初创公司尽调
尽调报告 Robotics / full-stack humanoid robot and embodied AI Late private super-unicorn fundraising 2026-08-30

Zhisquare

战略相关性强,已有真实工业场景验证,资本获取能力也极强——但当前估值已经计入了大量运营兑现,公开记录能清晰支撑的部分还远远不够。

Zhisquare 具备战略重要性,也有机会成长为国家级机器人龙头,但当前私人市场估值相对公开基本面显得偏贵。

封面要素

最新公开融资规模 02
5 RMB B [CO018, CV002]
具名工业锚点订单 03
1000+ robots [CO028, CU006]

公司概况

Zhisquare(智平方,AI² Robotics)是一家总部位于深圳的具身智能机器人创业公司,2023 年由郭彦东创立。它的公开策略,是搭建一个纵向一体化的生产力机器人平台:AlphaBot 硬件由 AlphaBrain / GOVLA / NeuroVLA 模型系统定义,再借助粤港澳大湾区的真实工业场景持续迭代。公司从早期融资走到超级独角兽估值的速度异常快,同时已在半导体 / 显示、汽车、生物科技、零售和公共服务工作流里露出可见落点。

成立时间
2023-04-17
创始人
Guo Yandong
创立地点
Shenzhen, Guangdong, China
总部
Shenzhen / Greater Bay Area, Guangdong, China
产品
面向工业生产力及相邻服务场景的 AlphaBot 通用机器人,以及 AlphaBrain / GOVLA / NeuroVLA 具身模型系统。
客户
制造业优先,尤其是半导体 / 显示、汽车、生物科技和其他工作流密集型工业场景,同时选择性延伸到零售和公共服务。
商业模式
硬件加部署模式,收入可能来自机器人销售或租赁、集成、支持和软件价值,但公开披露的已实现经济性有限。
阶段
Super-unicorn private company
融资情况
公开报道显示,2026 年 2 月 B 轮融资超过 RMB 1B,2026 年 6 月融资接近 RMB 5B,将估值推高至 RMB 20B 以上;此后有报道称公司最早可能在 2027 年筹备香港 IPO。
[CO018, CO019, CO024, CE037, CU001, CU035]

执行摘要

主要优势

  • Zhisquare 拥有中国具身 AI 里最强的公开融资曝光度之一,包括 2026 年近 RMB 5B 融资,以及跨区域国资背景投资人支持。
  • 公司已有真实工业验证,尤其是媒体报道中 HKC 相关的 1,000+ 台机器人半导体 / 显示项目,三年规模约 RMB 500M。
  • AlphaBot 加 AlphaBrain / NeuroVLA 策略有技术差异化,对全栈具身模型架构的表述也少见地清晰。
  • 深圳和更广的大湾区,为工业机器人迭代部署提供强制造和落地生态。

主要风险

  • 公开收入、毛利率、现金跑道和股权结构条款仍太不透明,支撑不了高确信估值判断。
  • 客户证明有意义但仍集中;最强的点名锚定客户,披露细节远多于其他公开客户。
  • 安全、认证、可靠性和现场支持证据,明显薄于架构和融资故事。
  • 更广的人形机器人市场已经被公开描述为过热,倍数压缩或 IPO 延后风险随之上升。
  • 创始人依赖和披露不足的治理,在公司试图最快扩张时增加执行风险。

未决问题

  • 按季度确认的收入、毛利率、积压订单转化、烧钱速度 / 现金跑道。
  • 股权结构表、优先股堆叠、治理权,以及近期融资中新股与老股的拆分。
  • 头部客户集中度、续约行为和跨多 site 重复部署证据。
  • 安全论证文档、认证包、在线率 / MTBF 数据和事故历史。
  • IPO 准备细节,包括审计状态、董事会委员会、报告控制和替代流动性路径。

目录

Chapter 01

01公司概况

1.1 身份、创始人与运营边界

从已审阅的公开记录看,Zhisquare 不是隐身实验室,也不是概念模糊的品牌。公司官方英文和中文页面把 AI² Robotics / 智平方描述为一家 AGI-native 通用机器人公司,2023 年 4 月创立,目标是把具身智能从数字世界带入物理工作流。法律身份线索也支撑这套叙事:第三方登记类页面列出深圳主体成立于 2023-04-17;公司 2025 年 5 月发布的反冒名声明,则给出了深圳南山的具体注册地址和经营地址、北京办公室以及官方渠道。公开材料里的领导层高度围绕创始人展开。郭彦东是清晰可见的经营者和发言人,公司反复强调他的 Microsoft、XPeng 和 OPPO 背景,证明自己能够把原创 AI 工作与大规模智能终端产品化结合起来。因此,公开侧面对身份、使命和创始人可信度的支撑很强,但对更广泛的高管梯队和正式治理机制仍然薄弱。[CO001, CO002, CO003, CO004, CO005, CO006]

快照 KPI 表
指标数值 / 状态日期 / 期间置信度缺口 / 备注
创立2023 年 4 月2023-04公司官方历史和英文 About 页面在创立月份上相互一致。
法律主体注册2023-04-172023-04-17第三方登记类页面,不是政府摘录。
创始人 / CEOGuo Yandong当前官方和媒体来源之间有强交叉印证。
总部中国广东深圳当前公司材料和法律声明都指向深圳南山。
其他运营存在北京和上海当前公司公开材料显示两个城市都有布局。
最新披露估值RMB 20B 以上2026-06 至 2026-08基于媒体报道,不是监管文件。
最新披露融资规模接近 RMB 5B2026-06多家媒体来源均有报道。
具名旗舰订单HKC 相关三年期 1,000+ 台机器人部署2025-09报道订单金额接近 RMB 500M。
关键未披露指标当前 ARR、经审计收入、员工数、董事会权利运行日期这些仍是重要尽调缺口。

混合使用公司官方披露、第三方融资报道和订单报道;未披露指标被明确保留为缺口。

[CO001, CO002, CO004, CO006, CO019, CO018]
领导层与创始人表
人物角色背景职能覆盖关键人影响
Guo Yandong创始人兼 CEOPurdue 博士;曾任 Microsoft、XPeng、OPPO战略、AI 路线图、商业化叙事很高;公司公开身份与创始人高度绑定。
Zhang Peng合伙人在 PKU 实验室揭牌材料中被具名合作伙伴关系与生态支持中等;在研究和生态活动中可见,但公开画像少于创始人。
未具名资深模型 / 硬件负责人公开披露不完整公司称团队包括来自 Microsoft、Google、OPPO、XPeng、Momenta 的人才创始人之下的执行深度重要治理和继任缺口,因为具名高管梯队披露不足。

公开领导层表面高度围绕创始人;更广泛的高管梯队和董事会结构并未完整披露。

[CO004, CO005, CO011, CO035]
FO002: 公司快照逻辑

身份、机器人“大脑”、制造、客户和资本共同强化当前公司叙事。

该图是逻辑图而非定量图;它展示公开公司叙事如何相互放大。

[CO004, CO009, CO008, CO010, CO021, CO037]

1.2 产品足迹与早期商业表面

非融资证据里最强的一点,是 Zhisquare 有一条连贯的产品故事,而且已经把故事推入真实用例。官方材料以 AlphaBot 硬件为中心,硬件由 AlphaBrain / AI2R Brain 模型栈定义;随后,公司又通过产品发布、会议演示和面向具体场景的合作,把叙事铺开。到 2025 年 4 月 AlphaBot 2 发布时,公司已经把汽车、半导体和生物技术制造列为核心工业三角,同时宣布计划延伸到机场和社区服务。2024 WRC 文章和 2026 WAIC 报道显示,公司参加公共活动不只是做品牌曝光,也在定义能力边界。这仍不能证明它已经拥有可持续的量产经济性,但确实说明公司具备一套整合的产品语言、具名机身形态,以及反复被描述的面向客户部署目标,而不只是研究论文叙事。[CO008, CO009, CO010, CO012, CO026, CO027]

里程碑表
日期事件类型金额 / 状态参与方含义
2023-04公司成立创立创立月份已披露Guo Yandong 与创始团队标志 Zhisquare 的 AGI 原生定位起步。
2023-04-17深圳主体注册治理主体有公开登记Zhisquare 法律主体为尽调提供法律锚点。
2024-08-25首次亮相世界机器人大会产品公开演示公司与生态合作伙伴显示早期商业化意图。
2025-01-07宣布 Pre-A 轮融资融资数亿 RMBFortune Capital、Dunhong、CStone 等释放早期机构背书信号。
2025-03-06宣布 Pre-A+ 轮融资融资数亿 RMBDunhong、Yunqi、SDIC 系资本支撑模型迭代和商业化。
2025-04-17AlphaBot 2 与 AGI 终端战略发布产品新旗舰平台公司、PKU 实验室、Bloomage 合作关系把产品与场景策略放进公开路线图。
2025-09-11据报道获得 HKC 关联 1,000+ 台机器人订单规模近 RMB 500MShenzhen Huizhi IoT / HKC 与 Zhisquare形成首个具名大型工业锚点。
2026-02据报道完成 B 轮系列融资融资RMB 1B+Baidu 战略资本等将公司推入 100B-RMB 估值梯队。
2026-06NeuroVLA 发布,并据报道融资近 RMB 5B产品NeuroVLA 加 RMB 5B 融资公司与融资财团技术里程碑叠加超级独角兽估值。
2026-08据报道筹备香港 IPO治理潜在 2027 时间表公司与顾问引入近期流动性叙事。

日期采用公开宣布日,而非内部董事会批准日;本表是唯一记录口径的时间线。

[CO001, CO002, CO012, CO013, CO014, CO008]
FO001: 公司里程碑时间线

从 2023 年成立到 2026 年超级独角兽与 IPO 准备信号的时间线。

时间线日期反映公开披露节点,而不是公司内部私有里程碑。

[CO001, CO012, CO013, CO028, CO016, CO019]

1.3 资本联盟与估值跃迁

Zhisquare 画像发生跃迁,来自资本形成。2025 年官方公告显示,公司已经以 Pre-A 和 Pre-A+ 轮快速融资,用于模型迭代和商业化。随后外部报道陡然提速:2025 年 A 系列融资、2026 年 2 月据报超过 RMB 1B 的 B 轮,以及 2026 年 6 月接近 RMB 5B 的融资;多家媒体称这轮把估值推高到 RMB 20B 以上。6 月融资团的重要性不只在规模,也在构成。Tencent News 描述的资本名单覆盖国家级战略基金、广东和深圳政策载体、保险资金、产业方、券商和财务投资人。正因如此,外部观察者更愿意把 Zhisquare 看作具身智能的区域乃至国家级旗手,而不是又一家拿了风险投资的机器人实验公司。2026 年 8 月 36Kr 报道又叠加了下一层叙事:公司据称启动股改并筹备香港 IPO,等于把 Zhisquare 从高速融资的私营公司,推成近期开启资本市场故事的标的。[CO013, CO014, CO015, CO016, CO017, CO018]

利益相关方与投资人图谱
利益相关方角色控制权 / 经济重要性尽调追问
创始人与运营团队控制权与产品方向运营杠杆高核实留任、归属期和关键人保护条款。
国家级政策基金战略资本国家支持信号价值高判断资本是否附带政策条件或治理权。
广东 / 深圳基金本地产业投资方在 GBA 制造业生态里撬动能力强厘清部署支持和隐含本地落地承诺。
产业战略方潜在渠道与场景提供方合同转化后商业杠杆高区分战略信号与合同收入。
财务投资人与券商定价资本未来流动性 / IPO 压力高测算老股供给和退出预期。

本轮投资方异常多元;核心待解问题是,其价值究竟多少来自战略准入,多少只是纯财务承销。

[CO021, CO022, CO024, CO038]
FO003: 快照 KPI

紧凑版尽调读数显示,Zhisquare 的身份和资本背书很强,但披露缺口也真实存在。

KPI 值混合了硬事实和公开信息明确缺失的指标;这些值用于尽调分诊,不用于估值计算。

[CO001, CO019, CO018, CO028, CO035, CO036]

1.4 事实底座、信号强度与未解缺口

总体看,已审阅证据足以让后文把几件事当作事实底座:Zhisquare 是真实公司,总部在深圳,创立于 2023 年,由郭彦东主导,正在讲一个全栈机器人“大脑”加硬件的故事,并且到 2026 年中已经获得超级独角兽级融资。记录也足以说明商业化并非纯愿景。与 HKC 相关的订单据报接近 RMB 500M、三年交付 1,000 多台机器人,给公司至少一个具名工业锚点,也给出行业里更清楚的规模化证明之一。与此同时,多项尽调关键问题仍未解决。已审阅公开来源没有看到经审计收入、ARR、当前员工数、正式董事会名单,也没有看到 2026 年 6 月融资后的精确股权结构表。因此,公司进入后文时,是一个高度可读的战略叙事,拥有有意义的产品和资本证明,但仍带着会影响承销、治理和下行情景分析的重大披露缺口。[CO028, CO029, CO030, CO034, CO035, CO036]

Chapter 02

02市场分析

2.1 市场边界与规模测算逻辑

要夸大 Zhisquare 的机会,最容易的做法是把它称为机器人公司,然后把整个自动化或 AI 市场都算进来。已审阅证据支持更窄、也更有用的边界:Zhisquare 卖入的是具身工作流,买方需要可重新编程的机器人身体、机器人“大脑”栈、集成和持续服务,才能在真实运营环境里替代或增强劳动力。这个边界排除了大量固定功能自动化和纯软件 AI。在这个更窄的范围内,公开估算仍有参考意义。CNBC 引述 Morgan Stanley 估计,中国人形机器人市场 2026 年为 $2 billion,到 2030 年为 $15 billion;TrendForce 则称中国进入关键商业化阶段时,已经是全球最大的人形机器人市场。这些估算更适合作为方向性的 TAM 视角,而不是精确的 Zhisquare 承销输入,因为它们仍是全国性、外部销售口径的观点,不是从买方预算自下而上搭出来的模型。[CM001, CM002, CM003, CM019, CM020, CM021]

市场定义表
细分 / 类别纳入支出排除支出买方 / 付款方重要性
工业具身机器人机器人本体、脑 / 模型、部署、服务已安装的固定功能自动化工厂业主 / 自动化预算Zhisquare 当下的核心市场。
工业机器人脑平台感知、规划、控制、数据闭环不含具身的通用 LLM 支出CTO / 智能制造预算支撑跨硬件平台经济性。
公共服务人形机器人部署、机队运营、维护消费级家用机器人政府或运营方预算目前证据较弱的邻近市场。
医疗 / 照护邻近场景工作流机器人、任务自动化、服务运营单纯医疗器械或医院 IT医院运营 / 创新预算中期邻近市场,不是核心 SAM。

边界行是基于所引市场和公司材料作出的分析定义;它们不是行业标准分类法。

[CM001, CM002, CM003, CM024]
规模测算视角表
发布方年份地域数值方法 / 单位置信度局限
Morgan Stanley 经 CNBC2026中国$2B 市场;50,000 台仅外部销售;不含原型机 / 内部使用单一分析师估计。
Morgan Stanley 经 CNBC2030中国$15B 市场;446,000 台前瞻出货预测文章未完整披露假设。
TrendForce2026中国年产出增长 94%生产 / 产出增长口径这是增长率,不是绝对 TAM。
深圳行动计划2027 目标深圳集群关联产业规模 RMB 100B+本地集群政策目标集群产值目标,不是厂商收入。
MIIT + SASAC2026 目标中国100+ 个高价值场景;10,000 台部署能力政策部署里程碑政策目标,不是已实现出货。

本表混合分析师估计和政策目标;各行不可相加,也不能合成一个 TAM。

[CM020, CM021, CM018, CM010, CM011]
FM001: 市场规模视角

Zhisquare 切入的是更大机器人市场中的一部分;当下真正可变现的核心层,是工业具身部署。

金字塔混合了市场、场景和集群口径;各层只是背景参照,不能相加。

[CM020, CM011, CM010, CM028]
FM002: 市场估计区间

公开估算都指向大市场,但口径和精度层级并不一致。

低 / 高值是在引用的基准估计周围转换出的敏感性边界。

[CM019, CM020, CM021]

