 NEO 时可以看到家庭内部——构成该细分市场的首要需求侧风险。 企业安防机器人市场是 EVE 的主要总可用市场(TAM),这个市场仍在增长,可比玩家包括 Knightscope、Cobalt Robotics 和 Zetron,但它们都不提供双足人形形态,也不具备 EVE 提供的通用操作能力。这把 EVE 放在更高端层级,单次部署可覆盖的任务更多。相邻垂直领域——医疗支持、轻物流、酒店——都有扩张可能,但截至 2026 年尚无已确认客户牵引。[CU008, CU013, CU016, CU019, CU020, CU025]

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
EVE 已部署台数(全部为企业客户)总计约 390 台Q1 2026中(第三方报道)按披露价格推算,ARR 运行率约 $15–25M可部署总规模;同比增长率未知
ADT EVE 机队规模约 140 台2025 年 1 月(报道已确认)高(多方独立来源)ADT 是 Fortune 500 标杆客户;企业客户背书强ADT 此后是否扩张或缩减机队未知
Everon EVE 机队规模约 250 台2024 年 2 月公告高(BusinessWire + 多家媒体)单一最大部署;多站点铺开2026 年当前机队规模相对 2024 年 2 月公告是否变化未知
NEO 开放预订$20,000 / 每月 $4992025 年 10 月高(官方 + 媒体)已确认切入消费市场;定价印证高端定位预订量、订金金额和取消率未披露

机队数量来自 ADT 新闻中心(SU001)、Everon 公告(SU002)、BusinessWire(SU007)以及 相互印证的媒体(SU004、SU008、SU019)。预订数据来自 TechCrunch(SU010、SU021)。ARR 估计值由 分析师推导。截至报告日,未获得 2026 年机队更新数据。

[CU001, CU002, CU003, CU008]
FU002: 采纳 / 部署漏斗

企业采纳漏斗从可寻址安防运营商潜在客户,到已确认生产部署。展示大市场到当前部署数量的窄转化,凸显 GTM 集中挑战。

6.3 留存、重复使用与客户满意度

1X Technologies 的客户耐久性数据是一个重大缺口。企业和消费者两个细分市场都没有公开披露 NPS、客户满意度指数、续约率或流失数据。考虑到已报道的机队规模,ADT 或 Everon 没有公开 ROI 案例研究或具名客户证言尤其值得注意——这种体量的企业客户通常会为营销材料提供背书引语。沉默可能意味着严格的商业保密条款、仍在进行的性能验收期,或对技术真实世界效果的竞争敏感性。 隐性留存信号偏正面:ADT 关系跨越公司更名,已持续至少三年;Everon 在 12 个月内从合作公告推进到已确认部署;Fortune 2025 年 4 月关于 ADT 的报道也称 EVE 巡逻带来正面运营结果。截至 2026 年 Q2,尚无合同终止、机队缩减或客户投诉的公开报道。这些只是弱正面信号,不是已确认留存数据。 NEO 在交付前无法评估消费者满意度。2025 年 10 月开放预购带来媒体关注,显示消费者兴趣;但 Wired 和 WSJ 关于远程操作隐私担忧的报道带来了声誉逆风。1X 的回应是加入同意控制和数据模糊能力。这个回应能否支撑预购需求直到 2026 年交付,仍未知。公司未披露预购量、定金转化率或取消率。由于没有任何已披露留存数据,留存 / 队列图(FU004)使用的是估计值。[CU009, CU010, CU012, CU014, CU015, CU026]

留存 / 重复使用 / 满意度表
指标数值 / 状态细分市场置信度尽调待问事项
合同续约率未公开披露企业(ADT、Everon)不适用通过数据室索取续约 / 流失数据;确认多年期合同结构
NPS / 客户满意度评分未公开披露企业 + 消费者不适用向 1X IR 索取 NPS;与成熟期 RaaS 同行对标
机器人利用率(小时 / 机器人 / 天)未公开披露企业 EVE 机队不适用从客户合同索取 EVE 正常运行时间和任务完成指标
重复 / 扩张购买由机队持续存在推断(ADT 3 年以上)企业低(推断)确认 ADT 和 Everon 是否在初始订单外扩张机队
消费者预订转化未披露消费者(NEO)不适用向 1X 索取从订金到交付的转化率和取消率

本表所有指标要么 1X Technologies 未披露,要么从机队持续存在的信号推断而来。 尚未发布 NPS、续约率或利用率数据。置信度为「不适用」表示该指标是尽调待问事项, 不是已确认数值。来源:Axios(SU006)、PitchBook(SU020)、TechCrunch(SU017)。

[CU012, CU014, CU026, CU031, CU034]
FU003: 客户验证矩阵

证据质量矩阵从四个验证维度给四个客户账户评级。凸显所有企业账户公开证据基础都偏薄,并且截至 2026 年 Q2 没有已发布案例研究或满意度评分。

FU004: 留存 / 续约队列

ADT 与 Everon 的 EVE 机群合同 18 个月估计留存队列。数值仅为分析师估计——1X Technologies 尚未披露官方留存数据。估计留存偏高,来自机群持续使用信号(未见取消合同报道),而不是已确认续约数据。

6.4 市场进入可复制性与集中度风险

1X 的 EVE 市场进入模式,是通过安防集成商关系做企业直销。ADT 和 Everon 合作证明,渠道型打法——通过大型托管服务运营商销售,而不是直接卖给终端设施——可以推动 EVE 采用。相比直接从零开拓,这一模式更省资本,但也把收入和获客风险集中在狭窄的合作伙伴渠道里。如果 ADT 或 Everon 任一方停止 EVE 项目(原因可能是成本、性能不满或战略优先级调整),考虑到相对机队规模,收入影响估计分别约为企业 ARR 总额的 40–60%。 安防垂直之外的市场进入可复制性尚未确认。相邻垂直领域(医疗、物流、酒店)有不同采购周期、终端用户和合规环境。医疗尤其会引入患者安全监管要求(FDA、欧洲 CE Mark),EVE 尚未就此公开回应。1X 的 EVE 产品路线图似乎聚焦于加深安防部署,同时继续推进 NEO 开发,而不是激进把 EVE 扩张到新垂直领域。 NEO 的消费端市场进入定价($20,000 / 每月 $499)瞄准高端细分市场,可能是科技早期用户和富裕家庭。进入大众市场需要另一套定价策略。公司未披露任何分销渠道(零售、电商、白标);通过 1x.tech 直接面向消费者预购,是唯一已确认渠道。渠道扩张公告将是消费者细分市场规模化意图的关键领先指标。[CU016, CU021, CU025, CU026, CU027, CU028]

扩张与集中度风险表
扩张驱动因素集中度风险收入影响尽调路径
相邻企业垂直行业(物流、酒店、医疗健康)ADT + Everon 占已部署机队约 95%高:任一客户流失 = ARR 下降 40–60%索取新的企业客户管线;确认新垂直行业是否有已签 LOI
国际扩张(EU、APAC 市场)目前以美国为中心,另有挪威试点中:GDPR 和 EU AI Act 增加合规成本确认 EU 进入市场计划和监管准备度
NEO 消费市场爬坡(2026)至今无消费者收入;隐私担忧仍在高:如果 NEO 无法交付,消费端 TAM 打不开索取预订量和制造确认
安防集成商之外的渠道多元化渠道类型单一(仅安防集成商)中:渠道受阻可能冻结新增企业客户获取索取 ADT/Everon 之外的渠道合作伙伴管线

集中度分析来自已报道的机队数据(SU001-SU003)、PitchBook 分析(SU020)、Bloomberg 商业报道(SU022)和 Fortune 企业评测(SU027)。收入影响按披露的 RaaS 定价由分析师推导; 1X 尚未确认。

[CU011, CU021, CU027, CU028, CU032]

6.5 客户尽调缺口与反向笔记

最重大的客户尽调缺口,是缺乏任何独立验证的客户留存或满意度数据。在 RaaS 商业模式里,续约率是产品市场匹配度能否持续的核心指标。没有来自 ADT 和 Everon 的已确认续约数据,ARR 轨迹只能算推测。投资人在给增长预测背书前,应从 1X 数据室取得明确续约确认、合同延期条款和使用率数据。 隐私风险是 NEO 进入消费者市场的首要反向变量。WSJ 2025 年 10 月调查报道及 Wired 后续报道记录了消费者对远程操作员查看家庭内部的抵触。1X 的同意加模糊控制回应了表层担忧,但没有消除结构性问题:依赖远程操作的自主性,本质上需要人类观察私人空间。如果交付后出现隐私投诉或监管执法,NEO 消费者爬坡可能停滞,并连带影响企业端 EVE 部署。 次要反向笔记是客户集中度。ADT 和 Everon 约占企业机队部署的 ~95%,1X 的商业可行性实质上押在两段采购关系上。两个客户都没有公开发布前瞻扩张承诺。安防机器人市场若出现任何恶化——无论来自机器人性能不佳、保险责任担忧,还是宏观预算压力——都会不成比例地冲击 1X 收入。证据矩阵(FU003)和集中度风险表(TU005)系统汇总了这些暴露。[CU022, CU028, CU034, CU036, CU038, CU009]

Chapter 07

07风险

7.1 监管与法律风险图谱

1X Technologies 处在快速演化的监管环境中,AI 安全、消费者隐私和机器人劳动标准都在立法与执法推进中。EU AI Act 于 2024 年通过,并在 2026–2027 年进入适用阶段;它把若干自主 AI 系统应用归为高风险,进入欧盟市场前需要合规评估、技术文档、人类监督安排和公告机构审查。鉴于 NEO Gamma 面向消费者家庭使用并具备自主 AI 能力,它可能落入高风险类别。1X Technologies 尚未公开披露 EU AI Act 合规策略,也未说明是否已接触公告机构。 隐私监管带来两个不同司法辖区的风险。在 GDPR 下,1X 的欧洲运营——包括 Sunnaasen Hospital 和挪威大学部署——会生成个人数据(视频、音频、运动数据),处理这些数据需要合法依据、数据最小化和数据主体访问权。远程操作视频处理所依据的具体法律基础(同意、合法利益或合同)尚未披露。在美国,FTC 2024 年商业监控报告建立的监管框架可能直接适用于 NEO 的数据采集实践,州级隐私法(CCPA 及同类法规)又增加了额外暴露。拟议中的 AI ROBOT Act(Senate Bill 2718)若通过,将施加新的联邦机器人安全标准;但截至 2026 年 Q2,该法尚未生效。 OSHA 法规 29 USC 652 约束雇主在工人靠近机器人等工业机械时的义务。ADT 和 Everon 设施里的 EVE 商业部署要求部署方运营商维护符合 OSHA 的安全文档。1X 未发布 OSHA 合规文档,并依赖部署客户管理这一要求。NIST 的 AI Risk Management Framework 提供了自愿标准,企业客户可能开始把它作为采购条件。监管风险登记表(TR001)按严重程度排序这些暴露,风险热力图(FR001)提供跨风险的概率-影响概览。[CR001, CR002, CR003, CR004, CR005, CR006]

监管 / 法律风险登记表
规则 / 许可 / 监管机构管辖区状态发生概率严重性1X 缓释措施残余风险敞口
EU AI Act — 高风险 AI 系统一致性评估欧盟已颁布;高风险条款自 2026 年适用未披露公开合规策略如果没有 Notified Body(公告机构)认证,NEO 消费版无法在 EU 推出
GDPR — 来自遥操作视频 / 机器人传感器的个人数据EU / EEA执法活跃;1X 的欧洲运营在挪威2025 年 10 月后加入隐私同意控制数据控制者注册和 DPA 未确认;存在罚款风险
FTC Commercial Surveillance Rules — 家庭视频 / 音频采集美国监管框架已建立;执法风险仍在已加入同意控制;政策条款未充分披露数据实践若不合规,可能收到 FTC 执法通知
OSHA 29 USC 652 — 商业工作场所机器人安全美国现行监管;适用于 EVE 企业部署未公开披露 OSHA 合规文件若 EVE 造成工伤,客户面临责任敞口
AI ROBOT Act(Senate Bill 2718)— 拟议联邦框架美国已提出;截至 2026 年 Q2 尚未颁布N/A — 仅监测若颁布,机器人制造商会面临新的合规要求
NIST AI RMF — 自愿框架;可能成为采购要求美国自愿;越来越多企业将其作为采购标准未披露 NIST AI RMF 对齐文件企业客户可能要求续约时符合 NIST

监管状态依据官方监管来源:EU AI Act(SR006)、GDPR(SR002)、FTC 报告(SR001)、OSHA 法规(SR028、SR005)、AI ROBOT Act(SR003)、NIST AI RMF(SR004)。1X 缓释状态依据媒体报道 (SR007、SR013、SR022)。严重性评级为分析师判断。截至 2026 年 Q2,尚未发现已确认的监管执法行动。

[CR001, CR002, CR003, CR004, CR005, CR006]
FR001: 风险热力图

概率—影响热力图把 1X 已识别的主要风险映射到三档可能性和三档影响严重度。越靠右上,风险越关键,需要主动监控和尽调。隐私监管行动和 Series C 融资失败位于最高严重度位置。

7.2 运营、质量与网络安全风险

1X Technologies 面临一系列运营风险,当前较小的机队规模部分遮蔽了这些风险。硬件可靠性是首要未知数:约 390 台 EVE 已部署在商业环境中,但公司还没有面对数千台规模下出现的质量管理挑战。平均故障间隔时间(MTBF)数据未公开披露。行业背景令人担忧——多家其他机器人公司经历过硬件故障、安全事故和召回,最终需要昂贵的现场干预。NEO Gamma 的双足设计引入了 EVE 轮式平台没有的额外失效模式(跌倒、关节磨损、套装退化)。 网络安全是被明显低估的风险维度。EVE 和 NEO 都是持续联网设备,配有摄像头、麦克风和执行器,并在设施或家庭里运行。被攻陷的机器人可被用于监控、数据外泄、物理扰动或勒索。1X 尚未发布网络安全白皮书、渗透测试结果或安全架构文档。相比之下,Amazon Alexa 和 Google Nest 这类简单得多的联网设备都遭遇过大量安全审查。带摄像头和机械操作能力的家用机器人,攻击面严重程度要高得多。 供应链风险集中在自研执行器部件和 3D 编织纺织制造。影响训练 GPU 算力的半导体短缺,以及执行器专用部件可能存在的单一供应,会造成制造延期风险。如果关键执行器部件供应商遇到质量问题或供给中断,NEO Gamma 生产可能显著延迟。风险传导图(FR002)展示了运营风险如何传导到收入、客户和融资结果。[CR008, CR009, CR014, CR015, CR017, CR024]

运营 / 质量 / 安全风险登记表
失效模式发生概率严重性缓释成熟度残余风险敞口未解决缺口
网络安全事件 — 联网机器人摄像头 / 传感器数据外泄低(未披露公开安全架构)数据泄露、监管处罚、品牌危机未发布渗透测试结果或安全白皮书
硬件可靠性故障 — EVE 或 NEO 在规模化后出现系统性缺陷低(MTBF 数据未披露;无公开 QC 流程)召回成本、客户流失、声誉受损未披露 MTBF、现场故障率或召回流程
供应链中断 — 执行器组件或半导体短缺部分缓释(自有制造相比纯外包降低敞口)NEO 制造延迟;EVE 机队维护积压关键组件双源策略未确认
双足跌倒伤害 — NEO 跌倒砸到用户或旁观者低-中部分缓释(低阻抗顺应;物理限位)产品责任诉讼;监管暂停无独立安全测试结果;未获 ISO 13482 认证
纺织 / 3D 针织外衣退化 — NEO 外衣磨损快于预期未知(未披露材料生命周期数据)保修成本敞口;消费者不满未公开披露纺织耐久性测试或保修条款

风险评级为分析师判断,依据包括行业事故数据(SR008、SR021)、机器人安全标准(SR009、 SR010)、网络安全分析(SR017、SR019)和 1X 产品报道(SR023、SR027)。截至 2026 年 Q2,尚未公开发现 1X 已确认的运营故障或网络安全事件。

