初创公司尽调
尽调报告 Robotics / Hardware (Embodied AI) Series B 2026-07-14

Stardust Intelligence

中国绳驱动人形机器人公司,面向零售、工业和教育场景

Stardust Intelligence 的绳索 / 线缆驱动人形机器人已跑出真实早期商业牵引力,但要撑起激进估值,还得证明制造规模、单位经济,以及这套非常规执行器路线的耐久性。

封面要素

估值 01
>RMB10B (~$1.38B) [CO008]
总融资额(Series B 轮) 02
>RMB1B [CO009]
T1 机器人价格 03
RMB 89,900 [CO015]
成立时间 04
December 2022 [CO001]

公司概况

Stardust Intelligence(Astribot,星尘智能)是一家总部位于深圳的人形机器人创业公司,2022 年 12 月由前 Tencent Robotics X 工程师 Lai Jie 创立。 公司以绳/缆驱动(肌腱驱动)执行器技术见长,并称该技术可让人形机器人更轻、更安全。2026 年中,公司通过连续三轮 Series B 融资跻身独角兽,估值超过 RMB10B(约 $1.38B),融资额超过 RMB1B;China Renaissance 担任顾问,投资方包括 Thundersoft、Kede Education、Liangxi fund 等。 继 S1 研发平台之后,公司于 2026 年 5 月推出 T1 人形机器人,起价 RMB89,900;据报道,工业、零售和教育客户已给出千台级订单,部署覆盖约六个中国城市。

官网
www.astribot.com
成立时间
2022-12-01
创始人
Lai Jie (赖杰), Dai Yuan (戴媛)
创立地点
Shenzhen, China
总部
Shenzhen (Nanshan district), China
产品
S1 双臂研发平台(2024)与 T1 全身人形机器人(RMB89,900,2026 年 5 月发布),建立在 Design-for-AI 技术栈之上,结合 Lumo 式 AI 模型、DuoCore 式 OS 和绳/缆驱动机体;已部署于工业、零售和教育场景。
客户
面向工业/物流、零售和教育领域的 B2B 企业客户;部署覆盖约六个中国城市。
商业模式
向企业和机构买家销售硬件,并借助合作伙伴/分销渠道;目标是在无锡工厂实现规模化制造。
阶段
Series B
融资情况
2026 年中连续三轮 Series B(B/B+/B++)融资超过 RMB1B;估值超过 RMB10B(约 $1.38B),跻身独角兽;累计融资约 $150M。
[CO001, CO004, CO008, CO009, CO015, CO032]

执行摘要

主要优势

  • 凭差异化绳索 / 线缆驱动执行器,率先推出具备商业定价的人形机器人(T1,RMB89,900)
  • 千台级采购订单(尤其 SEER)显示工业、零售和教育场景已有真实早期需求
  • Thundersoft 战略背书叠加 China Renaissance 顾问支持,补强制造、OS 和交易架构可信度
  • 已在约六个中国城市部署,在政策支持的国内市场获得早期真实场景验证

主要风险

  • 收入、利润率和单位经济未披露,约 $1.38B 估值更多押注可选性,而非基本面
  • 绳索 / 线缆驱动执行器非常规;长期可靠性、线缆磨损和规模化维护尚未验证
  • 中国人形机器人竞争极其激烈,Unitree-AgiBot 双寡头主导,全球资金充足的同行也在追赶
  • 监管和地缘政治暴露上升,包括中国 HEIS 2026 标准以及美国限制中国机器人立法动向

未决问题

  • 审计收入、利润率和单位经济未披露
  • 执行器可靠性(MTBF)和规模化制造良率缺少独立验证
  • 已交付数量与公告订单数量、具名终端客户均未披露
  • 除创始人外,技术团队履历和管理梯队深度记录较少

目录

Chapter 01

01公司概览

1.1 身份、产品与商业模式

Stardust Intelligence 以 Astribot 品牌运营,中文名为星尘智能(Xingchen Zhineng),是一家深圳人形机器人公司,由 Tencent Robotics X 首位工程师、前 Baidu Xiaodu 机器人团队负责人 Lai Jie 于 2022 年 12 月创立。法律实体为 Astribot (Shenzhen) Co., Ltd.。 公司的核心技术押注是绳/缆驱动(肌腱)执行机构;公司认为,相比传统电机驱动关节,该方案能做出更轻、更安全、更柔顺的人形机器人。Stardust 采用全栈 Design for AI(DFAI)架构,把自有 AI 模型、具身操作系统和绳驱动机体接在一起。商业上,公司向工业、零售、研究和教育领域的企业买家销售硬件,定位于「Physical AI」商业化,而不是纯研发演示。主打产品是 S1 双臂机器人,以及 2026 年 5 月发布的 T1 人形机器人。[CO001, CO002, CO003, CO004, CO006, CO017]

FO002: 公司快照逻辑

技术、产品、客户和资本如何咬合,以及最主要的依赖项。

[CO017, CO015, CO018, CO019, CO008, CO031]

1.2 创始人、领导层与治理

公司的公开形象几乎绑定在创始人、董事长兼 CEO Lai Jie 身上。他履历覆盖 Tencent Robotics X(首位工程师)和 Baidu Xiaodu 机器人项目,据报道拥有约 16 年机器人经验。创始工程团队很大程度上来自同一批 Tencent Robotics X 人才,Dai Yuan 被列为联合创始人。 这种集中度强化了技术可信度,但也带来明确的关键人物风险:几乎所有媒体报道、产品叙事和投资人沟通都经由 Lai Jie;公开资料没有继任者、具名 COO 或完整高管团队。治理披露很薄——即便 Series B 引入多家战略和国资背景投资方,公司也没有披露董事会名单、控制权安排或持股比例。尽调应优先确认股权结构、董事会构成,以及投资方是否拥有控制权或否决权;同时摸清创始人之下技术梯队的厚度。[CO004, CO005, CO030, CO031, CO032, CO035]

领导层与创始人表
人物角色背景创始人 - 市场匹配 / 职能覆盖关键人物依赖
Lai Jie创始人、董事长兼 CEOTencent Robotics X 首位工程师;曾带领 Baidu Xiaodu 机器人团队;约 16 年机器人经验机器人与具身 AI 深度匹配;掌握技术和公开叙事关键——主要公开面孔和技术负责人;未见具名继任者
Dai Yuan联合创始人据报道为创始团队联合创始人参与创始领导层高——独立公开履历有限
核心工程团队创始工程师主要为前 Tencent Robotics X 机器人专家硬件 + AI 系统深度中等——集中在前 Tencent 队伍

截至运行日期,仅列举公开具名领导者;完整高管名单和董事会未披露。

[CO004, CO005, CO030, CO031, CO032]

1.3 融资历史、估值与资本结构

Stardust Intelligence 在资本栈上推进很快。中文数据库显示,2023 年 Yunqi Partners 领投天使轮,2024 年 Matrix Partners China 领投 Pre-A 轮,随后 Jinqiu Capital 和 Ant Group 参与 A 轮;聚合平台估算累计融资约 $150M。最醒目的事件是 2026 年中的 Series B:中文财经媒体称,公司在约三个月内连续完成三轮融资,总额超过 RMB1B,China Renaissance 担任独家财务顾问。 已披露的 Series B 投资方包括 Thundersoft、Kede Education、Liangxi fund、Yangzhou Longtou 核心芯片基金和 Zhongbo Juli。该轮把公司估值推过 RMB10B(约 $1.38B),正式进入独角兽行列,也强化了深圳具身 AI 集群。Thundersoft 的参与尤其具战略性,不只是资本,还带来操作系统、制造和渠道协同,目标指向千台级规模。逐轮融资金额、持股比例和优先权条款仍未披露。[CO007, CO008, CO009, CO010, CO011, CO012]

利益相关方 / 投资人图谱
利益相关方角色控制 / 经济重要性尽调问题
ThundersoftSeries B 战略投资人 / 合作伙伴战略——OS、制造、渠道;千台级合作确认股权比例和董事 / 观察员权利
Kede EducationSeries B 战略投资人战略——职业教育渠道澄清商业承诺与股权投资的关系
Liangxi fundSeries B 国资基金国资;无锡 / 梁溪产业纽带确认投资金额和条件
Yangzhou Longtou 核心芯片基金Series B 国资基金区域产业资本确认分配金额和锁定期
Zhongbo JuliSeries B 投资人财务参与确认投资金额
Yunqi PartnersAngel 轮领投(2023)早期经济权益确认持股和按比例跟投权
Matrix Partners ChinaPre-A 轮领投(2024)早期股权确认持股和权利
Jinqiu Capital / Ant GroupA 轮投资人(2024)战略 + 财务持股确认与 Ant 的战略条款

投资人名单来自中文财经媒体披露;投资金额、持股比例和控制权未公开。

[CO011, CO012, CO013]

1.4 快照指标与证据缺口

可支撑的封面指标包括:估值超过 RMB10B,Series B 融资超过 RMB1B,旗舰 T1 定价 RMB89,900,成立于 2022 年 12 月,以及报道中的千台级订单。真正决定投资判断的数字却都缺席:收入、ARR、毛利率、单位经济、现金消耗、资金跑道、员工数和经核验的活跃客户数均未披露,这符合中国私营硬件公司的常见状态。 牵引用订单公告表达,而不是用确认收入表达,因此已签订单与已交付、已付款设备之间的差距很关键。公司估值把它放进 2025-2026 年形成的一批中国人形机器人独角兽之中,但同一时期,分析师和中国官方也明确提示泡沫与估值过高风险,这让快照必须带着谨慎解读。尽调需要把订单公告转成交付和收入时间表,并拿到财务与员工数的数据室视图。[CO008, CO009, CO015, CO018, CO023, CO025]

KPI 快照表
指标数值 / 状态日期置信度证据缺口
估值> RMB10B(约 $1.38B)2026-06第三方报道;无原始申报文件
Series B 融资额> RMB1B(3 轮)2026-06未逐轮列示拆分
累计融资(估算)~$150M2026聚合平台估算;未验证
阶段Series B / 独角兽2026-06无重大缺口
T1 价格RMB89,9002026-05配置 / 选项未详述
成立时间2022 年 12 月2022-12公司与登记信息确认
总部中国深圳2026-07公司官网确认
创始人 / CEOLai Jie2026团队名单未完全公开
收入 / ARR未披露2026-07n/a私营公司;无财务披露
员工数未披露2026-07n/a无公开员工数
客户数未披露(千台级订单)2026具名终端客户有限
主要产品S1、T1(绳驱)2026-05无重大缺口

截至 2026-07-14 运行日期,这些数值是公开资料中最有支撑的口径;这家私营公司的收入、ARR、毛利率和员工数均不可得。

[CO008, CO009, CO014, CO007, CO015, CO001]
FO003: 快照 KPI

紧凑的可投资性快照。

[CO008, CO009, CO015, CO001, CO018, CO023]

1.5 里程碑、产品与商业牵引

公司在 2022 年成立后,发展节奏迅速加快。绳驱动 S1 双臂机器人于 2024 年 8 月在 World Robot Conference 亮相,凭借手臂速度和灵巧性出圈,并在国际上建立品牌认知。商业证据随后出现:2025 年 9 月,SEER(Xiangong Intelligence)签下据称为中国首个千台级工业人形机器人订单,计划两年内交付;2025 年 11 月,公司又在 CHTF 与 Baidu、Jishu Diedai 签署具身 AI 订单和开发者平台合作。 T1 人形机器人于 2026 年 5 月发布,定价 RMB89,900,价格激进,意在打开商业、研究和教育市场;目标场景覆盖家庭服务、商业门店、科研/教育、文旅和轻工业五大类。上述里程碑支撑了 2026 年中的 Series B 和独角兽估值,但公司也处在极其拥挤的中国人形机器人市场,按出货量看,Unitree 和 AgiBot 领先。[CO016, CO019, CO022, CO015, CO028, CO018]

里程碑表
日期事件类型金额 / 估值 / 状态参与方影响
2022-12公司在深圳成立创立-Lai Jie + 前 Tencent Robotics X 团队确立绳驱人形机器人创业公司
2023Angel 轮融资未披露Yunqi Partners(领投)首笔机构资本
2024-08S1 在 World Robot Conference 首秀产品演示 / 出圈Astribot机械臂速度和灵巧度获得全球关注
2024-2025Pre-A 与 A 轮融资未披露Matrix China、Jinqiu Capital、Ant Group 参投推动研发走向产品化
2025-09SEER 千台级工业订单规模化约 1,000 台,2 年内SEER(Xiangong Intelligence)中国首个千台级工业人形机器人订单
2025-11CHTF 具身 AI 订单 + 开发平台合作订单 + 战略 MOUBaidu;Jishu Diedai开发者平台生态布局
2026-05T1 人形机器人发布产品RMB89,900Astribot用激进价格打开 B2B 市场
2026-06完成 Series B融资> RMB1B;估值 > RMB10BThundersoft;Kede Education;Liangxi;其他;China Renaissance 顾问进入独角兽状态
2026-06入选深圳 RMB10B 级具身智能独角兽治理 / 生态估值里程碑深圳具身 AI 集群强化区域龙头定位

单一记录口径;日期和金额采用公开披露中支撑最强的版本。未披露 = 金额未公开。

[CO001, CO013, CO016, CO019, CO022, CO015]
FO001: 公司里程碑时间线

从创立到 B 轮的时间线,并标出一个行业层面的负面信号。

[CO001, CO013, CO016, CO019, CO022, CO015]

1.6 图表

Chapter 02

02市场分析

2.1 市场边界与替代品

Stardust 竞争的是中国企业服务和轻工业人形机器人市场——面向组织销售、执行真实任务的双足或双臂系统——这是更广义具身 AI 版图中的一个子集。市场内支出包括通用和轻工业人形机器人、教育和研究设备,以及零售/文旅服务机器人。 边界之外包括四足机器人(例如 Unitree 的 Go 系列)、固定式工厂自动化和 PLC 驱动产线单元,以及独立协作机械臂;这些更应视为替代品或相邻市场,而不是直接市场。最重要的现状替代品仍然是人力,其次是在结构化任务中使用固定自动化和协作机器人。这个边界很重要,因为标题式的「人形机器人市场」数据常把这些类别混在一起;把 Stardust 的可服务切片定义为企业人形机器人,才能看清真正相关的需求和竞争者,也避免夸大可触达机会。实际边界并不硬:买家评估人形机器人时,也会为同一任务比较协作机器人或新增班次人力,因此竞争预算远超具名人形机器人厂商。[CM011, CM012, CM013, CM022, CM036]

市场定义表
细分 / 类别纳入支出排除支出买方 / 付款方相关性
服务 / 通用人形机器人用于任务的双臂和完整人形机器人固定自动化、PLC 产线企业、机构核心——Stardust S1/T1 目标市场
轻工业人形机器人物料搬运、分拣、装载重型固定机器人单元工厂、物流核心——SEER 类订单
教育 / 研究机器人培训平台、实验室设备K-12 玩具机器人高校、职业学校高——Kede Education 渠道
零售 / 旅游服务机器人导览、零售推车、接待服务自助终端、静态展示零售连锁、场馆相邻——需求初现
四足 / 移动机器人-四足机器人(Unitree Go)-排除——形态不同
协作机械臂(cobots)-独立 cobots-替代品,不属人形机器人

边界划在面向企业销售的双足 / 双臂人形机器人;四足、固定自动化和协作机器人视为替代品 / 相邻品类,不计入市场。

[CM011, CM012, CM013, CM022]
FM003: 买方 / 细分市场地图

目标细分市场里的买方、付款方与触发因素关系。

[CM013, CM016, CM019, CM022, CM026]

2.2 市场规模与增长

衡量中国人形机器人市场需要多重视角,因为公开估算差异极大。EmbodiedGlobal 给出的 2026 年中国规模接近 RMB104.7B,其他口径则高达 RMB1,047B;分析师还描述,未来十年 CAGR 超过 65%。全球看,Axis Intelligence 估算 2026 年行业收入约 $6.24B,Goldman Sachs 预计 2035 年全球 TAM 约 $38B,更长期的银行预测则设想本世纪中叶达到万亿美元级市场。 按单位计,Morgan Stanley 预测 2026 年中国人形机器人销量约 50,000 台。中国的结构性主导地位很清晰:2025 年约占全球产量 90%,2026 年约占出货量 85%,预计产量同比增长约 94%。放在这些数字面前,Stardust 的可服务可获取市场近期只是以数千台计的早期小切片。估算跨度很宽——其中保留了一个真实的 RMB1,047B 离群值——本身就是需要纳入尽调的重要发现。[CM001, CM002, CM003, CM004, CM005, CM006]