2.2 买方、细分场景与大湾区适配性

买方地图比一个原始 TAM 数字更清楚。制造业是第一个严肃的钱包。MIIT 2026 年真实场景训练通知,实际上给出了企业买方采购和降低具身机器人风险的蓝图:开放真实场景,和机器人厂商、供应商组建联合体,改造环境,在近似生产条件下训练,验证安全性和经济性,然后才进入常规部署。这套流程与大湾区优势匹配。深圳修订后的行动计划,明确要围绕核心部件、AI 芯片、测试、开放数据、制造和应用场景,打造密集的具身机器人集群;Foxconn 在 GTC 的披露也显示,大型电子制造商已经围绕物理 AI、混合机器人和智能工厂工作流布局。具体到 Zhisquare,汽车、半导体、生物科技、机场和社区场景的列表说明它走的是多细分策略,但短期最深的适配点仍在工业环境:预算负责人已经购买自动化,可重新编程劳动力也能抵消用工短缺、柔性需求或质量约束。[CM012, CM036, CM023, CM025, CM026, CM028]

细分市场与买方图谱
细分买方用户付款方工作流预算负责人采用触发点
电子 / 半导体制造工厂总经理 / 自动化负责人产线操作员与主管资本开支 / 智能制造预算物料搬运、检测、重复性任务运营 + 自动化劳动力替代或良率提升。
汽车制造工厂运营方装配 / 物流员工工业资本开支厂内物流、柔性装配支持制造工程多变环境需要可重新编程的劳动力。
生物技术 / 制药生产设施运营方实验室或生产技术员工厂现代化预算搬运、转运、重复性支持运营 + 质量安全且合规友好的自动化。
机场 / 社区服务运营方或公共机构服务人员运营预算巡逻、服务、物流支持运营 / 采购人员约束或公共服务创新。

买方、用户、付款方角色是根据政策部署逻辑和具名场景证据推断,并非来自已披露合同语言。

[CM023, CM027, CM028, CM029]
FM003: 买方与细分市场图

制造业细分市场主导近期买方准备度,服务业相邻场景仍更早期。

序数评分概括有证据支撑的准备度,并不主张精确市场份额。

[CM023, CM027, CM028, CM029, CM025]
FM004: 采用漏斗

公开采用路径从政策放开的场景开始,只有通过验证后才走到规模化部署。

[CM012, CM036, CM031, CM032]

2.3 增长驱动、约束与采用张力

这个赛道罕见地同时具备很强的自上而下和自下而上驱动。自上而下看,深圳目标到 2027 年培育 10 家以上 RMB-10B 级具身机器人企业、20 家以上 RMB-1B 营收企业、50 个以上十亿元级应用场景,并形成 RMB 100 billion 以上本地集群。MIIT 和 SASAC 则在 2026 年瞄准 100 个以上高价值场景,以及万台级部署能力。自下而上看,Zhisquare 现在通过与 HKC 相关的订单,在半导体显示制造里拿到了具名证据;更广泛的行业报道也说明,真正的启动区是工厂,而不是消费者家庭。但采用摩擦仍然真实存在。标准仍在制定,认证框架仍在演进,出货增长和融资总额也都不能证明工作流层面的 ROI。反向市场解读是:公开估值和创业公司数量,已经跑在单位经济和稳定客户预算的硬证据前面。决定性证明会是反复出现的、由企业自掏腰包的订单,并且这些订单在最初的政策推动或试点光环消退后仍能延续。[CM008, CM009, CM010, CM011, CM027, CM015]

增长驱动与约束表
驱动 / 约束方向时间含义尽调追问
政策支持的场景开放正向近期可加快客户触达和证据积累。衡量部署中补贴驱动与自筹资金各占多少。
GBA 制造业集群密度正向近期依托真实工厂和供应商加快迭代。确认具名本地买方和转化漏斗。
模型 / 数据飞轮正向中期部署越多,机器人性能可能越好。核实数据权利以及跨客户迁移能力。
安全、标准与认证摩擦负向近期可能拖慢大规模人机协作。确认产品已满足哪些标准。
ROI 不确定性负向近期没有回本证据,试点可能停滞。按部署项目索取已实现的用工 / 良率指标。
资本密集度负向持续需要持续融资,并守住制造规模纪律。压测烧钱速度与订单转化。

驱动与约束把已观察到的市场事实和明确尽调问题放在一起,因为公开 ROI 证据仍有限。

[CM025, CM030, CM031, CM032, CM034, CM033]

2.4 市场底线判断

公开证据支持一个强但克制的市场结论。Zhisquare 所处的是一个真实且快速扩张的中国人形机器人与具身智能市场;粤港澳大湾区拥有制造、供应商和地方政策支持的高密度组合,是中国最适合把市场叙事转成部署的地区之一。今天证据最扎实的需求,来自工业工作流:买方可以用劳动力替代、产能吞吐或质量改善来证明支出合理。服务类相邻场景在战略上重要,但货币化证据还没那么充分。尽调里最大的市场问题,不是中国是否在投资这个品类——答案显然是肯定的——而是 2026 年需求里有多少由补贴辅助,哪些买方预算能在没有政策支持时续上,以及标准和 ROI 证据要多快成熟,才能支撑规模化采用。[CM007, CM006, CM025, CM024, CM035, CM034]

Chapter 03

03竞争格局

3.1 格局与竞争类别

Zhisquare 面对的不是一家公司对另一家公司那种简单赛道。竞争场覆盖中国全栈创业公司、拥有相邻机器人业务的上市公司、全球工业自动化龙头,以及人工劳动力加任务专用自动化的现状。最直接的同行,是试图把机器人身体、自研大脑和工业部署故事结合起来的中国具身智能公司。Unitree 和 LimX 尤其相关,因为它们把中国本土产品速度与可见商业化信号放在一起;UBTech 则作为治理更可见的本土上市公司而重要。Boston Dynamics 和 Tesla 是两个最重要的全球参照点:Boston 已经在销售企业级人形机器人叙事,并发布了工业规格;Tesla 的资产负债表和 AI 基础设施,则意味着它能承受任何创业公司都难以匹配的迭代成本。在这个背景下,Zhisquare 最突出的仍是资本信号和机器人“大脑”定位,还不是公开证明过的规模领先。[CP001, CP002, CP008, CP009, CP012, CP014]

竞争对手画像表
竞争对手类别规模 / 融资信号目标细分差异化局限
Zhisquare全栈具身 AI 创业公司RMB 20B+ 估值;近 RMB 5B 融资工业生产力机器人机器人脑叙事叠加国家背书资本厚度规模经济和定价仍不透明。
Unitree硬件驱动的商业化人形机器人厂商H2 公开定价;分析师点名的领先者人形机器人硬件与广泛场景规格可见、标价透明公开证据偏向硬件与商业化头条。
LimX深圳具身 AI 竞争对手披露 2026 年 B 轮 + Pre-IPO 融资人形机器人与具身 OS / VLA 技术栈产品节奏快,开源叙事明确部署规模和定价公开度较低。
UBTech已上市机器人老牌厂商董事会 / 治理公开,方案多元服务、教育、养老、人形机器人治理透明度与品类宽度定位更宽,可能稀释工业聚焦。
Boston Dynamics全球老牌厂商企业品牌与 Hyundai 现场测试工业物料搬运企业集成与公开规格公开定价未知;足迹以美国为中心。
Tesla Optimus未来规模化进入者$43.5B 现金;2026 年资本开支计划 $25B+通用自主型人形机器人垂直整合与 AI 数据规模商业化时间表和产品契合度仍在变化。

各行只比较公开信号;多家私营竞争者仍未披露定价、客户数或单位经济性。

[CP004, CP006, CP010, CP012, CP016, CP018]
FP001: 竞争定位图

目前看,Zhisquare 属于高资本背书 / 强大脑叙事;Unitree 和 Boston Dynamics 的公开商业化证明更强,Tesla 的规模潜力最占优。

X 轴代表相对资本 / 战略背书;Y 轴代表公开商业化证明。评分是序数综合判断,不是审计指标。

[CP028, CP030, CP031, CP032, CP033, CP034]

3.2 能力、定价与证明对比

公开对比并不均衡,而这种不均衡本身就有信息量。Unitree 用 US$29,900 的 H2 给买方提供了最清楚的标价锚点,同时披露自由度、扭矩和算力等硬件细节。Boston Dynamics 披露 Atlas 的运行特征、企业集成,以及与 Hyundai 的客户测试路径。Tesla 几乎不披露产品层面的购买细节,但 SEC 文件清楚表明,Optimus 正在一个巨大的工业和 AI 基础设施机器内部获得资金支持。Zhisquare 介于这些模式之间。借助 AlphaBrain 和 NeuroVLA,它的公开 AI 叙事强于许多同行;它披露的具名工业订单,也强于许多创业公司的公开证明。但它在价格上仍以报价为主,毛利率、积压订单转收入和合同结构都不透明。对买方而言,选择不只是看谁的机器人最好,也要看谁的证明足够清晰、能支撑采购,以及谁能扛过这个赛道的资本强度。[CP006, CP007, CP015, CP016, CP018, CP019]

功能 / 能力矩阵
购买标准ZhisquareUnitreeLimXUBTechBoston DynamicsTesla Optimus
全栈机器人脑叙事
公开标价
具名工业证据低-中低-中
公开治理透明度
制造规模级资本厚度很高

单元格是有证据支撑的序位判断,不是基准评分;公开证据未知时按保守口径处理。

[CP003, CP024, CP005, CP012, CP020]
定价 / 打包对比
竞争对手公开价格 / 合同模式包含能力未知项含义
Zhisquare报价驱动 / 未披露暗含机器人本体、机器人脑和部署支持ASP、服务费、折扣、租赁条款定价不透明会拖慢外部投资测算。
Unitree H2US$29,900 标价,不含税和运费人形机器人本体,硬件规格已列明实际落地服务和企业定制透明标价可锚定买方预期。
LimX未披露 / 可能按报价人形硬件,加具身 OS / VLA 叙事标价、服务、机群软件定价竞争压力可能取决于捆绑软件。
UBTech未披露 / 企业销售主导人形服务、教育、养老照护解决方案工业定价和实际利润率品类宽度有助打包销售,但削弱可比性。
Boston Dynamics Atlas未披露 / 企业销售主导Atlas 机器人,加 Orbit 和工作流集成机器人租赁 / 购买价格和支持经济性价值主张可能靠企业 ROI,而不是标价。
Tesla Optimus未披露 / 尚未完全商业化在研通用人形机器人商业价格、部署条款、支持模式未来激进定价可能压缩行业利润率。

保留来源中只有 Unitree 公布标价;其余行故意保留定价不透明性,而不是猜测。

[CP006, CP025, CP024, CP038]
FP002: 功能广度 / 能力图

不同竞争者在不同采购标准上领先;价格、治理、规模和 AI 叙事合看,没有单一厂商全面占优。

单元格概括保留下来的公开证据;低分往往既反映披露限制,也反映产品短板。

[CP003, CP024, CP016, CP012, CP020]

3.3 护城河耐久性与 Zhisquare 真正取胜的位置

Zhisquare 真正的竞争论点,在政治、资本和模型野心交汇处最强。公司看起来与中国当前产业政策方向高度一致,而且对一家 2023 年创立的私营机器人创业公司来说,融资能力异常强。这有助于它招人、建设产能、赢得关注,并可能拿到试点入口。问题在于,这些优势能否滚成可持续护城河。公开证据还没有证明。Unitree 可能用透明定价和可见商业化给整个品类施压。LimX 正在软件和开源定位上快速推进。UBTech 提供公开市场治理可见度。Boston Dynamics 展示了当企业级就绪度把集成和可靠性放在前台时应是什么样子。Tesla 仍是长周期的生存性威胁,因为它能把真实世界 AI、资本开支和制造能力结合到别人难以匹敌的规模。若 Zhisquare 的机器人“大脑”栈能转化为实质更好的部署表现,或更快的工业扩张,它仍然能赢;但这个优势目前仍是承销问题,不是公开事实。[CP028, CP029, CP035, CP030, CP031, CP032]

护城河耐久性 / 竞争风险清单
护城河主张威胁严重性证据缓释 / 尽调问题
政府背书的资本厚度商业验证更充分的对手若部署更快,仍可能胜出Zhisquare 融资规模很大,但相对估值,公开规模化证据偏薄要求提供复购客户队列和生产经济性。
机器人脑差异化开源或竞品具身模型栈可能缩小差距LimX 开源叙事、Tesla / Boston / Unitree 的 AI 推进都在加剧竞争要求提供基准测试、可靠性和迁移学习证据。
工业场景立足点买方可能多家并用,或退回既有自动化方案中高标准和锁定效应仍不成熟衡量切换成本和合同排他性。
中国政策同频行业整合仍可能压低私募估值中高独角兽密集意味着后续洗牌把估值下调轮纳入估值测算。
早期具名订单证明一笔订单不等于持久分销能力HKC 订单金额可观,但集中度高验证 backlog 转化,以及向第二、第三个锚点客户扩张。

本清单聚焦护城河耐久性,不是一般经营风险;多项缓释动作需要私下尽调,而不是继续公开网页研究。

[CP028, CP026, CP027, CP036, CP005]
FP003: 护城河 / 准备度 KPI

Zhisquare 拿到顶级资本支持,AI 叙事也强,但公开准备度证明仍落后于最强的工业既有厂商。

[CP004, CP005, CP025, CP035, CP036]

3.4 竞争底线判断

综合来看,按资本信号和叙事野心,Zhisquare 属于中国具身智能竞争者的第一梯队;但按商业化证明,它还不是公开记录里最干净的领导者。如果委员会最看重价格透明度和可见硬件商业化,Unitree 领先。如果最看重工业企业级包装和已发布的性能特征,Boston Dynamics 目前更成熟。如果最看重压倒性的资产负债表实力,Tesla 无可匹敌。Zhisquare 的差异化道路,是一个中国工业冠军逻辑:强机器人“大脑”叙事、有分量的国资和战略资本联盟,以及至少一个大型具名工业订单。这个竞争起点很有吸引力,但仍需要更广分销、更清晰经济性,以及证明客户留下来是因为性能,而不是新鲜感或政策势能。下一个决定性竞争数据点,将是多地点重复部署,能证明买方粘性,而不是一次性的展示胜利。[CP004, CP005, CP024, CP014, CP020, CP036]

Chapter 04

04财务情况

4.1 收入模式与真正可见的部分

Zhisquare 的公开商业故事,在机制层面很清楚,在已实现经济性层面并不清楚。公司卖的不只是一具机器人身体:官方页面、发布材料和 WRC 画像都持续暗示,一个完整包会包括具身硬件、机器人“大脑”栈、部署到客户工作流,以及持续运营支持。这意味着收入很可能来自硬件销售或租赁、集成工作、软件 / 机队价值和部署后服务的某种组合。然而,保留下来的公开记录没有披露确认收入、ARR、ASP、毛利率,也没有披露这些元素的 attach rate。一个重要例外是商业需求证明:与 HKC 相关的订单给出了一个规模不小的具名积压订单信号,但即便如此,它也更适合理解为订单证明,而不是确认收入证明。换句话说,商业模式可读,利润表不可读。[CI001, CI002, CI003, CI011, CI012, CI013]

收入来源表
收入流机制单位当前数值 / 状态质量尽调问题
机器人销售向企业场景销售具身机器人按台未披露真实存在,但经济性不透明要求提供交付台数、ASP 和收入确认政策。
集成 / 部署场地适配、验证、安装按项目未披露早期部署中可能占比较大要求提供实施费用和利润率结构。
软件 / 机器人脑栈模型、控制、机群或工具链经济性许可 / 捆绑未披露战略重要,但公开未定价区分已变现软件与开源生态活动。
服务 / 维护支持、更新、保修、现场服务按合同 / 周期未披露可能形成经常性收入,但公开证据未验证要求提供附加率、服务毛利率和续约条款。
租赁 / RobotaaS政策鼓励使用或租赁模式按合同未来潜在路径Zhisquare 公开证据尚未显示询问是否已有签署的 RaaS 合同。

多数行故意保留无数值状态,因为公开资料讲机制和用例比讲已入账收入更清楚。

[CI001, CI002, CI019, CI021]
定价 / 变现表
产品方案公开价格 / 单位标价与实际未知项有来源支撑的含义
Zhisquare 具身机器人套装未披露UnknownASP、折扣、保修、安装费采购经济性无法清晰对标。
软件 / 机器人脑经济性未披露Unknown独立软件价格与捆绑销售的拆分战略意义可能很强,但尚未独立变现。
集成服务未披露Unknown日费率、里程碑计费、软硬件拆分实施收入在早期可能重要。
维护 / 支持未披露UnknownSLA 价格、备件、续约经常性利润率质量仍被遮住。
潜在租赁 / RaaS政策支持,但未披露Unknown残值、融资伙伴、使用量指标可降低买方阻力,但可能抬高资本需求。