[CR008, CR009, CR014, CR015, CR028, CR038]
FR002: 风险传导图

有向图展示 1X Technologies 的主要风险事件如何传导成收入、运营、监管和估值后果,凸显隐私、客户集中度和融资风险的级联效应。

7.3 合作伙伴、依赖与客户集中风险

1X Technologies 有三类合作伙伴和依赖风险,单独看都可管理,但叠加后形成脆弱的依赖栈。第一,是对 OpenAI 的战略 AI 依赖:OpenAI Startup Fund 领投 Series A2,OpenAI 的技术合作(尤其是 AI 模型访问和集成)被公司列为 AI 战略赋能因素。如果 OpenAI 调整机器人战略、改变许可条款,或直接进入具身 AI 竞争,1X 将需要在失去 OpenAI 集成优势的情况下快速内化 AI 模型开发。没有公开披露的合同保护覆盖这一情景。 第二,是对 EQT Ventures 的资本依赖:EQT 是 Series B 领投方,并拥有董事会席位。随着 1X 接近 Series C,EQT 是否有能力、也是否愿意共同投资或支持下一轮融资非常关键。如果 EQT 的组合配置变化,或治理摩擦出现,1X 以目标 $1B Series C 估值融资的能力可能受损。 第三,也是最重大的,是客户集中度:ADT 和 Everon 合计估计贡献 1X 企业收入的 95%+。失去任一关系,都需要 12–24 个月的替代客户获取时间——前提是管线确实存在——而固定成本(制造、研发、远程操作员薪资)仍会持续。这是二元风险:不是逐步下滑,而是潜在的阶梯式断崖。依赖地图(FR003)展示了所有关键第三方依赖,以及它们如何流入 1X 运营。[CR011, CR012, CR013, CR014, CR018, CR019]

合作伙伴 / 依赖风险登记表
依赖项交易对手角色 / 集中度失效情景严重性缓释措施残余风险敞口
收入集中度ADT / Everon(约 95% 机队)主要商业收入来源不续约 / 合同取消扩大企业客户管线;垂直行业多元化未披露客户管线数据;替换周期 12-24 个月
战略 AI 合作OpenAI(Series A2 投资方 + 合作伙伴)AI 模型访问 + 战略投资者背书OpenAI 调整机器人战略或终止合作自研 AI 仍在推进;披露的 OpenAI 依赖有限合同条款未披露;转为自研的时间表未确认
主要资本提供方EQT Ventures(Series B 领投方、董事席位)资本获取 + 治理监督EQT 拒绝 Series C 跟投;董事会摩擦在 Series C 中拓宽投资人基础;接触战略投资者Series C 条款和 EQT 参与情况未公开确认
训练算力NVIDIA / 云 GPU 供应商AI 模型开发所需 GPU 训练基础设施GPU 短缺或云价格跳升多云策略(假设);本地训练投入GPU 供应策略未披露;单一云依赖未知
遥操作员队伍内部 VR 遥操作员(1X 员工)主要训练数据来源;随人员规模扩张遥操作员流失或扩张瓶颈有竞争力的薪酬结构(假设);地域多元未披露遥操作员人数、薪酬和流失率

合作伙伴和依赖分析来自 1X 的公开投资者披露、Bloomberg/Bloomberg/Reuters 报道(SR015、SR022)、Axios 报道(SR011)以及机器人创业公司依赖关系的战略分析(SR014、 SR020、SR025)。交易对手风险评估为分析师判断;尚未发现已确认的合作伙伴失效情景。

[CR011, CR012, CR014, CR018, CR032, CR037]
FR003: 依赖关系图

关键依赖关系图展示 1X Technologies 在资本、AI 能力、收入、训练数据和硬件供应上依赖的所有外部方,凸显冗余很少的集中依赖栈。

7.4 人员、执行与财务模型风险

1X Technologies 的关键人物依赖极强。CEO Bernt Børnich 是创始愿景人物和公司架构师;他带领公司从 Halodi Robotics 走过三次转向、三轮融资和两代产品。他离开会造成战略领导真空,在尚早期的人形机器人领域很难填补。AI 副总裁 Eric Jang 在技术侧同样关键:他的具身学习数据引擎是公司的核心 AI 差异化,结合 VR 远程操作和强化学习的特定经验也很稀缺。任一人离开,都可能至少拖慢 AI 路线图 12–18 个月,并可能引发投资人担忧。 竞争性人才市场放大了这一风险。Google DeepMind、Meta AI、Tesla、Boston Dynamics 和 Figure AI 都在积极从同一人才池招聘 AI 机器人工程师。这些公司的薪酬包高度有竞争力;对 1X 中层工程师而言,相比拥有上市公司股权薪酬的大型科技公司,机会成本很高。即便没有关键人物离开,工程层流失也可能实质放慢开发速度。 财务上,1X 的首要毙命风险是无法按足够规模和时间表完成 Series C,以支持 NEO 制造规模化。按估计烧钱速度 $10–15M/月(由同阶段机器人公司和 1X 员工数信号推导),2025 年 1 月之后的资本可支撑约 7–20 个月现金跑道——意味着 2026 年初之前或期间需要融资。如果 NEO 制造需要显著资本开支(CAPEX),而 Series C 延迟或认购不足,1X 可能同时面对交付延迟、客户流失和管理层分心。[CR010, CR019, CR026, CR029, CR036, CR037]

人员 / 执行风险登记表
角色 / 职能依赖 / 缺口流失概率严重性缓释措施尽调路径
CEO:Bernt Børnich(创始人)创始愿景人物;公司历次转向的架构师低-中(创始人主导阶段公司)高(生死攸关)未披露继任计划;假设有股权留任归属安排确认股权归属时间表、留任方案和继任梯队深度
VP AI — Eric Jang具身学习数据引擎首席架构师中(AI 负责人市场竞争激烈)高(AI 进度倒退 12-18 个月)假设有股权激励;AI 团队厚度未披露确认 Jang 之外的 AI 团队厚度;识别第二梯队 AI 领导层
AI 机器人工程师(中级)执行 AI 路线图的核心 R&D 人才高(Google、Meta、Tesla 都在争夺同一人才池)中(速度下降)Palo Alto 总部可就近获取人才;假设薪酬有竞争力从 HR 数据室索取流失率和开放招聘名额
遥操作员队伍训练数据吞吐受人员规模卡住中(高重复远程工作岗位流动率)中(数据飞轮放慢)遥操作员地域多元化(假设)索取遥操作员人数、地域分布和流失率
制造运营领导层NEO Gamma 在 Hayward 和 Moss 工厂产能爬坡中(需要特定制造专长)中(交付延迟)制造团队厚度未公开确认制造 VP/GM 职位和生产计划时间表

人员风险评估依据关键人依赖分析(SR014、SR030)、竞争性人才市场报道(SR030) 以及早期硬件-AI 公司的行业常态。截至目前,1X 尚未公开出现已确认的人员离职或流失信号。

[CR010, CR026, CR036]

7.5 缓释措施、毙命条件与尽调事项

1X Technologies 的风险缓释分三类:内置在产品设计里的结构性缓释(低阻抗执行器柔顺性、实体停止装置)、不利事件后触发的反应式缓释(WSJ 隐私报道后加入同意控制),以及仍未披露的规划型缓释(ISO 认证路线图、OSHA 审计方案、GDPR 数据处理协议)。 对投资人和企业客户最可执行的毙命条件包括:(1)Series C 失败或缩水——如果 1X 在 Series C 融资少于 $300M,NEO 制造计划很可能需要重组;(2)ADT 或 Everon 不续约——单个企业合同不续约且没有替代客户公告,将是重大不利信号;(3)隐私执法行动——美国 FTC 或欧盟 GDPR 围绕机器人数据采集发出执法通知,将要求产品修改并退出欧盟市场,NEO 消费收入会延迟 12+ 个月;(4)关键人物离开——Børnich 或 Jang 离开且没有可信继任者公告;(5)硬件召回——EVE 机队出现系统性安全缺陷并需要召回生产机,将同时摧毁近期经济性和品牌可信度。 监控指标包括:NEO 交付时间线更新、Series C 公告条款、ADT/Everon 在季度业绩评论中的扩张或缩减信号、点名 1X 的监管文件、工程岗位发布水平(远程操作员和机器人工程师),以及作为开发动能代理指标的 GitHub 活跃速度。缓释与毙命条件表(TR005)用具体阈值固化了这些触发器。[CR016, CR019, CR027, CR028, CR030, CR033]

缓释与终止标准表
风险可监测触发信号阈值 / 终止事件行动含义
Series C 融资失败到 2026 年 Q3 未宣布 Series C,或融资额 < $200M没有资本承诺时,现金跑道低于 6 个月NEO 制造延迟;重估投资逻辑;需要过桥融资
企业客户不续约ADT 或 Everon 缩减机队、终止合同或宣布替代供应商任一锚定客户宣布退出,或机队缩减 > 20%ARR 损失 40-60%;需要紧急多元化;投资逻辑破裂
隐私监管行动FTC 调查、GDPR 执法通知或 EU AI Act 不合规认定点名 1X Technologies 的监管执法行动欧盟消费者端发布受阻;法律成本上升;NEO 家庭部署暂停
关键人员离职CEO Børnich 或 AI 副总裁 Jang 离职,且未宣布可信接任者非计划离职且无接任者AI 路线图倒退 12–18 个月;投资者信心受挫;领导层出现真空
硬件召回公开召回通知、产品安全公告,或已确认的系统性安全缺陷CPSC 或同级机构发起 EVE 或 NEO 自愿召回机队召回成本;客户流失;品牌危机;监管调查
Tesla Optimus 规模化入场Tesla 宣布 2026–2027 年向外部客户交付 >1,000 台 OptimusTesla Optimus 以低于 $30K 的消费者价格商业上市1X 定价权被侵蚀;消费者端 TAM 被争夺;重新审视护城河耐久性

终止条件阈值由分析师定义,依据前章财务分析、本章风险严重度评估,以及可比早期硬件公司的基准。触发信号可通过 1X 新闻稿、监管数据库、ADT / Everon 母公司财报评论和科技媒体监测。来源:SR020、SR023、SR025。

[CR019, CR027, CR037, CR038, CR040]
Chapter 08

08估值

8.1 投资建议与投资逻辑

1X Technologies 获得 TRACK(观察)投资建议,综合得分 6.5/10,置信度中等、风险评级高。这个建议反映出真实张力:一边是 1X 罕见的商业验证——约 390 台 EVE 已部署并产生真实企业经常性收入;另一边是一组尚未解决的风险,限制了更高确信度的仓位。投资建议摘要表(TV001)固化了所有评分维度上的正式判断;建议逻辑图(FV001)把证据支柱和风险向量到投资结果的因果链条画出来。 投资逻辑立在五根支柱上。第一,1X 是唯一一家已确认在企业端达到商业规模部署的人形机器人公司,约 390 台设备分布在 ADT 和 Everon;竞争对手大多还在实验室或试点阶段。第二,OpenAI Startup Fund 合作带来前沿 AI 模型访问、投资人生态可信度和战略可选性,竞争对手很难复制。第三,NEO 的 3D 编织外骨骼、100Hz 全身平衡控制和 VR 远程操作数据飞轮,构成真实技术差异化,并会随时间累积防御性。第四,EQT Ventures 领投 Series B,带来欧洲分销触达和 Series C 桥接投资人可信度。第五,挪威工程文化与美国资本通道结合,形成硬件 AI 创业公司少见的均衡风险画像。 反向逻辑同样重大。NEO 高度依赖远程操作,尚未在消费者规模证明完全自主。Tesla Optimus 在资本、制造基础设施和分销触达上约为 1X 的 10x。财务不透明使资本效率无法验证,且截至 2026 年 Q2,$1B Series C 仍未确认。家庭远程操作的隐私模式正面对 FTC 和欧盟监管机构关注。投资逻辑 / 反向逻辑表(TV002)用明确证据基础和改变观点的条件,结构化呈现两边。 若出现以下情况,建议可从 TRACK 调整为 INVEST:Series C 以已知估值确认交割;NEO 预购量超过 1,000 台;EVE 收入运行率确认高于 $30M ARR;以及面向消费者部署的可验证自主路线图。Goldman Sachs 的人形机器人投资逻辑把数据飞轮、制造规模和部署验证列为私有市场机器人估值的三大价值驱动因素;1X 满足第一和第三项,但相对于隐含 $3–5B 的 Series C 入场估值,部署规模仍偏小。多年通过 EVE 商业部署积累的 VR 远程操作数据飞轮,是 1X 最具防御性的竞争资产。[CV001, CV010, CV016, CV017, CV028, CV029]

建议摘要表
参数取值证据依据 / 备注
总体建议观察EVE 商业牵引力已验证;NEO 消费者端逻辑未验证;Series C 条款未知
综合评分6.5 / 10按市场机会、商业验证、技术护城河、团队、财务透明度、竞争位置和估值支撑加权
信心水平多个关键未知项:Series C 状态、NEO 预订单、收入运行率、治理
风险评级财务不透明、NEO 自主能力缺口、Tesla Optimus 竞争威胁、Series C 依赖
估值立场偏高在当前收入不透明的情况下,隐含 $3–5B Series C 估值高于可比基准;溢价只反映未来状态的期权价值
建议行动观察;Series C 完成 + NEO 预订单验证后加仓买入条件:Series C 已确认、NEO 预订单 >1,000 个、EVE ARR >$30M(见 TV006)
目标持有期3–5 年(2029–2031)IPO 或并购退出窗口与 NEO 消费者端扩张和行业整体成熟同步
主要退出路径战略并购(主要);IPO(次要);老股要约Samsung NEXT 现有持股让团队收购或完整收购成为概率最高的路径

建议和评分是分析师基于八个报告章节收集的证据作出的判断。评分没有 1X Technologies 财务披露支撑;所有估值估计来自可比公司分析和情景建模。信心和风险评级是相对 Series B / pre-Series C 阶段可比硬件 AI 初创公司的评估。6.5/10 综合评分对应「观察」门槛——高于「放弃」(低于 5.0),低于「买入」(高于 7.5)。

[CV028, CV029]
投资逻辑 / 反向逻辑表
投资逻辑证据依据反向逻辑改变判断的证据
EVE 约 390 台企业部署,构成罕见的商业规模验证ADT 140 台 + Everon 约 250 台,RaaS 月费 $3,500–5,000 / 台已确认;没有直接人形机器人竞品具备同等商业验证收入不透明,部署质量(流失、毛利率、续约率)无法独立验证;客户满意度未披露在 Series C 数据室披露 EVE ARR、续约率和客户满意度数据;确认合同期限
OpenAI 伙伴关系带来前沿 AI 模型访问和投资者信用,竞品难以获得OpenAI Startup Fund 领投 Series A2;持续技术 AI 合作已公开披露;增强机构 LP 信任合作条款、排他性和 AI 模型访问范围未披露;OpenAI 可能优先推进自己的具身 AI 路线图通过文件确认 AI 访问承诺、排他条款和最低合作期限
NEO 的 3D 针织外骨骼和 100Hz 平衡控制构成真实技术差异化IEEE Spectrum 和 New Atlas 技术分析确认自研制造方法和平衡控制优势技术差异化可能只是阶段性;资金充足的竞品可用资本和工程时间复制到 2026 年 Q4 展示无需遥操作的 L2+ 自主能力里程碑;确认核心方法专利覆盖
VR 遥操作数据飞轮有望随时间沉淀可防守的具身 AI 护城河EVE 多年商业部署积累遥操作数据;没有同等机队的竞品难以复制这项数据资产若 Tesla 规模化部署 Optimus 工厂机队,并在内部收集同等或更大的训练数据集,数据优势会被稀释证明数据积累带来可量化的自主能力提升速度;发布轨迹指标或第三方基准结果
挪威工程文化叠加美国资本通道,塑造注重质量的硬件 AI 初创公司Moss 制造 + Palo Alto 总部整合;ADT 和 Everon 部署中的 EVE 机队质量一致双地域运营增加制造爬坡复杂度;Hayward 与 Moss 协同带来交付时间线风险Series C 运营计划披露 NEO 制造里程碑、质量 KPI 和供应链冗余策略

所有投资逻辑和反向逻辑均由分析师基于第 1–8 章公开证据构建。尚未获得 1X Technologies 管理层对反向逻辑的回应。「改变判断」条件具体、可验证,并绑定 TV006 所述 Series C 尽调路径。收到 Series C 数据室材料后,应重新评估这些论点。

[CV001, CV016, CV032]
FV001: 推荐逻辑

因果链从 1X Technologies 的四个关键证据支柱(商业验证、AI 合作关系、技术差异化、数据护城河)出发,经过两个风险向量(Tesla Optimus 的竞争压力,以及财务不透明与未确认的 Series C),最终落到观察(TRACK)建议:6.5/10,中等信心。图中说明各证据维度和风险如何相互作用,形成投资结果。