TAM/SAM/SOM 或测算视角表
视角发布方年份地域数值CAGR置信度 / 局限
全球人形机器人收入Axis Intelligence2026全球~$6.24B>50%估算;收入基数仍早期
全球人形机器人 TAMGoldman Sachs2035全球~$38B-远期预测
中国人形机器人市场EmbodiedGlobal2026中国~RMB104.7B>65%不同来源差异很大
中国人形机器人市场(异常值)EmbodiedGlobal2026中国最高 RMB1,047B-异常值;方法不清
中国人形机器人销量Morgan Stanley2026中国约 50,000 台-投行估算;近期
Stardust SOM(近期)报告估算2026中国数千台-基于订单;低置信度

列示多个视角,是因为单点估算分歧很大;各行单位不同(收入美元、市场人民币、台数)——不要跨行加总。

[CM003, CM006, CM007, CM008, CM025, CM024]
FM001: 市场规模测算视角

从全球 TAM 到 Stardust 可服务可获得市场的分层测算。

[CM007, CM003, CM008, CM025, CM035]
FM002: 市场估计区间

关键市场变量的低 / 基准 / 高边界;注意中国 2026 年 RMB1,047B 这个离群值。

区间综合了分歧很大的第三方估计;CAGR 行把市场 CAGR 与 2026 年产值增速混用, 作为上界。

[CM003, CM004, CM006, CM005, CM002]

2.3 买家、细分市场与采用路径

需求来自多元买家。工业和物流运营方用企业资本开支采购,用于自动化重复或危险的物料搬运;零售、文旅和酒店场所采购面向客户的服务与差异化体验;高校和实验室用科研经费采购研究平台;职业学校和高校采购训练设备,资金常来自政府或机构预算,并经 Kede Education 等渠道。 各细分市场的采用路径通常从演示和展示兴趣开始,进入付费试点,再到有限生产,最后才是多站点规模化铺开和扩张——每一步都要用已验证 ROI 过关。许多已部署人形机器人仍只达到人类效率的 20-30%,因此从试点转向规模化生产才是关键瓶颈。CHTF、Hong Kong InnoEx 2026 等公开展示显示兴趣上升,政府部署项目也会缩短政策支持垂直领域的采购周期,但真正的生产经济性仍是闸门。[CM013, CM016, CM019, CM021, CM022, CM026]

细分 / 买方图谱
细分买方用户付款方工作流 / 采用触发
工业 / 物流工厂 / 运营负责人产线操作员企业资本开支自动化重复 / 危险物料搬运
零售 / 商业连锁运营负责人门店员工市场 / 运营预算面向客户的服务、新奇感、客流
旅游 / 酒店场馆经理客人 / 员工场馆资本开支体验差异化
研究PI / 实验室负责人研究人员课题经费 / 机构操作与具身 AI 研究
教育 / 职业培训院长 / 采购学生 / 教师政府 + 机构AI / 机器人训练课程(Kede 渠道)

各细分的预算归属不同;教育和工业买方常使用国资或机构资本,政策支持项目的采购周期会缩短。

[CM013, CM016, CM019, CM022]
FM004: 采用漏斗或价值链图

从演示到扩容的企业采用漏斗示意,ROI 验证是门槛。

漏斗比例用于示意典型企业机器人转化,不是公司特定数据。

[CM021, CM018, CM020]

2.4 增长驱动、约束与有争议的估算

市场顺风很强:低于 RMB90,000 的人形机器人代表成本快速下行,HEIS 2026 成为全球首个国家标准,MIIT/SASAC 部署项目创造国资背景需求。大量资本正在涌入——2025 年前 11 个月,具身 AI 融资超过 RMB33.5B,约为上一年的四倍;2026 年至今,机器人融资轮次超过 500 起,总额约 RMB96B,科技巨头和车企也在入场。 但约束同样真实且已有明确记录:ROI 尚未证明,可靠性、电池和安全限制仍在,半导体进口限制会封顶 AI 进展,Morgan Stanley 等银行和中国官方反复提示泡沫与产能过剩风险。国家标准会降低碎片化,但全行业采用需要时间。尽调中,市场估算分歧以及订单公告与已验证生产 ROI 之间的落差,是两个核心未解问题;最悲观的分析师观点应被加权,而不是被忽略。[CM009, CM010, CM014, CM015, CM017, CM018]

增长驱动与约束表
驱动 / 约束方向时点影响尽调问题
成本大幅下降(低于 RMB90k 的机型)驱动当前降低采用门槛,扩大买方范围核实该价格下的单机经济性
国家标准(HEIS 2026)驱动因素2026+安全与互操作性支撑规模化跟踪认证要求
工信部 / 国资委部署项目驱动因素2026国家支持的需求与试点确认项目准入资格
科技巨头 / 车企入局利弊并存2026+需求变多,竞争也更激烈评估竞争导致的份额流失
ROI 尚未证实(人效提升 20-30%)约束当前拖慢生产级铺开获取部署 ROI 数据
可靠性 / 安全 / 电池限制约束当前许多场景只能停留在试点现场可靠性证据
泡沫 / 产能过剩风险约束2026融资与估值风险压测需求假设
半导体进口限制约束持续存在限制 AI / 算力进展评估芯片供应敞口

表中既有需求拉力,也有采用阻力;ROI 和可靠性约束,是近期生产级落地最大的两道刹车。

[CM017, CM015, CM014, CM023, CM018, CM020]

2.5 图表

Chapter 03

03竞争对手

3.1 竞争格局与进入者

Stardust 所处的是 2026 年最拥挤的科技市场之一:中国已有 50 多家人形机器人开发商,还有美国主要玩家。中国市场实际上接近双寡头,Unitree 和 AgiBot 合计约占国内 80% 份额。Unitree 据称以近 $7B 估值冲刺 IPO,拥有品牌强度、病毒式演示和约 $16,000 的低成本 G1 人形机器人。AgiBot(Zhiyuan)凭借重度国资和投资人支持推进量产。 资金充足的同行还包括估值约 $3B 的 Galbot、上市公司 UBTech 及其 Walker 系列、以康复起家的 Fourier 及其 GR 系列,以及 EngineAI 等新兴创业公司。海外方面,Figure AI(约 $39B)、1X(Neo 约 $20,000)、Apptronik(约 $5.5B)和 Tesla Optimus 都掌握巨额资本。潜在新进入者——Alibaba、ByteDance 等科技巨头,以及 BYD、XPeng 等车企——会进一步加剧竞争。在这个背景下,Stardust 虽然融资速度很快,但仍是体量相对较小的挑战者。[CP001, CP002, CP004, CP005, CP006, CP007]

竞争对手画像表
竞争对手类别规模 / 融资目标客群差异化短板
Unitree中国直接竞品IPO 目标估值约 $7B科研、工业、消费级品牌、低价 G1(~$16k)、爆款演示四足机器人起家;人形灵巧度仍在成熟
AgiBot (Zhiyuan)中国直接竞品资金充足,国资背书工业、服务量产能力、A2 产品线产品线拥挤;毛利承压
Galbot中国竞品估值约 $3B操作 / 服务聚焦抓取 AI商业化牵引仍早期
Fourier中国竞品VC 支持康复、科研、医疗康复背景、GR 系列相比通用型更偏小众
UBTech中国上市竞品上市公司(HKEX)工业、教育Walker 系列、上市融资平台成本高;盈利能力受疑
Figure AI美国竞品估值约 $39B美国工业资本雄厚、绑定 OpenAI无中国布局;ROI 尚未证实
1X美国竞品资金充足西方家庭 / 企业Neo(~$20k)、发力消费级规模有限;聚焦西方市场
Stardust Intelligence本报告主体估值约 $1.38B,已融资 >RMB1B工业、零售、教育绳驱执行器、全栈 DFAI、T1(RMB89,900)规模较小;执行器可靠性未证实

画像综合截至 2026 年中公开估值与产品定位;私营公司估值来自第三方报道,均为近似值。

[CP002, CP004, CP005, CP006, CP007, CP008]
FP001: 竞争定位图

以可负担性(x,越高 = 越可负担)与能力 / 全栈强度(y)定位。

坐标轴为证据支持的序数评分(0-10):x = 价格可及性,y = 能力 / 全栈成熟度; 并非厂商基准测试。

[CP002, CP006, CP008, CP013, CP015, CP019]

3.2 能力与定价对比

能力上,Stardust 最强的牌是操控灵巧性:S1 展示的末端执行器速度最高约 10 m/s,重复定位精度约 0.1 mm,属于公开展示中最快的一档;其 Design-for-AI 全栈把 AI 模型、具身 OS 和绳/肌腱驱动机体组合起来。多数对手依赖传统电机或谐波减速关节,因此执行方案确实构成技术差异。 价格上,RMB89,900 的 T1 与 Unitree 约 $16,000 的 G1 同属可负担区间,明显低于 1X 约 $20,000 的 Neo,也低于 UBTech、Apptronik、Fourier 等仅面向企业且价格未披露的产品。企业买家会权衡灵巧性、自主性、价格、可靠性和部署支持;Stardust 在前三项得分不错,但大规模可靠性尚未验证,而 Unitree 和 AgiBot 在量产成熟度上领先。因此能力矩阵显示,Stardust 在窄维度强,但在生产和品牌上落后于头部厂商。[CP003, CP009, CP012, CP013, CP014, CP015]

功能 / 能力矩阵
采购标准StardustUnitreeAgiBotFigure AIFourier
操作灵巧度 / 速度高(S1 ~10 m/s)
可负担性(单价)高(RMB89,900)高(~$16k G1)
全栈 AI + OS高(DFAI/DuoCore)
量产成熟度中低
品牌 / 分销
可靠性履历未证实初步显现初步显现未证实

序数评级(高 / 中 / 低)是有证据支撑的判断,不是厂商基准测试;“未证实”表示缺少独立现场数据。

[CP003, CP013, CP015, CP016, CP017, CP027]
定价 / 包装比较
产品供应商价格模式包含内容未知项 / 含义
T1StardustRMB89,900 (~$12,500)硬件销售完整人形机器人,23 DoF服务 / 保修条款未披露
G1Unitree~$16,000硬件销售紧凑型人形机器人载荷低于工业机型
Neo1X~$20,000预订 / 硬件家用人形机器人西方市场;中国参考价值有限
Walker SUBTech企业级(未披露)B2B 合同工业人形机器人成本高;按项目定价
ApolloApptronik企业级(未披露)B2B / 租赁工业人形机器人聚焦美国;定价不透明
GR-seriesFourier未披露科研 / B2B科研人形机器人小众定价;非大众市场

价格口径混合了零售与企业级 / 未披露模式;T1 和 G1 定住了可负担价格带,美国和企业级机型定价更高且不透明。

[CP003, CP009, CP015, CP031]
FP002: 功能广度 / 能力地图

五家头部人形机器人厂商的能力覆盖与强弱。

[CP003, CP013, CP016, CP027, CP035]

3.3 护城河、切换成本与耐久性

Stardust 的护城河真实存在,但很窄。执行器技术诀窍和全栈集成今天有意义,但随着时间推移,资本更雄厚的对手可以复制;供应商共享执行器、传感器和 AI 工具包,国家标准也会降低全行业差异,赛道正在商品化。买家的切换成本中等——集成、安全认证和再培训会形成一定锁定,但企业可以多供应商并行,内部自建或继续用人力仍是成本敏感客户的默认选项。 可能持久的优势来自 DuoCore OS 周围的开发者生态,或来自已验证的大规模可靠性,但两者今天都未建立。关键在于,Stardust 在规模、融资和生产成熟度上落后于 Unitree 和 AgiBot,其灵巧性领先仍处演示阶段,尚未被现场验证。竞争结论是:Stardust 凭借可负担价格加全栈设计拥有可防守切口,但护城河能否持久,取决于执行,也取决于能否证明绳驱动执行器在真实、持续的商业部署中可靠,而不只是在展示中亮眼。[CP018, CP019, CP021, CP022, CP023, CP029]

护城河耐久性 / 竞争风险登记表
护城河主张威胁严重性缓释措施 / 尽调问题
绳驱执行器诀窍竞争对手复制,或用更好的电机绕开核实专利覆盖与可靠性数据
全栈 DFAI(模型 + OS + 机身)大公司在 AI 栈投入更猛对比同业 AI 模型基准
可负担性 / 成本领先Unitree / AgiBot 跟进低于 RMB100k 定价确认 BOM 成本与毛利率
S1 灵巧度 / 速度领先仍处演示阶段;规模化可靠性未证实获取现场可靠性和 MTBF 数据
开发者生态锁定生态尚未成形跟踪第三方应用和 SDK 采用
S1 爆款演示带来的品牌Unitree 品牌与份额大得多衡量主动需求与赢单率

行按严重性排序;每条护城河主张都可能被资本更足的对手复制,因此耐久性取决于执行力和可靠性证据。

[CP012, CP014, CP019, CP022, CP023, CP024]
FP003: 护城河 / 就绪度 KPI

紧凑的竞争耐久性快照。

[CP001, CP002, CP015, CP013, CP022]

3.4 分销力量、潜在进入者与替代风险

除了当下的竞争者,分销力量和新进入者也会塑造竞争轨迹。Unitree 的品牌认知和病毒式演示给它带来分销与心智优势;Stardust 虽然也凭借 WRC 2024 的 S1 首秀吸引注意,但尚未匹配这种优势。所有中国玩家都受益于密集的国内供应链,成本因此较低,所以单纯成本优势并非专属资产。 最关键的未来威胁来自资源雄厚的新进入者:Alibaba、ByteDance 等科技巨头,以及 BYD、XPeng 等车企正在进入人形机器人,带来制造规模、资本和既有企业客户关系。因此替代风险很高——随着低于 RMB100,000 的人形机器人扩散、共享工具包让功能趋同,Stardust 这样的小型专精公司可能被上方既有双寡头和一波深口袋进入者夹击。它最好的防守,是在同质化抹平切口之前,把早期灵巧性和成本领先转化为有黏性的开发者生态,或转化为可演示且经独立验证的可靠性优势。[CP020, CP021, CP024, CP025, CP033, CP034]

3.5 图表

Chapter 04

04财务

4.1 收入来源与变现

Stardust 主要靠一次性硬件销售变现:RMB89,900 的 T1 人形机器人、价格更高的 S1 研究系统(估计 $100,000-150,000)、多台套企业和工业合同、经 Kede Education 进入教育渠道,以及由 CHTF MOU 播种的早期开发者平台线索。关键在于,公司没有披露收入,因此每条收入流都只能用状态信号描述,不能用财务数据描述。 最清晰的牵引信号是 2025 年 9 月 SEER 的首个同类千台级工业人形机器人订单,加上 CHTF 的额外订单,以及在约六个中国城市部署。但订单公告不是确认收入:大型多台套交易的确认取决于交付、验收和部署里程碑,存在时间风险;企业销售周期长,也意味着签约订单可能缓慢转成收入。收入组合据推断今天高度偏向硬件,教育和软件收入并不重要。尽调第一优先级,是把头部订单信号转成经核验的发货收入和平均售价。[CI004, CI005, CI009, CI010, CI014, CI015]

收入来源表
收入流机制单位当前价值 / 状态质量尽调问题
T1 硬件销售一次性硬件销售按每台机器人(RMB89,900)2026 年 5 月发布;有订单报道收入未披露获取出货台数和实际 ASP
S1 科研系统向实验室销售硬件按每套系统(~$100-150k)2024 年起销售未披露确认 S1 销量
企业 / 工业订单多台合同按订单SEER 千台订单(2025)仅是订单信号核实订单转化为收入
教育渠道机构销售按单台 / 项目Kede Education 合作尚早期量化教育收入
开发者平台 / 软件平台 / MOUTBDCHTF MOU(2025)尚未产生收入评估软件变现计划

收入流根据公开产品、订单与合作信号推断;没有任何收入流披露了收入,因此所有数值都是状态指标,不是财务数据。

[CI004, CI005, CI009, CI010, CI014, CI015]
定价 / 变现表
产品标价实际成交价模式折扣 / 未知项来源
T1RMB89,900 (~$12,500)未披露一次性硬件销售批量 / 企业折扣未知Netease / 厂商
S1~$100,000-150,000(估算)未披露科研 / B2B 销售仅为估算;非官方标价第三方规格
企业多台订单协商定价未披露合同 / 项目订单定价不透明SEER 订单报道
教育机型未披露未披露机构 / 项目可能与补贴挂钩Kede 渠道报道