公开价格缺失具有分析意义,因为正常单位经济性三角测算因此被堵住。

[CI014, CI021, CI020]
FI001: 收入模型桥

Zhisquare 可能把部署通过硬件、集成和服务组合转成收入,但落地后的经济性没有公开披露。

[CI001, CI002, CI020, CI014]

4.2 资本历史与前瞻充足性

如果说收入表面很薄,资本表面则正好相反。Zhisquare 的融资节奏从 2025 年官方 Pre-A 和 Pre-A+ 公告,提速到 2026 年 2 月超过 RMB 1B 的 B 轮,再到 2026 年 6 月接近 RMB 5B 的融资,并把估值推高到 RMB 20B 以上。公开报道和公司表述显示,这些资金用于支持商业化、量产、机器人“大脑”研发和更广泛的工业落地。这是一份非同寻常的融资履历,也明显降低了短期出现简单流动性危机的概率。但承销边界同样显而易见:总融资额不等于可用现金。公开来源没有披露新股和老股交易拆分、交割后现金余额、受限现金、计划烧钱速度,或与工厂建设绑定的融资承诺。因此,正确结论不是“资金充足性已经解决”,而是 Zhisquare 为自己买到了时间和选择权,却没有买来公开透明度。[CI004, CI005, CI006, CI007, CI008, CI009]

资本充足性表
指标公开信号含义为什么重要缺口
最新大额融资2026 年 6 月近 RMB 5B融资渠道很强支撑制造、招聘和商业化推进募资用途拆分仍不清楚。
上一轮披露的大额融资2026 年 2 月 RMB 1B+ B 轮超级独角兽跃升前,资本获取能力已增强显示融资节奏很快不能用总融资额推断当前现金。
具名积压订单信号与 HKC 相关的近 RMB 500M 订单是商业信号,但单独看还不够可能支撑工厂规划和投资人信心收入转化时间未披露。
账上现金未披露重大缺口缺少它就无法判断现金跑道要求提供当前现金和受限现金。
月度烧钱 / 现金跑道未披露重大缺口决定对下一轮融资的依赖要求提供月度烧钱桥接表和现金跑道模型。

账面融资很强,但没有现金、烧钱速度和扩产计划背后的承诺,就无法为资本充足性下结论。

[CI008, CI006, CI011, CI030, CI038]
FI003: 财务估计区间

最干净的公开区间不是经营表现,而是资本信号:大额融资轮次和估值已知,现金和收入仍未知。

这是资本信号区间,不是收入预测。订单信号的低值偏保守,因为积压订单未必全部转化。

[CI006, CI008, CI011, CI007, CI009]
FI004: 资本密度 / 现金流图

即便融资巨额,在稳定经常性经济性可见之前,现金大概率先被制造、库存、算力、数据采集和现场支持吃掉。

[CI016, CI017, CI018, CI027]

4.3 单位经济与公开证明为什么不够

公开记录强烈提示了哪些单位经济变量重要,却无法有力说明这些变量的数值。人形机器人部署资本强度高,Zhisquare 的毛利路径将取决于实际 ASP、硬件 BOM、现场服务强度、保修成本、利用率,以及集成工作多快变得可复用。MIIT 2026 年部署指引和 Foxconn 的智能制造姿态,都指向技术要求高的企业买方和由集成牵引的长销售周期,这通常意味着更高的售前和交付成本。这个动态不会让模型失去吸引力;它只是说明,如果支持负担居高不下,或客户 ROI 证明得太慢,资本可能很快消耗掉。危险在于,巨额融资会暂时遮住这些低效。没有披露毛利率、回本周期或保修数据时,投资人看到的主要是融资动能和一个具名订单,而不是已经成型的硬件软件贡献毛利故事。[CI020, CI022, CI023, CI026, CI035, CI036]

单位经济性表
指标数值 / null置信度为什么重要尽调问题
已确认收入null / 未披露锚定估值和现金生成需要它要求按收入流提供月度收入。
毛利率null / 未披露硬件毛利率决定烧钱速度和规模化可行性要求提供产品和服务毛利率。
CAC / 销售周期成本null / 未披露企业部署销售成本可能很高要求提供漏斗转化率和售前人力负担。
回本周期 / 客户 ROInull / 未披露复购需求靠 ROI,不靠演示要求提供人工、良率和 uptime 改善。
保修 / 服务负担null / 未披露现场故障会击穿硬件经济性要求提供索赔、备件成本和现场服务率。

公开证据能支撑这些指标的重要性,但不支撑具体数值;null 才是正确答案。

[CI013, CI015, CI020, CI026]
FI002: 单位经济桥

缺失变量很清楚,尽管数值没有披露:硬件价格、服务负担、利用率和质保共同决定盈利能力。

[CI015, CI026, CI022, CI036]

4.4 财务结论

从财务角度看,Zhisquare 像是一家拥有顶级融资渠道和有意义早期商业信号的公司,但基本面距离可公开承销仍差得很远。这个结论不是看空业务本身,而是在评价披露质量。委员会可以合理推断,公司正在建设真实的硬件加部署业务,有足够资本支持继续推进,也存在具名工业需求。但委员会还不能推断确认收入是否已具规模、毛利率能否在规模化后转正、客户集中度是否可控,或下一阶段增长是否需要比 headline 轮次规模暗示的更多资本。据报道的香港 IPO 准备最终可能迫使更干净的指标浮出水面;在此之前,审慎判断应是:Zhisquare 的资本结构令人印象深刻,商业证明有希望,财务内部情况仍是最大的尽调阻塞点。在经审计收入、毛利率和现金指标出现之前,融资规模应被视为喘息空间,而不是运营质量证明。[CI032, CI033, CI034, CI037]

公开财务缺口表
缺失的私有指标影响具体尽调路径
按季度确认收入估值无法对产出或积压订单做基准对照要求提供月度 P&L 和收入确认政策。
按收入流拆分毛利率若服务负担高,规模化可能摧毁经济性要求提供单机 BOM、保修和服务利润率分析。
现金 / 烧钱 / 现金跑道资本充足性仍停留在推测要求提供资金报告和 18 个月运营计划。
新股与老股拆分融资质量和稀释压力仍不清楚要求提供已签署融资摘要和股权结构表。
债务 / 担保 / 项目融资工厂义务可能压缩灵活性要求提供贷款、租赁和担保明细表。

本表有意记录认真做投资测算时标准要求但当前缺失的财务包。

[CI013, CI015, CI030, CI029, CI031]
Chapter 05

05产品与技术

5.1 产品定义与模块地图

Zhisquare 不是在孤立销售单一硬件 SKU。已审阅公开记录描述的是一个分层产品:AlphaBot 硬件、从 AI2R Brain 演进到 Alpha Brain 和 NeuroVLA 的具身模型栈,以及让这些系统在真实环境里变得有用所需的部署工作流和数据闭环。这一点重要,因为判断产品成熟度,不能只看机器人能否行走或抓取物体。真正交付的系统还包括推理和控制软件、边缘部署、场景适配、客户工作流契合和现场运营。因此,公司的模块地图更像一套面向生产力机器人的纵向一体化操作系统,而不是一条裸机械臂或一个演示型人形机器人。这也是开放 AlphaBrain Platform 重要的原因:它把技术表面延伸到封闭设备之外,也支撑了公司关于模型、硬件和场景数据共同演进的说法。[CE001, CE002, CE004, CE007, CE009, CE010]

产品模块 / 资产矩阵
模块 / 资产用户状态 / 成熟度差异化尽调缺口
AlphaBot 硬件线工厂或服务运营方已公开,并有现场展示围绕 AlphaBrain 设计的通用本体详细 BOM 和子系统采购未披露。
AlphaBrain / GOVLA运营人员和机器人系统核心产品大脑全身 VLA 和长时程推理叙事基准测试和生产使用拆分不清楚。
NeuroVLA机器人脑升级路径2026 年公开发布类脑控制和故障恢复框架独立现场性能证据有限。
AlphaBrain Platform研究人员 / 开发者生态开放生态入口数据-训练-模型-评估闭环归于一体采用指标和贡献深度未披露。
场景工作流企业部署团队多个垂直场景活跃场景复利和数据闭环按场景拆分的结果指标仍稀疏。

各行把官方产品组件与其周边公开工作流层放在一起;成熟度按证据可见度判断,不按内部就绪度打分。

[CE001, CE002, CE007, CE009, CE017]
FE001: 产品架构图

Zhisquare 的公开架构把具身模型、机器人硬件、场景数据和部署系统拼成一个一体化产品叙事。

[CE001, CE002, CE003, CE007, CE009, CE016]

5.2 架构与预期工作方式

Zhisquare 公开产品故事里最有差异化的一点,是它愿意暴露相当多的架构语言。2025 年 AlphaBot 2 发布把 GOVLA 描述为全身 VLA 系统,结合广域感知、长程推理和全身输出,而不只是机械臂轨迹。架构被描述为用于规划的慢 System2,以及用于控制的快 System1。2026 年 NeuroVLA 叙事又加入类脑框架,围绕大脑皮层、小脑和脊髓功能展开,明确连接语义理解、动态纠偏、毫秒级执行和故障恢复。AlphaBrain 文档把这个故事拓展为更大的技术框架,覆盖 baseline VLA、世界模型、RL 微调和持续学习。对尽调而言,这强烈说明公司内部有一套连贯架构。但单靠这一点,不能证明该架构在生产环境里优于竞争对手。[CE003, CE005, CE006, CE007, CE008, CE011]

技术 / 运营架构表
层 / 组件角色依赖风险
感知与多模态输入识别场景、物体和指令摄像头 / 传感器栈与数据质量公开传感器物料清单尚未披露。
慢推理系统任务拆解与语言推理模型质量与算力在强实时环境下可能落后。
快控制系统动作生成与轨迹输出低时延控制器与硬件集成实时稳定性公开证明有限。
机器人本体与执行机构执行全身运动机械设计、执行器、电池可维护性和现场维护负担仍不清楚。
数据 / 训练 / 评估闭环在各类场景中改进模型场景准入与数据权利如果没有足够真实部署,泛化可能停滞。

这张表提炼的是公司与技术公开信息呈现的架构逻辑,不是完整工程物料清单。

[CE005, CE013, CE025, CE026, CE036]
FE002: 客户工作流 / 运行流程

产品流程从多模态感知开始,到场景执行结束,并把数据反馈回模型循环。

[CE017, CE013, CE005, CE033]
FE004: 产品成熟度 / 能力图

公开证明在架构叙事和工作流广度上最强,在可靠性、认证和支持细节上最弱。

单元格评分的是公开证据深度,不是公司的内部工程信心。

[CE002, CE017, CE034, CE031, CE032]

5.3 部署工作流与差异化

Zhisquare 把技术绑定到具体工作时,公开工作流证据最强。在工业侧,WAIC 和发布材料强调半导体、汽车和生物科技场景里的 PCB 上下料、跨线物料转运和多步骤制造作业。在服务场景里,公司也展示饮品制作和公共服务任务,但更有投资价值的故事仍是工业生产力。Zhisquare 声称的差异化,是全栈纵向一体化加场景复利式数据闭环:进入真实客户工作流,收集更好的数据,改进模型,再部署更强系统。这个逻辑可信,也有战略一致性,尤其是在深圳这样同时拥有测试基础设施和开放场景政策支持的生态里。挑战在于,公开证据告诉我们的仍然更多是场景广度,而不是每个场景里的运营表现深度。这个区别很重要,因为工作流广度早就能打动会议观众,却要很久以后才可能转化为稳定毛利或可预测的客户支持成本。[CE017, CE018, CE019, CE020, CE026, CE015]

工作流 / 用例表
用户任务当前工作流公司方案可量化收益限制
PCB 上下料人工或固定式半自动化AlphaBot 2 搭配 NeuroVLA 工作流演示可处理多样托盘和连续动作公开 ROI 指标缺失。
半导体产线物料转运工序之间人工搬运AlphaBot 系列用于上下料和转运任务跨任务灵活移动按晶圆厂拆分的规模广度未披露。
生物科技物料搬运和检测人工无菌或重复性工作流机器人执行转运、拆包、视觉检测、补给可能降低污染并提高自动化灵活性已验证质量结果未披露。
机场 / 社区服务人类服务人员AGI-terminal 服务机器人落地可能支撑人员配置和客户互动生产经济性和支持负担不清楚。
零售饮品制作门店人工操作员咖啡 / 冰淇淋 / 鸡尾酒机器人服务可重复服务演示和长营业时长演示成功不等于持久单位经济性。

收益来自公司或媒体叙事;公开记录仍缺少按工作流输出的标准化 KPI 读数。

[CE018, CE022, CE021, CE017, CE019]
路线图 / 发布表
日期 / 阶段功能 / 里程碑状态含义来源
2024-08WRC 真实场景交付首秀公开显示早期产品化姿态WRC 官方帖子。
2025-04AlphaBot 2 + GOVLA 发布公开硬件与大脑栈的重大版本升级官方发布帖子。
2025-04PKU 联合实验室公开加深世界模型与智能体研究闭环官方实验室公告。
2026-06NeuroVLA 发布公开类脑架构成为旗舰叙事APC 及后续 WAIC 报道。
2026-07WAIC 展示与新服务工作流公开覆盖工业以及零售 / 公共服务广度Gasgoo 报道。

路线图记录公开产品与研究里程碑,不代表内部冲刺或发布节奏。

[CE023, CE003, CE024, CE007, CE035]
FE003: 关键依赖图

公开产品故事不只靠模型质量,还依赖场景准入、安全标准、真实世界数据、芯片和现场支持。

[CE026, CE027, CE028, CE016, CE032]

5.4 信任、安全与产品技术结论

产品技术尽调中最尖锐的缺口不是想象力,而是信任证据。MIIT 2026 年部署行动和标准工作清楚表明,碰撞检测、力限制、急停行为、黑箱功能、接口、生命周期管理和评估标准,都正在成为真实人形机器人部署的常规要求。但已审阅的 Zhisquare 公开表面,没有完整回答认证、可靠性指标、事故历史、SLA 设计或正式安全论证。公司未必缺少这些控制;问题在于,公开证据里的架构故事明显跑在信任和运营故事前面。因此,正确的产品结论应当更有层次:Zhisquare 展示了中国具身智能里更丰富的架构叙事之一,也有很强的多场景部署野心;但谨慎投资人要把这个技术栈视为充分去风险化之前,还需要远多于现在的公开或私下证明,尤其是可靠性、合规和现场支持成熟度。严肃尽调应把该技术栈视为有前景但尚未完全工业化,直到公司拿出硬可靠性和认证证据。[CE029, CE030, CE031, CE032, CE033, CE037]

信任 / 质量 / 合规表
控制项 / 指标状态范围缺口
碰撞检测 / 力限制 / 紧急停止政策预期要求;Zhisquare 尚未完整披露真实场景工业部署需要模型到产品的证据和测试报告。
生命周期 / 接口标准国家框架正在成形人形与具身机器人市场需要把标准草案映射到产品状态。
认证包未披露产品与部署站点公开信息中没有完整证书组合。
可靠性 / 正常运行时间指标未披露现场运营需要 MTBF、故障恢复和事故数据。
操作员培训 / 支持 SLA未披露部署与售后服务需要支持手册和人员配置模型。

公开合规证据比架构证据更薄;这种失衡会影响真实部署风险。

[CE029, CE028, CE031, CE032, CE033]
Chapter 06

06客户情况

6.1 客户基础与分层

已审阅公开记录支持一条清楚的客户逻辑:Zhisquare 主要卖入企业和工业工作流,而不是消费需求。半导体显示制造是最强的具名细分市场,汽车和生物科技是接下来最可信的工业垂直,机场、社区和零售服务场景则把产品延伸到面向公众的环境。这个分层重要,因为每个场景里的买方、用户和付款方并不相同。在工业场景里,买方通常是运营方或工厂业主,用户是一线工人或主管,付款方则可能来自自动化、资本开支或运营改善预算的某种组合。在服务场景里,同一个产品可能更接近运营方的 opex 决策。最重要的结论是,Zhisquare 的客户经济性和采用质量,重心仍在制造业优先,而不是大众消费者采用。这个焦点也让客户尽调更可读,因为工厂买方通常比纯消费用例拥有更清晰的预算权限、工作流指标和重复扩张逻辑。[CU001, CU002, CU003, CU004, CU031]