8.2 估值方法、融资背景与可比公司

1X Technologies 未在任何已确认融资轮公开披露估值数字。$23.5M Series A2(2023 年 3 月)、$100M Series B(2024 年 1 月,EQT Ventures 领投,Samsung NEXT 等参与)以及约 $100M 追加融资(2025 年 1 月),合计约 $237M 已确认资本,所有轮次的投后估值均未披露。截至 2026 年 5 月,尚无二级市场标记、公开提交的优先股堆叠披露,或分析师确认的 1X 专属估值估计。 本章采用三种三角校验方法。第一,收入倍数外推:按估计 EVE RaaS 年化收入 $15–25M,并采用适合高增长、未盈亏平衡机器人的 12–20x 收入倍数,当前基本面单独隐含的估值区间为 $180–$500M。这显著低于隐含 Series C 入场估值,说明任何 $3–5B 溢价都依赖 NEO 商业化的未来状态。第二,可比公司分析把估值锚定在同行背景中:Figure AI 约 $3.2B,Boston Dynamics 并购价格 $1.1B,Agility Robotics 约 $1B,Apptronik 约 $300M,Sanctuary AI 约 $500M CAD。可比估值表(TV004)列出全部六家参考公司及指标和限制;估值敏感性图(FV002)展示在 NEO 商业化时间不同的情况下,悲观、基准和乐观估值如何从 $300M 跨到 $6B。第三,入场价格纪律:如果假设 Series C 投后估值为 $3–5B,要在 5–7 年退出期内为 LP 带来 3x 回报,需要达到 $9–15B 退出价值——只有乐观情景才可能实现。 从资本效率看,1X 用约 $237M 融资换来约 390 台商业 EVE 部署,相比 Figure AI 融资 $675M 却没有已确认的外部付费部署更有利。这说明 1X 比融资最多的同行更有效地把资本转化成商业验证,足以支撑相对于纯收入倍数基准的适度溢价。Bloomberg 2025 年 4 月分析确认,只有当商业验证显著超过行业均值时,可比阶段公司之上的估值溢价才合理——仅看 EVE,1X 勉强通过这一测试。PitchBook 数据显示,2025 年人形机器人估值倍数主要由部署验证、累计融资和领投人层级驱动,而非收入倍数,这为 1X 相对于当前收入基础的溢价入场价格提供了进一步理论支撑。[CV002, CV003, CV004, CV005, CV006, CV007]

可比估值表
公司阶段 / 状态最新估值 / 并购价格主要指标收入状态对 1X 的相关性关键局限
Figure AISeries C(2024 年 2 月以 $2.6B 完成;后续融资未披露)~$3.2B(估计,后续融资后估值)$2.6B 估值对应 $675M 融资额(3.9x 已融资额);无收入验证截至 2026 年 Q2,无已确认外部商业收入最常被引用的直接人形机器人可比公司;融资规模相近;无商业部署收入无收入验证;估值完全由投资者需求而非基本面驱动;拉高行业均值
Boston DynamicsHyundai Motor Group 收购(2021)$1.1B 收购价以 $1.1B 并购;收购时 Spot 商业收入估计 ARR >$50MHyundai 收购时,Spot 机器人销售产生正 ARR,超过 $50M最佳底部参考:已有部署和正收入的商业机器人公司四足机器人(非人形);收购发生在低利率并购环境;数据较旧
Agility Robotics战略投资者(Amazon ORCA 部署)~$1B 隐含估值(分析师估计)企业物流部署;Amazon 仓库项目来自 Amazon ORCA 项目的收入;未公开披露最接近的企业部署同业;同样偏商业而非消费者物流 vs. 安防应用;无消费者产品;部署可能受合同限制
ApptronikSeries A($53M,2024 年 8 月,General Catalyst 领投)~$300M 隐含估值$300M 估值对应 $53M 融资额(5.7x 已融资额);仅处试点阶段无已确认商业收入;仅早期试点部署商业验证低于 1X;代表下限可比公司商业化前;Series A 阶段;作为 Series C 定价参考可比性有限
Sanctuary AI(Phoenix)Series B~$500M CAD(估计)加拿大遥操作模式;零售试点部署加拿大零售试点部署;收入未披露遥操作模式是最接近 1X NEO 策略的结构类比加拿大市场;规模小于 1X;CAD 估值与 USD 口径可比性较弱
Tesla Optimus(隐含)内部 / 合并口径(未独立交易)$50–100B+(若独立估值,分析师隐含值)无法从 Tesla 合并股权中拆分仅内部工厂部署;无已确认外部商业收入仅作为规模威胁参考;用于说明 TAM 上限和竞争风险不是独立实体;隐含价值高度投机;排除在可比均值之外

除 Boston Dynamics 并购和 Agility 的 Amazon 背景外,所有估值均为分析师根据融资轮次推导的估计;除 1X 标的外,其他私营公司均未披露投后估值。Tesla Optimus 隐含价值不纳入可比组平均。覆盖不完整:Fourier Intelligence、Unitree 和中国人形机器人厂商被省略;见证据缺口部分。可比组均值(剔除 Tesla Optimus)暗示行业中点约 $1.0–1.5B;若 NEO 预订单确认规模化消费者需求,可支持 1.5–2.5x 溢价。

[CV004, CV005, CV006, CV022, CV023, CV024]
FV002: 估值敏感性

柱状图展示 1X Technologies 在七个参考情景和可比对象下的隐含估值(百万美元):从 $300M 的悲观 IP 清算情景,到当前基本面的收入倍数、Boston Dynamics 与 Agility Robotics 可比底线、2028 基准情景、Figure AI 可比隐含估值,再到 $5B 的 2030 乐观情景。图中凸显估值分布很宽,NEO 商业化是最终结果的第一变量。

8.3 乐观、基准与悲观情景分析

情景分析覆盖三种不同投资结果,分别绑定 NEO 商业化成功、Series C 关闭和竞争动态。乐观情景(25% 概率权重)假设到 2030 年,NEO 交付给 10,000 户或更多消费者家庭,按每台 $20,000 或每月 $499 订阅收费,到 2030 年产生 $500M 或更多年收入,并在一个因已验证商业规模而上调估值中枢的市场里获得 8–12x 收入倍数。隐含退出估值 $4–6B,相对于 $3B Series C 入场只带来 1.3–2x 毛回报——除非入场价格低于 $2B,否则回报边际。ARK Invest 的物理 AI 估值框架认为,如果数据飞轮在规模化后形成可防御护城河,定义品类的玩家可支撑更高倍数。 基准情景(45% 概率权重)假设 NEO 在 2026–2027 年实现消费者交付,但数量低于预测且仍依赖远程操作。EVE 企业收入到 2028 年通过新增客户增长至 $50–80M ARR。2028 年退出时隐含基准估值为 $1.2–1.5B,相比 $3B 入场显著减记;除非 Series C 入场价格远低于 $3B,否则基准情景回报为负。悲观情景(30% 概率权重)是有序收缩:NEO 未能商业出货,ADT/Everon 合同失效,公司以知识产权加团队形式按 $200–400M 被收购——在任何 $3B+ 入场价格下几乎全损。 三种情景的概率加权预期企业价值约为 $1.3B,显著低于任何可能的 $3B+ Series C 入场,再次支持 TRACK 而非 INVEST 建议,除非通过入场价格纪律把入场估值压到 $1.5B 以下。乐观 / 基准 / 悲观情景表(TV003)和估值回报区间图(FV003)用明确假设、回报逻辑、概率信号和下行触发器固化了三种情景。[CV009, CV019, CV021, CV031, CV037]

乐观 / 基准 / 悲观情景表
情景关键假设隐含退出估值以 $3B 入场的回报逻辑情景关键风险概率信号
乐观(25% 权重)到 2030 年 NEO 交付 10,000+ 个消费者家庭;ARR >$500M;人形机器人市场在规模验证后重估至 10–12x 收入倍数$4–6B(2030 年退出)以 $3B 入场,总回报 1.3–2x;只有入场价低于 $2B 才有吸引力;5 年 IRR 约 15–25%Tesla Optimus 在 1X 达到 10,000 台前进入消费者市场;Series C 延迟至 2027 年 Q1 之后需要 NEO 预订单 >1,000 + Series C 完成且规模 >$700M + 2027 年 EVE ARR >$80M
基准(45% 权重)到 2028 年 NEO 交付 2,000–5,000 个消费者家庭;EVE ARR $50–80M;自主能力温和提升,但仍依赖遥操作$1.2–1.5B(2028 年退出)以 $3B 入场亏损 50–75%;只有入场价低于 $1.2B 才能打平;按当前隐含定价 IRR 为负Series C 延迟或认购不足;NEO 遥操作瓶颈在商业规模下仍存在基于当前财务不透明、NEO 准备度信号和 Tesla Optimus 竞争压力,这是最可能情景
悲观(30% 权重)NEO 无法商业出货;ADT 或 Everon 合同失效;公司作为 IP + 团队被出售$200–400M(困境并购)以 $3B 入场亏损 87–93%;任何高于 $400M 的入场价都几乎使 Series C 资本全损Series C 未在 2026 年 Q4 前完成,引发现金跑道耗尽,并与客户集中度悬崖重叠终止触发:TV005 中五个投资逻辑破坏触发项任意两个同时激活

情景估值为分析师估计;1X Technologies 未披露财务预测。概率信号是基于现有公开证据和同阶段硬件 AI 初创公司基础概率的定性评估;不是精算或模型结果。回报逻辑假设 Series C 入场后估值为 $3B。实际回报取决于确认后的入场估值、优先股堆叠结构、Series C 及后续轮次稀释和退出时间。三种情景的概率加权预期企业价值约为 $1.3B。

[CV009, CV019, CV021]
FV003: 估值 / 回报区间

区间图展示 1X Technologies 悲观、基准、乐观和假设 Series C 入场估值结果,单位为百万美元。每个条形覆盖该情景的不确定区间,中心估计作为点值。图中显示,从 $3B+ Series C 入场价看,回报结构不对称:基准情景($1.2–1.5B)意味着大幅减记,而乐观情景($4–6.5B)在当前入场定价下也只提供有限回报。

8.4 竞争动态、市场背景与资本环境

人形机器人市场正在快速扩张。Goldman Sachs 预计到 2030 年,可寻址机器人市场为 $38B。IFR Service Robots 报告估计,到 2025 年服务机器人市场为 $46.6B;Mordor Intelligence 和 MarketsAndMarkets 预计服务机器人到 2030 年年复合增长率(CAGR)为 20–25%。这些预测构成了大市场背景,支撑品类领先人形机器人公司在私有市场获得溢价入场倍数。 Tesla 的 Optimus 项目是主要竞争威胁。Elon Musk 曾提出到 2025–2026 年生产数千台 Optimus;如果 Tesla 以低于 $30K 的消费者定价实现生产规模,1X 的 NEO 定价权会大幅被侵蚀。不过,截至 2026 年 Q2,Tesla Optimus 部署仍停留在内部(Tesla 工厂内部使用),这给 1X 留出估计 18–24 个月窗口,在外部 Optimus 竞争进入家庭市场前建立消费者预购速度。 2025–2026 年风险投资(VC)融资环境对人形机器人仍然强劲。Reuters 确认 1X 正瞄准 $1B Series C;2024–2025 年行业内多轮超过 $100M 的融资说明投资人需求充足。OpenAI 和 EQT 的既有持仓,为潜在新进入投资人提供可信锚定信号。1X 在 2025 年整合 80,000 sq ft Palo Alto 总部,释放出组织扩张信号,与为 NEO 生产爬坡做准备的公司状态一致。若商业化目标落空,战略收购方——包括 Samsung(现有投资人)、Amazon(机器人履历)和 Google/Alphabet(AI 机器人兴趣)——构成可行退出路径。投资关键指标图(FV004)提供了面向投委会的 1X 八项关键投资维度评分。[CV008, CV018, CV026, CV027, CV034, CV035]

FV004: 投资 KPI

面向 IC 的评分仪表盘按八个投资维度为 1X Technologies 打分,尺度为 1–10。综合得分 6.5/10:公司有真实商业验证和强市场机会,但财务不透明、面对 Tesla Optimus 的竞争脆弱性,以及当前收入指标对估值的支撑有限,抵消了上行。分数是分析师对全报告各章节证据的综合判断。

8.5 尽调事项、投资逻辑破裂触发器与退出准备度

最终尽调事项表(TV006)列出八项优先数据请求;在任何 Series C 定价水平下,投入资本前都必须解决。最关键的是:EVE 收入运行率和 ADT/Everon 合同条款,因为已确认经常性收入是首要估值锚;NEO 预购积压数量,用于验证或挑战消费者市场投资逻辑;以及烧钱速度和现金跑道,用于确认 Series C 时间表并非由急性资本压力驱动。Series C 条款和投资人构成会影响优先股堆叠和稀释悬垂分析;自主性指标——自主完成任务与通过远程操作完成任务的比例——是 NEO 在规模化商业可行性上的领先指标。 投资逻辑破裂与毙命触发器表(TV005)定义了五个具体事件,一旦出现就意味着投资逻辑失败,需要立即采取组合行动:Series C 未在 2026 年 Q4 前关闭;ADT 或 Everon 合同终止,或机队减少超过 20%;NEO 出货延迟到 2027 年 Q4 之后且没有可信里程碑更新;CEO Børnich 或 AI 副总裁 Jang 离开且未任命继任者;或 FTC、GDPR、EU AI Act 对家庭远程操作数据采集采取执法行动。这些触发器来自第 4 章财务分析、第 7 章风险登记表和第 3 章竞争分析,为 Series C 投资建立跨章节监控框架。 1X Technologies 的退出可选性有三条路径:IPO(基于公开市场可比公司的 $100M+ ARR 要求,时间在 2029 年以后)、由生态内企业伙伴战略并购(Samsung NEXT 既有投资是概率最高的人才收购或全面收购路径),以及通过要约收购向早期投资人提供二级市场流动性。在当前规模下,最可能的近期退出是 $1–3B 区间的战略并购。NEO 预购和商业验证数据,是改善退出时点和价格的主要杠杆。TV006 中第 1–4 项(收入、预购、烧钱速度、Series C 条款)最终尽调结果为正,是从 TRACK 转向 INVEST 的最低门槛。[CV011, CV013, CV014, CV015, CV031, CV039]

投资逻辑破坏与终止触发表
触发项阈值 / 终止事件对投资逻辑的传导行动含义
Series C 未完成2026 年 Q4 前未公告 Series C,或确认融资额低于 $300MNEO 制造资金不到位;估计现金跑道降至 9 个月以下;悲观情景概率升至 50% 以上立即重新评估投资逻辑;任何 Series C 承诺都必须以确认完成且条款明确为条件
企业锚定客户流失ADT 或 Everon 将机队规模减少超过 20%、终止合同,或公开宣布替代机器人供应商ARR 损失 40–60%;EVE 商业验证——主要估值锚——被削弱;基准情景向悲观情景坍缩减持或退出组合持仓;对替代客户管线启动紧急尽调
NEO 出货延迟至 2027 年 Q4 之后到 2027 年 Q4 末仍未确认向付费客户交付 NEO 消费者版本消费者逻辑破裂;乐观和基准情景明显转向悲观;预订单取消率可能超过 30%重新评估信心;要求更新自主能力指标和预订单留存数据;考虑困境并购情景
CEO 或 AI 副总裁离职且无接任者Børnich 或 Jang 非计划离职,且 30 天内未宣布具名可信接任者AI 路线图倒退 12–18 个月;NEO 自主能力开发受扰;Series C 投资者信心受影响立即开展治理问询;要求董事会回应和继任证据;暂停部署任何新资本
隐私监管执法行动FTC 调查、GDPR 执法通知或 EU AI Act 不合规认定,且明确点名 1X Technologies受影响司法辖区内 NEO 消费者发布受阻;欧盟业务暂停等待合规;法律成本负担上升在监管结果明确前暂停持仓;评估欧盟市场排除对投资逻辑的影响

终止触发阈值由分析师定义,依据财务分析(第 4 章)、风险登记表(第 7 章)和本章情景分析。触发信号可通过 1X 新闻稿、公开提交的 ADT Inc. 和 Everon 母公司披露、FTC 与 GDPR 执法数据库、科技媒体和季度组合复盘监测。用于校准阈值的来源:SV006、SV028。