公开标价只有 T1;所有实际成交价和企业级定价都未披露,因此变现经济性还无法核实。

[CI002, CI005, CI025, CI027]
FI001: 收入模型桥

客户活动如何转化为收入和毛利;金额未披露。

Stardust 未披露收入或成本数据,因此所有金额节点均为定性。

[CI004, CI014, CI015, CI016, CI034]

4.2 单位经济与利润率风险

Stardust 财务问题的核心,是一台人形机器人能否以 RMB89,900 盈利出售。公司没有披露毛利率、BOM 成本、CAC 或单台服务成本,因此几乎所有单位经济指标都是空白。低于 RMB90,000 的激进价格,是公司有意扩大市场的动作,也受益于中国密集且低成本的供应链;但它直接提出了利润率能否持续的问题,因为人形机器人 BOM 成本仍高,全球供应链约束还可能在扩产时推高零部件价格。 已部署机器人还需要持续服务和支持,交付成本会上升;行业分析师也强调,人形机器人的单位经济在全行业尚未证明,ROI 取决于真实世界有效产出。唯一外部锚点是上市同行 UBTech,其公开财务显示,在高估值之下,人形机器人厂商仍在亏损。合理的基准判断是,Stardust 近期贡献利润率很薄或为负,靠股权融资而非经营现金流支撑——这一假设必须用真实利润率桥来验证。[CI006, CI012, CI016, CI018, CI027, CI031]

单位经济性表
指标值 / null置信度重要性尽调问题
毛利率null(未披露)-决定 RMB89,900 价格下能否盈利索取 BOM 与毛利率桥接
T1 BOM 成本null-给激进定价划出成本底线获取组件成本拆分
CAC / 回本周期null-企业销售效率提供交易级经济性
单台服务成本null-售后支持拖累分享维护成本数据
单台机器人收入null-核心单机收入披露实际 ASP
可比公司(UBTech)毛利率亏损(上市公司)行业盈利能力基准对比同业披露

Stardust 几乎没有披露单位经济指标,因此这些字段基本都是 null;唯一锚点是上市同业 UBTech,显示人形机器人厂商仍在亏损。

[CI006, CI012, CI016, CI018, CI031, CI034]
FI002: 单位经济桥

从 T1 价格到贡献利润率的定性单位经济桥。

输入缺失;考虑到激进定价,节点保持定性,并标出利润率未验证。

[CI012, CI027, CI034, CI036]

4.3 资本充足性、现金消耗与融资依赖

Stardust 最强的财务事实是资本位置。如公司概览融资时间线所述,公司在 2026 年中约三个月内连续完成三轮 Series B(B、B+、B++),融资超过 RMB1B,估值超过 RMB10B(约 $1.38B);在此之前,累计融资估计为 $150M。投资团包括 Thundersoft、Kede Education、Liangxi Fund、Yangzhou core-chip fund 和 Zhongbo Juli,China Renaissance 担任顾问。资金用途指向量产、研发和市场拓展。 但现金余额、月度现金消耗和资金跑道均未披露,也未见债务或项目融资,因此公司看起来由股权资金支持。资本强度是真实的:无锡(梁溪)量产建设意味着可观资本开支和营运资金需求,可能部分由地方补贴抵消。三轮快速融资显示投资人需求强,也说明计划中的资本消耗激进;下一轮融资触发点很可能是无锡产能和收入里程碑。由于全行业泡沫和产能过剩警示可能收紧未来资本可得性,融资耐久性是实质且目前无法量化的风险。[CI001, CI003, CI007, CI008, CI011, CI013]

资本充足性表
项目数值 / 状态备注
累计融资总额~$150M(估算)天使轮至 Series B;聚合方估算
Series B 轮融资3 轮,>RMB1B,2026 年中约 3 个月内完成 Series B / B+ / B++
估值>RMB10B (~$1.38B)2026 年中达到独角兽状态
账面现金未披露未公开
月度现金消耗未披露未公开
现金跑道未披露未公开
拟定资金用途量产、研发、GTM据公司表述
下一轮触发点无锡产能 / 收入里程碑推断
债务 / 项目融资未见明显债务(股权融资支撑)未公开披露债务

参见「公司概览」中的融资时间线;资本基础强,但现金余额、月度消耗和跑道都未披露,融资韧性无法量化。

[CI001, CI003, CI007, CI008, CI021, CI022]
FI003: 财务估计区间

有来源支撑的资本与估值区间;收入和烧钱因不可估算而省略。

收入 / 烧钱区间有意排除 —— 没有公开依据可给出边界。

[CI001, CI003, CI029]
FI004: 资本强度 / 现金流图

股权资本流向:研发、制造资本开支、营运资本和部署。

现金流规模为定性;只有融资额已量化。

[CI007, CI011, CI013, CI028, CI036]

4.4 财务结论与尽调阻断项

财务结论是:Stardust 资本充足,但收入不透明。公开证据确认了强劲股权资金基础、激进 T1 定价和真实订单信号,却无法建立收入、利润率、现金消耗、资金跑道或客户级经济性——这些全部未披露。因此估值与收入的关系无法衡量,这是最重要的尽调阻断项。 融资事实截至 2026 年中,是最新可得信息,但它们描述的是融到多少钱,不是资本效率。投资逻辑几乎完全压在尚未证明的单位经济,以及公司能否在无锡执行制造扩张、同时守住低于 RMB90,000 价格下的利润率。尽调必须取得经审计财务、BOM 与毛利率桥、支撑订单信号的发货和验收数据、单笔交易 CAC/回本期,以及资本开支计划。在此之前,应把财务画像视为高风险、证据稀薄,符合一家按未来潜力定价而非按已验证经济性定价的早期硬件公司。[CI006, CI026, CI030, CI032, CI034, CI035]

公开财务缺口表
缺失的私营公司指标影响尽调路径
收入 / ARR无法判断规模或估值倍数索取经审计财务报表
毛利率无法判断 T1 定价下的盈利能力获取 BOM 和毛利桥
现金消耗 / 跑道无法评估融资风险索取现金流量表
出货量订单信号未验证为收入获取发货与验收数据
客户级经济性CAC/LTV 未知提供交易级 P&L
资本开支计划(无锡)资本强度不清提供制造资本开支计划

缺口很普遍;报告能验证融资额和定价,但若管理层不披露,就无法验证收入、利润率或现金动态。

[CI006, CI032, CI035, CI011, CI031]

4.5 图表

Chapter 05

05产品与技术

5.1 产品定义与组合

Stardust 打造能在客户工作流中执行物理任务的通用人形机器人。产品组合以两款硬件平台和一套软件栈为中心。S1 是双臂轮式系统,手臂 7-DoF,单臂载荷约 5 kg;它于 2024 年 8 月在 World Robot Conference 首秀,并以最高约 10 m/s 的末端执行器速度和约 0.1 mm 的重复定位精度吸引全球关注——以人形机器人级操控器而言,这是异常突出的灵巧性。 T1 于 2026 年 5 月发布,定价 RMB89,900,是一款全人形机器人,高约 1.53-1.55m,重约 66kg,23 个自由度,单臂载荷约 5 kg;定位是家庭、商业、科研/教育、文旅和轻工业场景中的轻量操控,而不是重工业负载。硬件之上是 Design-for-AI 软件栈——Lumo AI 模型负责感知、规划和操控,DuoCore 具身 OS 则被公司定位为第三方应用平台。产品成熟度最高的是已展示的 S1 和核心执行器;最早期的是 T1 的规模化可靠性,以及 AI 层经过基准测试的能力。[CE001, CE002, CE003, CE005, CE006, CE007]

产品模块 / 资产矩阵
模块 / 产品用户状态 / 成熟度差异化尽调缺口
S1 双臂系统研究人员、服务运营2024 年起已展示绳驱,~10 m/s,±0.1mm规模化可靠性数据
T1 人形机器人企业、教育、消费者2026 年 5 月发布23 DoF,售价 RMB89,90023-DoF 机身现场耐久性
Lumo AI 模型自主能力栈公司声称具身智能层基准测试能力未披露
DuoCore 具身 OS开发者、集成商公司声称 / 平台第三方应用 OS生态采用证据
绳驱执行器硬件子系统核心,已展示轻量、柔顺肢体线缆磨损 / MTBF 数据

成熟度从已展示的 S1 和执行器,到新发布的 T1,再到公司声称的 AI/OS;缺口集中在规模化可靠性和未披露的 AI 基准测试。

[CE001, CE002, CE003, CE005, CE008, CE009]
工作流 / 用例表
用户任务当前工作流Stardust 方案可衡量收益限制
轻量操作 / 搬运人工或固定机械臂T1/S1 灵巧操作~5 kg/臂,高速运动不适合重载
具身 AI 研究定制实验室装置S1 研究平台快速、精准的操作测试床成本高(~$100-150k)
教育 / 培训教材、仿真T1 教育机型动手式 AI/机器人学习课程整合工作量
零售 / 文旅服务静态终端、员工T1 商用模式客户互动、新鲜感ROI 未验证
轻工业任务人工 / 协作机器人T1 工业场景自动化重复任务规模化可靠性未验证

用例覆盖研究、教育、服务和轻工业;可衡量收益来自已展示的灵巧性,但多个部署的 ROI 未验证。

[CE006, CE007, CE016, CE022, CE028]
FE002: 客户工作流 / 运营流程

客户从需求界定到扩展部署,如何部署并运营一台 Stardust 人形机器人。

[CE007, CE022, CE009, CE016]

5.2 架构与绳驱动差异化

核心技术选择是绳/缆驱动(肌腱驱动)执行。Stardust 不把沉重电机和齿轮箱放在每个关节,而是把电机移到躯干,再通过缆绳向四肢传力。这样降低了肢体惯量,支撑 S1 展示出的高速和柔顺性;因为更轻、低惯量的肢体在接触时危险性更低,也可能提升人机交互安全。 大多数竞争者——包括 Unitree G1 和 Fourier GR 系列——依赖传统电机或谐波减速关节,因此该路径确实是差异化点,也是 Stardust IP 和技术诀窍的核心,尽管公开专利覆盖宽度并不详细。完整技术栈把 Lumo AI 模型和 DuoCore OS 叠在这一机体之上,公开 GitHub 存在则显示公司有意培育开发者生态。关键但书是可靠性:复杂肌腱走线和缆绳磨损是公认工程挑战,绳驱动执行器在持续商业负载下的长期耐久性和维护负担尚未被独立验证。灵巧性领先在演示中真实存在,但还没有在标准化、独立基准中证明。[CE003, CE004, CE008, CE009, CE010, CE011]

技术 / 运行架构表
层级 / 组件作用依赖风险
AI 模型(Lumo)感知、规划、自主AI 算力 / 芯片基准测试能力未披露
具身 OS(DuoCore)机器人控制 + 应用平台开发者生态采用未验证
绳驱执行器向关节传力线缆 / 执行器供应商线缆磨损、可靠性
传感套件视觉、力觉、本体感知传感器供应商供应与校准
制造(无锡)T1 量产资本开支、供应链爬坡执行
开发者工具(GitHub)SDK、示例社区参与早期生态

架构横跨 AI、OS、执行、传感和制造;最尖锐的风险是执行器可靠性,以及出口限制下对 AI 算力供应的依赖。

[CE008, CE009, CE011, CE017, CE018, CE023]
FE001: 产品架构图

DFAI 全栈从应用场景一直下探到绳驱硬件。

[CE003, CE004, CE009, CE017, CE027]
FE003: 关键依赖图

关键供应商、算力和制造依赖,支撑产品与客户部署。

[CE017, CE018, CE012, CE026]

5.3 部署、路线图与制造

部署 Stardust 人形机器人遵循可重复的运营流程:界定任务或场景,在现场集成并配置机器人,完成安全验证和操作员培训,通过 Lumo/DuoCore 技术栈运行自主作业,持续维护和保养执行器,并在价值被证明后扩展到更多设备或站点。 路线图从 S1 研究首秀(2024)推进到 T1 大众市场发布(2026 年 5 月),再到 2026 年无锡(梁溪)工厂扩产——这一步将检验由设计简化和中国密集零部件供应链支撑的 RMB89,900 低价,能否在规模化下维持。T1 的 23-DoF 全人形形态,天生比 S1 的轮式双臂底座更难可靠制造,执行风险更高。关键依赖覆盖执行器和缆绳供应商、传感器、国内机器人供应链,以及 AI 算力;最后一项受半导体出口限制影响,暴露中国机器人产业风险。因此,制造扩张与供应韧性是近期技术执行问题。[CE012, CE013, CE017, CE018, CE022, CE026]

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
Aug 2024S1 在 WRC 首秀已完成建立灵巧性声誉Pandaily
May 2026T1 大众市场发布(RMB89,900)已完成进入可负担人形机器人档位厂商 / 新闻
2026无锡量产爬坡推进中检验制造执行力Shenzhen News
2026HEIS 标准合规新兴塑造安全要求SESEC
持续开发者平台 / 生态早期若被采用,可能形成锁定GitHub / CHTF MOU

路线图从研究展示,走到大众市场发布,再到量产爬坡;短期考验是无锡制造执行和标准合规。

[CE006, CE012, CE013, CE014, CE023]
FE004: 产品成熟度 / 能力图

各产品模块的成熟度、差异化和可靠性证据。

[CE010, CE020, CE021, CE031, CE035]

5.4 信任、安全、合规与结论

信任与安全证据比能力叙事更薄。中国 HEIS 2026 国家标准为 Stardust 适用的安全和互操作要求定下新框架,深圳 2026 年推出的具身 AI 责任保险也显示,人形机器人部署的合规框架正在成熟。绳驱动设计借助低惯量肢体,合理上有助于交互安全;安全控制还必须覆盖碰撞避免和力限制。 但公司层面的证据稀缺:认证、安全测试结果、执行器 MTBF、数据/隐私政策均未披露,质量与合规披露构成实质缺口。因此产品/技术结论是双面的:Stardust 拥有真实且已展示的灵巧性领先,以及独特、看似可防守的执行方案;但大规模可靠性未证实,AI 基准记录未披露,合规透明度也很薄。技术尽调应优先获取独立可靠性和灵巧性基准、执行器耐久数据、AI 模型评估,以及标准认证证据,再把技术主张纳入投资判断。[CE014, CE015, CE021, CE024, CE029, CE034]

信任 / 质量 / 合规表
控制 / 标准状态范围缺口
HEIS 2026 国家标准新兴(2026)人形机器人安全 / 互操作性认证证据未披露
深圳具身 AI 责任保险已可用(2026)部署责任承保条款不清
避障 / 力限制设计意图人机交互未发布安全测试数据
执行器可靠性(MTBF)未披露绳驱关节无现场耐久性数据
产品认证未披露监管批准认证未公开
数据 / 隐私控制未披露机载传感 / 数据政策未公开

合规依托新兴国家标准和新的保险机制,但公司自身的认证、安全测试和可靠性证据未公开。

[CE014, CE015, CE029, CE030, CE034]

5.5 图表

Chapter 06

06客户

6.1 客户基础与细分

Stardust 的客户集中在三大主要细分市场,外加研究和开发者渠道。工业和物流买家——以 SEER 为代表——用企业资本开支采购,目标是物料搬运和重复任务自动化;这是目前最具体、规模最大的需求。零售和商业买家在多个城市部署 T1,用于面向客户的服务,但具名账户有限,报道更多停留在场景层面。 教育和职业培训买家通过 Kede Education 渠道采购训练设备,依托机构或政府预算,形成可重复但单台规模不大的需求;研究买家则采购 S1 作为操控测试平台。第五条早期渠道是通过 CHTF 开发者平台 MOU 触达的开发者和集成商。各细分市场中,买家、用户和付款方经常不同——例如学校采购办、学生和政府预算——这会拉长销售周期。无锡和深圳的政府与市政项目也通过场景资金塑造销售管线。公司报告已在约六个中国城市部署,说明运营覆盖面存在,即便单站点深度仍不清楚。[CU001, CU006, CU011, CU012, CU013, CU019]