客户分层表
分层买方 / 用户 / 付款方使用场景规模收入 / 战略价值缺口
半导体 / 显示制造工厂运营方 / 产线工人 / 自动化预算PCB 上料、搬运、贴合、测试已命名规模最高核心工业证明,也可能是主要收入驱动需要站点数量和收入拆分。
汽车制造工厂运营方 / 物流人员 / 工业资本开支物料搬运与柔性工厂任务已命名但未量化战略上重要的参考垂直行业需要客户名称和订单规模。
生物技术制造设施运营方 / 技术员 / 运营预算转运、拆包、检测、无菌工作流已命名合作阶段显示高价值、对污染敏感的使用场景需要生产结果指标。
公共服务 / 机场 / 社区运营方 / 服务人员 / 运营开支服务协助与面向公众的工作流计划中或部分落地扩大品牌与数据触面需要持续合同证据。
零售饮品运营运营方 / 服务人员 / 运营预算咖啡、冰淇淋、鸡尾酒服务声称已在多城运营证明面向消费者场景的可靠性需要经济性和留存证明。

这些行区分工业锚点和相邻场景;最强公开证据仍集中在制造业。

[CU003, CU005, CU010, CU015, CU017]
FU001: 客户旅程图

最可能的客户旅程是从认知、试点验证走到生产部署,再到多产线扩张;但首个生产落地案例之后,公开证据迅速变薄。

[CU020, CU028, CU023, CU026]

6.2 具名客户证明与采用轨迹

与许多可比机器人创业公司相比,Zhisquare 的公开具名客户证明更强,但并不均匀。最突出的锚点是深圳惠智物联 / HKC 相关项目:三年超过 1,000 台机器人,据报价值接近 RMB 500M,并且工作流细节覆盖仓储、上下料、组装、检测和 OLED 相关工序。这是一个有意义的生产级证明项,不是模糊的 logo 页。除 HKC 外,公司还为生物科技制造提到 Bloomage / 华熙生物,为半导体任务提到 Jingneng Microelectronics,并为汽车制造提到一家未具名的顶级国际车企。WAIC 报道还补充了 10 多个省市常态化零售类运营的证据。由此浮现的采用路径是可信的:先获得可见度,再验证场景,再进入生产部署,随后可能扩展到多条产线或多个地点。未解决的问题是,究竟有多少部署真的走完了这个闭环。[CU005, CU006, CU007, CU008, CU010, CU014]

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
HKC 相关部署规模3 年内 1,000+ 台机器人2025-09Securities Times / CNStock该领域的大型锚定订单没有总客户基数。
HKC 相关订单金额接近 RMB 500M2025-09Securities Times / CNStock有分量的商业信号没有已确认收入排期。
零售运营覆盖范围10+ 个省 / 市2026-07Gasgoo暗示服务场景触达更广没有站点或设备数量。
饮品出杯量主张每天数百杯饮品2026-07Gasgoo暗示有重复运营使用没有收入或正常运行时间细节。
活跃付费站点未披露运行日期公开缺口综合客户质量的主要盲区活跃站点总数未知。

轨迹证据最强的是一个工业锚点和一条零售运营主张;广义机器人总基数仍未披露。

[CU006, CU007, CU017, CU018, CU021]
已命名客户证据表
客户分层部署 / 使用场景生产部署 / 试点结果限制
Shenzhen Huizhi IoT / HKC半导体显示制造仓储、上料、组装、测试环节合计 1,000+ 台机器人生产规模项目接近 RMB 500M 订单;3 年期限收入转化和在线站点数量未披露。
Bloomage / Huaxi Biology生物技术制造物料搬运、拆包、检测、供应战略合作 / 早期部署可能降低污染并提升自动化没有公开生产量指标。
Jingneng Microelectronics半导体制造上下料与产线间物料转运已命名部署验证半导体工作流适配度订单规模和合同期限未披露。
顶级国际车企(未具名)汽车制造AlphaBot 承担工厂任务已提及但客户未披露释放汽车行业适配信号客户未公开具名。
机场 / 社区运营方公共服务计划 2025 年推出服务计划中 / 部分落地显示相邻场景野心没有清晰证据证明有持续生产合同。

已命名证据混合了硬订单、战略合作和计划推出;投资人不应把所有行视为同等质量证据。

[CU005, CU010, CU014, CU013, CU015]
FU002: 采用 / 部署流程

当 Zhisquare 能展示从场景适配到上线工作流、再到重复规模化的路径时,公开客户证明最强。

[CU028, CU020, CU027]
FU003: 客户证明矩阵

证明集在 HKC 锚点的点名和生产细节上最强,在其他客户的续约可见度上较弱。

单元格衡量公开证据质量,而不是收入绝对重要性。

[CU005, CU010, CU014, CU013, CU017]

6.3 留存、扩张与集中度

客户尽调的主要问题不是没有任何证明,而是缺少耐久性证明。公开来源没有披露活跃站点数、续约率、流失、NRR、GRR 或客户满意度数据。最清楚的合同期限证据,也只有 HKC 相关项目三年期。这意味着,当前对扩张的最佳判断仍是逻辑推断,而不是审计事实:如果 ROI 和可靠性站得住,一个成功的机器人工作流应当能扩散到更多产线、班次和工厂。政策支持和战略资本信号也可能打开入口,但它们本身不能证明终端客户会在补贴或关注退潮后续约。因此,集中度风险很高。当一个锚点客户比其他证明集细节丰富得多时,合理假设应是:在管理层证明之前,客户组合可能比叙事暗示的窄得多。这在机器人行业尤其重要,因为一个耀眼的首个客户,可能遮住浅得多的装机基础,以及通往大范围机队标准化更长的路。[CU021, CU022, CU023, CU024, CU025, CU026]

留存 / 重复使用 / 满意度表
指标数值 / 空值分层置信度尽调要求
合同期限可见度只有 HKC 相关锚点公开 3 年期限工业要求提供前 10 大合同期限和续约选项。
NRR / GRR空值 / 未披露全部要求按签约年份和分层提供队列留存。
流失 / 取消空值 / 未披露全部要求提供流失试点和原因代码。
客户满意度 / NPS空值 / 未披露全部要求提供客户访谈或调研结果。
重复站点扩张数量空值 / 部分可见工业要求提供第二站点或第二产线扩张数量。

留存问题的正确公开答案就是空值;本章记录应要求公司提供什么,而不是猜测。

[CU024, CU023, CU025, CU022]
扩张与集中度风险表
扩张驱动集中度风险影响尽调路径
单一工作流可扩展到多条产线HKC 可能主导已命名证据,也可能主导近期待交付订单要求提供头部客户收入和待交付订单占比。
场景复利数据闭环试点可能无法转化为客户自费生产部署审计试点到生产的转化率。
政策与投资人关注需求可能被信号价值而非 ROI 放大中高区分补贴驱动部署和商业合理部署。
零售与服务相邻场景相邻场景可能分散工业变现注意力扩张前按分层建模毛利率。
政府或集成商撮合间接渠道可能遮蔽真实终端用户留存要求提供直销与渠道收入拆分和终端客户图谱。

这张表关注增长质量,而不是需求是否存在;Zhisquare 显然有需求信号,但集中度可能扭曲解读。

[CU027, CU026, CU028, CU030, CU034]
FU004: 客户侧表现柱状图

公开采用指标并不多,但已披露的少数数字都指向有分量的工业或运营信号,并非没有牵引力。

多根柱采用下限,因为公开披露使用的是 1,000+、10+ 这类表述,而不是精确数量。

[CU006, CU007, CU017, CU018, CU024]

6.4 客户章节结论

Zhisquare 在客户侧看起来已经越过纯愿景阶段。公司有一个真实工业锚点、多个具名垂直参考,也有足够具体的工作流,说明客户不只是来观看 demo。这是真实优势。与此同时,客户章节还不足以支撑低风险规模化结论。公开证据仍混合着生产部署、战略合作、计划 rollout 和运营演示。缺失项经典但关键:活跃站点数、按客户拆分收入、续约行为、流失和头部客户集中度。在这些数据披露前,正确看法是 Zhisquare 拥有有意义的客户牵引和使用证明,但牵引的耐久性和多元化仍是关键承销未知数。因此,纪律严明的投资人应把客户质量视为管理层尽调中需要“再次证明”的领域,而不是已经解决的问题。[CU032, CU033, CU034, CU035]

Chapter 07

07风险

7.1 主要风险簇

Zhisquare 的公开风险表面不同寻常,因为公司同时处在极高估值、极快融资、早期但真实的工业部署,以及仍然稀薄的运营披露交叉点上。这带来四组主要风险。第一,安全和监管风险:人形机器人部署标准仍在变化,MIIT 明确强调安全控制、评估和生命周期管理。第二,集中度和执行风险:HKC 相关锚点是强证明,但细节远多于其他客户证据,比例过重。第三,资本市场风险:公司在一个被乐观观察者也形容为可能在 2027-2028 年洗牌的赛道里,估值已经快速上调。第四,治理和人员风险:公开身份围绕创始人展开,而董事会和高管梯队深度披露不足。真正重要的是相互作用。如果一家高速机器人公司披露了更干净的单位经济,类似估值风险可能更容易被解释;如果一家增长更慢但审计更充分的工业公司面对安全和集中度问题,估值受损可能较小。Zhisquare 今天既没有那么长的审计记录,也没有公开毛利透明度,所以不能把这些风险彼此割裂来看。[CR001, CR002, CR004, CR010, CR025, CR015]

监管 / 法律风险登记表
规则 / 案件司法辖区状态可能性严重性缓释措施剩余敞口尽调路径
人形机器人安全与部署规则中国全国演进中 / 从草案走向实施跟踪 MIIT 标准和部署协议获取产品到标准的映射以及任何站点审批。
数据、隐私与跨客户学习条款中国客户 / 合同公开信息未披露用合同控制和客户专属数据隔离墙约束审查实际客户数据权利和留存条款。
诉讼 / 执法状态中国法律公开信息未发现公开重大案件中高维护合规与品牌控制中高按实体和产品名称检索法院、监管机构和知识产权局记录。
品牌冒用与身份滥用中国已在 2025 年声明中出现强化商标与渠道控制审查商标、纠纷以及经销商 / 渠道滥用控制。

这张表覆盖保留公开来源中可见的监管和法律风险;其穷尽性只限于本次审阅的公开信息面。

[CR004, CR022, CR009, CR008]
运营 / 质量 / 安全风险登记表
故障模式可能性严重性缓释成熟度剩余敞口未解决缺口
实时部署中的安全控制缺口没有公开的安全论证材料。
制造爬坡不达预期中高工厂经济性和良率未披露。
现场服务或备件瓶颈中低SLA 和支持网络未披露。
异常工况密集流程下的可靠性中高没有 MTBF 或干预率数据。
关键子系统供应链集中中高中高子系统归属和采购来源不透明。

运营风险更多来自可靠性和支持体系披露不足,而不是产品野心不够。

[CR006, CR012, CR014, CR031, CR013]
FR001: 风险热力图

剩余风险最高的环节是安全 / 合规披露、客户集中、规模化执行和估值依赖。

序数单元格反映已留存的公开记录,而不是统计损失模型。

[CR006, CR010, CR012, CR002, CR015]

7.2 这些风险为什么有经济意义

这些风险并不抽象。安全控制或认证缺口会拖慢部署,集中度问题会让收入看起来比实际更强,规模化失误会通过返工和服务负担侵蚀毛利,IPO 撤回或 down round 会重定价整个融资故事。由于公开收入、毛利和烧钱速度仍不透明,外部投资人更缺少硬数字来抵消这些下行路径。这意味着风险传导异常直接:运营证明一旦走弱,资本信心可能迅速走弱;资本信心一旦走弱,估值耐久性立刻成问题。Zhisquare 确实有缓释因素:深厚的本地制造生态、国资支持的资本渠道,以及如果旗舰部署成功转化就可能滚起复利的场景化学习闭环。但这些缓释因素只能降低、不能取消硬运营证据的必要性。经济传导之所以严重,还有另一个原因:机器人失败往往会在多个维度同时变贵。一次部署可能错过运营目标、需要更多现场工程师、制造客户不满,并吃掉稀缺管理层注意力,而这些都发生在确认收入追上来之前。在资本充裕阶段,这些问题可以被覆盖一阵子;但一旦公开或私募市场收紧,同样的问题就会迫使估值快速重定价。[CR003, CR011, CR012, CR034, CR035, CR036]

合作伙伴 / 依赖风险登记表
依赖项对手方角色集中度失效情景严重性缓释措施剩余敞口
头部工业锚点客户HKC / Huizhi IoT需求、数据、标杆客户积压订单延期或站点扩张停滞拓展第二、第三个锚点客户
政策基金 / 国资投资方国家级 + 地方基金资本通道和信号背书政策优先级转向或支持降温分散商业验证和私营资本基础
大型买方 / 集成商工厂、运营方、集成商工作流验证和采购中高买方压价或抬高支持负担中高标准化部署并证明 ROI中高
算力 / 组件生态芯片和子系统供应商边缘部署和制造关键部件短缺或成本飙升中高关键子系统多源化中高

依赖风险不只是供应商集中;资本、采购和旗舰客户依赖也在其中。

[CR010, CR001, CR021, CR013]
FR002: 风险传导图

最重要的风险会传导至客户采用、收入质量、融资信心和估值。

[CR011, CR034, CR037, CR039]

7.3 缓释、监控与否决标准

正确的承销姿态不是把 Zhisquare 视为不可投,而是定义哪些证据能降低风险,哪些事件会击穿逻辑。去风险证据包括正式安全和认证映射、超越首个锚点的多地点重复部署、披露现金和烧钱指标、更广客户基础,以及 IPO 准备推进时更清晰的治理。反过来,几类事件应被视为否决标准:旗舰站点发生重大安全事故或召回,积压订单明显无法转成重复生产部署,或 2026 年融资激增后出现可见资本市场压力。这些是可衡量阈值,不是感觉。只要发生其中任何一项,当前超级独角兽叙事都需要大幅修正,因为眼下案例很大一部分押注在未来运营证明追上当前资本信心。本章因此建议,像盯运营里程碑一样盯证据节奏。如果管理层仍依赖宏大的战略叙事,而具体风险指标持续不披露,这本身就应被视为警讯,而不是中性的“没有数据”。实践中,这意味着要求月度部署里程碑复盘、事故报告、集中度 dashboard 和融资应急计划,而不是等待年度叙事更新。公司扩张这么快,如果运营地基真的像资本故事暗示的那样强,就应该能拿出这些材料。[CR038, CR037, CR039, CR040]

人员 / 执行风险登记表
角色 / 职能依赖或缺口可能性严重性缓释措施尽调路径
创始人 / CEO外部身份和战略高度围绕创始人扩充管理层梯队并覆盖接班安排审查留任计划和接班结构。
高级工程负责人公开信息对梯队深度披露不足招聘并披露更强的领导层厚度索取组织架构图和关键招聘留任数据。
现场部署组织支持和运营成熟度不清楚中高投入培训、备件和现场支持按区域和站点审查部署组织。
财务 / IR / 治理IPO 路径会迅速抬高报告负担中高推动内控和董事会委员会专业化审查审计准备、内部控制和董事会材料。

人员风险的核心,是公司能否在扩张速度内完成专业化。

[CR015, CR016, CR014, CR025]
缓释措施和终止标准表
风险可监测触发因素阈值 / 事件行动含义
客户集中最大客户占比持续过高最大客户 >40% 积压订单或收入不要按溢价倍数承销。
安全 / 质量严重事故或召回工人受伤、监管行动或召回事件暂停或拒绝,直到问题修复。
执行质量积压订单转化停滞旗舰订单明显错过部署里程碑转为观察 / 继续研究立场。
资本市场信心降价轮或 IPO 撤回新融资隐含估值更低,或发行推迟大幅重置估值判断。
治理成熟度IPO 准备阶段披露仍然单薄没有经审计指标、董事会可见度弱,或存在内控疑虑视为治理红旗。

这些终止标准把宽泛风险转成可衡量的决策规则。

[CR010, CR038, CR037, CR039, CR025]
FR003: 依赖关系图

Zhisquare 同时依赖政策资本、旗舰客户、监管进展和运营成熟。

[CR001, CR010, CR004, CR012, CR015]
Chapter 08

08估值

8.1 投资逻辑与反向逻辑

Zhisquare 的投资逻辑很容易讲清楚。中国正在加速采用人形机器人和具身智能,粤港澳大湾区是全球最适合把这种采用转化为真实工作流部署的制造集群之一,而 Zhisquare 围绕机器人“大脑”主导的产品故事,组建了一支异常强的资本联盟。这个组合可能让它成为工业具身智能里的国家冠军式赢家。反向逻辑同样清楚:公开估值扩张快于公开财务披露,客户证明基础仍然集中,整个赛道也被广泛描述为过热。换句话说,Zhisquare 完全可能在战略上变得非同寻常,同时以当前私募估值来看却不是一笔风险调整后划算的买入。价格敏感性就是本建议的核心。[CV005, CV006, CV007, CV008, CV009, CV010]