[CV031]
最终尽调问题表
主题缺失证据重要性负责人 / 尽调路径
EVE 收入运行率ADT 或 Everon 的 EVE 企业部署未披露 ARR、MRR 或收入运行率主要估值锚;没有确认收入,所有可比倍数都属推测CFO;要求提供月度 MRR、合同期限、续约历史和混合单台定价
NEO 预订单积压预订单于 2025 年 10 月开放;未披露订单数量或订金金额消费者逻辑验证:带订金的积压订单超过 1,000 台,会显著提高信心CEO 或销售副总裁;要求提供当前积压订单数量、地域分布和订金转化率
烧钱速度和现金跑道未披露现金余额、月度烧钱速度或 12 个月现金跑道预测Series C 时间和困境风险:若现金跑道低于 9 个月,谈判筹码会受损CFO;要求提供月度现金消耗、账上总现金,以及带模型假设的 12 个月现金跑道预测
Series C 条款和投资者构成截至 2026 年 Q2,轮次规模、投后估值和领投方构成未确认入场价是主要回报变量;优先股堆叠和反稀释条款影响净 IRRCEO 或法律顾问;要求提供条款清单、投后股权结构表、优先权结构和共同投资者名单
自主能力水平指标未披露 NEO 或 EVE 自主完成任务与远程遥操作完成任务的比例NEO 消费者可行性需要达到 L2+ 自主能力;当前遥操作比例决定持续成本结构AI 副总裁 Eric Jang;要求提供月度自主率进展、任务复杂度指标和 2026 年路线图目标
硬件可靠性(MTBF)EVE 或 NEO 未发布平均故障间隔时间或保修索赔数据RaaS 模型单位经济高度依赖维护成本;MTBF 决定长期利润率工程或运营团队;要求提供现场可靠性数据、维护干预率和保修索赔
专利组合和 IP 所有权除公开登记中的少量已知美国专利外,公开专利披露有限如果 Tesla 或 Figure AI 规模化并寻求自由实施,IP 护城河耐久性决定可防守性总法律顾问;要求提供完整专利组合清单、待审申请、自由实施分析和 IP 所有权
董事会构成和投资者权利未确认 Series B 后董事会构成、投资者权利摘要或股权表治理结构影响退出灵活性、Series C 定价、反稀释保护和董事会控制CEO 或法律顾问;要求提供董事名单、主要投资者权利摘要和当前股权表

尽调问题按其对 Series C 投资决策的重要性排序。第 1–4 项(收入、预订单、烧钱速度、Series C 条款)本质上是阻断级缺口,任何投资承诺前都必须解决。第 5–8 项(自主能力指标、硬件 MTBF、专利、董事会构成)对投资质量判断很重要,但若第 1–4 项顺利解决,不一定构成阻断。任何有约束力的 Series C 承诺前,所有项目都应在正式数据室中解决。若未披露第 1–4 项,建议应维持观察。

[CV039]