客户分层表
细分购买者 / 用户 / 付款方用例规模收入 / 战略价值缺口
工业 / 物流工厂运营 / 操作员 / 企业资本开支物料搬运、重复任务SEER 千台订单战略价值高终端用户结果未披露
零售 / 商业运营负责人 / 员工 / 营销预算面向客户的服务多城市、场景级中等;品牌 / 新鲜感已披露具名客户有限
教育 / 职业培训院长 / 学生 / 政府+机构AI/机器人培训Kede 渠道机构经常性需求单台收入有限
研究PI / 研究人员 / 经费操作研究(S1)小众标杆价值体量小
开发者 / 合作伙伴通过 MOU 接入的企业 / 集成商 / 合作方平台应用CHTF MOU(Baidu、Jishu Diedai)合作线索生态选项采用未验证

分层显示工业、零售、教育、研究和合作伙伴渠道的购买者 / 用户 / 付款方各不相同;工业战略价值最高,但结果证据薄。

[CU001, CU006, CU011, CU019, CU029, CU033]
FU004: 客户牵引力 KPI

客户牵引力快照:订单信号强,留存和成果证据弱。

[CU001, CU003, CU002, CU009, CU027]

6.2 采用轨迹与具名证据

采用进程已经从可见度推进到订单,再推进到产品发布。S1 在 WRC 2024 首秀后引发全球兴趣;2025 年 9 月,SEER(Xiangong)下达据称为中国首个千台级人形机器人订单,这是一个标志性的工业需求信号;2025 年 11 月 CHTF 上,Stardust 又获得额外订单,并与 Baidu、Jishu Diedai 相关方签署开发者平台 MOU;2026 年 5 月,T1 以 RMB89,900 发布。 因此最强的具名证据是 SEER 订单,Kede Education 渠道和 CHTF 合作作为补充。但这些证据的参考质量为中等:它们是媒体报道的订单和 MOU,而不是经独立验证、带量化结果的生产部署。头部订单数字背后的具名终端客户身份、生产还是试点状态,以及可衡量结果(生产率、成本节约)大多未披露。谨慎尽调应把千台数字视为签约订单信号,等待交付和验收;还要查明报道需求中有多少已转为已部署、能产生收入的设备。[CU002, CU003, CU004, CU005, CU007, CU015]

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
S1 公开首秀WRC 展示Aug 2024Pandaily兴趣,不是订单
SEER 工业订单~1,000 台Sep 2025Zhongzheng/Aibang订单;交付未确认
CHTF 订单 + MOU订单 + 平台 MOUNov 202521jingji订单金额未披露
T1 发布RMB89,900May 2026Netease销量未披露
城市部署~6 个城市2026Leaderobot单点利用率未知

轨迹从 2024 年展示,升到 2025 年订单,再到 2026 年发布和多城市部署;每个指标都缺少收入或已交付台数分母。

[CU002, CU003, CU005, CU007, CU024]
具名客户证据表
客户细分部署 / 用例生产还是试点结果限制
SEER(Xiangong)工业 / 物流~1,000 台人形机器人订单,用于物料搬运订单(交付推进中)中国首个千台级人形机器人订单终端用户结果未披露
Kede Education教育教育单位分销渠道渠道合作伙伴机构需求可重复单台收入有限
Baidu(经 CHTF MOU)开发者平台开发者平台合作 MOUMOU / 合作生态建设意向不具约束力;范围不清
Jishu Diedai开发者平台开发者平台 MOU 参与方MOU / 合作平台生态信号早期阶段

具名证据主要是媒体披露的订单、渠道合作和 CHTF 开发者 MOU;生产状态和落地结果尚未获得独立验证。

[CU003, CU005, CU006, CU016, CU017]
FU001: 客户旅程图

客户旅程从展会发现,到多台订单,再到扩展。

[CU009, CU014, CU015, CU022, CU024]
FU003: 客户证据矩阵

按具名证据梳理证据质量、生产成熟度和成果具体度。

[CU003, CU008, CU017, CU018, CU027]

6.3 留存、扩张与集中度

客户画像中,耐久性证据最弱。公开资料没有留存、复购、流失、利用率或满意度指标;这对一家年轻私营硬件公司并不意外,但仍是实质缺口。初始公告之外,复购和扩张订单尚无记录,因此先落地再扩张的潜力仍是假设:如果试点能转成多站点生产,潜力存在,但转化证据仍待补。 同时,集中度风险真实存在。少数标志性订单,尤其 SEER,主导可见需求,让早期收入波动大,并依赖单个交易完成转化和交付。市场拓展也集中在合作伙伴渠道——SEER 分销关系和 Kede 教育渠道——带来渠道依赖;开发者 MOU 不具约束力,采用情况也未证明。整个客户基础最关键的未证实一步,是从展示兴趣和试点转向持续、重复的生产部署;这正是行业分析师警告人形机器人订单可能跑在已证明 ROI 前面的地方。[CU009, CU010, CU016, CU021, CU022, CU023]

留存 / 重复使用 / 满意度表
指标数值 / null细分置信度尽调待索取
净收入留存率(NRR)null全部-索取队列收入数据
总留存率 / 流失率null全部-获取续约 / 流失日志
复购null(未披露)工业-确认后续订单
单台利用率null全部-提供部署遥测数据
客户满意度 / NPSnull全部-安排参考客户访谈

留存和满意度指标完全未披露;公司还早、且为非上市公司,未展示队列或续约证据,这是重大尽调缺口。

[CU009, CU021, CU022, CU027]
扩张与集中度风险表
扩张驱动因素集中度风险影响尽调路径
试点转量产少数标杆订单(SEER)占主导收入不平滑,依赖单笔交易核验管线多样性
多站点铺开单一分销渠道(SEER)渠道依赖评估直销与渠道占比
教育渠道放量依赖 Kede 合作需求依赖合作伙伴审阅渠道协议
开发者生态MOU 不具约束力平台采纳不确定跟踪 SDK / 应用采纳
政府项目政策驱动需求受项目变化影响确认项目资格

能否扩张,取决于把试点和渠道转成可重复的多站点需求;但订单和合作伙伴集中在少数主体,早期收入因此不平滑、依赖度高。

[CU011, CU022, CU028, CU036]
FU002: 采用 / 部署漏斗

示意性采用漏斗:从展示会兴趣到扩展,ROI 证明是关口。

漏斗百分比是企业机器人转化的示意假设,不是公司披露数据。

[CU023, CU024, CU034, CU010]

6.4 客户结论

客户结论是:Stardust 确实产生了强劲早期需求信号,但耐久、有效采用的证据很薄。积极面看,SEER 千台级订单是真实且同类首创的工业里程碑,多城市部署显示运营覆盖面,教育和开发者渠道也在政策支持的中国需求环境中提供多元化选项。 谨慎面看,具名生产部署、留存、复购和量化客户结果均未披露;需求集中在少数标志性交易和合作伙伴渠道;从试点到生产的关键转化仍未证明。客户证据很新——截至 2025 年末和 2026 年中——但它更多描述签约订单和公告,而非已交付、可持续价值。尽调应优先核对交付设备与订单标题,获取具名客户背书和结果数据,评估销售管线与渠道多元化,并在充分认可牵引叙事前拿到任何可用的分群或复购证据。[CU008, CU024, CU026, CU027, CU031, CU032]

6.5 图表

Chapter 07

07风险

7.1 监管、法律与地缘政治风险

监管和法律暴露是最高层级风险,并带有鲜明地缘政治边缘。中国国内发布了全球首个面向人形机器人和具身 AI 的 HEIS 2026 国家标准,既让赛道获得合法性,也在扩产期间带来认证工作量和合规成本;中国《民法典》带来缺陷机器人产品责任,深圳新的具身 AI 责任保险可部分缓释;包括 AI 法草案在内的 AI 治理制度演进,未来可能增加义务。 更尖锐的风险来自外部:美国立法者提出法案——包括 House Select Committee 法案和 American Security Robotics Act——禁止或限制政府采购中国人形机器人;拟议中的 GUARD Act 则会把 FCC 覆盖名单框架扩展到机器人。这些措施反映出针对中国机器人的战略安全审查上升,可能封顶 Stardust 的国际扩张,收窄其可触达市场。中外标准碎片化还可能进一步提高出口合规摩擦。本评估截至 2026 年中,反映最新标准与立法活动。[CR001, CR002, CR003, CR015, CR016, CR017]

监管 / 法律风险台账
规则 / 牌照 / 案件司法辖区状态可能性严重性缓释措施剩余暴露
美国拟禁中国机器人(众议院法案 ASRA)美国2026 年提议聚焦中国市场高——限制美国扩张
GUARD Act(将 FCC 涵盖清单扩展至机器人)美国2026 年提议跟踪立法高——市场准入
HEIS 2026 国家标准中国2026 年形成中对齐标准中——合规成本
产品责任(《民法典》)中国已生效责任保险中——缺陷索赔
AI 法草案 / 治理中国草案跟踪合规中——未来义务
深圳具身 AI 责任保险中国(深圳)2026 年可用投保低——已有缓释工具

这些规则和案例覆盖中国的标准 / 责任制度与美国立法提案;排序按市场准入和合规影响的严重性。

[CR001, CR002, CR003, CR015, CR016, CR017]
FR003: 依赖图

Stardust 周边的关键供应商、投资方和客户 / 渠道依赖。

[CR006, CR010, CR011, CR020]

7.2 运营、质量与安全风险

运营风险集中在 Stardust 用来差异化的同一项技术上。绳/缆驱动执行器在持续商业规模下的可靠性未被证明,是最重要的单一技术风险;缆绳磨损和复杂肌腱走线是公认工程挑战,可能推高维护和故障率,公开资料也没有 MTBF 或现场故障数据。无锡新工厂扩产带来执行、良率和营运资金风险,且连续性依赖单一旗舰工厂,风险进一步叠加。 公共和商业场景中的人机交互安全——碰撞和力风险——需要稳健控制;低惯量柔顺绳驱动肢体可部分帮助,但没有公开安全测试数据支撑。质量与认证证据同样未披露。按严重性排序,大规模执行器故障(关键影响,缓释成熟度低)和无锡爬坡不达预期(高影响)是主要运营暴露;两者缓释都不成熟,残余风险为实质到高。[CR004, CR005, CR012, CR018, CR033]

运营 / 质量 / 安全风险台账
故障模式可能性严重性缓释成熟度剩余暴露未解决缺口
绳驱执行器规模化失效致命低(仅演示)无 MTBF / 现场数据
无锡制造爬坡不及预期低(新设施)无良率数据
线缆磨损 / 维护成本高中高耐久性未证实
人机交互安全事故中低中(柔顺肢体)无安全测试数据
单一设施连续性无备份产能
质量 / 认证不足认证未披露

各行按严重性排序;执行器可靠性和制造爬坡是最严重的运营暴露,缓释也最不成熟。

[CR004, CR005, CR012, CR018, CR033]
FR001: 风险热力图

主要风险的发生概率、影响和缓释成熟度。

[CR004, CR009, CR007, CR023, CR010, CR013]

7.3 财务、合作伙伴与人员风险

财务模型风险很高。收入和利润率未披露,而价格低于 RMB90,000,意味着单位经济可能很薄或为负;全行业泡沫和产能过剩警示——投资人明确把人形机器人列为可能的下一个 AI 泡沫——会威胁未来资本可得性,并可能引发压缩利润率的价格战。 合作伙伴和依赖风险集中:少数标志性订单,尤其 SEER,主导可见需求;市场拓展依赖合作伙伴渠道(SEER、Kede Education);AI 算力和芯片供应暴露于出口限制;对战略投资方 Thundersoft 的依赖也带来连续性风险。人员和执行风险包括创始人关键人物集中在 Lai Jie 身上,他的机器人愿景和 Tencent 背景可信度支撑公司;技术团队履历公开记录很薄;量产执行未证明,也缺少团队厚度或继任缓释的公开证据。上述风险会通过收入偏弱、利润率受压、融资受限,最终传导为估值倍数压缩。[CR007, CR008, CR009, CR010, CR011, CR013]

合作伙伴 / 依赖风险台账
依赖项交易对手角色集中度失败场景严重性剩余暴露
标杆客户SEER(Xiangong)最大订单 / 渠道订单未交付 / 未复购
教育渠道Kede Education分销渠道表现不达预期
战略投资方Thundersoft资金 + OS / 制造支持支持撤回
AI 算力 / 芯片芯片供应商AI 技术栈算力出口限制产生实质影响
执行器 / 线缆供应商零部件供应商核心硬件供应中断

各行按严重性排序;在地缘政治供应约束下,客户集中(SEER)和 AI 芯片依赖是最尖锐的伙伴风险。

[CR006, CR010, CR011, CR020]
人员 / 执行风险台账
角色 / 职能依赖或缺口可能性严重性缓释措施尽调路径
创始人 / CEO(Lai Jie)关键人集中中低公开未见证据评估留任 / 接班
技术领导层公开资料薄弱Unknown核验团队履历
制造执行大规模量产仍是新课题无锡建设审查运营领导力
商业化 / GTM 团队依赖渠道合作伙伴渠道评估直销能力
梯队厚度公开未见证据Unknown审阅组织架构图

各行按严重性排序;创始人关键人风险和未经验证的大规模制造执行,是主要人员 / 执行暴露,公开可见的缓释证据有限。

[CR013, CR014, CR035]
FR002: 风险传导图

技术、制造、融资和监管风险如何传导到收入、利润率和估值。

[CR021, CR009, CR007, CR032]

7.4 缓释措施、止损标准与风险判断

缓释手段有,但仍在早期。Stardust 刚完成的新股权融资、战略股东、与新兴标准接轨、可获得责任保险,以及产品布局分散到多个场景,能压低部分风险;但没有一项能中和核心技术和财务风险,因此多数类别的剩余敞口仍处于重大到关键水平。投资人应盯住一组可监测触发项和止损标准:执行器现场失效率、无锡量产计划下的实际交付量、毛利率披露、融资续航和新轮融资进展、美国立法状态,以及客户订单多元化。最核心的推翻投资假设事件包括:出现系统性执行器故障证据、无锡产能爬坡失败,或融资市场冻结导致公司无力为扩产供血;美国市场准入禁令会显著压缩可服务市场。总体风险判断为很高,主要由未经验证的技术可靠性、财务不透明,以及上升中的监管和地缘政治敞口驱动。剩余未解缺口是:公司几乎没有披露运营、可靠性和团队数据,无法更精确量化风险。[CR022, CR026, CR027, CR028, CR029, CR036]

缓释措施和否决标准表
风险可监控触发项阈值 / 事件行动含义
执行器可靠性现场故障率出现系统性故障报告退出 / 暂停——投资逻辑失效
制造爬坡交付台数对比计划无锡爬坡未达目标重新评估——执行失败
单位经济性毛利率披露毛利持续为负降低风险敞口——模型失效
融资现金跑道 / 轮次进展融资失败或市场冻结退出——资本风险
监管 / 地缘政治美国立法状态美国禁令落地重新评估 TAM——市场准入丧失
客户集中订单多元化SEER 订单未交付 / 未复购重新评估牵引力

这些触发项是头部风险的可监控代理指标;执行器系统性失效、无锡爬坡失败或融资冻结,是主要投资逻辑破裂事件。

[CR027, CR028, CR029, CR021]

7.5 展示材料

Chapter 08

08估值

8.1 投资建议与论点

我们的建议是跟踪,整体评分 6.2/10,置信度低到中等。判断链条很清楚:公司确有早期牵引力——千台级订单、多城市部署、旗舰产品已按商业价格销售——也有差异化的绳 / 缆驱动架构;但风险极高、财务披露很薄,我们认为现估值相对已验证基本面偏贵。多头逻辑是,绳驱差异化叠加先发商业定价和政策支持的国内需求,可能在未来大市场里换来超额份额。反面论点同样站得住:核心执行器技术尚未在规模化场景验证,订单更像集中签单,而非已交付价值,单台经济性未披露,来自 Unitree-AgiBot 双寡头的竞争激烈,地缘政治限制可能给全球扩张封顶。两套叙事都合理;要判断孰是孰非,所需披露仍然缺位,因此现在承诺资本不如先监控明确触发项。若披露的单台经济性转正,且执行器可靠性在规模上获得独立验证,可上调至买入。[CV014, CV015, CV016, CV024, CV025, CV030]

建议摘要
维度评估
建议观察
综合评分6.2 / 10
置信度中低
风险评级极高
估值立场昂贵
当前估值>RMB10B(~$1.38B)
累计融资~$150M(>RMB1B B 轮)