投资逻辑 / 反向逻辑表
论点什么会改变判断
政策背书资本叠加 GBA 制造业密度,可能在工业具身 AI 里跑出全国冠军。如果多客户工业部署反复落地,并披露经济性,投资逻辑会更强。
如果 AlphaBrain / NeuroVLA 能明显改善部署效果,机器人“大脑”差异化可能带来平台溢价。独立现场指标或客户案例显示任务成功率、正常运行时间更高,或集成成本更低。
当前估值跑在公开基本面前面,可能反映泡沫状态。收入、利润率、留存和积压订单转化证据会削弱反向逻辑。
客户集中和治理不透明削弱承销质量。最大客户多元化和 IPO 级治理披露会提高信心。

投资逻辑表围绕能推动决策改变的证据展开,而不是罗列静态观点。

[CV026, CV008, CV027, CV034]
FV001: 推荐逻辑

委员会决策应从真实客户验证和市场顺风出发,穿过经济性缺口与泡沫背景,落到价格敏感的观察立场。

这条流程说明,立场并非看空公司,而是在当前估值标记下保持怀疑。

[CV022, CV025, CV038]

8.2 当前价格语境与可比公司纪律

估值超过 RMB 20B,且刚完成接近 RMB 5B 的融资后,Zhisquare 已经不再按一家有前景创业公司定价。它被定价成一个很可能的未来领导者。这不意味着估值标记一定错误,但意味着证明负担很高。可比公司集合强化的是纪律,而不是确定性。Unitree 用公开硬件定价和据报道的 IPO 进展提供透明度。LimX 是一个快速推进的本土全栈同行,CNBC 报道中披露了定价。UBTech 提供治理可见的上市机器人参照。Boston Dynamics 展示企业级部署成熟度的样子。Tesla 则是规模标杆,也暴露出长期游戏可能有多高资本强度。没有一家是完美可比公司,但它们合在一起传递同一件事:Zhisquare 值得严肃关注,但当前估值要求投资人在公开记录还没充分解释经济性之前先付高价。正因如此,委员会应把可比公司当作三角定位工具,而不是精确计算的虚假邀请。一个可比公司给透明度,另一个给规模,另一个给企业成熟度;它们的重叠部分才是关键信号:Zhisquare 当前定价已经假设若干难事会按顺序做成。[CV001, CV002, CV012, CV013, CV014, CV015]

可比估值表
可比对象指标倍数 / 估值 / 状态参考价值局限
Zhisquare最新私募估值> RMB 20B;近 RMB 5B 融资直接的当前入场参考只有标称估值,没有公开经济性。
Unitree公开价格 + IPO 动能H2 标价 US$29,900;IPO 进展据报道推进价格透明度和资本市场信号更强不是同业务的直接估值倍数。
LimX私募融资和产品动能2026 融资节奏快;CNBC 披露 Oli 定价中国本土全栈同行这里没有公开锚定估值。
UBTech上市且治理可见的同行上市公司,董事会和 IR 信息可见可用于校准治理和披露预期业务组合比 Zhisquare 更宽。
Boston Dynamics / Atlas企业部署准备度客户试点和企业级规格叙事工业准备度的最佳参考所有权和私募估值背景不可直接比较。
Tesla / Optimus全球规模基准$43.5B 现金;2026 年 $25B+ 资本开支;公开销售更晚资本强度和长期规模的最佳参照机器人业务只是 Tesla 的一部分。

可比组按商业模式匹配,而不是追求倍数纯净;每一行回答不同的估值问题。

[CV001, CV012, CV013, CV014, CV015, CV016]
FV002: 估值敏感性

最大的敏感项不是市场规模,而是证据质量:客户多元化和已披露经济性,对公允价值信心的影响大于叙事动能。

柱形为 1-10 的序数敏感性评分,不是统计弹性。

[CV032, CV036, CV034, CV010]

8.3 情景视角与改变建议的因素

基准情景不是崩盘,而是谨慎推进。Zhisquare 可以继续把工业可信度转成更广部署,继续吸引资本,并走向 IPO 窗口。但若要在今天价格上显著改善投资判断,公司需要的不只是动能。它需要首个锚点之外的重复部署、更分散的客户,以及经审计或至少达到管理层级别的收入、毛利率和集中度证据。这些才是能支撑溢价进入、而不是旁观等待的里程碑。因此,乐观情景由里程碑和运营驱动。悲观情景不是具身智能消失,而是一张强订单和强融资观感,在赛道遭遇第一次真正修正、或 IPO 准备要求比当前叙事更干净的数字时,被证明并不够用。[CV028, CV029, CV030, CV031, CV032, CV036]

乐观 / 基准 / 悲观情景表
情景假设估值 / 回报逻辑关键风险概率信号
乐观多站点订单反复赢单、客户更多元、经济性更干净、IPO 准备就绪当前估值可能站得住,甚至显得偏早执行和安全仍然关键有可能,但需要证据
基准强叙事和部分牵引力延续,但透明度改善缓慢没有折扣或结构保护,新资金仍难证明当前估值合理倍数压缩和集中度问题延续从公开证据看最可信
悲观积压订单转化不及预期、融资降温、IPO 延后、泡沫泄气私募估值较当前标称价值明显压缩集中度、单位经济性和情绪同时受冲击不能排除

这些情景偏定性,因为公开记录太薄,不足以支撑诚实的 DCF 式精确计算。

[CV029, CV028, CV030, CV036]
投资逻辑破裂和终止触发因素表
触发因素阈值如何冲击投资逻辑行动含义
积压订单转化不及预期旗舰订单错过部署或扩张里程碑削弱客户验证支柱转为回避 / 重新定价立场。
进入 IPO 窗口后指标仍不透明没有可信的收入、利润率或集中度披露削弱治理和估值支撑不要承销溢价入场。
降价轮或 IPO 撤回资本市场信号反转削弱估值和资本通道两大支柱重置可比组和下行情景。
安全或质量事件重大事故、召回或监管挫折直接削弱部署逻辑事实清楚前完全暂停。
同行重估潮可比龙头估值大幅下修削弱行业倍数支撑立即采用更低入场纪律。

终止触发因素是为投委会监测而设计,不只是事后解释。

[CV036, CV025, CV034]
FV003: 估值 / 回报区间

核心经济性缺失,公开证据只能支撑宽估值区间,而不是精确目标价。

这个区间是克制的情景视角,不是基于审计财务的模型估值结果。

[CV001, CV028, CV029, CV030]

8.4 建议

建议立场是观察 / 继续研究,置信度中等,风险高。这不是否定公司。Zhisquare 拥有真实战略相关性、同类公司里更强的公开工业证明之一;如果运营证据快速追上,仍有明确上行空间。问题在价格和可见度。公开来源仍未披露收入包、毛利率轮廓、股权结构表、优先股堆叠、集中度明细或退出机制,无法支撑在当前 headline 估值下激进投入新钱。纪律严明的投资人可以贴近公司,定义基于里程碑的进入条件,并在公司拿出缺失证明时准备行动。在此之前,最合适的判断是:估值偏高,选择权真实,尽调负担仍未解决。新钱投资人应坚持让运营证据,而不是市场兴奋,决定下一步定价。[CV022, CV023, CV024, CV025, CV033, CV035]

建议摘要表
建议信心风险评级估值立场决策含义
观察 / 继续研究偏高 / 激进没有更清晰的经营证据前,不要按当前估值买入。

建议明确对价格敏感;这不是对公司质量的否定。

[CV022, CV023, CV024, CV025]
最终尽调要求表
主题缺失证据重要性负责人 / 尽调路径
收入和利润率资料包已确认收入、毛利率和积压订单转化数据可信估值的核心输入索取管理层资料包和审计材料。
股权结构表和优先权稀释、优先级、老股交易、治理权利影响真实入场价格和下行保护索取已签署融资摘要和章程文件。
客户集中最大客户占比和续约行为单一锚点客户可能夸大牵引力质量索取前 10 大客户明细和队列数据。
安全和部署证明认证、事故和正常运行时间指标支撑规模化工业逻辑所必需索取站点级运营报告。
退出准备度IPO 工作流、审计师、董事会委员会、报告控制决定实际流动性时间表索取 IPO 准备清单。

这些要求定义了最低资料包;只有拿到它,今天的故事才能转成可投资的定价决策。

[CV009, CV033, CV006, CV034]
FV004: 投资 KPI

Zhisquare 在市场和战略相关性上得分高,验证程度中等,公开经济性可见度低。

[CV007, CV005, CV006, CV009, CV025]

免责声明

本报告是基于截至 2026-08-30 公开信息、由 AI 辅助完成的尽调摘要,不构成投资建议。Zhisquare 是一家私营机器人公司,在收入、利润率、治理、客户集中度和运营可靠性上存在重大披露缺口,因此多项对投资判断关键的结论仍基于情景,而非达到审计级别。