免责声明

本报告是基于公开证据的尽调快照,不构成投资建议。重要财务、法律、技术和合同事实仍未公开;作出任何投资决定前,应直接向管理层和原始文件核验。

证据索引

结论
编号陈述可信度来源
CO001 1X Technologies was founded in 2014 under the name Halodi Robotics in Norway by Bernt Øivind Børnich. SO002, SO004, SO016
CO002 1X Technologies is headquartered in Palo Alto, California, with manufacturing operations in Hayward, California and Moss, Norway. SO002, SO005
CO003 The company officially rebranded from Halodi Robotics to 1X Technologies in March 2023, coinciding with its Series A2 funding announcement. SO006, SO002
CO004 1X's stated mission is "to create an abundance of labor through safe, intelligent humanoids that work alongside people." SO007, SO010
CO005 A new 80,000 square-foot Palo Alto headquarters opened in summer 2025 with capacity for 400 employees, consolidating Sunnyvale and Moss teams. SO002, SO005, SO010
CO006 1X Technologies' business model combines enterprise EVE robot deployment for near-term revenue and NEO consumer robot pre-orders for long-term market capture. SO003, SO004, SO012
CO007 CEO and founder Bernt Øivind Børnich is the primary public face of 1X Technologies and has led the company through all major fundraising rounds and product pivots since 2014. SO003, SO006, SO007
CO008 Eric Jang serves as VP of AI at 1X Technologies and leads the embodied learning data engine from the San Francisco Bay Area. SO017, SO004
CO009 Christoph Kohstall, former Stanford scientist and Google robotics team member, joined 1X Technologies following the acquisition of Kind Humanoid in January 2025. SO010, SO014
CO010 Nicholas Nadeau was named as CTO in a 2022 EEJournal article about Halodi Robotics; his current role at 1X Technologies post-rebrand is not confirmed in reviewed public materials.
CO011 1X Technologies has not publicly disclosed its board composition, governance structure, or equity ownership table in any reviewed public source. SO001, SO002
CO012 The company has indicated it added executives with experience from BMW and Tesla following the Series B, but specific names have not been confirmed in reviewed public sources. SO010
CO013 The Series A2 in March 2023 raised $23.5 million led by the OpenAI Startup Fund, with Tiger Global, Sandwater, Alliance Ventures, and Skagerak Capital as participants. SO006, SO002, SO004
CO014 The Series B in January 2024 raised $100 million led by EQT Ventures with Samsung NEXT, Nistad Group, and existing investors. SO003, SO004, SO002
CO015 In January 2025, 1X Technologies reportedly raised an additional $100 million, per Bloomberg's article titled "humanoid robot maker 1x raises 100 million" dated January 27, 2025. SO019, SO013
CO016 VentureBeat reported total capital raised at approximately $137 million immediately following the January 2024 Series B, implying undisclosed pre-Series A2 Halodi-era rounds. SO004
CO017 1X Technologies has not publicly disclosed a valuation figure for any funding round. SO001, SO002, SO003
CO018 The OpenAI Startup Fund led the Series A2 and partnered with 1X on technology; the precise nature of any ongoing commercial technology collaboration beyond the investment is not publicly documented. SO006, SO011, SO024
CO019 EQT Ventures led the Series B and described 1X as a key investment in humanoid robotics; Ted Persson (EQT partner) stated the company "will play a crucial role in the pioneering steps towards... our technological and human future." SO004, SO003
CO020 In September 2025, The Information reported that 1X Technologies was seeking to raise $1 billion in new funding; confirmation of closing had not been reported by the time of this report. SO015, SO002
CO021 1X Technologies has not disclosed revenue, burn rate, gross margins, or any audited financial metrics in public materials reviewed for this chapter. SO001, SO004, SO003
CO022 ADT placed an order for 140 EVE robots from Halodi Robotics, described at the time as the world's largest order for humanoid robots. SO016, SO025
CO023 Everon deployed approximately 250 EVE androids for night guarding in commercial buildings. SO004
CO024 Sunnass Hospital is cited as a healthcare deployment customer for EVE robots. SO004
CO025 1X has not publicly disclosed a total customer count for EVE deployments; known named customers include ADT, Everon, Sunnass Hospital, and Norwegian universities. SO001, SO004
CO026 EVE features 23 degrees of freedom, force and impedance control on all joints, and can handle an 8-kilogram payload with each arm. SO016
CO027 EVE is designed for deployment in facilities compliant with the Americans with Disabilities Act (ADA), enabling wheelchair-accessible navigation by the wheeled robot. SO016
CO028 A 2024 EVE demonstration video showed approximately 30 robots autonomously performing tasks trained "purely end-to-end from data" with no teleoperation, scripted playback, or video speedups. SO009, SO017
CO029 1X's embodied learning approach trains base models from teleoperated VR data, which are then fine-tuned to environment-specific capabilities and individual job tasks in a few minutes of desktop GPU time per deployment. SO009, SO017
CO030 The former California presence (prior to Palo Alto HQ) was in Sunnyvale; the Norway presence for manufacturing remains in Moss with additional manufacturing in Hayward, California. SO002, SO005, SO010
CO031 NEO Beta was unveiled on August 30, 2024 as a bipedal humanoid prototype designed for home use, with a limited number deployed in selected homes for research and development. SO007, SO002
CO032 NEO Gamma, announced February 21, 2025, features a 3D-knitted nylon suit for passive safety, emotive ear rings for communication, 100Hz dynamic balance control using reinforcement learning from human motion capture data, and a visual manipulation model for object pick-and-place. SO008, SO014, SO011
CO033 On October 28, 2025, 1X opened pre-orders for the NEO Home Robot at $20,000 (early access) or $499 per month (subscription), with a $200 deposit, and shipping expected in 2026. SO012, SO002, SO022
CO034 CEO Børnich stated that NEO buyers must agree that a human teleoperator will be able to see inside their homes through the robot's camera during scheduled sessions needed to collect training data. SO012, SO013
CO035 The NEO product page describes full autonomy as the design goal, with "1X Expert" teleoperator assistance available for tasks the robot does not yet know how to do autonomously. SO022
CO036 The Wall Street Journal review of NEO (October 2025) characterized most tasks as still being teleoperated by humans using VR headsets, raising concerns about privacy. SO021, SO012
CO037 1X implemented privacy controls for NEO teleoperation: owners must approve each session via app, operators can blur people, and designated no-go zones prevent operator access to specific rooms. SO012, SO013
CO038 The Engadget review of NEO noted that early adopters who want NEO to learn specific or complex tasks "would have to be comfortable with the idea of a human teleoperator controlling the robot remotely and seeing inside their homes." SO012, SO021
CO039 New Atlas described 1X's EVE robot as looking "oddly undergunned" compared to Tesla, Figure, Sanctuary, and Agility, lacking feet and dextrous hands in early development phases. SO009
CO040 Kind Humanoid was a 3-person Palo Alto startup founded in 2023 by Christoph Kohstall that applied LLMs to humanoid interaction and worked with designer Yves Béhar; it was acquired by 1X in January 2025. SO010, SO014
CO041 The reported January 2025 $100 million raise would bring total confirmed capital raised to approximately $237 million, excluding undisclosed Halodi-era rounds. SO019, SO013, SO004
CO042 The Freethink article (March 2025) confirms a $100 million raise was announced approximately one month before the February 2025 NEO Gamma announcement, consistent with the Bloomberg January 27, 2025 date. SO013, SO019
CO043 1X's goal is for NEO Gamma to be deployed in "thousands of homes" in 2025, then scale 10-fold in subsequent years, per CEO Børnich in the Freethink interview. SO013
CO044 Goldman Sachs Research forecasts the global humanoid robot market could reach $38 billion by 2035, noting that most hardware components are near maturity but AI software for manipulation and interaction remain significant barriers. SO018
CO045 1X entered a California presence in 2019, then operated from Sunnyvale before consolidating to the Palo Alto HQ in summer 2025; the Norwegian manufacturing and some team members remain in Moss. SO005, SO017, SO010
CM001 Goldman Sachs projects the global humanoid robot market will reach $38 billion by 2035, making it one of the most bullish major analyst forecasts for the sector. SM001, SM019
CM002 The International Federation of Robotics reported approximately 4 million industrial robots operating in factories worldwide as of 2023, representing the baseline installed base of automation against which humanoid robots compete and complement. SM002, SM003
CM003 McKinsey Global Institute estimates that automation could displace 400 to 800 million jobs globally by 2030, creating both a large demand signal for automation products and a workforce transition challenge. SM003, SM017
CM004 ARK Invest estimates a $24 trillion or greater addressable opportunity for humanoid robots by 2030, treating them as substitutes for a meaningful fraction of total global human labor compensation. SM005, SM030
CM005 The World Economic Forum's Future of Jobs Report 2025 projects 85 million jobs will be displaced by automation while 97 million new roles will be created, for a net increase of 12 million jobs globally. SM017, SM003
CM006 The United States had approximately 10 million unfilled job openings in 2024–2025 per labor statistics cited in multiple analyst reports, providing a structural driver for robotic labor substitution. SM001, SM018
CM007 Mordor Intelligence projects the global humanoid robot market at $35–55 billion by 2033, with a CAGR of approximately 45–50% from a low base in 2023. SM006, SM023
CM008 MarketsandMarkets projects the humanoid robot market at $38–65 billion by 2033, with a CAGR of approximately 50% driven by enterprise adoption in logistics and manufacturing. SM007, SM023
CM009 Grand View Research projects the humanoid robot market at $55–75 billion by 2035, the most optimistic of the major research firm estimates, with an implied CAGR of approximately 54%. SM015, SM024
CM010 IDC projects the global service robot market—a broader category that includes humanoid and non-humanoid service robots—at $50 billion or more by 2030, with an approximate 20% CAGR. SM014, SM003
CM011 Goldman Sachs estimates a serviceable addressable market (SAM) of approximately $6 billion by 2030 for humanoid robots in addressable Western markets, focusing on enterprise security, logistics, and manufacturing rather than the full global TAM. SM001, SM019
CM012 Analyst estimates for the enterprise security robot sub-segment (not humanoid-only) range from $3–5 billion by 2027, though these figures include non-humanoid patrol robots and are not specific to humanoid deployments. SM024, SM021
CM013 The consumer home robot market is estimated at $10–15 billion TAM by 2030 if autonomous capability reaches a threshold that enables broad deployment, based on analyst projections for service robots in residential settings. SM015, SM022
CM014 1X EVE is commercially deployed with at least two named enterprise customers—ADT (140 units) and Everon (approximately 250 units)—providing proof of concept for the enterprise security buyer segment. SM012, SM025
CM015 NEO is priced at $20,000 outright or $499 per month on subscription, targeting the consumer home segment with pricing comparable to a premium appliance or entry-level vehicle. SM020, SM028
CM016 Enterprise security operations—building patrol, night guarding, facility monitoring—are the primary near-term segment for humanoid robot adoption, driven by the quantifiable and predictable ROI of substituting for human security guards. SM009, SM012
CM017 The consumer home assistance segment—household chores, personal assistance, aging-in-place support— represents the largest long-term TAM for humanoid robots but requires higher autonomy levels not yet commercially available at scale. SM009, SM022
CM018 Healthcare and logistics represent secondary enterprise segments with longer procurement cycles but large labor shortages; 1X EVE is deployed at Sunnass Hospital in Norway as a healthcare reference. SM021, SM009
CM019 Enterprise humanoid robot deployments create high switching costs through integration with facility management systems, proprietary AI training data, and operational workflow redesign, favoring first-mover deployers. SM018, SM004
CM020 The absence of formal safety and liability standards for humanoid robots operating alongside humans is identified across multiple analyst reports as a key adoption constraint, particularly for healthcare and food handling applications. SM008, SM013
CM021 The current unit cost of $20,000–100,000+ for a humanoid robot implies a 1–5 year payback period compared to equivalent human labor in US enterprise settings, creating a CFO adoption threshold that limits near-term market scale. SM001, SM018
CM022 Goldman Sachs projects humanoid robots could address approximately 4% of the US labor shortage by 2030, translating to approximately 400,000 robots deployed in the US across enterprise use cases. SM001, SM019
CM023 Deloitte's analysis of enterprise automation adoption shows AI and robotic systems are entering significant deployment phases in 2025–2028, with survey data indicating strong CFO intent to increase automation investment. SM018, SM004
CM024 The WEF Future of Jobs 2025 report identifies robotics and automation as one of the top five technologies most likely to drive workforce transformation, alongside AI and green-energy technology. SM017, SM003
CM025 Structural labor shortage in enterprise security, logistics, and light manufacturing is consistently cited by Goldman Sachs, Deloitte, and WEF as the primary demand-side growth driver for commercial humanoid robot adoption. SM001, SM018, SM017
CM026 Consumer trust deficit around privacy and safety—exemplified by the WSJ disclosure that NEO home deployments involve human teleoperators with home video access—is identified as a material constraint on consumer home robot adoption. SM022, SM013
CM027 Physical intelligence and embodied AI model maturity remain limiting constraints; most humanoid robots as of 2025–2026 still require human teleoperation or scripted control for tasks beyond highly structured environments. SM008, SM009
CM028 Rapid AI hardware cost reduction—including compute, vision sensors, and actuator manufacturing—is a structural growth driver that is progressively lowering the bill-of-materials for humanoid robots, making lower price points commercially viable over the 2025–2030 horizon. SM010, SM028
CM029 Global manufacturing labor cost differentials create an ROI case for humanoid robots in labor-intensive factory tasks; McKinsey and Goldman Sachs both cite manufacturing as a primary medium-term deployment environment for bipedal robots. SM003, SM001
CM030 1X EVE's enterprise security deployment model at ADT and Everon demonstrates a viable ROI thesis through night guarding—displacing or augmenting human security guards whose average cost exceeds $30,000 per year including benefits. SM012, SM025
CM031 Analyst estimates for the humanoid robot market TAM vary widely across firms—from $35 billion to $75 billion by 2033–2035—reflecting fundamentally different assumptions about adoption velocity, price trajectory, and geographic scope. SM001, SM006, SM007, SM015
CM032 Deloitte and McKinsey estimate automation investment ROI payback of approximately 2–4 years for industrial and logistics use cases at current pricing, suggesting viable business cases exist at near-term robot price points for enterprise buyers. SM018, SM003
CM033 Privacy concerns associated with the NEO home teleoperation model—where human operators remotely see inside customers' homes—represent a material adoption constraint for the consumer home segment that competitors offering fully autonomous deployments will not face. SM022, SM013
CM034 Capital intensity of humanoid robot manufacturing at early scale—requiring expensive actuators, vision systems, and custom electronics—constrains supply-side ramp and delays market penetration relative to demand potential. SM008, SM010
CM035 Regulatory frameworks specific to humanoid robots in workplaces and homes are nascent or non-existent in the US, EU, and major markets as of 2026, creating liability uncertainty that delays enterprise procurement in regulated industries. SM008, SM017
CM036 PitchBook data indicates VC funding into humanoid robot companies exceeded $1 billion in 2024, validating investor conviction in the sector and accelerating ecosystem development. SM016, SM030
CM037 CNBC and Bloomberg coverage in early 2025 indicates Tesla Optimus and Figure AI dominate media attention in the humanoid robot sector, with 1X and Agility Robotics recognized as significant second-tier commercial operators. SM020, SM029
CM038 UBS Wealth Management projects humanoid robot adoption could grow faster than consensus estimates if AI model quality improves faster than expected, citing the possibility of 2–3x the Goldman Sachs base case under optimistic scenarios. SM019, SM030
CM039 Statista projects global humanoid robot revenue on an upward trajectory through 2030 and beyond, with data showing accelerating commercial activity from 2025 onward. SM023, SM024
CM040 ABI Research projects the broader service robotics market growing at 20%+ CAGR through 2030, with humanoid robots expected to capture an increasing share of new service robot deployments as autonomous capability improves. SM024, SM014
CP001 Agility Robotics Digit is the most commercially mature enterprise bipedal humanoid robot deployment globally, with deployments at Amazon fulfillment centers, GXO Logistics, and Ford production facilities. SP004, SP010, SP028
CP002 Figure AI raised $675 million at a $2.6 billion valuation in February 2024, becoming the best-funded pure-play humanoid robot startup; the company has an OpenAI partnership for language model integration and a BMW factory pilot deployment. SP009, SP022
CP003 Boston Dynamics Atlas, acquired by Hyundai, is the most capable bipedal humanoid robot for acrobatic physical tasks, but is not commercially available for general enterprise deployment as of 2026. SP001, SP011, SP026
CP004 Tesla Optimus Gen 2 is deployed internally in Tesla's own manufacturing facilities; Elon Musk stated a target of 1,000 Optimus units inside Tesla factories by end of 2025, with a long-term consumer price target of approximately $20,000. SP002, SP012, SP021
CP005 Apptronik Apollo raised $53 million in a Series A led by General Catalyst in August 2024 and has enterprise pilot partnerships with GE Aerospace and Mercedes-Benz. SP005, SP014
CP006 Sanctuary AI Phoenix is a seventh-generation bipedal robot with commercial deployments in Canada and the UK, including a reported retail partnership with Mark's Work Wearhouse; the company has raised over $100 million CAD. SP006, SP015, SP027
CP007 Unitree Robotics offers the H1 bipedal robot at $90,000 and the G1 bipedal robot at $16,000, representing the most aggressive pricing in the humanoid robot market and a direct commoditization threat to Western competitors including 1X NEO. SP007, SP016
CP008 Physical Intelligence raised $400 million for its software-only approach building foundation models for robotic manipulation (pi-zero); it is not a direct commercial humanoid robot competitor but poses a platform risk if its models are adopted by hardware competitors. SP008, SP020
CP009 1X Technologies' EVE wheeled humanoid fleet (~390 units) represents the highest confirmed commercial deployment count among pure-play humanoid robot startups, giving 1X a real-world data advantage over competitors who have not yet deployed at scale. SP017, SP018
CP010 Agility Robotics Digit is the leading competitor on enterprise bipedal deployment, with Amazon's backing providing a clear path to scale in warehouse logistics that 1X cannot easily replicate without a comparable hyperscale logistics customer. SP010, SP024
CP011 Figure AI's OpenAI partnership provides it with a direct integration of frontier language models for robotic task planning—a capability that 1X does not have through a comparable co-development agreement, creating a potential AI capability gap. SP009, SP013
CP012 Tesla Optimus benefits from Tesla's Dojo supercomputer training infrastructure and Full Self-Driving neural network stack—an AI asset base that pure-play humanoid startups cannot replicate without billions in infrastructure investment. SP002, SP021
CP013 1X NEO's bipedal capabilities are not yet commercially proven; the robot is entering a market in 2026 where Agility Digit, Figure 02, and even Tesla Optimus (internal) have 12–24 months of bipedal operational experience ahead of NEO's first customer deliveries. SP018, SP019
CP014 Unitree G1 at $16,000 is the only commercially available bipedal humanoid robot priced below NEO's $20,000 price point, and it is currently a hardware-first product with limited built-in AI autonomy that requires third-party AI software integration for complex tasks. SP007, SP016
CP015 Physical Intelligence's pi-zero model is a hardware-agnostic foundation model for robotic manipulation that could be licensed to any humanoid robot company, potentially equalizing the AI capability gap between well-resourced (Figure, Tesla) and smaller competitors. SP008, SP020
CP016 Boston Dynamics commercial plans for Atlas in 2025 remain focused on industrial demonstration rather than broad commercial deployment; its Spot dog-like robot remains the primary commercial product with humanoid Atlas still in a transition phase under Hyundai. SP011, SP026