摘要判断:观察。真实牵引力被披露薄弱和极高风险抵消。

[CV001, CV002, CV003, CV014, CV015, CV016]
投资逻辑 / 反向逻辑表
投资逻辑(乐观视角)反向逻辑(悲观视角)
绳驱差异化让人形机器人更轻、更安全、成本更低非常规执行器尚未经历规模化验证;可靠性未知
早期千台级订单显示真实商业需求订单集中,且只是预订不是交付;ROI 未证实
未来 TAM 大(2035 年约 $38B),中国有产出主导地位泡沫 / 产能过剩警示意味着估值倍数会被压缩
战略支持方(Thundersoft)加上标准对齐财务披露薄弱;单位经济性未披露
先发者,T1 商业定价为 RMB89,900激进定价可能意味着利润率很薄或为负
政策托底的内需环境地缘政治限制(美国限制)压住全球扩张

多空论证相对平衡;每个看多点都有可信的看空反证。

[CV024, CV025, CV012, CV021]
FV001: 建议逻辑

从牵引力、差异化、风险、披露和估值,推导出跟踪结论的证据路径。

[CV014, CV036, CV012]
FV004: 投资 KPI

Stardust Intelligence 的核心投资 KPI。

[CV001, CV002, CV014, CV015, CV016]

8.2 估值、情景与敏感性

Stardust 估值在 2026 年中超过人民币 100 亿元(约 $1.38B),约三个月内连续完成三轮 Series B 融资,合计超过人民币 10 亿元,累计融资约 $150M,由 China Renaissance 担任顾问。关键问题是,公司没有公开收入、ARR 或毛利率,因此无法计算基于基本面的倍数;这一定价反映的是前沿龙头可选性,而非已验证经济性。估值敏感性主要受假设销量、毛利率和终局市场份额支配,TAM 兑现节奏是次要变量,因为机会很大但仍很远。不同情景下结果区间很宽:牛市情景假设无锡扩产成功、执行器可靠性获得验证、多行业订单转成交付,可产生高回报倍数;基准情景假设商业化逐步推进,并在增长但竞争激烈的中国市场拿到适度份额,对应中等结果;熊市情景——执行器或制造失败、融资冻结,或被双寡头挤出——意味着大幅减值。若人形机器人热度降温,行业泡沫和产能过剩预警会增加倍数压缩风险,进一步支持估值偏贵的判断。本数据截至 2026 年中。[CV001, CV002, CV003, CV013, CV017, CV018]

乐观 / 基准 / 悲观情景表
情景关键假设结果隐含回报
乐观无锡产能爬坡跑通,执行器得到验证,多行业订单转化成为前沿龙头,拿下增长中 TAM 的大份额高估值倍数
基准商业化逐步推进,中国份额有限,市场竞争激烈可行的细分玩家;温和增长中等
悲观执行器 / 制造失败,融资冻结,被双寡头挤出减值或下轮折价融资大幅亏损

情景结果跨度很大,反映早期不确定性。

[CV017, CV018, CV019, CV020]
FV002: 估值敏感性

估值对关键假设的相对敏感性估计。

权重是定性尽调估计,不是拟合模型。

[CV029, CV010, CV011]
FV003: 估值 / 回报区间

按情景示意的投资资本倍数区间。

区间是情景示意,不是预测。

[CV020, CV017, CV018, CV019]

8.3 可比估值与市场背景

可比人形机器人估值同时勾勒机会和泡沫。中国方面,Unitree 正以约 $7B 估值推进 IPO,Galbot 估值约 $3B;UBTech 作为上市公司,提供了公开市场的现实校准,其交易所估值通常低于热门私募轮次暗示的倍数。美国方面,Figure AI(约 $39B)和 Apptronik(约 $5.5B)说明,在另一套资本市场和炒作环境里,私募估值可以跑到多远。对照这些公司,Stardust 约 $1.38B 的估值低于头部公司,符合其更早的商业化阶段。市场侧,Goldman Sachs 预计 2035 年全球人形机器人 TAM 约 $38B,中国占全球人形机器人产量的绝大多数,支撑可信的国内基准增长路径;更投机的万亿美元叙事也存在,但距离太远,不应锚定近期价值。AgiBot、Fourier 等更多国内玩家进入,扩大了可比样本,也加剧了同一批资本和客户的竞争,压低相对估值。分析师还警告,中国产量增速快于已验证需求,这会冲击任何乐观可比估值里隐含的增长假设。[CV005, CV006, CV007, CV008, CV009, CV010]

可比估值表
公司地区阶段估值依据对 Stardust 的参照
Unitree中国IPO 前~$7BIPO 报道中国龙头;估值标杆更高
Figure AI美国未上市~$39BVC 轮次报道美国私募市场极端定价
Apptronik美国未上市~$5.5BVC 轮次报道美国头部可比公司
Galbot中国未上市~$3BVC 轮次报道近似国内可比公司
UBTech中国上市公开市场市值交易所上市公开市场现实校验

可比估值来自 IPO、VC 与交易所来源;Stardust(约 $1.38B)低于同业,反映阶段更早。

[CV005, CV006, CV007, CV008, CV009, CV037]

8.4 推翻投资假设的触发项与尽调问题

投资逻辑取决于尚未披露的事实,因此我们把风险转成可监控触发项和必须过关的尽调问题。主要推翻投资假设的触发项包括:系统性执行器故障证据、无锡产能爬坡失败、财务披露后持续负毛利,或融资市场冻结;被双寡头挤出、美国禁令落地是次级触发项,但会迫使 TAM 重新评估。需要跟踪的投资 KPI 是实际交付量、毛利率、订单积压转化、融资续航和现场失效率。投资前必须完成的尽调问题包括:经审计财务、按产品拆分的单台经济性、执行器可靠性和 MTBF 数据、订单到交付的转化证据、股权结构和融资条款,并审阅团队与组织架构以评估执行纵深。若财务和可靠性问题给出正面结果,可支持从跟踪上调到买入;若未通过,或触发任一推翻投资假设的信号,则应放弃。关键未解缺口在于,几乎所有基本估值输入——收入、毛利和可靠性数据——都未披露,估值判断精度因此被封顶。[CV026, CV027, CV028, CV030, CV031, CV023]

论点破裂与止损触发表
触发项信号阈值 / 事件建议影响
执行器可靠性现场故障数据系统性故障退出 —— 论点破裂
制造爬坡实际交付 vs 计划无锡爬坡失败下调
单位经济毛利率披露持续负毛利下调
融资融资进展市场冻结 / 融资失败退出
竞争份额指标被双寡头挤出下调
监管美国立法禁令落地重新评估 TAM

可监控触发项已对应到建议影响;执行器系统性失效和融资冻结是主要的 论点破裂事件。

[CV026, CV027, CV031]
最终尽调索取清单
索取项目的优先级
经审计财务报表确立收入 / 毛利率基准
按产品拆分的单位经济验证定价可行性
执行器可靠性 / MTBF 数据降低核心技术风险
股权结构表与融资条款评估稀释 / 清算优先权
订单到交付转化核验业务牵引质量
团队 / 组织架构图评估执行纵深

尽调索取项按决策影响排序;财务、单位经济和执行器可靠性是门槛项。

[CV028, CV030, CV036]