证据索引

结论
编号陈述可信度来源
CO001 AI² Robotics says Zhisquare was founded in April 2023. SO002, SO003
CO002 Baidu Baike states that Zhisquare was established on 2023-04-17 and lists Guo Yandong as legal representative. SO011
CO003 The company describes itself as an AGI-native general-purpose robot developer, manufacturer, and service provider. SO001, SO002
CO004 The company says Zhisquare was founded by Dr. Yandong Eric Guo (郭彦东). SO002, SO003
CO005 Company and media materials describe Guo as a former Microsoft researcher and former XPeng and OPPO AI leader. SO002, SO008, SO022
CO006 Public company materials show a Shenzhen headquarters footprint with Beijing and Shanghai operating presence. SO002, SO004
CO007 The May 2025 anti-impersonation statement lists a Nanshan Zhiyuan address in Shenzhen as the registered and operating address. SO004
CO008 The company positions AlphaBot as its mass-production general-purpose robot series. SO002, SO005, SO025
CO009 The company positions AlphaBrain or AI2R Brain as the embodied model stack that defines AlphaBot hardware and behavior. SO002, SO005, SO023
CO010 Official company materials say Zhisquare already has deployments or applications across automotive, semiconductor, biotech, public-service, and retail settings. SO002, SO005, SO009
CO011 Zhisquare and Peking University formally unveiled a joint embodied-intelligence laboratory in April 2025. SO006, SO011
CO012 The company publicly exhibited Alpha Bot at the 2024 World Robot Conference as an early commercialization showcase. SO009
CO013 Zhisquare announced a Pre-A strategic financing in January 2025 led by Fortune Capital and Dunhong Asset with CStone participation. SO007
CO014 Zhisquare announced a follow-on Pre-A+ financing in March 2025 backed by Dunhong Asset, Yunqi Capital, and SDIC-linked capital. SO008
CO015 36Kr reported that a later 2025 A-series round was led by Shenzhen Capital with more than RMB 100 million from that investor alone. SO012
CO016 36Kr reported that Zhisquare completed a B-round series above RMB 1 billion in February 2026. SO013
CO017 36Kr reported that Zhisquare's valuation exceeded RMB 10 billion at the B-round stage. SO013
CO018 36Kr Europe reported that Zhisquare completed a new financing round of nearly RMB 5 billion in June 2026. SO014, SO016, SO017
CO019 Multiple June 2026 reports placed Zhisquare's valuation above RMB 20 billion after the near-RMB 5 billion round. SO014, SO016, SO018
CO020 Chinese media described Zhisquare as the first Greater Bay Area embodied-AI unicorn to clear the RMB 20 billion threshold. SO016, SO015, SO018
CO021 Tencent News described the June 2026 syndicate as spanning national funds, regional government capital, insurers, brokers, industrial investors, and financial investors. SO016
CO022 Tencent News named national SME-system and China cultural-system funds plus Guangdong AI and Nanshan strategic-emerging funds among major backers. SO016
CO023 Public funding coverage implies Zhisquare reached a rare mix of Guangdong/Shenzhen and Beijing-oriented state capital. SO016, SO019
CO024 The August 2026 36Kr feature said Zhisquare had completed shareholding reform and was considering a Hong Kong IPO as early as 2027. SO015
CO025 The August 2026 36Kr feature said Zhisquare had completed 12 funding rounds within roughly one year. SO015, SO013
CO026 The April 2025 AlphaBot 2 launch said the company had already built a three-pronged manufacturing layout across automotive, semiconductor, and biotechnology. SO005
CO027 The AlphaBot 2 launch announced 2025 plans for airport and community-service rollouts beyond factory use. SO005
CO028 Securities Times reported that HKC subsidiary Shenzhen Huizhi IoT would deploy more than 1,000 embodied robots with Zhisquare over three years. SO024
CO029 Securities Times reported that the HKC-linked order value was close to RMB 500 million. SO024
CO030 Tencent News said Zhisquare had built semi-automated annual capacity above 2,000 units and planned larger lines. SO016
CO031 Official company materials frame Zhisquare as already in the industrialization phase rather than a lab-only robotics team. SO010, SO002
CO032 APC Reports said Zhisquare launched NeuroVLA in June 2026 and open-sourced it through AlphaBrain Platform. SO023
CO033 Gasgoo reported that Zhisquare used WAIC 2026 to showcase NeuroVLA, AlphaBrain Platform, and AlphaBot 2 to customers and investors. SO020
CO034 No reviewed public source disclosed a precise current company-wide headcount for Zhisquare. SO002, SO003, SO015
CO035 No reviewed public source disclosed a standalone board roster or investor-governance rights schedule for Zhisquare. SO002, SO015, SO016
CO036 Public sources mention historical revenue confirmation and order values but do not disclose a current audited revenue run-rate or ARR. SO008, SO024, SO015
CO037 External coverage consistently presents Zhisquare as a brain-first or robot-brain-led competitor rather than as a purely hardware vendor. SO018, SO016, SO020
CO038 The speed of financing, IPO preparation, and valuation step-up creates visible risk that capital-market expectations run ahead of fully disclosed economics. SO015, SO019, SO018, SO026
CO039 The public evidence supports classifying Zhisquare as a late-private super-unicorn rather than an early prototype-only robotics startup. SO016, SO015, SO024
CO040 The HKC-linked order gives Zhisquare a stronger named industrial anchor than most early robotics startups disclose publicly. SO024, SO015
CM001 For Zhisquare, the relevant market is not all robotics spend but deployable embodied robots and software sold into repeat industrial and service workflows. SM016, SM017, SM018
CM002 Included spend covers robot bodies, model stacks, deployment services, data capture, maintenance, and factory integration tied to embodied workflows. SM016, SM007, SM015
CM003 Excluded spend includes legacy fixed automation, pure software AI without robot embodiment, and research prototypes not sold externally. SM012, SM013, SM007
CM004 Manual labor remains the status-quo substitute in many manufacturing and service scenarios. SM007, SM014, SM015
CM005 Task-specific industrial automation is the other main substitute where variability is low and ROI is already proven. SM008, SM015, SM013
CM006 36Kr Europe said 19 robotics or embodied-intelligence companies reached unicorn status in H1 2026, showing extreme investor appetite. SM001, SM002
CM007 Embodied Global reported RMB 93.5 billion of disclosed China embodied-AI funding in H1 2026. SM022
CM008 Shenzhen's 2026 revised plan targets more than 10 embodied-robot companies valued above RMB 10 billion by 2027. SM003, SM004
CM009 The same Shenzhen plan targets more than 20 companies with revenue above RMB 1 billion and more than 50 billion-yuan-class application scenarios by 2027. SM003, SM004
CM010 Shenzhen aims for an embodied-robot industry cluster exceeding RMB 100 billion with more than 1,200 related companies by 2027. SM003, SM004
CM011 MIIT and SASAC set a 2026 goal of more than 100 high-value humanoid or embodied-AI scenarios and the capacity for 10,000-unit-scale deployment. SM007
CM012 The MIIT program explicitly pushes user units, robot makers, model vendors, and suppliers into joint deployment consortia, clarifying the adoption path for buyers. SM007
CM013 The MIIT notice encourages robot-as-a-service, leasing, and utility-style payment models to lower buyer adoption friction. SM007
CM014 The AI+Manufacturing implementation opinion ties AI deployment to manufacturing, logistics, and industrial upgrading rather than consumer novelty. SM008
CM015 The 2026 draft national humanoid-robot standards guide shows that testing, interfaces, safety, and lifecycle management are moving into a more formal regime. SM009, SM010
CM016 The presence of central legal and policy publications indicates that embodied robotics is moving from permissive experimentation toward more codified governance. SM011, SM009
CM017 TrendForce called China the world's largest humanoid-robot market in 2026. SM013
CM018 TrendForce said Chinese annual humanoid output could grow 94% in 2026 as use cases and production scale improve. SM013
CM019 CNBC reported Morgan Stanley lifted China's 2026 humanoid shipment forecast to 50,000 external-sales units from 28,000. SM012
CM020 Morgan Stanley estimated China's humanoid robot market would reach $2 billion in 2026. SM012
CM021 Morgan Stanley estimated China's humanoid robot market could reach $15 billion by 2030, with 446,000 annual units. SM012
CM022 China Daily reported that more than 60 humanoid robots were on display at WAIC 2025 and that the market was moving toward mass production and commercial deployment. SM014
CM023 Policy and industry sources consistently point to manufacturing as the earliest large-budget humanoid adoption path. SM007, SM008, SM015, SM014
CM024 Service settings like healthcare, retail, and public-service operations are adjacent markets but likely later or more fragmented than industrial deployments. SM007, SM014, SM018
CM025 The Greater Bay Area offers a dense manufacturing base, supplier network, and electronics ecosystem that reduces deployment iteration time for companies like Zhisquare. SM003, SM015, SM006
CM026 Foxconn's 2025 GTC disclosure shows a major electronics manufacturer publicly investing in humanoid, hybrid, and AI-workforce systems, validating buyer-side interest. SM015
CM027 The HKC-linked order shows semiconductor and display manufacturing can be a real embodied-robot buyer segment rather than a hypothetical one. SM020
CM028 Official Zhisquare materials position automotive, semiconductor, biotech, airport, and community settings as the company's target scenario map. SM018, SM017
CM029 The buyer is typically an enterprise operator or plant owner, the user is a line worker or supervisor, and the payer may be an automation, capex, or digital-transformation budget owner. SM007, SM015, SM020
CM030 Both policy documents and Zhisquare's own messaging emphasize that real deployment data is an adoption driver because it improves robot brains and reduces failure risk. SM007, SM027, SM019
CM031 Incomplete standards and certification still constrain adoption because buyers need testing, safety, and lifecycle rules before large-scale human-robot collaboration. SM009, SM010, SM007
CM032 The largest unresolved adoption constraint is workflow-level ROI because shipment and funding data do not automatically prove payback. SM023, SM012, SM013
CM033 Local policy support can accelerate pilots, but it may also distort true willingness-to-pay if deployments depend on subsidies or state sponsorship. SM003, SM006, SM023
CM034 Manufacturing-scale humanoid deployment remains capital-intensive because the market requires bodies, sensors, compute, testing, and service operations at once. SM015, SM013, SM023
CM035 Switching costs are moderate rather than absolute because customers can multi-source bodies, models, or integrators until standards and data moats harden. SM009, SM013, SM023
CM036 The public adoption path runs from scenario selection and environment adaptation to consortium build, training, validation, and scaled deployment. SM007
CM037 Zhisquare's ability to attract both Guangdong/Shenzhen and Beijing-oriented state capital matters because local funds usually follow geographic binding logic. SM001, SM021
CM038 The market has a clear tension between aggressive valuation marks and still-emerging commercialization reality. SM023, SM022, SM012
CM039 Public sources provide national shipment and funding estimates, but they do not isolate a clean Zhisquare-specific SAM or buyer-level budget pool. SM012, SM013, SM007
CP001 Zhisquare presents itself as a full-stack embodied-AI company combining robot bodies with an in-house robot brain. SP002, SP001, SP020
CP002 Zhisquare's public scenario map skews toward industrial and productivity use cases rather than pure consumer robotics. SP003, SP005, SP019
CP003 Independent coverage often frames Zhisquare as a brain-first humanoid competitor. SP018, SP019
CP004 Zhisquare entered the market with unusual funding depth, including a near-RMB 5B round and valuation above RMB 20B. SP004, SP016
CP005 The HKC-linked order gives Zhisquare a named industrial proof point that many private competitors do not disclose publicly. SP005, SP017
CP006 Unitree publicly lists the H2 humanoid at US$29,900 before tax and shipping. SP006
CP007 The same Unitree page advertises 31 degrees of freedom, 360 N·m leg torque, and a 2070 TOPS chip. SP006
CP008 TrendForce identifies Unitree as one of the two commercialization leaders in China. SP015, SP014
CP009 LimX is another Shenzhen-based embodied-robot competitor with a 2022 founding and rapid financing cadence. SP007, SP026
CP010 LimX says it completed a US$200M B round in February 2026 and a Pre-IPO round in July 2026. SP007
CP011 LimX says it launched LimX COSA and open-sourced FluxVLA Engine in 2026, highlighting direct competition in embodied-model stacks. SP007
CP012 UBTech is a listed Shenzhen-headquartered public robotics company with disclosed board and governance structure. SP008
CP013 UBTech markets humanoid service, education, and elderly-care solutions, signaling broader service and education exposure than Zhisquare. SP008
CP014 Boston Dynamics markets Atlas directly as an enterprise-grade industrial humanoid robot. SP011, SP010
CP015 Boston Dynamics discloses Atlas operating specs including 4-hour battery life, 50kg instant payload, 30kg sustained payload, and IP67 durability. SP011
CP016 Boston Dynamics says Atlas is already in a Hyundai customer facility for field testing on real-world tasks. SP011
CP017 Tesla's 2026 Q2 filing says it is developing and commercializing Optimus as a general-purpose autonomous humanoid robot. SP012
CP018 Tesla disclosed $43.52B of cash and short-term investments at June 2026, giving it a radically stronger balance sheet than any private humanoid startup. SP012
CP019 Tesla expects more than $25B of 2026 capex driven partly by AI, compute infrastructure, and robotics. SP012
CP020 Tesla is the most dangerous long-term competitor because it combines real-world AI data, manufacturing scale, and balance-sheet depth. SP012, SP017, SP027
CP021 Foxconn's GTC disclosure shows large manufacturers are building around AI workforce and physical-AI concepts, strengthening the bargaining power of enterprise buyers. SP013, SP022
CP022 Manufacturing is the central battleground where Zhisquare, Unitree, LimX, Boston Dynamics, and Tesla can all converge. SP005, SP013, SP011, SP012, SP015
CP023 Chinese competition is unusually intense because Shenzhen and broader China host multiple well-funded full-stack and component players at once. SP026, SP021, SP016
CP024 Unitree's transparent list price creates a procurement anchor that can pressure opaque quote-led competitors. SP006, SP014
CP025 Most other competitors do not publish comparable list pricing, leaving realized price and service margins opaque. SP011, SP008, SP001
CP026 Data and model moats are still immature because the sector is early, standards are evolving, and many workflows remain pilot-heavy. SP023, SP017, SP020
CP027 Enterprise buyers can likely multi-home across vendors or component stacks before platform lock-in hardens. SP023, SP006, SP011
CP028 Zhisquare's standout advantage is strategic-capital depth and cross-regional government alignment rather than clear public proof of superior deployed scale. SP004, SP026, SP016
CP029 Zhisquare's second public advantage is its robot-brain narrative through AlphaBrain and NeuroVLA. SP020, SP019, SP018
CP030 Unitree's strongest public advantages are price transparency, visible hardware specs, and perceived commercialization leadership. SP006, SP015
CP031 LimX differentiates with aggressive product cadence, open-source language around FluxVLA, and Pre-IPO financing momentum. SP007
CP032 UBTech differentiates through public-company governance visibility and diversified robot solution categories. SP008
CP033 Boston Dynamics differentiates through enterprise readiness, workflow integrations, and disclosed reliability-oriented specs. SP011, SP010
CP034 Tesla differentiates through manufacturing scale, capital depth, and vertically integrated AI infrastructure. SP012
CP035 Public evidence does not yet prove that Zhisquare has a durable moat on deployed scale, channel power, or standard-setting. SP001, SP005, SP023
CP036 The sector's dense unicorn field implies future consolidation or down-round pressure when repeat deployment economics become clearer. SP017, SP016
CP037 The status quo competitor is still manual labor plus task-specific automation, which means robot vendors compete not only with one another but with no-purchase decisions. SP022, SP013
CP038 Across the landscape, public proof is uneven: some players show list pricing or product specs, some show governance or filings, and many remain quote-led. SP006, SP008, SP012, SP001
CI001 Zhisquare's public business model centers on selling embodied robots plus the software and deployment stack required to operate them. SI001, SI002, SI025
CI002 The product story implies revenue from hardware, integration, support, and ongoing model-improvement services rather than one-off software subscriptions alone. SI002, SI005, SI018
CI003 Official materials repeatedly frame Zhisquare as a productivity robot company aimed at real industrial scenarios. SI001, SI005, SI025
CI004 The January 2025 Pre-A announcement said proceeds would be used for embodied-brain R&D and commercialization. SI003
CI005 The March 2025 Pre-A+ announcement tied financing directly to commercial deployment of end-to-end VLA-powered robots. SI004
CI006 36Kr reported a February 2026 B-round series above RMB 1B. SI006
CI007 The same 36Kr report placed valuation above RMB 10B in February 2026. SI006
CI008 Multiple June 2026 reports said Zhisquare closed a new round of nearly RMB 5B. SI007, SI009, SI010
CI009 Multiple June 2026 reports placed valuation above RMB 20B after that round. SI007, SI009, SI014
CI010 The August 2026 36Kr feature said valuation doubled within about four months and IPO preparation was underway. SI008, SI007, SI009, SI014
CI011 Securities Times reported a near-RMB 500M HKC-linked order for 1,000+ robots over three years. SI011
CI012 That order is the clearest public commercial proof and likely represents backlog rather than recognized revenue. SI011, SI008
CI013 No reviewed public source discloses a current recognized-revenue or ARR figure for Zhisquare. SI002, SI008, SI011
CI014 No reviewed public source discloses Zhisquare list pricing, realized ASP, or discount policy. SI002, SI025, SI015
CI015 No reviewed public source discloses Zhisquare gross margin, contribution margin, or service-margin profile. SI002, SI008, SI011
CI016 A hardware-plus-deployment business implies meaningful working-capital needs in inventory, receivables, field service, and manufacturing ramp. SI005, SI010, SI019
CI017 Tencent News and CNStock described semi-automated annual capacity above 2,000 units and plans for larger lines. SI009, SI010
CI018 CNStock said the June 2026 financing would accelerate scaled mass production. SI010