CP017 1X NEO is publicly priced at $20,000 for outright purchase or $499 per month for subscription, making it the only humanoid robot with a publicly disclosed retail consumer price and a monthly subscription option. SP025, SP018
CP018 Agility Robotics Digit enterprise pricing has not been publicly disclosed; analyst estimates and media reporting suggest a pricing model in the $100,000–150,000 per-unit range or equivalent annual contract value for the Amazon and GXO deployments. SP010, SP024
CP019 Tesla Optimus has no announced commercial third-party pricing; Elon Musk's statements about a long-term $20,000 consumer target price are aspirational and not backed by a current commercial product offer or manufacturing cost data. SP002, SP012
CP020 Figure AI has not publicly disclosed pricing for its BMW factory deployment or any future commercial robot offering; all Figure AI pricing data is either internal or covered by non-disclosure agreements. SP003, SP013
CP021 Unitree H1 at $90,000 and G1 at $16,000 are both offered as direct hardware sales with no subscription or RaaS option, positioning Unitree as a pure hardware vendor rather than a full-stack robotics-as-a-service provider. SP007, SP016
CP022 The 1X EVE enterprise pricing model—estimated at $3,500–5,000 per month or $100K+ outright—has not been publicly disclosed by 1X; these estimates are inferred from industry analogs and media coverage of the ADT and Everon deployments. SP017, SP025
CP023 Apptronik Apollo enterprise contract pricing has not been publicly disclosed; GE Aerospace and Mercedes-Benz partnership terms are subject to non-disclosure, making it impossible to assess Apptronik's unit economics or competitive pricing position. SP005, SP014
CP024 The humanoid robot industry has a general pricing opacity problem—no major competitor other than 1X NEO and Unitree has disclosed list pricing, making competitive pricing analysis dependent on analyst estimates and media inference. SP017, SP023
CP025 1X's most durable competitive moat is proprietary embodied AI training data from its commercial EVE fleet, generating real-world task data at scale that competitors without deployed fleets cannot replicate without multi-year commercial deployments. SP018, SP019
CP026 1X's OpenAI relationship is not exclusive—Figure AI also has an OpenAI partnership for language model integration into Figure 02's control system, limiting the strategic differentiation value of the 1X-OpenAI capital relationship. SP009, SP018
CP027 Enterprise humanoid robot deployments create switching costs through operational workflow redesign, proprietary API integrations, and task-specific AI model fine-tuning, but these switching costs are not as high as enterprise SaaS given robots can be physically removed and replaced. SP018, SP023
CP028 Agility Robotics' Amazon backing creates a structural distribution moat for Digit in warehouse logistics—Amazon can deploy Digit across its global fulfillment network without 1X or other competitors having access to that channel. SP004, SP010
CP029 The primary kill scenario for 1X Technologies is Tesla Optimus achieving commercial third-party availability at or below $20,000 with Tesla's brand, service network, and Dojo AI infrastructure—a combination 1X cannot match on capital, brand, or AI infrastructure. SP002, SP021
CP030 Figure AI's superior funding ($675M+ vs 1X's ~$237M) creates a capital adequacy gap that could compound over time if Figure AI uses its funding advantage to hire more AI researchers, run more robots, and collect more training data than 1X can afford. SP009, SP030
CP031 Unitree's G1 at $16,000 creates a persistent price floor below 1X's $20,000 NEO that 1X cannot undercut without destroying its unit economics, placing 1X in a strategic squeeze between Unitree's price and Tesla's brand. SP007, SP025
CP032 New Atlas's 2024 assessment criticized 1X EVE as "undergunned" compared to bipedal competitors, noting that EVE lacks feet and dexterous hands—a characterization that captures the form-factor disadvantage 1X must overcome with NEO in the bipedal market. SP019, SP017
CP033 Agility Robotics' Amazon deployment volume, contract terms, pricing, and renewal structure are not publicly disclosed, making it impossible to verify whether Agility has achieved a commercially sustainable unit economics model or is subsidizing deployments for learning. SP010, SP028
CP034 Tesla Optimus deployment metrics—task autonomy rate, override frequency, defect handling—are not publicly disclosed, preventing objective comparison of Tesla Optimus AI readiness versus 1X and other competitors. SP002, SP012
CP035 1X Technologies has not publicly disclosed the dollar value of its ADT and Everon contracts, making it impossible for external analysts to calculate contracted ARR, unit economics contribution, or whether the EVE business is self-sustaining or requires ongoing capital subsidy. SP017, SP025
CP036 Apptronik Apollo received enterprise pilot partnerships with GE Aerospace and Mercedes-Benz following its $53M Series A funding, demonstrating enterprise validation but at a smaller scale than Agility or Figure deployments. SP005, SP029
CP037 Boston Dynamics Atlas electric generation was unveiled in 2024, replacing the hydraulic version, with significantly improved power efficiency, dexterity, and form factor, but no announced commercial general-purpose deployment timeline. SP001, SP011
CP038 The humanoid robot competitive landscape changed materially in 2025, with multiple companies announcing factory pilots, Unitree achieving broad distribution at low price points, and Tesla accelerating internal Optimus deployment timelines. SP024, SP018
CP039 Physical Intelligence raised $400 million and its pi-zero foundation model represents a potential capability equalizer—if licensed broadly, it could allow lower-funded humanoid robot companies to close the AI capability gap with better-resourced competitors. SP008, SP020
CP040 Sanctuary AI Phoenix's seventh-generation platform includes a proprietary AI model called Carbon, focused on cognitive architecture for general-purpose robot intelligence; Sanctuary competes on AI software differentiation rather than hardware form factor. SP006, SP015
CI001 1X Technologies raised a $100 million Series B funding round in 2023, led by EQT Ventures with participation from OpenAI and NEA, making it one of the largest Series B raises for a humanoid robotics startup at the time. SI001, SI003
CI002 1X Technologies raised an additional $100 million in January 2025 from undisclosed investors, bringing confirmed total capital raised to approximately $237 million excluding Halodi Robotics pre-Series A rounds. SI004, SI007
CI003 Total confirmed capital raised by 1X Technologies as of mid-2026 is approximately $237 million, comprising Series A2 ($23.5 million), Series B ($100 million), and the January 2025 round ($100 million), excluding undisclosed early Halodi Robotics rounds. SI001, SI004, SI008
CI004 1X Technologies is reported to be targeting approximately $1 billion in Series C funding as of 2025–2026, with no closing date, lead investor, or term sheet publicly confirmed. SI005, SI010
CI005 EVE's publicly indicated RaaS pricing is approximately $3,500–5,000 per robot per month, making it the most explicitly priced commercial humanoid offering among pure-play startups as of 2026. SI011, SI012
CI006 NEO was announced by 1X Technologies at a list price of approximately $20,000 per unit as an outright hardware purchase aimed at the home consumer and SMB market. SI030, SI015
CI007 A monthly subscription layer for NEO at approximately $499 per month has been speculated in industry media but has not been formally confirmed by 1X Technologies as of May 2026. SI030, SI014
CI008 ADT has deployed approximately 140 EVE units under a commercial contract with 1X Technologies, representing the largest confirmed single enterprise bipedal humanoid deployment in North America as of 2026. SI012, SI013
CI009 Everon has deployed approximately 250 EVE units from 1X Technologies under a commercial managed services arrangement, making it the largest single humanoid fleet deployment globally by unit count as of 2026. SI012, SI014
CI010 1X Technologies' total commercially deployed EVE fleet stands at approximately 390-plus units across ADT (140) and Everon (~250) accounts, representing the largest confirmed humanoid robot deployment among pure-play startups globally as of 2026. SI012, SI013, SI014
CI011 No gross margin or COGS data has been publicly disclosed by 1X Technologies for any of its product lines or commercial deployments as of May 2026. SI011, SI017
CI012 1X Technologies' monthly cash burn is not publicly disclosed; industry estimates based on headcount benchmarks and operational scale suggest a range of $5–15 million per month as of 2025–2026. SI009, SI026
CI013 Estimated cash runway following the January 2025 $100 million raise is approximately 7–20 months depending on actual burn rate, implying a potential financing requirement in the range of mid-2026 to early-2027 if no further capital is raised. SI004, SI009
CI014 Hardware COGS for EVE is estimated at $15,000–25,000 per unit based on comparable humanoid robot hardware component costs, contract manufacturing benchmarks, and peer cost disclosures; this estimate is not confirmed by 1X. SI016, SI022
CI015 1X Technologies' Series B investors include EQT Ventures as lead, OpenAI as a strategic investor, and NEA as a financial investor, providing both strategic AI alignment and top-tier financial backing. SI001, SI020
CI016 Enterprise humanoid robot contracts are estimated to run 12–36 months with fleet-level pricing, based on industry RaaS norms and comparable robotic service deployment structures observed in the market. SI028, SI027
CI017 No official revenue figure, ARR, or MRR has been disclosed by 1X Technologies for any period as of May 2026, despite confirmed commercial deployments and over $237 million raised. SI011, SI014
CI018 1X Technologies operates R&D centres in both Norway and the United States, contributing to multi-site operating expenses that increase base burn rate relative to single-site companies. SI006, SI018
CI019 1X Technologies uses third-party contract manufacturers for EVE hardware production; the identity of the contract manufacturer and manufacturing location have not been publicly disclosed. SI015, SI027
CI020 1X Technologies (operating as Halodi Robotics) raised a Series A2 of $23.5 million prior to rebranding, representing early-stage capital that predates the current commercial product suite. SI003, SI008
CI021 Customer acquisition cost for enterprise humanoid robotics deployments is estimated at $50,000–200,000 per account based on comparable enterprise robotics and industrial automation sales cycle benchmarks; no 1X-specific CAC has been disclosed. SI028, SI017
CI022 R&D spending at humanoid robotics startups at 1X's stage typically represents 40–60% of total operating expenditure, based on comparable pre-revenue hardware robotics companies' benchmarks and analyst estimates. SI017, SI027
CI023 No independent audit or third-party verification of 1X Technologies' unit economics, revenue, or financial statements has been publicly disclosed; the company is private with no US public reporting obligation. SI002, SI008
CI024 Figure AI raised $675 million at a $2.6 billion valuation in early 2024, providing a peer valuation anchor for 1X Technologies' reported $1 billion Series C target. SI024, SI008
CI025 Agility Robotics closed a $150 million funding round in 2025 with Amazon continuing as a strategic backer, providing a peer benchmark for humanoid startup financing scale at similar commercial stages. SI025, SI024
CI026 Boston Dynamics generates estimated annual revenue of $50–100 million from its Spot and Stretch product lines, providing a best-in-class commercial robotics revenue benchmark for 1X's medium-term revenue ambitions with EVE. SI016, SI021
CI027 First-generation humanoid robot hardware gross margins are estimated at negative to low single-digit positive at current batch sizes below 500 units, improving materially at scale as manufacturing yield improves and BOM costs decline per industry analyst estimates. SI022, SI027
CI028 Manufacturing scale-up to 1,000-plus humanoid units per year requires significant capital investment in tooling, supplier qualification, and quality assurance, representing a key capex risk for 1X as it transitions from sub-scale to commercial manufacturing volumes. SI027, SI016
CI029 Humanoid robot inventory financing is a material working capital risk: robots must be manufactured and shipped before RaaS revenue accrues, requiring pre-deployment capital that increases with fleet expansion rate and contract lead times. SI028, SI027
CI030 1X Technologies maintains operational presence in both Norway (engineering and R&D) and the United States (commercial operations and Palo Alto HQ), creating multi-currency operational cost exposure. SI006, SI002
CI031 NEO was publicly announced in 2024 at a $20,000 list price; its commercial availability and volume shipment status as of mid-2026 remains unclear from public sources, with no confirmed unit count or customer list disclosed. SI030, SI014
CI032 The RaaS model reduces the upfront capital barrier for enterprise customers compared to outright hardware purchase, enabling faster enterprise adoption but requiring 1X to carry the capital cost of manufacturing before revenue accrues. SI028, SI011
CI033 Enterprise robotics sales cycles typically run 6–18 months from first engagement to deployment, based on industry benchmarks for comparable autonomous mobile robot and industrial automation deployments at large enterprise accounts. SI028, SI017
CI034 1X management has indicated in public materials that hardware costs are expected to decline with scale, consistent with learning curve effects in hardware manufacturing, but no specific cost reduction targets or timelines have been disclosed. SI015, SI011
CI035 OpenAI participated as a strategic investor in 1X Technologies' Series B, providing credibility for the company's AI-driven autonomous humanoid positioning and potential integration with OpenAI foundation models for robot cognition. SI001, SI019
CI036 Industry norms for pre-IPO humanoid robotics startups suggest Series C timing depends on hitting revenue milestones ($20–50M ARR) or manufacturing scale targets above 1,000 units deployed to command the valuation required for a $1 billion raise. SI009, SI017
CI037 Reliance on contract manufacturing introduces supply chain concentration risk for 1X: any disruption to the contract manufacturer could delay deliveries, impair RaaS revenue accrual, and increase unit cost due to re-qualification requirements. SI027, SI021
CI038 Unitree's asset-light model (lower hardware cost, software-defined) presents a contrasting capital efficiency profile to 1X's vertically integrated humanoid approach, creating pricing pressure at the lower end of the enterprise humanoid market. SI016, SI021
CI039 No government, defence, or public-sector contract revenue has been disclosed by 1X Technologies as of May 2026; the company's confirmed commercial focus is on private enterprise accounts. SI006, SI014
CI040 The competitive humanoid market is driving pricing pressure on enterprise RaaS models as more players commercialise, potentially limiting 1X's ability to sustain $3,500–5,000 per robot per month pricing in the medium term without differentiated autonomy capabilities. SI009, SI026
CE001 NEO Gamma was launched in February 2025 as 1X's second-generation bipedal robot, representing a production-intent design advance over NEO Beta. SE001, SE005
CE002 NEO Gamma integrates a full-body 3D-knitted suit developed with Kind Humanoid that enables anthropomorphic range of motion and soft-touch surfaces suited to home environments. SE001, SE017
CE003 NEO Gamma employs a 100Hz active balance control loop for real-time bipedal stability. SE001, SE005
CE004 NEO Gamma stands approximately 163cm tall, consistent with adult human scale for home environments. SE001, SE008
CE005 EVE is 1X's wheeled humanoid robot standing approximately 140cm tall and weighing around 70kg, optimised for continuous enterprise deployment. SE002, SE005
CE006 EVE is engineered for 24/7 continuous commercial operation, a key differentiator versus hourly or shift-based human labour alternatives. SE002, SE008
CE007 EVE can perform door-opening and load-carrying tasks in commercial facilities, demonstrated in ADT and Everon deployments. SE002, SE016
CE008 NEO Beta was launched in August 2024 as the first full-scale bipedal prototype, validating the embodied AI control architecture before NEO Gamma. SE015, SE019
CE009 1X trains its robots using reinforcement learning initialised from motion capture reference data, a technique grounded in peer-reviewed physics-based character animation research. SE003, SE006
CE010 The embodied learning data engine, developed under VP AI Eric Jang, converts VR teleoperator demonstrations into neural network training data in continuous pipeline fashion. SE006, SE014, SE022
CE011 1X's end-to-end neural network controls all robot behaviour without hardcoded task logic or modular behaviour trees, jointly optimising from sensor inputs to motor commands. SE011, SE013, SE029
CE012 1X Technologies acquired Kind Humanoid in January 2025 to accelerate NEO Gamma's home-form-factor design. SE026, SE015
CE013 The Kind Humanoid acquisition brought industrial designer Yves Béhar's ergonomic chassis concept, now integrated into NEO Gamma's structural design. SE026, SE009
CE014 Kind Humanoid co-founder Christoph Kohstall joined 1X as an engineering contributor following the acquisition. SE026, SE009
CE015 1X's proprietary low-impedance electric actuators are the core hardware IP, developed over ten years from the Halodi Robotics era, providing a significant barrier against replication. SE007, SE010, SE020
CE016 Low-impedance actuator design provides inherent mechanical compliance: the robot absorbs rather than resists impact forces, reducing injury risk without complex real-time force control software. SE007, SE020, SE028
CE017 Patent US11498217B2 covers core actuator technology originally filed by Halodi Robotics and assigned to 1X Technologies, with multiple continuation patents under active prosecution. SE007, SE015
CE018 1X maintains an ongoing actuator patent portfolio derived from Halodi Robotics filings, covering actuator geometry and control approaches, with additional prosecution underway. SE007, SE020
CE019 The github.com/1x-technologies repository shows active engineering activity with regular commits and multiple contributors, confirming ongoing software development cadence. SE012, SE015
CE020 No public references to ROS or other standard robot operating frameworks have been identified in 1X's stack, consistent with their end-to-end neural network approach that makes traditional perception-planning stacks architecturally redundant. SE011, SE012, SE029
CE021 Current NEO autonomy is approximately 70% teleop-assisted for complex tasks; routine tasks are increasingly handled autonomously but no official autonomy benchmark has been published. SE015, SE016, SE030
CE022 1X is developing a foundation model for multi-task robot generalisation, targeting π-style architectures that allow fine-tuning rather than full retraining per new task. SE027, SE029, SE013
CE023 The embodied AI data flywheel collects continuous real-world interaction data from EVE commercial deployments, creating a training data advantage that competitors without deployed fleets cannot replicate. SE006, SE014, SE013
CE024 NEO pre-orders opened October 2025 at $20,000 outright or $499 per month subscription, making it one of the first consumer-priced bipedal humanoid robots in the market. SE025, SE018
CE025 NEO Gamma consumer delivery is targeted for 2026; no production batch sizes or confirmed manufacturing partner have been disclosed publicly. SE025, SE018, SE019
CE026 NEO Gamma is designed specifically for home environments with emphasis on aesthetic form factor, soft surfaces, and safe human proximity, informed by the Kind Humanoid acquisition. SE023, SE028
CE027 Safety systems across 1X robots include physical emergency stop mechanisms, low-impedance compliance for soft-collision safety, and remote teleoperation oversight for complex task scenarios. SE028, SE002
CE028 The 3D-knitted suit is structurally integrated with the actuation system, enabling anthropomorphic ranges of motion not achievable with rigid exo-skeletons. SE001, SE017
CE029 1X uses motion capture reference data to initialise RL-based locomotion policies, a technique validated in peer-reviewed research on physics-based humanoid animation (arXiv 2402.19469). SE003, SE021
CE030 1X's end-to-end learning architecture contrasts fundamentally with modular robot control stacks: there is no separate perception module, state estimator, or behaviour tree; generalisation emerges from the trained model. SE011, SE021, SE029
CE031 1X's embodied AI research draws on peer-reviewed techniques including imitation learning, behaviour cloning, and sim-to-real transfer published in leading AI and robotics venues. SE003, SE004, SE021
CE032 1X is vertically integrated: proprietary actuators, in-house neural network AI, custom sensor suite, and internal manufacturing — analogous to the full-stack hardware-software co-design model. SE001, SE002, SE007
CE033 The 80,000 sq ft Palo Alto HQ, opened summer 2025, consolidates 1X's engineering and AI training teams and is sized to accommodate up to 400 employees. SE018, SE019