8.5 展示材料

免责声明

本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;任何投资决策前, 都应直接向管理层和一手文件核验。

证据索引

结论
编号陈述可信度来源
CO001 Stardust Intelligence (Astribot) was founded in Shenzhen in December 2022. SO001, SO002, SO003
CO002 The company is headquartered in Shenzhen (Nanshan district), China. SO001, SO002
CO003 The company operates under the Astribot brand (Chinese name 星尘智能, Xingchen Zhineng). SO001, SO004
CO004 Founder and CEO Lai Jie was the first engineer at Tencent Robotics X and previously led Baidu's Xiaodu robot team. SO005, SO004
CO005 The core founding team was drawn largely from Tencent Robotics X. SO005, SO004
CO006 The company pioneered a software-hardware integrated 'Design for AI' (DFAI) full-stack architecture coupling AI models, an embodied OS, and a rope-driven body. SO001, SO006
CO007 Stardust Intelligence is at Series B stage and reached unicorn status in mid-2026. SO007, SO003
CO008 The company's valuation surpassed RMB10 billion (~$1.38 billion) in mid-2026. SO007, SO008, SO003
CO009 The Series B totaled more than RMB1 billion, raised across three consecutive rounds within roughly three months. SO009, SO010, SO011
CO010 China Renaissance (华兴资本) served as the sole financial advisor for the Series B. SO010
CO011 Disclosed Series B investors include the Liangxi fund, Yangzhou Longtou core-chip fund, Zhongbo Juli, Thundersoft, and Kede Education. SO012, SO013
CO012 Thundersoft made a strategic investment and is pursuing thousand-unit commercial cooperation with Stardust Intelligence. SO013, SO012
CO013 Earlier financing included an angel round led by Yunqi Partners, a Pre-A led by Matrix Partners China, and an A round with Jinqiu Capital and Ant Group. SO014, SO004
CO014 Third-party trackers estimate lifetime capital raised in the range of roughly $150 million. SO014
CO015 The T1 rope-driven humanoid launched in May 2026 priced from RMB89,900. SO015, SO016
CO016 The S1 dual-arm robot debuted at the World Robot Conference in August 2024 and drew global attention for its speed and dexterity. SO017, SO006
CO017 Rope/cable-driven (tendon) actuation is the company's core mechanical differentiator versus conventional motor-driven joints. SO001, SO016
CO018 The company reports thousand-unit orders spanning industrial, retail/commercial, and research/education sectors. SO018, SO019
CO019 In September 2025 SEER (Xiangong Intelligence) signed a thousand-unit industrial humanoid order, described as China's first at that scale. SO019, SO018
CO020 The company frames itself around 'Physical AI' commercialization for the embodied-intelligence era. SO011
CO021 Reported commercial deployments span multiple Chinese cities across retail, tourism, research, and industrial settings. SO009, SO011
CO022 At CHTF in November 2025 the company signed an embodied-AI order and dev-platform partnership with Baidu and Jishu Diedai. SO020
CO023 Revenue, ARR, gross margin, and headcount are not publicly disclosed. SO001
CO024 The company operates in an intensely competitive Chinese humanoid market led by volume players Unitree and AgiBot. SO021, SO022
CO025 Analysts and Chinese officials have flagged bubble and overvaluation risk across the humanoid-robot sector. SO023
CO026 The company's ~RMB10B valuation places it among a cohort of Chinese humanoid unicorns formed in 2025-2026. SO024, SO003
CO027 The legal entity is Astribot (Shenzhen) Co., Ltd., with Chinese ICP registration 粤ICP备2025360189号. SO001, SO002
CO028 T1 targets five scenario families: home service, commercial stores, research/education, tourism, and light industry. SO016, SO015
CO029 Early S1 research units were positioned as premium, with third-party price estimates around $100,000-$150,000. SO006
CO030 Founder Lai Jie is reported to have roughly 16 years of robotics experience and a master's from Wuyi University. SO005
CO031 Public sources concentrate the company's narrative on founder-CEO Lai Jie, creating key-person dependence. SO005, SO004
CO032 Dai Yuan (戴媛) is identified as a co-founder of the company. SO004
CO033 The vision stated by the company is to bring versatile AI robots into billions of people's lives and to promote human-robot co-existence. SO001
CO034 Milestone chronology spans a 2023 angel round, the 2024 S1 debut, 2024-2025 Pre-A/A rounds, the May 2026 T1 launch, and the mid-2026 Series B. SO014, SO003, SO007
CO035 No public board roster, control rights, or ownership percentages have been disclosed. SO009
CO036 The Series B propelled Shenzhen's embodied-AI cluster, adding another RMB10B-valued company. SO008, SO025
CO037 Independent robotics trackers profile Astribot as a leading tendon-driven humanoid maker. SO006, SO004
CM001 China accounted for roughly 90% of global humanoid-robot output in 2025. SM001, SM002
CM002 China's humanoid-robot output is forecast to surge about 94% in 2026, led by Unitree and AgiBot. SM001, SM003
CM003 EmbodiedGlobal projects China's 2026 humanoid market at roughly RMB104.7 billion, with some sources citing figures up to RMB1,047 billion. SM002, SM004
CM004 Published China humanoid market estimates for 2026 diverge widely by methodology, from tens of billions to over one trillion RMB. SM004, SM005, SM006
CM005 The China humanoid market is projected to grow at a CAGR exceeding 65% over the forecast horizon. SM002, SM005
CM006 Global humanoid-robot sector revenue in 2026 is estimated at roughly $6.24 billion. SM007, SM008
CM007 Goldman Sachs projects a global humanoid TAM around $38 billion by 2035. SM007
CM008 Morgan Stanley forecasts roughly 50,000 humanoid-robot sales in China for 2026. SM009
CM009 Embodied-AI financing in China reached over RMB33.5 billion in the first eleven months of 2025, about four times the prior year. SM005
CM010 China's 2026 robotics sector has seen over 500 financing rounds worth roughly RMB96 billion, with tech giants and carmakers entering. SM010, SM011
CM011 The addressable market for Stardust is enterprise service and light-industrial humanoids in China, excluding quadrupeds and fixed automation. SM012, SM013
CM012 Status-quo substitutes include human labor, fixed industrial automation, collaborative arms (cobots), and quadruped robots. SM013, SM014
CM013 Buyers span industrial/manufacturing, logistics, retail/commercial, tourism/hospitality, research, and education segments. SM015, SM011
CM014 MIIT and SASAC launched a 2026 nationwide initiative to deploy humanoid robots across industrial and service scenarios at scale. SM016
CM015 China released the world-first national standard system (HEIS 2026) for humanoid robots and embodied AI to support scale-up. SM017, SM018
CM016 Government and municipal programs (e.g., Shenzhen, Wuxi/Liangxi) provide procurement demand and scenario funding. SM016, SM019
CM017 Cost declines - exemplified by sub-RMB90,000 humanoids - are a primary demand driver lowering the adoption barrier. SM015, SM020
CM018 Near-term ROI is uncertain: many deployed robots operate at only 20-30% of human efficiency in trials. SM021
CM019 Analysts warn the sector runs ahead of real commercial traction, flagging bubble and overcapacity risk. SM009, SM022
CM020 Reliability, dexterity, battery, and safety limitations constrain industrial deployment beyond pilots. SM021, SM020
CM021 The enterprise adoption path runs from demo to pilot to limited production to scaled deployment, gated by ROI proof. SM021, SM011
CM022 Education-sector procurement is driven by vocational schools and universities, often with state or institutional budgets. SM015, SM016
CM023 Tech giants (Alibaba, ByteDance) and carmakers (BYD, XPeng) are entering humanoid robotics, intensifying competition and demand signals. SM023, SM010
CM024 China is projected to account for roughly 85% of global humanoid shipments in 2026. SM001, SM024
CM025 Serviceable obtainable market for Stardust is a small early slice of Chinese enterprise humanoid demand, measured in thousands of units near-term. SM025, SM007
CM026 Public showcases such as InnoEx 2026 and CHTF signal rising buyer and consumer interest in humanoids. SM026, SM027
CM027 The 2026 estimates are current as of the run date but derive from divergent third-party methodologies rather than audited data. SM004, SM007
CM028 Morgan Stanley and other banks caution the humanoid boom is further from mass commercialization than market enthusiasm implies. SM009
CM029 National standards reduce fragmentation and improve safety, but industry-wide adoption of the standards will take time. SM028, SM029
CM030 The value chain spans actuator/sensor/chip suppliers, body integrators, AI-model developers, and deployment/service partners. SM014, SM030
CM031 Switching costs for enterprise buyers include integration, retraining, and safety-certification effort per deployment. SM021, SM017
CM032 Semiconductor import restrictions constrain Chinese robots' AI-software and compute progress. SM021
CM033 Humanoid robotics attracted over $5 billion of cumulative industry investment globally since 2020. SM012, SM007
CM034 China's humanoid market is widely described as entering its 'mass production year' in 2026. SM004, SM011
CM035 Long-run analyst forecasts (e.g., Morgan Stanley) envision multi-million-unit and trillion-dollar-scale humanoid markets by mid-century. SM009, SM007
CM036 Company positioning targets the 'Physical AI' commercialization segment within embodied intelligence. SM031
CP001 Unitree and AgiBot together hold roughly 80% of China's humanoid-robot market. SP001, SP002
CP002 Unitree is targeting an IPO reportedly valuing the company near $7 billion, far above Stardust's ~$1.38B valuation. SP003, SP004
CP003 Unitree's G1 humanoid retails around $16,000, competing directly with Stardust's affordable positioning. SP003, SP005
CP004 AgiBot (Zhiyuan Robotics) is a leading Chinese humanoid maker pursuing mass production with strong state and investor backing. SP001, SP006
CP005 Fourier Intelligence, a rehabilitation-robotics pioneer, fields the GR-series humanoids targeting research and healthcare. SP007, SP008
CP006 Galbot reached a valuation around $3 billion, exceeding Stardust's, positioning it as a well-funded China peer. SP009, SP010
CP007 UBTech is a publicly listed Chinese humanoid maker with its Walker series targeting industrial and service deployments. SP005, SP006
CP008 US rival Figure AI has been valued around $39 billion, dwarfing all Chinese humanoid startups on paper. SP011, SP005
CP009 1X Technologies' Neo humanoid is priced near $20,000 for Western consumer/enterprise markets. SP005, SP011
CP010 Apptronik, maker of Apollo, has been valued around $5.5 billion with major industrial partnerships. SP011, SP005
CP011 Tesla's Optimus is a high-profile entrant leveraging Tesla's manufacturing and AI scale. SP005, SP011
CP012 Stardust's core differentiation is rope/tendon-driven actuation enabling lightweight, compliant, high-speed manipulation. SP012, SP013
CP013 Stardust's S1 demonstrated end-effector speeds up to ~10 m/s and ~0.1 mm repeatability, among the fastest shown publicly. SP014, SP015, SP016, SP017
CP014 Stardust pursues a 'Design-for-AI' full-stack strategy combining an AI model, embodied OS, and rope-driven body. SP012, SP013
CP015 Stardust's T1 at RMB89,900 undercuts many humanoid rivals, positioning on affordability. SP018, SP019
CP016 Most humanoid competitors use conventional motor-driven or harmonic-drive joints rather than tendon actuation. SP007, SP005
CP017 Key enterprise buying criteria are payload/dexterity, autonomy, price, reliability, and deployment support. SP006, SP020
CP018 Switching costs are moderate: integration, safety certification, and retraining create some lock-in but buyers can multi-home across vendors. SP020, SP021
CP019 Stardust trails Unitree and AgiBot on scale, funding, and mass-production maturity. SP003, SP001
CP020 Tech giants (Alibaba, ByteDance) and carmakers (BYD, XPeng) are likely entrants intensifying competition. SP022, SP023
CP021 China's humanoid firms benefit from a dense domestic supply chain, aiding cost competitiveness across all players. SP006, SP024
CP022 Stardust's moat rests on actuator know-how and full-stack integration, both replicable by better-funded rivals over time. SP012, SP013
CP023 Commoditization risk is high as sub-RMB100k humanoids proliferate and standards reduce differentiation. SP025, SP019
CP024 Unitree's brand recognition and viral demos give it distribution and mindshare advantages Stardust lacks. SP003, SP026
CP025 Stardust's S1 gained global attention after its WRC 2024 debut, building brand among researchers. SP027, SP013, SP028
CP026 Competitive intelligence here is current to mid-2026 but valuations and share shift rapidly. SP003, SP002
CP027 AgiBot and Unitree are furthest along in scaled production and government-backed deployments. SP001, SP006
CP028 Fourier targets rehabilitation and research niches, a different beachhead than Stardust's general-purpose commercial push. SP007, SP008
CP029 Ecosystem lock-in could emerge if Stardust's DuoCore OS and developer platform attract third-party applications. SP012, SP029
CP030 Internal build and continued human labor remain default alternatives for cost-sensitive buyers. SP020, SP030
CP031 Chinese humanoids increasingly compete on price, with several models now below RMB100,000. SP019, SP005
CP032 Stardust's dexterity and speed leadership is demo-stage; peers dispute real-world reliability claims across the sector. SP020, SP013
CP033 US players hold valuation and capital leads but face China's manufacturing-cost and speed advantages. SP011, SP024
CP034 EngineAI and other emerging Chinese startups add to a crowded field of 50+ humanoid developers. SP005, SP031
CP035 Stardust's affordability plus full-stack approach is a genuine but narrow wedge against larger incumbents. SP018, SP012
CP036 Feature parity across humanoids is converging as suppliers share actuators, sensors, and AI toolkits. SP006, SP031
CI001 Stardust Intelligence raised more than RMB1 billion across three consecutive Series B rounds in mid-2026, reaching a valuation above RMB10 billion (~$1.38B). SI001, SI002, SI003
CI002 Stardust's T1 humanoid is priced at RMB89,900 (~$12,500), launched in May 2026. SI004, SI005, SI002
CI003 Stardust's estimated lifetime funding is around $150 million across angel, Pre-A, A, and Series B rounds. SI006, SI007
CI004 Primary revenue streams are hardware sales of humanoid robots (T1, S1) to enterprise, research, and education buyers. SI002, SI005
CI005 The S1 research/dual-arm system is estimated to sell in the ~$100,000-150,000 range for research customers. SI008, SI009
CI006 Stardust has not disclosed revenue, gross margin, burn, or runway; it is a private company with undisclosed financials. SI010, SI006
CI007 Series B proceeds are earmarked for mass production, R&D, and commercialization/go-to-market scale-up. SI002, SI011
CI008 Stardust's investor syndicate includes Thundersoft, Kede Education, Liangxi Fund, a Yangzhou core-chip fund, and Zhongbo Juli, with China Renaissance as sole advisor. SI012, SI003
CI009 Stardust received a first-of-its-kind thousand-unit industrial humanoid order from SEER (Xiangong) in September 2025. SI013, SI014
CI010 Additional orders and a developer-platform MOU were secured at CHTF in November 2025. SI015, SI016
CI011 Stardust is building mass-production capacity at a Wuxi (Liangxi) facility, implying meaningful manufacturing capex. SI011, SI002
CI012 The sub-RMB90,000 T1 price raises questions about gross-margin sustainability given humanoid bill-of-materials costs. SI017, SI018
CI013 Humanoid hardware manufacturing imposes significant working-capital and inventory demands during scale-up. SI019, SI018
CI014 Enterprise humanoid sales cycles are long, involving pilots, integration, and safety validation before production orders. SI020, SI021
CI015 Revenue recognition on large multi-unit orders carries timing risk tied to delivery, acceptance, and deployment milestones. SI013, SI021
CI016 Deployed robots require ongoing service, maintenance, and support, adding to cost of delivery. SI021, SI020
CI017 Sector-wide financing-bubble and overcapacity warnings could tighten Stardust's access to future capital. SI022, SI023
CI018 Listed comparable UBTech's public financials show humanoid makers face heavy losses despite high valuations. SI024, SI025
CI019 Stardust's capital raised trails better-funded rivals such as Unitree (~$7B IPO target) and Figure AI (~$39B). SI026, SI027
CI020 China's embodied-AI financing surged past RMB33.5B in the first eleven months of 2025, easing near-term capital access. SI028
CI021 No public evidence indicates material debt or project-finance obligations; capital appears equity-funded. SI006, SI002
CI022 The next-round trigger is likely tied to hitting mass-production and revenue milestones from the Wuxi facility. SI011, SI002
CI023 Financing facts are current to mid-2026 (Series B closed ~June 2026), the most recent available as of the run date. SI001, SI029
CI024 Stardust is deployed across roughly six Chinese cities, a public operational-scale signal absent revenue disclosure. SI005, SI030
CI025 Revenue mix is inferred to skew toward hardware today, with education and developer-platform revenue nascent. SI005, SI015
CI026 Estimated near-term revenue and burn cannot be reliably bounded from public data; ranges are speculative. SI006, SI031
CI027 Cost declines across the China supply chain support Stardust's ability to price aggressively while pursuing scale. SI019, SI018
CI028 Government and municipal programs (Wuxi/Liangxi, Shenzhen) may offset capex through subsidies or scenario funding. SI032, SI011
CI029 The rapid three-round Series B in ~three months signals strong investor demand but also aggressive capital consumption plans. SI001, SI033
CI030 Financial verdict: capital-rich but revenue-opaque; the investment case rests on unproven unit economics and manufacturing execution. SI006, SI021
CI031 CAC/payback proxies are unavailable; enterprise deal economics must be obtained directly in diligence. SI020, SI021
CI032 Public financial gaps are extensive: revenue, margin, burn, runway, and customer-level economics are all undisclosed. SI010, SI006
CI033 Thundersoft's strategic investment may provide in-kind manufacturing, OS, and supply-chain support beyond cash. SI034, SI002
CI034 Humanoid unit economics are widely reported as unproven sector-wide, with ROI dependent on effective output. SI021, SI020
CI035 Stardust's valuation-to-revenue relationship is unmeasurable without revenue disclosure, a key diligence blocker. SI006, SI031
CI036 Global humanoid supply-chain constraints could raise component costs and pressure margins during scale-up. SI019
CE001 The T1 humanoid launched in May 2026 at RMB89,900 with 23 degrees of freedom, standing ~1.53-1.55m and weighing ~66kg. SE001, SE002, SE003
CE002 The S1 uses rope/cable-driven (tendon) actuation and demonstrated end-effector speeds up to ~10 m/s with ~0.1 mm repeatability. SE004, SE002, SE005
CE003 Stardust follows a 'Design-for-AI' (DFAI) full-stack strategy combining an AI model, an embodied OS (DuoCore), and a rope-driven body. SE002, SE006
CE004 Rope/cable-driven actuation relocates motors to the torso and transmits force via cables, yielding lighter, faster, more compliant limbs. SE004, SE007
CE005 The S1 is a dual-arm wheeled system with 7-DoF arms and ~5 kg payload per arm, aimed at manipulation research and service tasks. SE004, SE008
CE006 The S1 debuted at the World Robot Conference in August 2024, drawing global attention for its speed and dexterity. SE009, SE006
CE007 The T1 supports multiple deployment scenarios including home, commercial, research/education, tourism, and light industry. SE010, SE002
CE008 Stardust maintains a public GitHub organization providing developer-facing repositories and tools. SE011