CI019 NeuroVLA and AlphaBrain platform visibility support ecosystem building, but open-source technology distribution does not itself prove monetized software revenue. SI016, SI015
CI020 Enterprise robot sales likely require long integration-led cycles because MIIT deployment logic centers on training, validation, and scenario adaptation. SI018, SI011, SI017
CI021 MIIT explicitly encouraged robot-as-a-service and leasing structures, implying future monetization may mix capex sales with usage-oriented financing. SI018
CI022 Large manufacturers like Foxconn illustrate that enterprise buyers in this category can be technically demanding and commercially powerful. SI017, SI018
CI023 Humanoid robotics remains a capex-intensive category because compute, bodies, components, and testing all scale together. SI019, SI021, SI023
CI024 Tesla reported $43.52B of cash and short-term investments in June 2026, illustrating how undercapitalized startups look against global scale entrants. SI021
CI025 Tesla expects more than $25B of 2026 capex, underscoring how expensive large-scale humanoid and AI manufacturing could become. SI021, SI022
CI026 Public sources do not disclose CAC, payback, utilization, or warranty-cost data for Zhisquare. SI002, SI008, SI011
CI027 Because private operating metrics remain opaque, Zhisquare's underwriting currently leans heavily on its ability to keep financing growth and scale-up. SI007, SI009, SI010, SI002, SI008, SI011, SI005, SI019
CI028 The syndicate breadth suggests capital adequacy is stronger than that of typical startups, at least on headline fundraising access. SI009, SI026, SI012
CI029 Public sources do not reveal how much of the 2026 financing represented primary capital versus secondary liquidity. SI007, SI009, SI008
CI030 Public sources do not disclose cash on hand, monthly burn, or runway post-financing. SI008, SI009, SI010
CI031 Public sources do not disclose debt facilities, guarantees, or project-finance obligations tied to scale-up. SI010, SI008, SI009
CI032 Reported Hong Kong IPO preparation could improve financing flexibility but also raises pressure to show cleaner economics and governance. SI008, SI024
CI033 The public valuation mark is far more visible than public revenue or margin disclosure, which is a core diligence imbalance. SI007, SI009, SI014, SI002, SI008, SI011
CI034 Revenue quality cannot be called strong yet because public proof is skewed toward order announcements rather than recognized revenue, repeat cohorts, or gross margin. SI011, SI008, SI002
CI035 Heavy funding can mask poor unit economics for longer than in software markets. SI013, SI012, SI021, SI022
CI036 The named public commercial proof is concentrated enough that customer concentration remains a financial risk. SI011, SI008
CI037 The financial verdict is that Zhisquare has exceptional access to capital and credible backlog signal, but insufficient public evidence on revenue quality, margin path, and runway to underwrite cleanly. SI009, SI026, SI012, SI007, SI014, SI002, SI008, SI011
CI038 The public investor surface confirms that Zhisquare's capital stack sits alongside national, Guangdong/Shenzhen, and district-level state-backed capital platforms, reinforcing access to strategic financing even though exact ownership remains undisclosed. SI027, SI028, SI029, SI030, SI031, SI032, SI033
CE001 AlphaBot is the company's main general-purpose robot product line. SE001, SE002, SE024
CE002 AlphaBrain is positioned as the embodied foundation-model stack that defines robot behavior. SE002, SE004, SE008
CE003 The April 2025 launch introduced GOVLA, a global and omni-body VLA model aimed at full-body control and long-horizon reasoning. SE004
CE004 The company said AI2R Brain was upgraded and renamed Alpha Brain in 2025. SE004
CE005 Launch materials describe a slow System2 for reasoning and a fast System1 for robot control inside the GOVLA architecture. SE004
CE006 The April 2025 launch said DeepSeek techniques were integrated into the VLA-model build process to improve reasoning. SE004
CE007 WAIC 2026 coverage described NeuroVLA as a brain-inspired control stack using cortex, cerebellum, and spinal-cord analogies. SE007, SE008
CE008 The same coverage said NeuroVLA aims to support active perception, fault self-recovery, and temporal memory. SE007, SE008
CE009 Zhisquare publicly associated AlphaBrain Platform with an open ecosystem for data, training, models, and evaluation. SE007, SE014
CE010 The AlphaBrain docs describe an all-in-one open-source community unifying multiple VLA, world-model, RL, and continual-learning approaches. SE011
CE011 The docs explicitly list Baseline VLA, NeuroVLA, RL-Token, World Model, and Continual Learning as capabilities. SE011
CE012 Public GitHub surfaces show that Zhisquare or related AlphaBrain assets maintain at least a minimal public developer footprint rather than zero community surface. SE012, SE013, SE011
CE013 Launch materials say AlphaBot 2 adds 360°×360° sensing, 34+ total degrees of freedom, 0-240 cm vertical work range, and 6h+ continuous work. SE004
CE014 The same launch disclosed roughly 700 mm single-arm reach and a waist-leg lift architecture. SE004
CE015 Zhisquare frames its core differentiator as early commitment to full-stack vertical integration across models, hardware, deployment, and edge inference. SE004, SE003
CE016 The company says it can deploy across multiple chip environments and has experience supporting stable on-device operation across prior intelligent terminals. SE004
CE017 Official materials map the product across automotive, semiconductor, biotech, public service, retail, airport, and community workflows. SE002, SE004, SE007
CE018 WAIC 2026 coverage highlighted PCB loading and unloading in semiconductor-display manufacturing as a benchmark industrial workflow. SE007
CE019 WAIC 2026 coverage also described beverage, coffee, ice cream, and cocktail preparation as service-workflow demonstrations. SE007
CE020 The April 2025 launch said Zhisquare had already built a three-core manufacturing layout in automotive, semiconductor, and biotechnology. SE004
CE021 The launch announced a strategic cooperation with Bloomage / Huaxi Biology for biotech manufacturing scenarios. SE004
CE022 The launch said AlphaBot had entered Geely Tech Jingneng Microelectronics semiconductor workflows for loading and transfer tasks. SE004
CE023 The 2024 WRC post explicitly positioned Zhisquare around real-scene delivery rather than lab-only demos. SE006
CE024 The PKU joint-lab announcement shows an explicit research-to-product loop around 4D world models and end-to-end agents. SE005
CE025 Zhisquare is trying to run both an open ecosystem narrative and a proprietary product narrative at the same time. SE011, SE007, SE004
CE026 Foxconn and MIIT evidence suggests buyers want workflow-integrated robots rather than flashy general demos, which raises the bar for productization. SE021, SE016, SE022
CE027 Shenzhen's plan emphasizes testing, evaluation, open data, and industrial platforms that should help product iteration in the local cluster. SE020
CE028 MIIT's standards work shows that interfaces, safety, and lifecycle management are becoming formal product requirements. SE018, SE019
CE029 MIIT's 2026 deployment action expects collision detection, force limits, emergency stop, and black-box capabilities in real-scene deployments. SE016, SE017
CE030 Zhisquare's reviewed public materials do not fully disclose whether those expected controls are implemented and certified in current products. SE004, SE007, SE016
CE031 No reviewed source disclosed a complete certification or formal safety-case package for AlphaBot or NeuroVLA. SE004, SE024, SE019
CE032 No reviewed public source disclosed uptime, MTBF, field-failure, or incident-rate statistics. SE007, SE004, SE025
CE033 The public surface does not explain deployment support SLAs, spare-parts handling, or operator training depth. SE002, SE007, SE024
CE034 The model-centric product story is strong, but independent public verification of field-performance deltas remains limited. SE026, SE007, SE008
CE035 The public roadmap includes AlphaBot 2, NeuroVLA, AlphaBrain Platform, expanded manufacturing scenarios, and airport/community rollouts. SE004, SE007, SE005
CE036 The public developer signal exists, but it is much thinner than mature open-source or platform companies and should not be over-read. SE012, SE013, SE026
CE037 The product is best understood as a full-stack industrial embodied-AI system with unusually rich public architecture language but incomplete public proof on reliability, safety certification, and support maturity. SE004, SE002, SE007, SE016
CE038 Additional 2026 conference and regional-government coverage indicates Zhisquare's product story was being amplified beyond its own website, although those mentions still do not substitute for reliability proof. SE027, SE028
CU001 Zhisquare's customer base is centered on industrial buyers rather than households. SU001, SU003, SU007
CU002 The most credible public customer geography is China, especially Greater Bay Area manufacturing. SU002, SU007, SU013
CU003 Official and media materials point to semiconductor/display, automotive, biotech, public service, and retail as key customer verticals. SU001, SU003, SU005
CU004 The buyer is usually an enterprise operator, the user is a line or service worker, and the payer is an automation or operations budget owner. SU012, SU015, SU007
CU005 The clearest named customer proof is the Shenzhen Huizhi IoT / HKC-linked deployment into semiconductor display manufacturing. SU007, SU008, SU009
CU006 Public reporting said the HKC-linked program covers more than 1,000 robots across three years. SU007, SU008
CU007 The same reports put the order value near RMB 500M. SU007, SU008
CU008 The HKC deployment covers warehousing, loading/unloading, component assembly, quality testing, OLED vacuum lamination, and consumables management. SU007
CU009 The HKC cooperation also includes a joint technical team and factory-data co-development of new industrial VLA models. SU007
CU010 The AlphaBot 2 launch named Bloomage / Huaxi Biology as a strategic cooperation for biotech manufacturing workflows. SU003
CU011 The announced Bloomage scenarios include material transfer, unpacking and sterilization, visual inspection, and intelligent feeding. SU003
CU012 In biotech, Zhisquare says robots can reduce contamination risk and adapt faster to process changes. SU003
CU013 The launch materials say AlphaBot had already entered automotive manufacturing and won an order from a top international automaker. SU003
CU014 The launch materials also named Jingneng Microelectronics under Geely Tech for semiconductor material-handling tasks. SU003
CU015 The April 2025 launch said airport deployments were planned for one-line-city airports in Q3 2025. SU003
CU016 The same launch said community deployments were planned for demo neighborhoods in Q4 2025. SU003
CU017 WAIC 2026 coverage said Zhisquare's retail robot service had reached regular operation in more than ten provinces and cities. SU005
CU018 That coverage said robot retail operators were independently producing hundreds of drinks per day with zero-error operation claims. SU005
CU019 Sina's WAIC coverage highlighted PCB loading/unloading as a real industrial proof case rather than a generic demo. SU006, SU005
CU020 Customer adoption appears to move from showcase visibility to pilot/validation to production workflow replication. SU004, SU012, SU005
CU021 Public sources do not disclose the current count of active paying customer sites. SU001, SU017, SU007
CU022 Public evidence supports at least some multi-customer or repeat demand in semiconductor/display, but the breadth of repeat purchase is not quantified. SU005, SU007
CU023 No reviewed source discloses NRR, GRR, churn, or contract renewal rates. SU001, SU017, SU007
CU024 The clearest public contract duration is the three-year HKC-linked program. SU007
CU025 No reviewed public source provides customer-satisfaction or NPS evidence. SU001, SU005, SU017
CU026 Customer concentration risk is high because the largest named anchor is disproportionately more detailed than the rest of the public proof set. SU007, SU017, SU016
CU027 Industrial customers may offer land-and-expand potential because one successful workflow can spread across multiple lines and plants. SU007, SU012, SU014
CU028 Procurement friction remains real because industrial deployments require validation, site adaptation, and reliability proof. SU012, SU015, SU016
CU029 Conference and media visibility do not automatically prove production deployment. SU019, SU020, SU004, SU016
CU030 Cross-regional policy and investor support likely helps Zhisquare win pilot access and procurement attention. SU018, SU024, SU025
CU031 Despite service and retail adjacencies, the strongest public buyer mix remains manufacturing-first. SU001, SU007, SU006, SU005
CU032 The public proof set includes named customers, scenario details, and one major order, but it is still mixed across demos, plans, and operating deployments. SU007, SU008, SU003, SU005
CU033 Public evidence does not quantify how many named logos progressed from pilot to fleet expansion. SU005, SU017, SU007
CU034 The public record does not clearly separate direct sales from channel, integrator, or government-mediated demand. SU001, SU012, SU016
CU035 Zhisquare has stronger named adoption proof than many robotics startups, but customer durability and concentration remain materially under-disclosed. SU007, SU008, SU001, SU017, SU016
CR001 Zhisquare's capital stack and market narrative are unusually dependent on policy-aligned funds and government-backed industrial momentum. SR006, SR010, SR021
CR002 Multiple adverse sources frame the 2026 humanoid financing boom as potentially bubble-like or headed for consolidation. SR011, SR012, SR014
CR003 Zhisquare carries a super-unicorn valuation despite limited public revenue, margin, and customer-diversification disclosure. SR009, SR008, SR004
CR004 National standards, interfaces, and safety rules are still evolving, which creates regulatory and deployment risk. SR017, SR019, SR020
CR005 MIIT explicitly expects collision detection, force control limits, emergency stop, and black-box capabilities in deployments. SR015, SR016
CR006 Zhisquare has not publicly disclosed a full safety-case or certification package matching those expectations. SR003, SR002, SR015
CR007 No reviewed public source surfaced a product recall history, but neither did the public record offer an incident database or safety audit trail. SR002, SR003, SR019
CR008 The company has already had to publish a formal anti-impersonation statement, showing that legal/brand-protection issues are not purely hypothetical. SR001
CR009 No reviewed public source disclosed material litigation or enforcement against Zhisquare, but the court and regulator surface was not exhaustively proved clean. SR001, SR008, SR020
CR010 The HKC-linked program is so much more detailed than other public customer proofs that concentration risk is presumptively high. SR004, SR008, SR011
CR011 Backlog-to-revenue conversion risk is material because named order size is public but recognized revenue and delivery cadence are not. SR004, SR005, SR008
CR012 Scale-up risk is high because Zhisquare is trying to move from promising deployments to mass production quickly. SR003, SR005, SR006
CR013 Public sources do not disclose the full supplier and subsystem concentration behind AlphaBot production. SR003, SR002, SR005
CR014 Public sources do not disclose field-service staffing, spare-parts coverage, or SLA maturity. SR002, SR003, SR004
CR015 The public company identity is unusually founder-centric, creating key-person and succession risk. SR002, SR008
CR016 The broader executive bench and board rights remain under-disclosed, making execution depth hard to assess. SR002, SR008, SR006
CR017 Public architecture language is richer than public reliability or benchmark disclosure, creating execution risk that the technology story outruns operational proof. SR003, SR002, SR011
CR018 Robots deployed in variable physical settings face generalization and exception-handling risk even when demos look strong. SR015, SR003, SR011
CR019 Open-source or competing embodied-model stacks may erode differentiation faster than current capital markets assume. SR027, SR030, SR011
CR020 Tesla's balance sheet and capex intensity show how vulnerable startup economics could be once fully scaled entrants commit seriously. SR024, SR025
CR021 Large industrial buyers and integrators can force reliability, pricing, and support burdens onto robot vendors. SR023, SR015, SR026
CR022 The public record does not disclose customer data-rights terms or how cross-customer learning is governed. SR004, SR003, SR020
CR023 Cross-regional state support may help in China but can raise geopolitical scrutiny or limit some international commercial options. SR010, SR006, SR024
CR024 A capital-intensive robot model with opaque unit economics remains dependent on continued financing confidence. SR011, SR012, SR014, SR009, SR008, SR004
CR025 Reported Hong Kong IPO planning raises pressure to professionalize governance, disclosure, and metrics quickly. SR008, SR006
CR026 District and city-level policy support can create implicit location or deployment obligations that reduce strategic flexibility. SR021, SR022
CR027 The regulatory stack spans standards, testing, deployment approvals, labor safety, data governance, and possible local policy conditions. SR015, SR017, SR018, SR020
CR028 Even competitor-surface opacity is a risk signal: a crowded sector with uneven disclosure makes benchmarking difficult and can hide sudden repricing. SR027, SR028, SR029, SR030
CR029 Expansion into service and retail scenarios can dilute focus from the higher-value industrial core if not controlled. SR003, SR004, SR011
CR030 As robots become more capable and policy-relevant, export controls and foreign procurement scrutiny may intensify. SR024, SR020, SR017
CR031 No reviewed source disclosed MTBF, failure rate, incident rate, or warranty claims. SR003, SR004, SR019
CR032 No reviewed public source disclosed current burn, cash runway, or debt structure post-super-unicorn round. SR008, SR006, SR005
CR033 A single blockbuster order can create a false sense of market readiness if later conversions disappoint. SR004, SR011, SR012
CR034 Large rounds can temporarily mask poor gross margins or heavy service burden. SR011, SR012, SR024
CR035 Some risks are partly mitigated by state support, local manufacturing ecosystem density, and scenario-specific learning loops. SR006, SR021, SR004
CR036 Even after those mitigants, residual execution and disclosure risk remains high because the company is scaling faster than its public evidence base. SR006, SR021, SR004, SR009, SR008
CR037 A failure to convert the headline backlog into repeat multi-site revenue by 2027 would be a thesis-breaking signal. SR004, SR026, SR011
CR038 A serious safety incident, recall, or failed deployment in a flagship industrial site would be a thesis-breaking signal. SR015, SR016, SR003, SR002, SR019, SR004