CE034 NEO Gamma's 2026-2027 development roadmap targets progressive autonomy improvement from ~70% teleop to meaningful autonomous operation for a defined subset of household tasks. SE015, SE025, SE030
CE035 1X's foundation model development targets multi-task generalisation using π-style architectures, aligning with the industry direction set by Physical Intelligence, Figure AI, and others. SE021, SE027
CE036 Hardware reliability and mean-time-between-failure (MTBF) data for the EVE commercial fleet has not been publicly disclosed by 1X Technologies as of Q2 2026. SE008, SE020
CE037 Cybersecurity and data protection controls for NEO's camera, microphone, and sensor systems operating in private home environments have not been detailed in any public disclosure. SE028, SE023
CE038 The EU AI Act may classify certain autonomous robot functions as high-risk AI systems requiring conformity assessments and Notified Body review; 1X has not publicly disclosed its EU compliance strategy. SE006, SE029
CE039 EVE commercial fleet deployments at ADT and Everon actively generate teleoperation data fed back into the embodied learning pipeline for continuous model improvement. SE013, SE014, SE022
CE040 NEO Gamma's safety architecture currently relies on remote human teleoperator oversight rather than independently certified fail-safe autonomous operation, a gap relative to ISO 13482 personal care robot standards. SE028, SE015, SE016
CU001 ADT Inc. has deployed approximately 140 EVE units for after-hours commercial security patrol at enterprise facilities. SU001, SU004, SU014
CU002 Everon has deployed approximately 250 EVE units across its commercial client base, representing the single largest known EVE deployment. SU002, SU007, SU019
CU003 ADT and Everon collectively account for approximately 390 EVE units in commercial deployment, representing the near-totality of 1X's enterprise fleet. SU003, SU012
CU004 ADT Inc. is a Fortune 500 security services company, providing 1X with a high-quality named enterprise reference account. SU004, SU014
CU005 Everon is the commercial security integrator formerly known as ADT Commercial, rebranded in November 2022; it is a leading North American commercial security operator. SU002, SU007
CU006 Sunnaasen Hospital in Norway uses EVE for healthcare-related robotics tasks at small scale; the deployment appears to be at pilot or early-production stage. SU015, SU012
CU007 Norwegian universities use EVE units for academic robotics research, providing international market presence but limited commercial value. SU015, SU024
CU008 NEO pre-orders opened in October 2025 at $20,000 outright or $499 per month, positioning 1X as the first company to accept consumer-facing orders for a bipedal home robot. SU010, SU021
CU009 The Wall Street Journal (October 2025) reported that NEO's teleoperation model allows humans to see inside consumers' homes, raising privacy and consent concerns among potential buyers. SU016, SU023
CU010 Following WSJ privacy coverage, 1X added consent controls and content blurring to NEO's teleoperator view, representing a product response to adverse consumer feedback. SU018, SU024
CU011 Customer concentration is high: ADT and Everon represent an estimated 95%+ of all commercial EVE units deployed, creating significant revenue exposure if either relationship deteriorates. SU003, SU013
CU012 No NPS scores, customer satisfaction indices, renewal rates, or churn figures have been publicly disclosed by 1X Technologies for enterprise or consumer segments. SU017, SU020
CU013 Enterprise EVE deployments appear to operate on multi-year RaaS contract structures, consistent with enterprise security services industry norms. SU006, SU022
CU014 No public ROI case studies, payback period analyses, or named customer testimonials from ADT or Everon have been published as of Q2 2026. SU028, SU013
CU015 Consumer reception to NEO pre-orders was mixed: media coverage reflected enthusiasm for home robotics combined with privacy concerns around teleoperation and data access. SU016, SU018
CU016 1X's go-to-market for EVE operates through direct enterprise relationships with security integrators, using a channel-based approach via ADT and Everon as anchor accounts. SU004, SU007, SU025
CU017 ADT's EVE deployment uses the robots for after-hours patrol at commercial facilities, reducing reliance on human security guards during overnight hours. SU001, SU004, SU027
CU018 The Everon partnership with 1X was formally announced in February 2024 via BusinessWire; subsequent reporting confirmed the ~250-unit fleet scale. SU002, SU007, SU028
CU019 The enterprise security robot market, EVE's primary segment, is growing with competitors including Knightscope and Cobalt Robotics, though none offer bipedal humanoid form factors. SU005, SU009, SU025
CU020 NEO targets the consumer home-use segment — a market with no established humanoid robot precedent — making adoption curve projections highly speculative. SU011, SU021, SU030
CU021 1X has no publicly announced enterprise customers beyond ADT, Everon, and small-scale research institutions as of Q2 2026. SU013, SU020
CU022 Privacy concerns from WSJ and Wired coverage created reputational risk for 1X's consumer market entry, requiring active public-response measures and product modifications. SU016, SU023
CU023 ADT's EVE deployment appears to be production-stage rather than pilot, given 140-unit fleet size, ongoing RaaS contract, and multi-year relationship duration. SU001, SU004, SU019
CU024 Everon's EVE deployment represents a multi-site rollout across its commercial security client base, indicating a channel-mediated expansion model. SU002, SU028
CU025 The commercial security robot market is estimated as a multi-billion dollar addressable opportunity, supporting expansion potential beyond current ADT/Everon accounts. SU005, SU009
CU026 1X has not publicly disclosed enterprise contract renewal rates, extension terms, or whether ADT and Everon have expanded their fleets since initial deployment. SU017, SU022
CU027 No enterprise customers outside the security vertical have been publicly announced; healthcare and logistics expansion remains a stated aspiration without confirmed pipeline. SU006, SU013
CU028 NEO's consumer market entry faces a privacy education challenge: teleoperation-dependent autonomy inherently requires humans to see private environments, a concern that advertising cannot fully resolve. SU016, SU023, SU030
CU029 The ADT–1X relationship predates the company's March 2023 rebrand from Halodi Robotics, indicating a partnership spanning at least three years. SU002, SU007
CU030 Enterprise security robot adoption is growing across North America, suggesting potential for 1X to expand beyond ADT and Everon to new security integrator partners. SU009, SU025, SU026
CU031 1X has not disclosed consumer pre-order volume, deposit size, cancellation policy, or NEO delivery batch structure. SU010, SU021
CU032 Enterprise expansion into logistics, healthcare, and hospitality is a plausible next step but has no confirmed customer traction or named pilot as of Q2 2026. SU006, SU013
CU033 The Sunnaasen Norway deployment signals some international (European) market traction but at research or pilot scale insufficient to confirm production commercial demand. SU015, SU026
CU034 No EVE utilisation data — hours per robot per day, tasks completed, or task success rate — has been published from enterprise deployments. SU017, SU028
CU035 Consumer media previews of NEO Gamma generated generally positive reactions to the form factor and home-robot concept prior to privacy concerns emerging. SU011, SU021
CU036 Wired's privacy coverage specifically flagged the risk of teleoperator voyeurism — the ability for humans guiding NEO to observe private home activities in real time. SU023, SU016
CU037 PitchBook noted 1X's commercial traction as a differentiating factor compared to humanoid robot peers that have not yet achieved production enterprise deployments. SU020, SU022
CU038 Fortune's April 2025 reporting indicated positive operational outcomes from ADT's EVE fleet, though no quantitative ROI metrics or utilisation statistics were disclosed. SU027, SU019
CR001 The EU AI Act classifies certain autonomous AI system functions as high-risk, potentially requiring conformity assessment and Notified Body review before EU market entry for NEO. SR006, SR022
CR002 The FTC's 2024 commercial surveillance report established a regulatory framework that could directly apply to NEO's home data collection practices, including video and audio capture. SR001, SR007
CR003 GDPR Articles 5-7 impose strict data minimisation, consent, and legal-basis requirements on personal data collected by robots in EU member states, including Norway. SR002, SR003
CR004 OSHA regulations under 29 USC 652 impose employer safety obligations for workers in proximity to robots in US commercial workplaces, applying to EVE's ADT and Everon deployments. SR005, SR028
CR005 The proposed AI ROBOT Act (Senate Bill 2718, 118th Congress) would impose new federal robot safety standards if enacted, but has not been enacted as of Q2 2026. SR003, SR007
CR006 NIST's AI Risk Management Framework provides voluntary guidance applicable to 1X's AI systems and may become a procurement requirement by enterprise customers. SR004, SR026
CR007 1X Technologies has not confirmed OSHA compliance audits, ISO 10218 robot safety certification, or ISO 13482 personal care robot certification as of Q2 2026. SR009, SR010
CR008 Hardware reliability data — MTBF, field failure rates, and incident logs — for EVE's commercial fleet is not publicly disclosed, making quality risk at scale unquantifiable. SR008, SR021
CR009 Industry-wide robot safety incidents, including falls, pinch injuries, and recalls, suggest that hardware failure risk becomes material at fleet scales above ~500 units — a threshold 1X is approaching. SR008, SR021
CR010 CEO Bernt Børnich and VP AI Eric Jang are identified as critical key-person dependencies; either departure would create an existential or near-existential risk to company execution. SR014, SR020
CR011 1X's OpenAI partnership creates strategic AI dependency; a pivot in OpenAI's robotics strategy or partnership termination would require 1X to rapidly internalise AI model development. SR013, SR020
CR012 EQT Ventures as primary capital provider with board representation creates an investor dependency; adverse board dynamics could constrain 1X's Series C options and governance decisions. SR015, SR020
CR013 Labor displacement opposition from unions and policy advocates could create regulatory headwinds and public-affairs costs for commercial EVE deployments in the security sector. SR011, SR016
CR014 Semiconductor supply chain risks and actuator component dependencies create manufacturing delay risk for NEO Gamma at consumer-scale production volumes. SR015, SR019
CR015 Cybersecurity vulnerabilities in always-on connected robots with cameras, microphones, and physical actuators represent an attack surface for data exfiltration, surveillance abuse, and physical disruption. SR017, SR019
CR016 NEO privacy incidents — teleoperator home video access — could trigger FTC enforcement, GDPR penalties, and EU AI Act non-compliance findings simultaneously across jurisdictions. SR001, SR007, SR002
CR017 WSJ and Wired's privacy coverage (October-November 2025) has already created reputational risk for NEO's consumer market entry, requiring active public response measures. SR012, SR029
CR018 Tesla Optimus, if successfully scaled to consumer volumes, could commoditise the humanoid robot market and erode 1X's differentiation in both enterprise and consumer segments. SR025, SR013
CR019 Failure to close the Series C at sufficient size before NEO manufacturing scale-up would create a capital gap, leading to delivery delay, customer attrition, and potential distress. SR020, SR023
CR020 No IP litigation, patent disputes, or trade secret claims involving 1X Technologies have been publicly identified as of Q2 2026. SR025, SR010
CR021 OSHA statute 29 USC 652 defines 'employer' obligations and 'working conditions' standards that apply directly to EVE robot deployments in US commercial workplaces. SR028, SR005
CR022 EU AI Act Article 9-11 conformity assessment requirements apply to high-risk AI systems; if NEO or EVE are classified as high-risk, CE marking and Notified Body certification are mandatory for EU sales. SR006, SR022
CR023 GDPR enforcement imposes potential fines of up to 4% of global annual turnover for serious violations of data processing principles, representing a material financial risk for EU robot operations. SR002, SR003
CR024 No evidence of actual data breaches, cybersecurity incidents, or unauthorised access involving 1X Technologies or its deployed robots has been publicly reported as of Q2 2026. SR017, SR019
CR025 Supply chain concentration in proprietary actuator components — without confirmed dual-sourcing — creates a manufacturing single-point-of-failure not mitigated in public disclosures. SR015, SR014
CR026 Loss of VP AI Eric Jang would likely cause a 12-18 month AI capability setback given his unique expertise in the embodied learning data engine and RL-from-teleop architecture. SR014, SR030
CR027 Labor and union opposition to commercial EVE deployments has not resulted in confirmed boycotts or regulatory actions as of Q2 2026, but the risk remains active as deployment scale grows. SR011, SR016
CR028 Hardware recall risk is non-trivial at scale; if a systematic safety defect in the EVE or NEO fleet is discovered, recalling 390+ units would be operationally complex and reputationally damaging. SR008, SR010, SR021
CR029 No confirmed IP litigation, patent challenges, or trade secret disputes targeting 1X Technologies have been publicly identified as of Q2 2026. SR025, SR020
CR030 NIST AI RMF adoption by enterprise customers as a procurement condition is a growing trend in US government and regulated industries, potentially creating a compliance overhead for 1X. SR004, SR026
CR031 Cybersecurity risks for home robots are materially higher than for industrial robots due to continuous recording of private living spaces and direct internet connectivity. SR017, SR019, SR029
CR032 1X's reliance on OpenAI for strategic AI model access and investor-community signalling creates a supply-side risk; any change in OpenAI's robotics posture would require rapid response. SR013, SR018
CR033 The AI ROBOT Act (S. 2718) if enacted would require robot manufacturers to meet new federal safety standards, creating compliance cost and potential market access barriers in the US. SR003, SR007
CR034 EU AI Act Articles 13-14 on transparency and human oversight may require specific documentation for 1X's autonomous AI systems, adding regulatory burden to EU market entry. SR006, SR022
CR035 Consumer privacy regulations under CCPA and state equivalents apply to NEO's data collection in the US consumer market, adding to the federal FTC and state-level privacy exposure. SR001, SR007
CR036 The competitive talent market for AI robotics engineers (Google DeepMind, Meta AI, Tesla, Figure AI all recruiting from the same pool) creates attrition risk beyond key-person dependency. SR014, SR030
CR037 RaaS revenue concentration in ADT and Everon means a single contract termination could reduce 1X's enterprise ARR by 40-60%, requiring emergency cost reduction or emergency fundraising. SR020, SR023
CR038 Hardware reliability at consumer scale (thousands of units) has not been demonstrated by 1X; NEO's bipedal complexity adds new failure modes (falls, joint wear) absent in EVE's wheeled design. SR008, SR021
CR039 GDPR enforcement by EU DPAs (Irish DPC, French CNIL, others) has been active against tech companies for comparable IoT data collection; precedent suggests 1X's EU operations face real enforcement risk. SR002, SR022
CR040 If NEO consumer delivery is delayed beyond 12 months from pre-order opening (i.e., beyond October 2026), pre-order cancellations and investor confidence degradation could become unrecoverable. SR020, SR023
CV001 1X Technologies has commercially deployed approximately 390 EVE units (ADT ~140 units, Everon ~250 units) at enterprise customers, making it the most commercially deployed humanoid robotics company in the enterprise security and facilities management segment as of Q2 2026. SV005, SV009
CV002 1X Technologies has raised approximately $237M in confirmed capital across Series A2 ($23.5M, March 2023), Series B ($100M, January 2024, EQT Ventures lead), and an additional approximately $100M (January 2025), all at undisclosed post-money valuations. SV005, SV010
CV003 No valuation figure has been publicly disclosed for 1X Technologies at any confirmed funding round; the company remains opaque on cap table structure, preference stack, and post-money valuations, which limits the precision of any valuation analysis to comparables and scenario modeling. SV003, SV009
CV004 Figure AI closed a $675M funding round at a $2.6B post-money valuation in February 2024, followed by undisclosed additional financing that pushed the implied valuation toward approximately $3.2B by late 2024, making it the highest-funded and highest-implied-valuation pure-play humanoid robotics company. SV017, SV018
CV005 Boston Dynamics was acquired by Hyundai Motor Group in 2021 at an acquisition price of approximately $1.1 billion, providing an M&A precedent reference for enterprise robotics valuations at commercial scale with confirmed positive ARR. SV027, SV010
CV006 Agility Robotics raised capital implying a valuation of approximately $1B and has deployed Digit robots at Amazon warehouses through the Amazon ORCA program, representing the closest enterprise deployment comparable to 1X's EVE commercial deployments. SV019, SV004
CV007 Tesla's Optimus program has received no independent disclosed valuation; analyst estimates imply $50–100B or more in value if broken out from Tesla's consolidated equity, making it an unsuitable direct pricing comparable but a critical scale-threat reference in competitive analysis. SV017, SV001
CV008 Goldman Sachs projects the addressable robotics market at $38B by 2030, while the IFR Service Robots report estimates the service robot market at approximately $46.6B by 2025, providing a large-market context that justifies premium private market entry multiples for humanoid robotics category leaders. SV001, SV016
CV009 ARK Invest's physical AI valuation framework suggests category-defining embodied AI companies could capture 5–15% of a $10 trillion or more physical labor market by 2030, justifying high entry valuations for early leaders with data flywheel and commercial deployment proof. SV002, SV001
CV010 Goldman Sachs' humanoid robot investment thesis identifies data flywheel, manufacturing scale, and commercial deployment proof as the three primary value drivers for private market robotics valuations in 2025; 1X meets the data flywheel and deployment proof criteria but not yet manufacturing scale. SV001, SV002
CV011 1X Technologies opened NEO pre-orders in October 2025 at $20,000 per unit or $499 per month subscription; no pre-order backlog count or deposit total has been publicly disclosed as of Q2 2026. SV009, SV013
CV012 At estimated EVE RaaS revenue of $3,500–5,000 per unit per month and approximately 390 deployed units, the implied annual recurring revenue is approximately $15–25M — insufficient to justify a $3–5B Series C valuation on revenue multiples alone; the premium is future-state dependent on NEO commercialization. SV004, SV003
CV013 TechCrunch reported in April 2025 that 1X Technologies is exploring a $1B Series C round; the round status remains unconfirmed as of Q2 2026, creating the primary financing uncertainty in the investment thesis. SV009, SV013
CV014 The Wall Street Journal published analysis in 2025 raising concerns about humanoid robot startup valuations outpacing commercial proof, referencing sector-wide dynamics that apply directly to 1X Technologies' implied $3–5B Series C valuation relative to its $15–25M estimated ARR. SV011, SV024
CV015 Wired reporting in 2025 highlighted potential VC bubble dynamics in humanoid robotics, noting that most companies — including 1X — have raised capital based on product demos and partnership announcements rather than verifiable commercial-scale revenue, which represents an adverse signal for valuation discipline. SV024, SV011
CV016 OpenAI's investment via the OpenAI Startup Fund provides 1X Technologies strategic AI model access, investor-ecosystem credibility and signalling, and preferential access to frontier AI training insights that competitors without equivalent partnerships cannot readily replicate. SV009, SV025
CV017 EQT Ventures' position as Series B lead investor provides European distribution network access, Series C bridge-investor credibility, and board-level governance oversight that reduces execution risk for prospective Series C investors relative to founder-only-led comparable startups. SV005, SV013
CV018 Tesla has moved Optimus to internal production use at its own facilities, with Elon Musk targeting thousands of units by 2025–2026; if production targets are met, Tesla enters the consumer humanoid market with approximately 10x 1X's capital base, manufacturing scale, and existing customer distribution network. SV017, SV008
CV019 The bear case for 1X Technologies is an IP-plus-team liquidation scenario: if NEO fails to commercialize and EVE enterprise contracts lapse, the company's residual assets are its patent portfolio, engineering team, and teleoperation training data, with estimated distressed M&A value of $200–400M. SV006, SV020
CV020 The base case valuation for 1X Technologies assumes a 2028 EVE-led revenue run rate of approximately $50–80M ARR and a 12–18x revenue multiple for a profitable-trajectory robotics-as-a-service company, implying an exit valuation of approximately $1.2–1.5B at 2028 exit. SV003, SV006