CE009 The embodied-intelligence layer integrates perception, planning, and manipulation for autonomous task execution. SE002, SE006
CE010 Rope-driven actuation is an unconventional approach; most humanoid competitors use motor or harmonic-drive joints. SE012, SE013
CE011 Long-term reliability and maintenance of rope/cable actuators under sustained commercial load are not independently validated. SE014, SE015
CE012 Stardust plans mass production of the T1 at a Wuxi (Liangxi) facility, targeting scale-up in 2026. SE016, SE017
CE013 The product roadmap runs from S1 (2024 research debut) to T1 mass-market launch (May 2026) to production scale-up (2026). SE009, SE017
CE014 China's HEIS 2026 national standards for humanoid robots and embodied AI set safety and interoperability requirements applicable to Stardust. SE018, SE019
CE015 Shenzhen introduced embodied-AI liability insurance in 2026, part of an emerging safety and compliance framework for humanoid deployment. SE020
CE016 Compliant, lower-inertia limbs from rope actuation can improve human-robot interaction safety versus rigid high-torque joints. SE004, SE007
CE017 Critical dependencies include actuator components, sensors, AI compute/chips, and the domestic robotics supply chain. SE021, SE022
CE018 Semiconductor and AI-compute access is a dependency risk given export restrictions affecting Chinese robotics. SE015, SE022
CE019 T1 offers ~5 kg per-arm payload and 23 DoF, positioning it for light manipulation rather than heavy industrial loads. SE001, SE010
CE020 Product maturity is highest for the demonstrated S1 platform and earlier-stage for T1 at-scale reliability. SE004, SE017
CE021 Stardust's dexterity and speed leadership is validated in demos but not in independent, standardized benchmarks. SE006, SE014
CE022 Deployment requires integration, safety validation, operator training, and ongoing maintenance per site. SE015, SE014
CE023 The DuoCore embodied OS is positioned as a platform for third-party application development. SE002, SE023
CE024 Product/technology information is current to mid-2026, reflecting the T1 launch and Series B disclosures. SE024, SE002
CE025 IP and know-how center on actuator design and full-stack integration; patent breadth is not publicly detailed. SE002, SE006
CE026 The T1's affordable price is enabled by design simplification and China's dense component supply chain. SE024, SE021
CE027 Manipulation performance (speed, precision) is the product's headline technical claim and demo differentiator. SE005, SE004
CE028 T1 targets both consumer/commercial and research/education buyers, broadening the addressable use base. SE010, SE002
CE029 Safety controls must address collision avoidance, force limiting, and compliance with emerging standards. SE018, SE020
CE030 Reliability of complex tendon routing and cable wear is a recognized engineering challenge for rope-driven designs. SE007, SE014
CE031 The AI model layer (embodied intelligence) is central to autonomy but its benchmarked capability is undisclosed. SE002, SE022
CE032 Stardust competes technically with Unitree's G1 and Fourier's GR-series on capability and price. SE013, SE025
CE033 Developer signals (GitHub repositories) indicate an intent to build an application ecosystem around the platform. SE011, SE002
CE034 Quality and compliance evidence is thin publicly; certifications and MTBF data are not disclosed. SE020, SE018
CE035 T1's 23-DoF, ~66kg full-humanoid form is more complex to manufacture reliably than S1's wheeled dual-arm base. SE001, SE004
CE036 Product/tech verdict: genuine, demonstrated dexterity leadership offset by unproven at-scale reliability and thin compliance disclosure. SE004, SE014
CU001 Stardust is deployed across roughly six Chinese cities, spanning industrial, retail/commercial, and education settings. SU001, SU002, SU003
CU002 The T1 launched in May 2026 at RMB89,900, targeting retail, industrial, and education buyers. SU004, SU001, SU005
CU003 SEER (Xiangong) placed a first-of-its-kind thousand-unit industrial humanoid order with Stardust in September 2025. SU006, SU007, SU008, SU009
CU004 The SEER order is reported as China's first thousand-unit-scale humanoid procurement, a landmark industrial demand signal. SU007, SU010
CU005 At CHTF in November 2025, Stardust secured additional orders and a developer-platform MOU involving Baidu and Jishu Diedai. SU011, SU012
CU006 Kede Education, a strategic investor, provides an education-sector distribution channel for Stardust's robots. SU013, SU014
CU007 Reported thousand-unit purchase orders span retail, industrial, and education sectors. SU001, SU006
CU008 Named production deployments and specific end-user identities behind the order signals are largely undisclosed. SU007, SU015
CU009 No public retention, repeat-purchase, churn, or satisfaction (NRR/GRR/NPS) metrics are available. SU015, SU016
CU010 Sector analysts caution that humanoid orders often outrun proven, sustained commercial ROI. SU017, SU016
CU011 Customer concentration is a risk: a small number of large orders (e.g., SEER) dominate visible demand. SU006, SU007
CU012 Education-sector procurement typically draws on institutional or state budgets, aiding pipeline stability. SU013, SU018
CU013 Government and municipal programs (Wuxi/Liangxi, Shenzhen) support the customer pipeline via scenario funding. SU019, SU018, SU020
CU014 The customer journey runs from showcase discovery to pilot to multi-unit order to multi-site expansion. SU011, SU015
CU015 Public showcases (WRC, CHTF, InnoEx, Tokyo Humanoids Summit) are key customer-discovery surfaces. SU021, SU011, SU022
CU016 The developer-platform MOU signals a partner/channel strategy to expand reach beyond direct sales. SU011, SU012
CU017 Reference quality of named proofs is moderate: orders are press-reported but lack independent deployment outcomes. SU007, SU015
CU018 Customer evidence is current to late-2025/mid-2026, the freshest available as of the run date. SU011, SU004
CU019 Industrial buyers (via SEER) represent the most concrete, largest-scale demand to date. SU006, SU007
CU020 Retail/commercial deployments are reported at the scenario level but with limited named-account detail. SU001, SU002
CU021 Repeat and expansion orders beyond the initial announcements are not yet publicly documented. SU015, SU007
CU022 Land-and-expand potential exists if pilots convert to multi-site production, but conversion evidence is pending. SU015, SU011
CU023 Adoption-funnel conversion from showcase interest to production is the key unproven step. SU015, SU016
CU024 Deployment across six cities suggests operational breadth but not necessarily deep per-site utilization. SU001, SU003
CU025 Order headlines (thousand units) should be treated as bookings signals pending delivery and acceptance. SU006, SU015
CU026 China's humanoid deployment momentum broadly supports Stardust's customer-acquisition environment. SU018, SU003, SU023, SU024
CU027 Customer outcomes (productivity, cost savings) from Stardust deployments are not publicly quantified. SU015, SU007
CU028 Partner dependence on distributors (SEER) and education channels (Kede) concentrates go-to-market risk. SU007, SU013
CU029 The education channel offers recurring institutional demand but modest per-unit revenue. SU013, SU001
CU030 Industrial deployment (SEER) targets material handling and repetitive tasks in manufacturing/logistics. SU007, SU006
CU031 Customer verdict: strong early demand signals, weak evidence on production, retention, and outcomes. SU007, SU015
CU032 Freshness risk is low for headline orders but outcome data may lag behind announcements. SU011, SU015
CU033 Buyer, user, and payer often differ (e.g., school procurement vs students vs state budget), affecting sales cycles. SU013, SU018
CU034 Global attention from showcases builds a top-of-funnel that must still convert to paid production deployments. SU021, SU015
CU035 China's broader humanoid adoption (retail, industrial, education pilots) validates the target-segment demand. SU018, SU002, SU025
CU036 Concentration in a few marquee orders makes early revenue lumpy and dependent on individual deals converting. SU006, SU007
CR001 China released the world-first HEIS 2026 national standards system for humanoid robots and embodied AI, setting safety and interoperability requirements Stardust must meet. SR001, SR002
CR002 US lawmakers introduced legislation (including a House Select Committee bill and the American Security Robotics Act) to ban or restrict government purchases of Chinese humanoid robots. SR003, SR004
CR003 The proposed GUARD Act would extend the FCC covered-list framework to robotics, potentially restricting Chinese robots in the US market. SR005, SR006
CR004 Rope/cable-driven actuator reliability at sustained commercial scale is unproven, the single most material technology risk. SR007, SR008
CR005 Cable wear and complex tendon routing are recognized engineering challenges that could raise maintenance and failure rates. SR007, SR008
CR006 Semiconductor and AI-compute export restrictions constrain Chinese robotics firms' software and compute progress. SR009, SR010
CR007 Sector-wide humanoid financing-bubble and overcapacity warnings threaten future capital access. SR011, SR012
CR008 Investors explicitly warn that humanoid robotics may be the next AI bubble, inflating valuations ahead of commercial reality. SR013, SR008
CR009 Undisclosed revenue and margins against a sub-RMB90,000 price create high financial-model risk of thin or negative unit economics. SR009, SR007
CR010 Customer concentration in a few marquee orders (notably SEER) makes early revenue lumpy and deal-dependent. SR007, SR009
CR011 Go-to-market depends on partner channels (SEER distributor, Kede education), concentrating channel risk. SR007, SR008
CR012 Manufacturing scale-up at the Wuxi facility carries execution, yield, and working-capital risk. SR014, SR010
CR013 Key-person risk attaches to founder Lai Jie, whose robotics vision and Tencent-derived credibility anchor the company. SR015, SR016
CR014 Technical-team depth and credentials are thinly documented in English-language media, limiting external verification. SR015, SR017
CR015 China's Civil Code imposes product-liability exposure for defective robots causing harm, a legal risk for humanoid makers. SR018, SR019
CR016 China's evolving AI governance (including a draft AI law) may impose future compliance obligations on embodied-AI firms. SR018, SR001
CR017 Shenzhen's 2026 embodied-AI liability insurance signals rising regulatory attention to deployment safety and liability. SR019, SR001
CR018 Human-robot-interaction safety risks (collision, force) require robust controls, especially in public/commercial settings. SR019, SR007
CR019 Intense competition from the Unitree-AgiBot duopoly and new entrants poses displacement risk. SR020, SR021
CR020 Dependency on strategic investor Thundersoft for manufacturing/OS support is a concentration and continuity risk. SR022, SR016
CR021 Risks transmit into the thesis via revenue (weak ROI), margin (aggressive pricing), financing (bubble), and valuation (multiple compression). SR011, SR009
CR022 Residual exposure after early-stage mitigations remains material-to-critical across technology, financial, and regulatory categories. SR008, SR007
CR023 US market-access restrictions could cap Stardust's international expansion and narrow its addressable market. SR003, SR023
CR024 The risk assessment is current to mid-2026, reflecting the latest standards, legislation, and financing signals. SR001, SR003
CR025 HEIS 2026 compliance will require certification effort and may raise costs during scale-up. SR002, SR001
CR026 Mitigations include equity capital raised, strategic backers, standards alignment, and multi-scenario diversification. SR016, SR022
CR027 Monitorable triggers include delivered-unit conversion, margin disclosure, actuator field-failure rates, and financing runway. SR007, SR008
CR028 A thesis-break trigger is evidence of systemic actuator failures or a failed Wuxi production ramp. SR007, SR014
CR029 Another thesis-break trigger is a financing-market freeze that leaves Stardust unable to fund scale-up. SR011, SR012
CR030 Geopolitical risk is rising: Chinese robots face growing scrutiny as strategic/security concerns in Western markets. SR003, SR004
CR031 Regulatory standards are double-edged: they legitimize the sector but raise compliance costs for scaling firms. SR002, SR024
CR032 Overcapacity in China's humanoid sector could trigger price wars, compressing Stardust's already-thin margins. SR011, SR008
CR033 Operational reliance on a single flagship facility (Wuxi) concentrates manufacturing continuity risk. SR014, SR010
CR034 Standards fragmentation between China (HEIS) and international regimes could raise export-compliance friction. SR024, SR001
CR035 People-risk mitigation (deepening the leadership bench beyond the founder) is not publicly evidenced. SR015, SR017
CR036 Risk verdict: very high overall, driven by unproven technology reliability, financial opacity, and regulatory/geopolitical exposure. SR008, SR003
CR037 Overcapacity and low-cost competition in China's humanoid sector could trigger price wars that compress Stardust's thin margins. SR025, SR026
CR038 Analysts including Morgan Stanley caution that Chinese humanoid output is running ahead of demonstrated end-demand, raising demand-supply-gap risk. SR027, SR028
CR039 Multiple Chinese standards and government bodies (MIIT, SAC/SESEC) govern humanoid safety and interoperability, defining Stardust's compliance surface. SR029, SR030, SR031
CR040 Aggressive low-cost pricing across Chinese humanoid makers pressures unit economics and heightens Stardust's financial-model risk. SR025, SR009
CV001 Stardust Intelligence reached a valuation of over RMB10 billion (~$1.38B) in mid-2026, achieving unicorn status. SV001, SV002
CV002 Stardust raised over RMB1 billion across three consecutive Series B rounds (B, B+, B++) in about three months in 2026. SV003, SV004
CV003 Stardust's cumulative capital raised is approximately $150M across its financing history. SV005, SV003
CV004 China Renaissance (Huaxing) advised the Series B financing, signaling institutional-grade deal structuring. SV006, SV002
CV005 Unitree, the China humanoid leader, is pursuing an IPO at a reported valuation of roughly $7B. SV007, SV008
CV006 US humanoid peer Figure AI has been valued at roughly $39B, illustrating extreme private-market pricing. SV009, SV010
CV007 Apptronik has raised at a valuation of roughly $5.5B, a leading US humanoid comparable. SV009, SV010
CV008 Chinese peer Galbot has reached a valuation of roughly $3B, a close domestic comparable. SV011, SV009
CV009 UBTech is a publicly listed Chinese humanoid maker, providing a market-priced comparable reference. SV012, SV013
CV010 Goldman Sachs projects a global humanoid TAM of roughly $38B by 2035, framing the long-run opportunity. SV014, SV015
CV011 China accounts for the overwhelming majority (~90%+) of global humanoid output, supporting the domestic growth case. SV016, SV017
CV012 Stardust's valuation implies frontier-leader pricing despite undisclosed revenue and unproven unit economics. SV018, SV019
CV013 No public revenue, ARR, or margin figures exist to anchor a fundamentals-based valuation multiple. SV005, SV018
CV014 The recommendation is track: real traction and differentiation are offset by thin disclosure and very high risk. SV018, SV020
CV015 The diligence assigns an overall score of 6.2 with low-to-medium confidence. SV005, SV018
CV016 The valuation stance is expensive relative to demonstrated fundamentals but defensible on optionality. SV021, SV018
CV017 A bull scenario assumes successful Wuxi scale-up, actuator reliability proven, and multi-sector order conversion. SV022, SV014
CV018 A base scenario assumes gradual commercialization with modest share in a growing but competitive China market. SV017, SV014
CV019 A bear scenario assumes actuator/manufacturing failure, financing freeze, or displacement by the Unitree-AgiBot duopoly. SV018, SV020
CV020 Implied returns span a wide range - large multiples in the bull case to substantial impairment in the bear case. SV014, SV018
CV021 Sector bubble warnings imply valuation-multiple compression risk if humanoid hype cools. SV020, SV021
CV022 Analysts caution that Chinese humanoid output is outpacing demonstrated demand, a risk to growth assumptions. SV021, SV023
CV023 The humanoid financing environment remains active but selective, favoring demonstrated commercialization. SV024, SV010
CV024 Thesis: rope-driven differentiation plus early orders could yield outsized share in a large future market. SV022, SV014
CV025 Anti-thesis: unproven technology, thin disclosure, intense competition, and geopolitical limits could impair value. SV018, SV025
CV026 Thesis-break triggers include systemic actuator failure, a failed production ramp, or a financing freeze. SV018, SV020
CV027 A key kill trigger is persistent negative margins revealed once financials are disclosed. SV019, SV026
CV028 Final diligence asks include audited financials, unit economics, actuator reliability data, and cap-table detail. SV005, SV018
CV029 Valuation sensitivity is dominated by assumed unit volume, gross margin, and terminal market share. SV014, SV016
CV030 Conditions to upgrade from track to buy: disclosed positive unit economics and proven actuator reliability at scale. SV022, SV005
CV031 Key investment KPIs: delivered units, gross margin, order backlog conversion, financing runway, and field-failure rate. SV027, SV018
CV032 The Western-peer valuations (Figure, Apptronik, 1X) sit far above Stardust, reflecting different capital markets and hype. SV009, SV010
CV033 The analyst TAM range (from tens of billions to ~$38B+ by 2035) bounds plausible long-run enterprise value. SV014, SV015
CV034 China's output dominance and policy support underpin a credible domestic base-case growth path. SV028, SV017
CV035 Valuation and comparable data are current to mid-2026, reflecting the latest funding rounds and peer marks. SV001, SV007
CV036 The recommendation logic path runs from traction and differentiation, through risk and disclosure gaps, to a track verdict. SV018, SV022
CV037 Listed-peer UBTech's public market cap offers a sobering read on how public markets price humanoid makers versus private rounds. SV012, SV029
CV038 AgiBot and other Chinese entrants intensify competition for the same capital and customers, pressuring relative valuation. SV030, SV031
CV039 Fourier and other domestic humanoid firms broaden the comparable set and the competition for market share. SV032, SV033
CV040 Trillion-dollar long-run humanoid market narratives exist but remain speculative and distant, not a near-term anchor. SV034, SV014
来源
编号出版方标题引文
SO001 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SO002 Baidu Baike Astribot (Shenzhen) Co., Ltd. (enterprise registration)
SO003 Pandaily Astribot Hits Billion-Dollar Valuation After Rapid Series B Close
SO004 AI Wiki Astribot
SO005 RobotTesters Who Is Lai Jie? The Tencent Roboticist Behind Astribot
SO006 RobixOne Astribot Humanoid Robot Company: Brand Review
SO007 36Kr Shenzhen's Stardust Intelligence Secures Over 1B Yuan in Series B
SO008 10jqka (Tonghuashun) Stardust Intelligence completes >RMB1B Series B; Shenzhen adds another RMB10B-valued embodied firm
SO009 36Kr (硬氪) Shenzhen embodied AI firm Stardust Intelligence completes >RMB1B Series B, valuation past RMB10B
SO010 Sina Finance Stardust Intelligence completes >RMB1B Series B; China Renaissance sole financial advisor
SO011 Shenzhen News Over RMB1B Series B: Stardust Intelligence builds commercialization standard for the Physical AI era
SO012 Sina Finance Exclusive: Stardust Intelligence valuation past RMB10B; full investor list revealed