CR039 A down-round, pulled IPO, or visible financing difficulty after the 2026 bubble wave would challenge the valuation premise sharply. SR008, SR011, SR012, SR024
CR040 Zhisquare's top risks cluster around policy dependence, safety/compliance disclosure, concentration, scale-up execution, and valuation ahead of fundamentals. SR006, SR021, SR004, SR009, SR008, SR015
CV001 Public reports place Zhisquare above RMB 20B valuation after the June 2026 financing. SV001, SV002, SV004
CV002 The same public reports place the round size near RMB 5B. SV001, SV002, SV005
CV003 A February 2026 36Kr report had already put Zhisquare above RMB 10B after its B-round. SV003, SV002
CV004 The August 2026 36Kr feature said Zhisquare completed shareholding reform and could target a Hong Kong IPO as early as 2027. SV003
CV005 Zhisquare has unusual capital access for a young robotics company. SV004, SV007, SV010
CV006 The HKC-linked order is the strongest public commercial proof behind the valuation. SV006, SV005
CV007 China's humanoid market has strong policy and shipment tailwinds in 2026. SV011, SV012, SV028, SV029
CV008 Zhisquare differentiates through a robot-brain-first narrative rather than pure hardware commoditization. SV030, SV031, SV010
CV009 Public revenue, margin, burn, and active customer metrics remain too thin for clean valuation support. SV030, SV003, SV006
CV010 The sector is widely described as overheated or vulnerable to correction. SV008, SV009, SV024, SV025
CV011 Part of Zhisquare's valuation premium likely reflects policy-champion status and investor signaling value. SV004, SV001, SV010
CV012 Unitree's public H2 pricing and IPO momentum provide a transparency anchor the market can use against more opaque peers. SV013, SV014
CV013 LimX offers a fast-moving China-local comparable with product and financing momentum. SV015, SV021
CV014 UBTech provides a governance-visible listed robotics comparable, even if its business mix is broader than Zhisquare's. SV016
CV015 Boston Dynamics provides an enterprise-readiness comparable for industrial humanoid deployment. SV017, SV018
CV016 Tesla is the long-horizon global scale comparable because it combines real-world AI, manufacturing, and deep capital. SV019, SV020, SV023
CV017 Tesla reported $43.52B of cash and short-term investments in June 2026. SV019
CV018 Tesla expects over $25B of 2026 capex, highlighting the capital appetite of scaled AI-robotics strategies. SV019
CV019 CNBC's January 2026 humanoid newsletter said business sales were expected to become the key driver of Chinese humanoid demand in 2026. SV021
CV020 CNBC's January 2026 newsletter said LimX's base Oli model cost about RMB 158,000 and the developer version about RMB 290,000. SV021
CV021 Zhisquare does not offer similar price transparency publicly. SV030, SV031
CV022 At the current mark, the evidence supports a track / research-more stance rather than an enthusiastic buy. SV001, SV030, SV003, SV006, SV008, SV009
CV023 Confidence should be medium because the story has real proof items but still major disclosure gaps. SV006, SV030, SV003, SV008, SV009
CV024 Risk rating should be high because valuation, concentration, and scale-up execution all matter simultaneously. SV008, SV009, SV006, SV030, SV003
CV025 The valuation stance is rich / aggressive relative to public fundamentals. SV001, SV030, SV003, SV008, SV009
CV026 The strongest bull argument is the combination of policy support, GBA manufacturing density, and robot-brain differentiation. SV004, SV007, SV010, SV011, SV012, SV028, SV029, SV030, SV031
CV027 The strongest anti-thesis is that the valuation has outrun public evidence on revenue quality and scalable unit economics. SV030, SV003, SV006, SV008, SV009
CV028 The base case is continued traction and funding support, but with valuation discipline until repeat industrial deployments and cleaner metrics emerge. SV006, SV030, SV003
CV029 The bull case requires repeat multi-site industrial wins, broader customer diversification, and credible IPO readiness. SV006, SV003, SV011, SV012, SV028
CV030 The bear case is that one strong order and strong capital signaling fail to convert into diversified revenue before the market cools. SV006, SV008, SV009, SV030, SV003
CV031 Any investor entry discipline should be milestone-based rather than purely momentum-based. SV008, SV009, SV003, SV011, SV012, SV028
CV032 A reasonable threshold for paying up would be proof of repeat deployments beyond the first anchor plus disclosed revenue and margin evidence. SV006, SV030, SV003
CV033 Public sources do not disclose current preference stack, dilution overhang, or liquidation rights. SV003, SV004
CV034 IPO preparation improves exit plausibility but does not prove offering readiness. SV003, SV016
CV035 Public sources do not show how much of recent financing was primary versus secondary. SV001, SV004, SV003
CV036 A delayed IPO, down-round, or weak backlog conversion would likely compress private marks quickly. SV008, SV009, SV003, SV006
CV037 Every current comparable is imperfect because each emphasizes a different mix of transparency, scale, or product maturity. SV013, SV015, SV016, SV017, SV019
CV038 The recommendation is highly price-sensitive because Zhisquare may still become strategically important even if today's mark is too rich. SV004, SV007, SV011, SV012, SV030, SV003
CV039 The realistic exit window is 2027-2028 through Hong Kong IPO or later private-market liquidity if milestones are met. SV003, SV016, SV008, SV009
CV040 The overall valuation verdict is track / research-more at current price, with upside optionality if operating proof catches up to narrative and capital. SV001, SV030, SV003, SV006, SV008, SV009
来源
编号出版方标题引文
SO001 AI² Robotics AI² Robotics | A Real-World AGI Company
SO002 AI² Robotics 关于我们 - 智平方科技
SO003 AI² Robotics About Us - AI² Robotics
SO004 AI² Robotics 智平方公司关于企业名称被仿冒的严正声明 - 智平方科技
SO005 AI² Robotics 智平方发布全新一代智能机器人AlphaBot 2,开启AGI终端新时代! - 智平方科技
SO006 AI² Robotics 强强联合!“北大—智平方具身智能联合实验室”正式揭牌成立 - 智平方科技
SO007 AI² Robotics 2025第一“融” ▏具身智能赛道喜迎“开门红” 智平方宣布完成Pre-A轮融资 - 智平方科技
SO008 AI² Robotics 智平方 ▏完成数亿元Pre A+轮融资, 推动搭载世界领先水平端到端VLA模型具身机器人商业落地 - 智平方科技
SO009 AI² Robotics 首秀WRC | 智平方开启具身AGI在真实场景交付新篇章 - 智平方科技
SO010 AI² Robotics 智平方荣登多项重磅榜单 引领具身智能“产业化时代” - 智平方科技
SO011 Baidu Baike 智平方(深圳)科技有限公司
SO012 36Kr 具身智能企业“智平方”完成新一轮A系列融资-36氪
SO013 36Kr 智平方完成B轮系列超10亿元人民币融资,公司估值正式超过百亿-36氪
SO014 36Kr Europe Zhihui Fangzhou has completed a new round of financing of nearly 5 billion yuan, with a valuation exceeding 20 billion yuan-36氪
SO015 36Kr 四个月估值翻倍,200亿“独角兽”智平方拟赴港IPO-36氪
SO016 Tencent News 智平方完成50亿融资,估值超200亿_腾讯新闻
SO017 CNStock 智平方完成近50亿元融资 将加速推进规模化量产-上海证券报·中国证券网
SO018 ChinaBizInsider China Robot AI Startup X Square Robot Hits $2.8B Valuation
SO019 Embodied Global China Embodied AI Funding Hits ¥93.5B in H1 2026, 22 Unicorns Emerge
SO020 Gasgoo WAIC 2026|智平方展示全球首个原创类脑大模型,让机器人更加聪明能干,开启通用机器人生产力时代
SO021 Sina Finance 智平方郭彦东:机器人大脑不应只是算力和数据竞赛,当前产业需跨越应用鸿沟
SO022 Sina Finance 智平方创始人郭彦东:世界模型与VLA天然统一 类脑架构将成下一代机器人大脑重要演进方向
SO023 中国报道 智源大会传来重磅成果 智平方发布全球首个类脑式具身智能系统NeuroVLA
SO024 Securities Times 智平方签下近5亿元人形机器人大单 机器人将大规模进入半导体显示行业
SO025 World Robot Conference 智平方(深圳)科技股份有限公司-展商信息-2026世界机器人大会
SO026 ChinaBizInsider 15 Embodied AI Unicorns in 6 Months: China's Robot Race Hits a Reality Check
SM001 36Kr Europe Capital landscape of 19 H1 2026 robotics unicorns
SM002 KuCoin News 19 new robot startups reach unicorn status in 2026
SM003 Shenzhen Science Innovation Bureau Shenzhen embodied-intelligence robot action plan (2025-2027, revised 2026)
SM004 Shenzhen Municipal Government Shenzhen Government Gazette 2026 issue with embodied-intelligence policy
SM005 Shenzhen DRC Shenzhen 15th Five-Year development planning consultation materials
SM006 Guangdong Government Guangdong investment and industrial policy support update
SM007 MIIT 2026 humanoid robot and embodied-intelligence real-scene training action notice
SM008 MIIT AI+Manufacturing action implementation opinions
SM009 MIIT National humanoid robot industry standards guide draft 2026
SM010 SAMR National standards portal
SM011 NPC National-level legal and policy publication relevant to embodied robotics governance
SM012 CNBC Morgan Stanley doubles China humanoid robot shipment forecast as commercialization accelerates
SM013 TrendForce China humanoid robot commercialization accelerates in H2 2026
SM014 China Daily Humanoid robots move closer to mass production and deployment in China
SM015 Foxconn Foxconn unpacks AI and humanoid robotics progress at NVIDIA GTC
SM016 AI² Robotics AI² Robotics home page
SM017 AI² Robotics AI² Robotics about page
SM018 AI² Robotics AlphaBot 2 launch post
SM019 Gasgoo WAIC 2026 coverage of Zhisquare and NeuroVLA
SM020 Securities Times Zhisquare signs near-RMB 500M humanoid robot order for semiconductor display sector
SM021 36Kr Europe Zhisquare nearly RMB 5B round and valuation over RMB 20B
SM022 Embodied Global China embodied AI funding hits RMB 93.5B in H1 2026
SM023 ChinaBizInsider 15 embodied AI unicorns in 6 months: China's robot race hits a reality check
SM024 ChinaBizInsider Zhisquare hits RMB 20B valuation on brain-first bet
SM025 Sina Finance Guo Yandong on crossing the application gap
SM026 Sina Finance Guo Yandong on world models and VLA
SM027 APC Reports NeuroVLA launch coverage
SP001 AI² Robotics AI² Robotics about page
SP002 AI² Robotics AI² Robotics home page
SP003 AI² Robotics AlphaBot 2 launch post
SP004 36Kr Europe Zhisquare near-RMB 5B round and valuation over RMB 20B
SP005 Securities Times Zhisquare HKC-linked semiconductor deployment order
SP006 Unitree Unitree H2 product page
SP007 LimX Dynamics LimX Dynamics home page
SP008 UBTech UBTech investor relations and governance page
SP009 Fourier Fourier Robotics home page
SP010 Boston Dynamics Boston Dynamics home page
SP011 Boston Dynamics Atlas enterprise humanoid page
SP012 SEC Tesla 2026 Q2 filing mentioning Optimus
SP013 Foxconn Foxconn GTC humanoid and AI workforce release
SP014 CNBC Morgan Stanley doubles China humanoid forecast
SP015 TrendForce China humanoid commercialization accelerates
SP016 Embodied Global China embodied AI funding hits RMB 93.5B in H1 2026
SP017 ChinaBizInsider 15 embodied AI unicorns in 6 months: China's robot race hits a reality check
SP018 ChinaBizInsider Zhisquare hits RMB 20B valuation on brain-first bet
SP019 Gasgoo WAIC 2026 Zhisquare coverage
SP020 APC Reports NeuroVLA launch coverage
SP021 Shenzhen Science Innovation Bureau Shenzhen embodied-robot action plan
SP022 MIIT 2026 real-scene humanoid deployment action
SP023 MIIT Humanoid robot standards guide draft 2026
SP024 SAMR National standards portal
SP025 China Daily Humanoid robots move closer to mass production
SP026 36Kr Europe Capital landscape of 19 H1 2026 robotics unicorns
SP027 SEC Tesla current report filing dated 2026-06-17
SI001 AI² Robotics AI² Robotics home page
SI002 AI² Robotics AI² Robotics about page
SI003 AI² Robotics Pre-A financing announcement
SI004 AI² Robotics Pre-A+ financing announcement
SI005 AI² Robotics AlphaBot 2 launch post
SI006 36Kr B-round above RMB 1B and valuation above RMB 10B
SI007 36Kr Europe Near-RMB 5B round and valuation over RMB 20B
SI008 36Kr Zhisquare considering Hong Kong IPO after valuation doubles
SI009 Tencent News Zhisquare completes near-RMB 5B financing; valuation above RMB 20B
SI010 CNStock Zhisquare completes nearly RMB 5B financing to accelerate mass production
SI011 Securities Times Near-RMB 500M HKC-linked humanoid order
SI012 Embodied Global China embodied AI funding hits RMB 93.5B in H1 2026
SI013 ChinaBizInsider Robot race hits a reality check
SI014 ChinaBizInsider Zhisquare hits RMB 20B valuation on brain-first bet
SI015 Gasgoo WAIC 2026 Zhisquare coverage
SI016 APC Reports NeuroVLA launch coverage
SI017 Foxconn Foxconn GTC humanoid and AI workforce release
SI018 MIIT 2026 humanoid deployment action
SI019 TrendForce China humanoid commercialization accelerates
SI020 CNBC Morgan Stanley doubles China humanoid forecast
SI021 SEC Tesla 2026 Q2 filing mentioning Optimus and capex
SI022 SEC Tesla 2026 current report filing
SI023 Foxconn IR Foxconn annual reports page
SI024 UBTech UBTech investor relations page
SI025 World Robot Conference WRC 2026 Zhisquare exhibitor page
SI026 36Kr Europe Capital landscape of 19 H1 2026 robotics unicorns
SI027 National SME Development Fund National SME Development Fund home page
SI028 National SME Development Fund National SME Development Fund company profile
SI029 SDIC Chuangyi SDIC Chuangyi home page
SI030 SDIC Chuangyi SDIC Chuangyi company profile
SI031 Shenzhen Capital Group Shenzhen Capital Group home page
SI032 Nanshan District Government Nanshan District strategic emerging industry update
SI033 Nanshan District Government Nanshan District innovation industry update
SE001 AI² Robotics AI² Robotics home page
SE002 AI² Robotics AI² Robotics about page
SE003 AI² Robotics 关于我们 - 智平方科技
SE004 AI² Robotics AlphaBot 2 launch post
SE005 AI² Robotics Peking University and Zhisquare embodied-intelligence lab announcement
SE006 AI² Robotics WRC 2024 debut post
SE007 Gasgoo WAIC 2026 Zhisquare coverage
SE008 APC Reports NeuroVLA launch coverage
SE009 Sina Finance Guo Yandong on crossing the application gap
SE010 Sina Finance Guo Yandong on world models and VLA
SE011 AlphaBrain AlphaBrain documentation
SE012 GitHub AlphaBrain repository
SE013 GitHub AI2-Robotics organization page
SE014 Baidu Baike AlphaBrain Platform entry
SE015 HKC HKC official article / customer context page
SE016 MIIT 2026 humanoid deployment action
SE017 MIIT Mobile MIIT humanoid deployment action page
SE018 MIIT Humanoid robot standards guide draft 2026
SE019 SAMR National standards portal
SE020 Shenzhen Science Innovation Bureau Shenzhen embodied-robot action plan
SE021 Foxconn Foxconn GTC humanoid and AI workforce release
SE022 China Daily Humanoid robots move closer to mass production
SE023 TrendForce China humanoid commercialization accelerates
SE024 World Robot Conference WRC 2026 Zhisquare exhibitor page
SE025 Securities Times Near-RMB 500M HKC-linked order
SE026 ChinaBizInsider Robot race hits a reality check
SE027 CENA CENA robotics coverage referencing Zhisquare at WAIC 2026
SE028 Guizhou Government 2026 digital-economy or robotics coverage mentioning Zhisquare
SU001 AI² Robotics AI² Robotics about page
SU002 AI² Robotics 关于我们 - 智平方科技
SU003 AI² Robotics AlphaBot 2 launch post
SU004 AI² Robotics WRC 2024 debut post
SU005 Gasgoo WAIC 2026 Zhisquare coverage
SU006 Sina Tech WAIC 2026 Zhisquare on factory and commerce deployment
SU007 Securities Times Near-RMB 500M HKC-linked humanoid order
SU008 CNStock Shanghai Securities Journal follow-up on Zhisquare and Huizhi IoT cooperation
SU009 HKC HKC official article / customer context page
SU010 APC Reports NeuroVLA launch coverage
SU011 China Daily Humanoid robots move closer to mass production
SU012 MIIT 2026 humanoid deployment action
SU013 Shenzhen Science Innovation Bureau Shenzhen embodied-robot action plan
SU014 TrendForce China humanoid commercialization accelerates
SU015 Foxconn Foxconn GTC humanoid and AI workforce release
SU016 ChinaBizInsider Robot race hits a reality check
SU017 36Kr Zhisquare IPO-prep feature
SU018 36Kr Europe Capital landscape of 19 H1 2026 robotics unicorns
SU019 CENA CENA robotics conference coverage
SU020 Guizhou Government Regional government technology coverage mentioning Zhisquare
SU021 China Development Network China Development embodied-intelligence industry article Jan 29 2026
SU022 China Development Network China Development embodied-intelligence industry article Jan 22 2026
SU023 RobotScope Zhisquare valuation and market article
SU024 National SME Development Fund Fund investment case page
SU025 SDIC Chuangyi Investment case page
SU026 Unitree Unitree home page
SR001 AI² Robotics Anti-impersonation statement
SR002 AI² Robotics AI² Robotics about page
SR003 AI² Robotics AlphaBot 2 launch post
SR004 Securities Times HKC-linked order coverage
SR005 CNStock Zhisquare nearly RMB 5B financing to accelerate mass production
SR006 Tencent News Zhisquare completes near-RMB 5B financing; valuation above RMB 20B
SR007 36Kr Chinese 36Kr flash on near-RMB 5B round
SR008 36Kr IPO-prep feature
SR009 36Kr Europe Near-RMB 5B round and valuation over RMB 20B
SR010 36Kr Europe Capital landscape of 19 H1 2026 robotics unicorns
SR011 ChinaBizInsider 15 embodied AI unicorns in 6 months: reality check
SR012 ChosunBiz China humanoid boom mints high-valuation startups, sparks bubble warnings
SR013 RobotScope Zhisquare 50bn / super-unicorn pulse
SR014 Embodied Global China embodied AI funding hits RMB 93.5B in H1 2026
SR015 MIIT 2026 humanoid deployment action
SR016 MIIT Mobile MIIT deployment action page
SR017 MIIT Humanoid robot standards guide draft 2026
SR018 MIIT CQCA 2026 MIIT / CQCA regulatory notice
SR019 SAMR National standards portal
SR020 NPC National policy/legal publication relevant to robotics governance
SR021 Shenzhen Science Innovation Bureau Shenzhen embodied-robot action plan
SR022 Shenzhen Government Gazette Shenzhen embodied-intelligence plan gazette text
SR023 Foxconn Foxconn GTC humanoid and AI workforce release
SR024 SEC Tesla 2026 Q2 filing mentioning Optimus
SR025 Tesla Tesla AI page (access-restricted public surface)
SR026 TrendForce China humanoid commercialization accelerates
SR027 AgiBot AgiBot public website
SR028 AgiBot AgiBot products page
SR029 Astribot Astribot public website
SR030 LimX Dynamics LimX products page
SV001 36Kr Europe Near-RMB 5B round and valuation over RMB 20B
SV002 36Kr Chinese 36Kr flash on near-RMB 5B round
SV003 36Kr IPO-prep feature
SV004 Tencent News Zhisquare completes near-RMB 5B financing; valuation above RMB 20B
SV005 CNStock Zhisquare nearly RMB 5B financing to accelerate mass production
SV006 Securities Times HKC-linked order coverage
SV007 Embodied Global China embodied AI funding hits RMB 93.5B in H1 2026
SV008 ChinaBizInsider 15 embodied AI unicorns in 6 months: reality check
SV009 ChosunBiz China humanoid boom sparks bubble warnings
SV010 RobotScope Zhisquare super-unicorn pulse
SV011 CNBC Morgan Stanley doubles China humanoid forecast
SV012 TrendForce China humanoid commercialization accelerates
SV013 Unitree Unitree H2 product page
SV014 The Robot Report Unitree IPO resets China robot valuations in 2026
SV015 LimX Dynamics LimX home page
SV016 UBTech UBTech investor relations page
SV017 Boston Dynamics Atlas enterprise humanoid page
SV018 TechCrunch Boston Dynamics Atlas pilot launches
SV019 SEC Tesla 2026 Q2 filing mentioning Optimus
SV020 Tesla Optimus page
SV021 CNBC China Connection newsletter on Chinese humanoids and Optimus
SV022 Axios Musk says Optimus public sales not until end of 2027
SV023 Reuters Tesla to begin commercial trials for Optimus in 2026
SV024 RobotToday Bloomberg warns of bubble risk as China humanoid boom overheats
SV025 TechBuzz China warns of humanoid robot bubble amid investment frenzy
SV026 Foxconn IR Foxconn annual reports page
SV027 Foxconn Foxconn GTC humanoid and AI workforce release
SV028 MIIT 2026 humanoid deployment action
SV029 Shenzhen Science Innovation Bureau Shenzhen embodied-robot action plan
SV030 AI² Robotics AI² Robotics about page
SV031 AI² Robotics AlphaBot 2 launch post