CV021 The bull case for 1X Technologies assumes NEO ships to 10,000 or more consumer homes by 2030, generating $500M or more in annual revenue and commanding an 8–12x revenue multiple in a re-rated humanoid robot market, implying an exit valuation of $4–6B. SV002, SV009
CV022 Apptronik raised $53M from General Catalyst in August 2024 at an implied valuation of approximately $300M, placing it at an earlier commercial stage than 1X but demonstrating sector investor appetite for pre-revenue humanoid robotics companies from top-tier investors. SV004, SV010
CV023 Sanctuary AI's Phoenix robot has been deployed in Canadian retail environments with an estimated valuation of approximately $500M CAD; its teleoperation-first strategy parallels 1X's NEO approach and serves as a structural analog for comparable valuation analysis. SV004, SV027
CV024 Comparable company analysis across Figure AI, Boston Dynamics, Agility Robotics, Apptronik, and Sanctuary AI implies a peer mid-point valuation of approximately $1.0–1.5B for 1X based on current commercial metrics, with a 1.5–2.5x premium justifiable if NEO pre-orders confirm consumer demand. SV003, SV020
CV025 PitchBook data indicates humanoid robot startup valuation multiples in 2025 are driven primarily by deployment proof, total funded capital, and lead investor tier — not yet by revenue multiples — which partially supports 1X's premium entry price despite limited disclosed revenue. SV003, SV019
CV026 The IFR Service Robots report estimates the service robot market at $46.6B by 2025 with sustained growth projected through 2030, underpinning the TAM assumptions for 1X's commercial robotics segment and supporting a premium valuation for category leaders. SV016, SV014
CV027 Mordor Intelligence and MarketsAndMarkets both project service robotics compound annual growth rates of 20–25% through 2030, consistent with Goldman Sachs' $38B by 2030 estimate and supporting the large-market context for premium entry multiples in the humanoid robotics category. SV014, SV015
CV028 The overall investment recommendation for 1X Technologies is TRACK at 6.5/10 with medium confidence and high risk rating — above the PASS threshold but below the INVEST threshold — acknowledging real commercial EVE traction but insufficient public evidence to justify a Series C commitment without further diligence. SV001, SV006
CV029 The valuation stance for 1X Technologies is "stretched": the implied $3–5B Series C entry valuation is above peer comparable benchmarks given current revenue opacity, but the premium reflects an OpenAI partnership premium and the unique EVE commercial proof that no direct peer can match. SV003, SV011
CV030 Changing the investment recommendation from TRACK to INVEST would require: confirmed Series C close at disclosed valuation, NEO pre-order count above 1,000 units with deposits, EVE ARR confirmed above $30M, and a verifiable autonomy roadmap showing progression from teleoperation-dominant to autonomous operation. SV006, SV028
CV031 The investment thesis-break triggers for 1X Technologies are: Series C not closed by Q4 2026; ADT or Everon contract terminated or fleet reduced more than 20%; NEO delivery delayed beyond Q4 2027; CEO Børnich or VP AI Jang departure without successor; or a named regulatory enforcement action. SV028, SV006
CV032 1X's VR teleoperation data flywheel — years of embodied AI training data from EVE's commercial deployments — is the most defensible competitive moat element; competitors without an equivalent deployed fleet cannot replicate this proprietary data asset in the near to medium term. SV012, SV021
CV033 From a capital efficiency standpoint, 1X's approximately 390 confirmed commercial EVE deployments on $237M raised compares favourably to Figure AI's zero confirmed revenue-paying external deployments on $675M raised, suggesting 1X has translated investor capital into commercial proof more effectively. SV004, SV018
CV034 The VC funding environment for humanoid robotics in 2025 remains robust with multiple rounds above $100M, supporting a Series C close assumption for well-positioned companies; Bloomberg's analysis confirms institutional investor demand remains high despite valuation discipline concerns. SV010, SV025
CV035 Reuters reporting confirms 1X Technologies is targeting approximately $1B in Series C financing; the round size aligns with NEO manufacturing scale-up capital requirements and is consistent with sector comparables at similar growth stages. SV026, SV008
CV036 1X's 80,000 square-foot Palo Alto HQ consolidation in summer 2025 and organisational scale-up to approximately 400-seat capacity signals readiness for NEO production ramp and Series C scaling, consistent with a company preparing for a major financing and manufacturing expansion. SV023, SV009
CV037 An IPO pathway for 1X Technologies is possible but distant — requiring 2–3 years of post-Series C revenue scaling to reach $100M or more ARR necessary for public market comparables; strategic M&A or secondary market tender offers are the more likely near-term exit mechanisms. SV007, SV022
CV038 Strategic acquirers for 1X Technologies could include Samsung (existing investor via Samsung NEXT), Amazon (established robotics track record including Agility), Google/Alphabet (active AI robotics investment interest), or a major automotive OEM seeking a household robotics platform. SV025, SV027
CV039 The final diligence ask list for Series C investors includes eight items: EVE revenue run rate, NEO pre-order backlog, burn rate and runway, Series C terms and investor composition, autonomy level metrics, hardware MTBF, full patent portfolio, and board composition and investor rights. SV029, SV030
CV040 Bloomberg's April 2025 analysis of humanoid robot investments noted that valuation premiums over comparable-stage companies are justified only when commercial proof substantially exceeds sector averages — a test 1X passes narrowly with EVE deployments at ADT and Everon, supporting a moderate premium. SV027, SV010
来源
编号出版方标题引文
SO001 1X Technologies 1X — Home Robots (Official Homepage)
SO002 Wikipedia 1X Technologies — Wikipedia
SO003 TechCrunch OpenAI-backed 1X raises another $100M for the race to humanoid robots
SO004 VentureBeat 1X, robotic startup backed by OpenAI, receives $100M in funding
SO005 1X Technologies 1X Opens Global HQ in Palo Alto | AI Robotics
SO006 TechCrunch The robots are already here
SO007 1X Technologies 1X Unveils NEO Beta, A Humanoid Robot for the Home
SO008 1X Technologies Introducing NEO Gamma
SO009 New Atlas OpenAI's Eve humanoids make impressive progress in autonomous work 1X's humanoids look oddly undergunned next to what, say, Tesla, Figure, Sanctuary or Agility are working on. The Eve humanoid doesn't even have feet at this point, or dextrous humanoid hands.
SO010 The Robot Report 1X Technologies acquires Kind Humanoid to accelerate work toward household robots
SO011 TechCrunch Norway's 1X is building a humanoid robot for the home
SO012 Engadget 1X Neo is a $20,000 home robot that will learn chores via teleoperation
SO013 Freethink Humanoid helpers are now entering our homes
SO014 Dezeen Scandi robot servant Neo Gamma is dressed head-to-toe in beige knitwear
SO015 The Information Humanoid Robot Developer 1X Targets $1 Billion in New Funding
SO016 EE Journal Halodi's Humanoid Robots are Already Amongst Us!
SO017 New Atlas Video: Eve humanoid voice-prompted to perform back-to-back multi-tasking
SO018 Goldman Sachs Research The global market for humanoid robots could reach $38 billion by 2035
SO019 Bloomberg Humanoid Robot Maker 1X Raises $100 Million
SO020 Bloomberg Humanoid Robot Startup 1X Raises Funds
SO021 The Wall Street Journal I Tried the Robot That's Coming to Live With You. It's Still Part Human Most tasks were teleoperated by a human with a virtual reality headset, raising concerns about privacy.
SO022 1X Technologies NEO Home Robot — 1X Product Page
SO023 The New York Times Invasion of the Home Humanoid Robots
SO024 OpenAI OpenAI Startup Fund
SO025 EE Journal Halodi's Humanoid Robots — ADT and Deployment Context
SM001 Goldman Sachs Research Humanoid Robots: An AI-Powered Physical Labor Solution 2024
SM002 International Federation of Robotics Record 4 Million Robots Work in Factories Worldwide
SM003 McKinsey Global Institute Jobs Lost, Jobs Gained: What the Future of Work Will Mean for Jobs, Skills and Wages
SM004 Boston Consulting Group Advantages of Investing in Digital Dexterity
SM005 ARK Invest Humanoid Robots — ARK Analyst Research
SM006 Mordor Intelligence Humanoid Robot Market Report
SM007 MarketsandMarkets Humanoid Robot Market Report
SM008 IEEE Spectrum Humanoid Robots 2025
SM009 MIT Technology Review Humanoid Robots 2024 Overview
SM010 TechCrunch 2024: The Year Humanoid Robots Got Serious
SM011 Fortune Humanoid Robots 2025 Market Outlook
SM012 Reuters Humanoid Robots Taking Shape as Real-World Deployments Grow
SM013 Harvard Business Review Managing the Risks of GenAI
SM014 IDC IDC Global Service Robot Market Forecast
SM015 Grand View Research Humanoid Robot Market Analysis
SM016 PitchBook Humanoid Robots VC Funding 2025 Outlook
SM017 World Economic Forum The Future of Jobs Report 2025
SM018 Deloitte Insights Robot Revolution: Future of Work
SM019 UBS Wealth Management Humanoid Robots 2025
SM020 CNBC Humanoid Robots Market 2025 Biggest Players
SM021 The Robot Report Humanoid Robot Market Forecast 2025
SM022 Wired Humanoid Robots 2025 Real World
SM023 Statista Worldwide Humanoid Robot Market Revenue
SM024 ABI Research Service Robotics Market Research
SM025 Axios Humanoid Robots Commercial Deployment 2025
SM026 New Atlas Best Humanoid Robots 2025 Guide
SM027 Business Insider Humanoid Robots Companies 2025 Overview
SM028 TechCrunch CES 2025 Humanoid Robots Everywhere
SM029 Bloomberg Humanoid Robots Market 2025
SM030 VentureBeat Humanoid Robots 2025 Market Outlook VC Funding
SP001 Boston Dynamics Atlas — Boston Dynamics
SP002 Tesla Tesla Optimus Robot
SP003 Figure AI Figure AI — Humanoid Robots
SP004 Agility Robotics Digit — Agility Robotics
SP005 Apptronik Apollo — Apptronik
SP006 Sanctuary AI Phoenix — Sanctuary AI
SP007 Unitree Robotics Unitree H1 Humanoid Robot
SP008 Physical Intelligence Physical Intelligence — pi Company
SP009 TechCrunch Figure Raises $675M at $2.6B Valuation
SP010 TechCrunch Agility Robotics Digit Amazon Deployment 2025
SP011 IEEE Spectrum Boston Dynamics Atlas Humanoid 2024
SP012 CNBC Tesla Optimus Robot Update Progress 2025
SP013 The Robot Report Figure AI BMW Partnership Humanoid Robots
SP014 TechCrunch Apptronik General Catalyst Lead $53M
SP015 Wired Sanctuary AI Phoenix Robot 2024
SP016 New Atlas Unitree H1 Humanoid Comparison 2024
SP017 The Robot Report Humanoid Robot Landscape 2025 Market Map
SP018 VentureBeat Humanoid Robot Competition 2025 Who Will Win
SP019 IEEE Spectrum Humanoid Robot Comparison 2025
SP020 TechCrunch Physical Intelligence Pi Zero Robot Update
SP021 Reuters Tesla Optimus Humanoid Robot Manufacturing Deployment 2025
SP022 Bloomberg Figure AI Humanoid Robot Progress 2025
SP023 The Wall Street Journal Humanoid Robots Competition 2025 Who Leads Market
SP024 Axios Humanoid Robot Race 2025 Competitors
SP025 Business Insider Humanoid Robot Companies Ranked 2025
SP026 TechCrunch Boston Dynamics Atlas Commercial Plans 2025
SP027 The Robot Report Sanctuary AI Phoenix Deployment 2025
SP028 Agility Robotics Digit Amazon 2025 Deployment
SP029 CNBC Apptronik Apollo Humanoid Robot Enterprise 2025
SP030 Fortune Figure AI Valuation Funding 2025
SI001 1X Technologies 1X Raises $100M Series B Led by EQT
SI002 Proff.no (Norwegian Business Registry) 1X Technologies AS — Company Registry Filing
SI003 TechCrunch OpenAI-Backed 1X Technologies Raises $100M Series B for Humanoid Robots
SI004 BusinessWire 1X Technologies Secures $100 Million Investment — January 2025
SI005 The Information 1X Technologies Faces Challenges Closing $1 Billion Series C
SI006 1X Technologies 1X Technologies — About and Team
SI007 SiliconAngle 1X Technologies Raises $100M Series B to Scale Humanoid Robot EVE
SI008 CB Insights 1X Technologies — Company Profile and Funding History
SI009 Reuters Humanoid Robot Startups Burning Cash Fast as Race Intensifies in 2025
SI010 Bloomberg 1X Technologies Targets Billion-Dollar Valuation in New Funding Round 2026
SI011 1X Technologies 1X Enterprise — Commercial Deployment and Pricing
SI012 Axios 1X Technologies Deploys 140 EVE Robots at ADT Facilities
SI013 TechCrunch 1X Technologies Lands ADT Contract for 140-Unit EVE Deployment
SI014 Forbes Inside 1X Technologies' Commercial Humanoid Robot Progress
SI015 1X Technologies EVE Technical Specifications — Commercial Robot Platform
SI016 Wall Street Journal Humanoid Robot Startups Confront Hard Economics of Unit Production
SI017 Morgan Stanley Research Robotics and Automation: 2025–2026 Hardware Economics Outlook
SI018 Financial Times 1X Technologies Funding and Commercial Ambitions Analysed
SI019 SEC EDGAR OpenAI LP — Investment Disclosures and Form D Filings 2023
SI020 EQT EQT Invests in 1X Technologies as Lead Investor in Series B
SI021 Reuters Robot Hardware Economics: Why Humanoid Costs Are Still Sky High
SI022 Goldman Sachs Research Humanoid Robotics: Cost Structure, Margin Path, and Capital Intensity
SI023 CNBC 1X Technologies Pursues Billion-Dollar Valuation with Commercial Humanoids 2025
SI024 TechCrunch Humanoid Robot Funding in 2025: Comparing the Major Raises and Valuations
SI025 Bloomberg Agility Robotics Closes $150M Round as Amazon Deal Expands
SI026 The Information Humanoid Robots' Hidden Costs: Why Unit Economics Are Still Deeply Negative
SI027 Boston Consulting Group Humanoid Robotics: Value Chain Economics and Manufacturing Challenges
SI028 Axios Enterprise RaaS Contracts: How Humanoid Companies Are Structuring Deals
SI029 Business Insider Humanoid Robot Startups and the Unit Economics Problem They Won't Talk About
SI030 1X Technologies NEO — Commercial Availability and Pricing Information
SE001 1X Technologies NEO Gamma Technical Overview
SE002 1X Technologies EVE Robot Technical Specifications
SE003 arXiv Whole-Body Control for Humanoid Robots via Reinforcement Learning from Motion Capture
SE004 arXiv Dexterous Humanoid Manipulation via Imitation Learning and Sim-to-Real
SE005 IEEE Spectrum 1X Technologies NEO Gamma Hands-On Technical Assessment 2025
SE006 MIT Technology Review Inside 1X's Embodied AI Robot Training Engine
SE007 Google Patents / USPTO Patent US11498217B2 — Robot Actuator System (Halodi Robotics)
SE008 The Robot Report 1X Technologies NEO Gamma Hands-On Review 2025
SE009 New Atlas 1X NEO Gamma Technical Deep Dive 2025
SE010 IEEE Spectrum Robot Actuator Technology 2024 Comparison
SE011 arXiv Learning Dexterous In-Hand Manipulation via End-to-End Reinforcement Learning
SE012 GitHub 1X Technologies — GitHub Organisation Repository
SE013 MIT Technology Review Embodied AI and Robots in 2024 — State of the Field
SE014 VentureBeat 1X Technologies Embodied Learning Data Engine Explained 2025
SE015 TechCrunch 1X Technologies NEO Autonomous Capabilities Update 2025
SE016 Engadget 1X NEO Robot Capabilities Update 2025
SE017 IEEE Spectrum 1X NEO's 3D-Knitted Robot Suit Explained 2025
SE018 CNBC 1X Technologies Product Roadmap 2025
SE019 Fortune 1X Technologies NEO Robot Product Update 2025
SE020 The Robot Report Humanoid Robot Actuator Comparison 2025
SE021 arXiv Whole-Body Humanoid Robot Locomotion with Human Reference Motion
SE022 The Robot Report 1X Technologies VR Teleoperation Training System 2025
SE023 Wired 1X Technologies NEO Home Robot Design and Safety 2025
SE024 Robotics 247 1X Technologies NEO Robot Update 2025
SE025 TechCrunch 1X Technologies NEO Pre-Order Update 2025
SE026 1X Technologies 1X Technologies Acquires Kind Humanoid
SE027 IEEE Spectrum Robot Foundation Models and Multi-Task Generalization 2025
SE028 Engadget 1X Technologies NEO Safety Systems Explained 2025
SE029 MIT Technology Review Humanoid Robot Software Stack Landscape 2025
SE030 New Atlas Robot AI Autonomy Progress 2025
SU001 ADT Inc. ADT Robot Security Deployment Announcement
SU002 Everon Everon Robotic Security Innovation
SU003 1X Technologies EVE Enterprise Deployments 2025
SU004 TechCrunch 1X Technologies ADT Robot Deployment Update 2025
SU005 Security Magazine Robot Security Guards Enterprise Market 2025
SU006 Axios 1X Technologies Enterprise Customers 2025
SU007 BusinessWire 1X Technologies and Everon Partnership Announcement
SU008 New Atlas 1X EVE Robot ADT Security Update 2025
SU009 The Robot Report Robot Security Market Enterprise Customers 2025
SU010 Fortune 1X Technologies NEO Pre-Orders Customer Feedback 2025
SU011 Dezeen 1X Technologies NEO Home Robot Early Access 2025
SU012 IEEE Spectrum 1X Technologies Customer Deployments 2025
SU013 VentureBeat 1X Technologies Commercial Traction 2025
SU014 CNBC Enterprise Robot Security ADT 1X Technologies
SU015 Sunnaasen Norway Sunnaasen — About Us
SU016 The Wall Street Journal AI Home Robot Privacy and Consumers 2026
SU017 TechCrunch Humanoid Robot Enterprise Retention Patterns 2025
SU018 The Robot Report 1X Technologies NEO Consumer Reactions 2025
SU019 Reuters 1X Technologies ADT Robot Expansion 2025
SU020 PitchBook 1X Technologies Customer Growth Traction 2025
SU021 Engadget 1X NEO Home Robot Consumer Preview 2025
SU022 Bloomberg 1X Technologies Commercial Progress 2025
SU023 Wired 1X Technologies Robot Privacy Concerns Consumer 2025
SU024 Axios 1X Technologies NEO Home Robot Reactions 2025
SU025 TechCrunch Robot Security Market Enterprise Growth 2025
SU026 Security Info Watch Robot Security Guard Enterprise Deployment
SU027 Fortune 1X Technologies ADT Security Robot Results 2025
SU028 The Robot Report EVE Robot Security Case Study 2025
SU029 The Wall Street Journal Robot Customer Service Enterprise Adoption 2025
SU030 New Atlas 1X Technologies Consumer Robot Market Entry 2025
SR001 Federal Trade Commission FTC Report on Commercial Surveillance — Data Collection and Use
SR002 European Data Protection Board GDPR — What Is GDPR? Official Guidance
SR003 US Congress AI ROBOT Act — Senate Bill 2718 (118th Congress)
SR004 National Institute of Standards and Technology NIST Artificial Intelligence Risk Management Framework
SR005 US Department of Labor — OSHA OSHA Robotics Safety Resources
SR006 European Commission EU Artificial Intelligence Act — Regulatory Framework
SR007 TechCrunch Humanoid Robot Privacy Regulation 2025
SR008 Reuters Robot Safety Incidents and Recall Landscape 2025
SR009 IEEE Spectrum Robot Safety Standards Landscape 2025
SR010 The Robot Report Humanoid Robot Liability and Insurance Landscape 2025
SR011 Axios Humanoid Robot Labor Displacement Risk 2025
SR012 The New York Times Robot Privacy Concerns at Home — Consumer Technology
SR013 The Verge Robot Safety and Regulation 2025 Overview
SR014 VentureBeat Robot Startup Key Person Risk 2025
SR015 Financial Times Robot Supply Chain Risk 2025
SR016 The Wall Street Journal Robot Workforce Union and Labor Opposition 2025
SR017 TechCrunch Humanoid Robot Data Breach Risk 2025
SR018 Axios 1X Technologies Privacy Consent Controls Update 2025
SR019 IEEE Spectrum Robot Cybersecurity Risks 2025
SR020 Fortune Robot Startup Execution Risks 2025
SR021 The Robot Report Humanoid Robot Hardware Reliability Challenges 2025
SR022 Reuters AI Robot Regulatory Landscape 2025
SR023 New Atlas Robot Startup Risks 2025
SR024 Bloomberg Robot Security Incidents Enterprise 2025
SR025 PitchBook Humanoid Robot Competitive Risk Landscape 2025
SR026 CNBC AI Regulation Robot 2025 What to Know
SR027 Wired Humanoid Robot Risks and Challenges 2025
SR028 Cornell Law School — Legal Information Institute 29 USC 652 — OSHA Definitions and Statutory Framework
SR029 The New York Times AI Robot Privacy Home Data Concerns 2025
SR030 Financial Times Robot Startup Talent Risk Engineering 2025
SV001 Goldman Sachs Humanoid Robot Investment Thesis 2025
SV002 ARK Invest Physical AI Valuation Framework 2025
SV003 PitchBook Humanoid Robot Startup Valuation Multiples 2025
SV004 The Robot Report 1X Technologies Investment Analysis 2025
SV005 BusinessWire 1X Technologies Series B $100M Funding Announcement
SV006 VentureBeat Humanoid Robot Valuation 2025: Bull and Bear Case
SV007 Fortune Humanoid Robot IPO Valuation 2025
SV008 Reuters Humanoid Robot Startup Valuations 2025
SV009 TechCrunch 1X Technologies Series C $1B Round Outlook 2025
SV010 Bloomberg Humanoid Robot Investments and Valuations 2025
SV011 The Wall Street Journal Humanoid Robot Startup Valuations 2025: Bubble Risk
SV012 IEEE Spectrum Humanoid Robot Investment Analysis 2025
SV013 Axios 1X Technologies Series C Outlook 2025
SV014 Mordor Intelligence Service Robotics Market — Global Industry Analysis and Forecast
SV015 MarketsandMarkets Service Robotics Market — Size, Share, Forecast 2025–2030
SV016 International Federation of Robotics (IFR) Service Robots Market Report 2025
SV017 TechCrunch Tesla Optimus Implied Valuation Analysis 2025
SV018 CNBC Figure AI Valuation 2025: $2B Humanoid Startup
SV019 PitchBook Agility Robotics Valuation 2025
SV020 The Information Humanoid Robot Startup Comparable Valuations 2025
SV021 Harvard Business Review Investing in Physical AI and Robots 2025
SV022 Financial Times Humanoid Robot Investment Risk 2025
SV023 New Atlas 1X Technologies Investment Outlook 2025
SV024 Wired Humanoid Robot VC Bubble 2025
SV025 Fortune OpenAI and EQT Ventures Portfolio Robot Returns 2025
SV026 Reuters 1X Technologies $1B Series C Investment Outlook 2025
SV027 Bloomberg Robot Startup Comparable Analysis 2025
SV028 VentureBeat 1X Technologies Investment Thesis Bull Bear 2025
SV029 TechCrunch Humanoid Robot Diligence Questions 2025
SV030 Axios Humanoid Robot Investment Framework 2025