SO013 Sina Tech Stardust Intelligence completes >RMB1B Series B; Thundersoft makes strategic investment
SO014 InforCapital Astribot - Robotics Startup, $150M Raised
SO015 Red Star News (via NetEase) From RMB89,900: Stardust Intelligence unveils rope-driven AI robot T1 that fries steak and runs experiments
SO016 Leaderobot (Robot Lecture Hall) RMB90k manipulation ceiling: rope-driven AI robot T1 launches, cracking 5 scenarios
SO017 Pandaily Astribot's New AI Robot Assistant S1 Officially Debuts at the World Robot Conference
SO018 Securities Daily Stardust Intelligence wins thousand-unit order; humanoid commercialization accelerates
SO019 NetEase SEER (Xiangong Intelligence) and Stardust Intelligence reach thousand-unit cooperation
SO020 21st Century Business Herald First embodied-AI order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai build dev platform
SO021 CNBC China's Unitree plans $7 billion IPO valuation as humanoid robot race heats up
SO022 South China Morning Post Robotics - China Future Tech
SO023 KrASIA Bubble or breakthrough? China's humanoid robotics race faces reality check
SO024 36Kr Another Robotics Unicorn Emerges After Raising 1 Billion Yuan
SO025 Sina Finance Stardust Intelligence completes >RMB1B Series B; Shenzhen adds RMB10B embodied enterprise
SM001 TrendForce China's Humanoid Robot Output to Surge 94% in 2026; Unitree and AgiBot Lead
SM002 EmbodiedGlobal China Humanoid Robot Market Projected 104.7B Yuan in 2026
SM003 Tech Times Unitree IPO Cleared, AgiBot Hits 10,000 Units: China Humanoid Robot Duopoly Takes Shape
SM004 EmbodiedGlobal China Humanoid Robot Market to Reach 1047 Billion RMB in 2026
SM005 36Kr Research Institute 2026 Research Report on the Development of Embodied Intelligence
SM006 AI Robotic Daily 2026 Embodied AI Market: Valuations and Humanoid Robot Leaders
SM007 Axis Intelligence Humanoid Robot Statistics 2026: Market Size, Deployments, and Who's Winning
SM008 Axis Intelligence Humanoid Robot Statistics 2026 - Dashboard
SM009 China Biz Insider Reality Check: Morgan Stanley Warns China's Humanoid Robot Boom Is Further Away Than You Think
SM010 Pandaily Tech Giants Flood Into Embodied Intelligence as the Battle Heats Up
SM011 Pandaily China's Robotics Industry Sees Multiple Breakthroughs Across Humanoid Robots
SM012 HumanoidRobots.com Humanoid Robot Companies in 2026: The Full List
SM013 GrabARobot Humanoid Robot Market 2026: Chinese Manufacturers Leading the Race
SM014 RoboZaps Humanoid Robot Industry Report 2026
SM015 Leaderobot (Robot Lecture Hall) RMB90k manipulation ceiling: rope-driven AI robot T1 launches, cracking 5 scenarios
SM016 Pandaily MIIT and SASAC Launch 2026 Humanoid Robot Real-Scene Training Initiative
SM017 Robotics & Automation News China sets national standards for humanoid robots to support industry scale-up
SM018 SCIO (State Council Information Office) China releases national standard system for humanoid robotics and embodied AI
SM019 Shenzhen News Over RMB1B Series B: Stardust Intelligence builds commercialization standard for the Physical AI era
SM020 Economy.ac Humanoids Accelerated Into the Field by Low-Cost Chinese Competition
SM021 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SM022 KrASIA Bubble or breakthrough? China's humanoid robotics race faces reality check
SM023 South China Morning Post Chinese carmakers race Tesla on humanoid robots; BYD launches humanoid push
SM024 NewsGlobeNow China Humanoid Robot Startups Race Past 100 Billion Yuan
SM025 Securities Daily Stardust Intelligence wins thousand-unit order; humanoid commercialization accelerates
SM026 South China Morning Post In pictures: robots steal the show at Hong Kong's InnoEx 2026
SM027 21st Century Business Herald First embodied-AI order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai build dev platform
SM028 Robotics & Automation News Why China's new humanoid robot standards could change the industry
SM029 RobotToday China's Humanoid Robot & Embodied Intelligence Standard System (HEIS 2026)
SM030 Tech Buzz China China Humanoid Robotics Tracker
SM031 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SP001 Tech Times Unitree IPO Cleared, AgiBot Hits 10,000 Units: China Humanoid Robot Duopoly Takes Shape
SP002 TrendForce China's Humanoid Robot Output to Surge 94% in 2026; Unitree and AgiBot Lead
SP003 CNBC China's Unitree plans $7 billion IPO valuation as humanoid robot race heats up
SP004 36Kr Another Robotics Unicorn Emerges After Raising 1 Billion Yuan
SP005 HumanoidRobots.com Humanoid Robot Companies in 2026: The Full List
SP006 RoboZaps Humanoid Robot Industry Report 2026
SP007 BotInfo Fourier Humanoid Robots: GR-2, GR-3 Specs & Pricing (2026)
SP008 Humanoid Index Fourier Intelligence: Funding, Valuation, Robot Specs
SP009 PR Newswire Galbot Secures Over $300 Million in New Funding, $3 Billion Valuation
SP010 NewMarketPitch Humanoid Robotics: 56 Funding Deals (Full List 2024-2026)
SP011 Axis Intelligence Humanoid Robot Statistics 2026: Market Size, Deployments, and Who's Winning
SP012 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SP013 RobixOne Astribot Humanoid Robot Company: Brand Review
SP014 Origin of Bots Astribot S1 by Stardust Intelligence: Specs & Review
SP015 Origin of Bots Stardust Unveils Astribot S1 Speed Demon
SP016 AI Wiki Astribot S1
SP017 RoboZaps Astribot S1 Review: Price & Specs
SP018 Red Star News (via NetEase) From RMB89,900: Stardust Intelligence unveils rope-driven AI robot T1 that fries steak and runs experiments
SP019 Leaderobot (Robot Lecture Hall) RMB90k manipulation ceiling: rope-driven AI robot T1 launches, cracking 5 scenarios
SP020 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SP021 Robotics & Automation News China sets national standards for humanoid robots to support industry scale-up
SP022 South China Morning Post Chinese carmakers race Tesla on humanoid robots; BYD launches humanoid push
SP023 Pandaily Tech Giants Flood Into Embodied Intelligence as the Battle Heats Up
SP024 NewsGlobeNow China Humanoid Robot Startups Race Past 100 Billion Yuan
SP025 Robotics & Automation News Why China's new humanoid robot standards could change the industry
SP026 Pandaily China's Robotics Industry Sees Multiple Breakthroughs Across Humanoid Robots
SP027 Pandaily Astribot's New AI Robot Assistant S1 Officially Debuts at the World Robot Conference
SP028 Humanoid.guide Astribot S1 Specs & Price
SP029 21st Century Business Herald First embodied-AI order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai build dev platform
SP030 GrabARobot Humanoid Robot Market 2026: Chinese Manufacturers Leading the Race
SP031 Tech Buzz China China Humanoid Robotics Tracker
SI001 36Kr Shenzhen's Stardust Intelligence Secures Over 1B Yuan in Series B
SI002 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SI003 Sina Finance Stardust Intelligence completes >RMB1B Series B; China Renaissance sole financial advisor
SI004 Red Star News (via NetEase) From RMB89,900: Stardust Intelligence unveils rope-driven AI robot T1 that fries steak and runs experiments
SI005 Leaderobot (Robot Lecture Hall) RMB90k manipulation ceiling: rope-driven AI robot T1 launches, cracking 5 scenarios
SI006 InforCapital Astribot - Robotics Startup, $150M Raised
SI007 AI Funding Tracker Top Humanoid Robotics Startups Funded in 2026
SI008 Origin of Bots Astribot S1 by Stardust Intelligence: Specs & Review
SI009 AI Wiki Astribot S1
SI010 Baidu Baike Astribot (Shenzhen) Co., Ltd. (enterprise registration)
SI011 Shenzhen News Over RMB1B Series B: Stardust Intelligence builds commercialization standard for the Physical AI era
SI012 Sina Finance Exclusive: Stardust Intelligence valuation past RMB10B; full investor list revealed
SI013 Securities Daily Stardust Intelligence wins thousand-unit order; humanoid commercialization accelerates
SI014 NetEase SEER (Xiangong Intelligence) and Stardust Intelligence reach thousand-unit cooperation
SI015 21st Century Business Herald First embodied-AI order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai build dev platform
SI016 Leaderobot (Robot Lecture Hall) First AI+robot order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai sign order and strategic cooperation
SI017 Economy.ac Humanoids Accelerated Into the Field by Low-Cost Chinese Competition
SI018 RoboZaps Humanoid Robot Industry Report 2026
SI019 McKinsey & Company Turning humanoid supply-chain constraints into billion-dollar wins
SI020 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SI021 Robotics & Automation News Humanoid robots show clearer ROI but commercial success depends on effective output
SI022 Robotics & Automation News Investors warn AI hype is fueling a bubble in humanoid robotics
SI023 Futurism Experts warn of a humanoid robot financial bubble
SI024 UBTech Robotics UBTech Robotics investor / financial information
SI025 CompaniesMarketCap UBTech Robotics annual reports and market cap
SI026 CNBC Chinese humanoid startups race to IPO: Unitree and peers
SI027 The Robot Report Figure AI raises $1B in Series C toward humanoid robot development
SI028 36Kr Research Institute 2026 Research Report on the Development of Embodied Intelligence
SI029 Pandaily Astribot Hits Billion-Dollar Valuation After Rapid Series B Close
SI030 Pandaily China's Robotics Industry Sees Multiple Breakthroughs Across Humanoid Robots
SI031 Axis Intelligence Humanoid Robot Statistics 2026: Market Size, Deployments, and Who's Winning
SI032 Pandaily MIIT and SASAC Launch 2026 Humanoid Robot Real-Scene Training Initiative
SI033 10jqka (Tonghuashun) Stardust Intelligence completes >RMB1B Series B; Shenzhen adds another RMB10B-valued embodied firm
SI034 Sina Tech Stardust Intelligence completes >RMB1B Series B; Thundersoft makes strategic investment
SE001 Baidu Baike Astribot T1
SE002 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SE003 Red Star News (via NetEase) From RMB89,900: Stardust Intelligence unveils rope-driven AI robot T1 that fries steak and runs experiments
SE004 Origin of Bots Astribot S1 by Stardust Intelligence: Specs & Review
SE005 Origin of Bots Stardust Unveils Astribot S1 Speed Demon
SE006 RobixOne Astribot Humanoid Robot Company: Brand Review
SE007 AI Wiki Astribot S1
SE008 Humanoid.guide Astribot S1 Specs & Price
SE009 Pandaily Astribot's New AI Robot Assistant S1 Officially Debuts at the World Robot Conference
SE010 Leaderobot (Robot Lecture Hall) RMB90k manipulation ceiling: rope-driven AI robot T1 launches, cracking 5 scenarios
SE011 GitHub Astribot organization on GitHub
SE012 BotInfo Fourier Humanoid Robots: GR-2, GR-3 Specs & Pricing (2026)
SE013 Unitree Robotics Unitree G1 humanoid robot (official product page)
SE014 Robotics & Automation News Humanoid robots show clearer ROI but commercial success depends on effective output
SE015 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SE016 Shenzhen News Over RMB1B Series B: Stardust Intelligence builds commercialization standard for the Physical AI era
SE017 SMZDM Stardust Intelligence T1 humanoid to mass-produce in 2026, accelerating home scenarios
SE018 SESEC (Seconded European Standardization Expert in China) China's first standards system for humanoid robots and embodied intelligence
SE019 The AI Insider China Releases National Standards for Humanoid Robotics and Embodied AI
SE020 Amazing Shenzhen Shenzhen issues first legal liability insurance for an embodied AI robot
SE021 McKinsey & Company Turning humanoid supply-chain constraints into billion-dollar wins
SE022 Edge AI and Vision Alliance Humanoid Robots 2026-2036: technologies, markets, opportunities
SE023 21st Century Business Herald First embodied-AI order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai build dev platform
SE024 SMZDM Stardust Intelligence T1 from RMB89,900, deliveries begin June 1
SE025 Fourier Intelligence Fourier GR-1 humanoid (official product page)
SU001 Leaderobot (Robot Lecture Hall) RMB90k manipulation ceiling: rope-driven AI robot T1 launches, cracking 5 scenarios
SU002 Pandaily China's Robotics Industry Sees Multiple Breakthroughs Across Humanoid Robots
SU003 The Diplomat How China's AI-powered robots could reshape the global order
SU004 Red Star News (via NetEase) From RMB89,900: Stardust Intelligence unveils rope-driven AI robot T1 that fries steak and runs experiments
SU005 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SU006 Securities Daily Stardust Intelligence wins thousand-unit order; humanoid commercialization accelerates
SU007 AiBang Robotics SEER and Stardust Intelligence reach thousand-unit cooperation for industrial/logistics deployment
SU008 NetEase SEER (Xiangong Intelligence) and Stardust Intelligence reach thousand-unit cooperation
SU009 Sohu Thousands over two years! World's largest rope-driven humanoid order emerges
SU010 10jqka (Tonghuashun) China's first thousand-unit industrial humanoid order lands
SU011 21st Century Business Herald First embodied-AI order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai build dev platform
SU012 Leaderobot (Robot Lecture Hall) First AI+robot order at CHTF: Stardust Intelligence, Baidu, Jishu Diedai sign order and strategic cooperation
SU013 Sina Tech Stardust Intelligence completes >RMB1B Series B; Thundersoft makes strategic investment
SU014 Sina Finance Exclusive: Stardust Intelligence valuation past RMB10B; full investor list revealed
SU015 Robotics & Automation News Humanoid robots show clearer ROI but commercial success depends on effective output
SU016 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SU017 Robotics & Automation News Investors warn AI hype is fueling a bubble in humanoid robotics
SU018 Robotics Center State of Robotics 2026: China
SU019 Shenzhen News Over RMB1B Series B: Stardust Intelligence builds commercialization standard for the Physical AI era
SU020 Fortune Business Insights Humanoid Robots Market Size, Share & Forecast (110188)
SU021 BusinessMirror China's robotics push shines at Tokyo Humanoids Summit
SU022 There's a Robot For That Humanoid robot companies 2026: complete guide
SU023 Curionic BYD AgiBot vs Unitree vs UBTech: China humanoid comparison 2026
SU024 AgentMarketCap TrendForce: China humanoid 94% surge; Unitree/AgiBot 80% share
SU025 ETCJournal The widening gap: China's humanoid robotics dominance (May 2026)
SR001 SESEC (Seconded European Standardization Expert in China) China's first standards system for humanoid robots and embodied intelligence
SR002 RobotToday China's Humanoid Robot & Embodied Intelligence Standard System (HEIS 2026)
SR003 U.S. House Select Committee on the CCP Lawmakers introduce legislation to ban dangerous Chinese robots
SR004 U.S. News & World Report US lawmakers to introduce bill to ban government use of Chinese robots
SR005 DroneLife Congress introduces GUARD Act extending FCC covered-list framework to robotics
SR006 Ripon Advance GUARD Act targets robot imports
SR007 Robotics & Automation News Humanoid robots show clearer ROI but commercial success depends on effective output
SR008 RoboHorizon Beyond the hype: a brutally honest guide to investing in robotics in 2026
SR009 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SR010 Edge AI and Vision Alliance Humanoid Robots 2026-2036: technologies, markets, opportunities
SR011 Robotics & Automation News Investors warn AI hype is fueling a bubble in humanoid robotics
SR012 Futurism Experts warn of a humanoid robot financial bubble
SR013 The AI Insider Humanoid robotics is the next AI bubble, investors warn
SR014 McKinsey & Company Turning humanoid supply-chain constraints into billion-dollar wins
SR015 RobotTesters Who Is Lai Jie? The Tencent Roboticist Behind Astribot
SR016 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SR017 AI Wiki Astribot
SR018 Rimon Law China AI Regulatory Developments | July 2026 Analysis
SR019 Amazing Shenzhen Shenzhen issues first legal liability insurance for an embodied AI robot
SR020 Curionic BYD AgiBot vs Unitree vs UBTech: China humanoid comparison 2026
SR021 CNBC Chinese humanoid startups race to IPO: Unitree and peers
SR022 Sina Tech Stardust Intelligence completes >RMB1B Series B; Thundersoft makes strategic investment
SR023 The AI Insider US lawmakers to introduce American Security Robotics Act to ban federal purchases of Chinese humanoids
SR024 Robotics & Automation News Why China's new humanoid robot standards could change the industry
SR025 Economy.ac Humanoids Accelerated Into the Field by Low-Cost Chinese Competition
SR026 KrASIA Bubble or breakthrough? China's humanoid robotics race faces reality check
SR027 China Biz Insider Reality Check: Morgan Stanley Warns China's Humanoid Robot Boom Is Further Away Than You Think
SR028 CNBC Morgan Stanley: China humanoid robot market forecast
SR029 SCIO (State Council Information Office) China releases national standard system for humanoid robotics and embodied AI
SR030 Pandaily MIIT and SASAC Launch 2026 Humanoid Robot Real-Scene Training Initiative
SR031 Robotics & Automation News China sets national standards for humanoid robots to support industry scale-up
SV001 36Kr Another Robotics Unicorn Emerges After Raising 1 Billion Yuan
SV002 Pandaily Astribot Hits Billion-Dollar Valuation After Rapid Series B Close
SV003 36Kr Shenzhen's Stardust Intelligence Secures Over 1B Yuan in Series B
SV004 Sohu Financing: Stardust Intelligence completes >RMB1B Series B
SV005 InforCapital Astribot - Robotics Startup, $150M Raised
SV006 Sina Finance Stardust Intelligence completes >RMB1B Series B; China Renaissance sole financial advisor
SV007 CNBC Chinese humanoid startups race to IPO: Unitree and peers
SV008 CNBC China's Unitree plans $7 billion IPO valuation as humanoid robot race heats up
SV009 ValueAdd VC Humanoid robots in 2026: Figure, Apptronik, 1X, Tesla Optimus compared
SV010 NewMarketPitch Humanoid Robotics: 56 Funding Deals (Full List 2024-2026)
SV011 PR Newswire Galbot Secures Over $300 Million in New Funding, $3 Billion Valuation
SV012 CompaniesMarketCap UBTech Robotics annual reports and market cap
SV013 UBTech Robotics UBTech Robotics investor / financial information
SV014 Goldman Sachs Research Global automation: the humanoid robot as an AI accelerant
SV015 IDTechEx Humanoid Robots 2026 market research report
SV016 TrendForce China's Humanoid Robot Output to Surge 94% in 2026; Unitree and AgiBot Lead
SV017 Robotics Center State of Robotics 2026: China
SV018 RoboHorizon Beyond the hype: a brutally honest guide to investing in robotics in 2026
SV019 CKGSB Knowledge Humanoid Robots in China: Progress, Limits and Reality
SV020 Robotics & Automation News Investors warn AI hype is fueling a bubble in humanoid robotics
SV021 China Biz Insider Reality Check: Morgan Stanley Warns China's Humanoid Robot Boom Is Further Away Than You Think
SV022 Astribot Astribot | 星尘智能 — About Us Astribot was founded in Shenzhen in December 2022.
SV023 ABS-CBN (AFP) China says humanoid robot buzz carries bubble risk speed and bubble have always been issues
SV024 AI Funding Tracker Top Humanoid Robotics Startups Funded in 2026
SV025 U.S. House Select Committee on the CCP Lawmakers introduce legislation to ban dangerous Chinese robots
SV026 Economy.ac Humanoids Accelerated Into the Field by Low-Cost Chinese Competition
SV027 Robotics & Automation News Humanoid robots show clearer ROI but commercial success depends on effective output
SV028 Faxiangongchang China Humanoid Robot 2026 report
SV029 Robotics 24/7 UBTech Walker S humanoid in automotive manufacturing
SV030 AgiBot (Zhiyuan Robotics) AgiBot official site
SV031 Tech Times Unitree IPO Cleared, AgiBot Hits 10,000 Units: China Humanoid Robot Duopoly Takes Shape
SV032 Fourier Intelligence Fourier GR-1 humanoid (official product page)
SV033 Humanoid Index Fourier Intelligence: Funding, Valuation, Robot Specs
SV034 CNBC Humanoid robots and the trillion-dollar AI market