Haomo Zhixing Technology Co., Ltd.
曾有真实商业牵引力的中国自动驾驶独角兽,但 2025-2026 年价值严重受损
Haomo 仍有战略和技术残值,但截至 2026 年 8 月的公开证据支持回避:运营、偿付能力和重启经济性都没解开,历史独角兽估值也明显过时。
封面要素
公司概况
Haomo Zhixing Technology 2019 年从 Great Wall Motor 分拆出来,想把乘用车辅助驾驶系统和末端自动配送车商业化。公司曾用 HPilot 在 20 多款车型上的部署、北京零售生态里的配送路线运营,以及 2021 年独角兽融资,拼出过有分量的历史验证;2024 年又拿到 B1 和 B2 轮后续资金。但到 2026 年 8 月尽调时,公开记录指向的已不是一家正在放量的成长公司,而是一个承压的独立资产:2025 年末报道提到运营停摆和欠薪,Great Wall 仍是控制性生态,公开材料里也没有一条清晰链路能把过往牵引力接到当前偿付能力或重启状态。
- 成立时间
- 2019-11-01
- 创始人
- Great Wall Motor intelligent-driving unit, Gu Weihao
- 创立地点
- Beijing, China
- 总部
- Beijing, China
- 产品
- 面向乘用车的 HPilot 及相关 ADAS 软件、HPark / 城市导航功能、Little Camel / Xiaomotuo 自动物流车、HUANMO 自动驾驶平台,以及面向智能驾驶工作流的 DriveGPT。
- 客户
- 优先服务 Great Wall 乘用车,同时围绕 Wumart、Dmall、Meituan、Dada 做物流 / 零售部署,并有限尝试向外部车企扩展。
- 商业模式
- 面向 OEM 辅助驾驶的软件和系统集成,加上自动配送车部署与平台商业化。
- 阶段
- Distressed late-stage private
- 融资情况
- 2021 年曾为独角兽;公开数据库可支持的披露融资约 $245.3M;2024 年宣布新增超过 CNY400M 的 B1/B2 轮融资;当前救助融资状态未披露。
执行摘要
主要优势
- Haomo 曾跑出真实商业化证据,包括 ADAS 覆盖 20 多款车型,以及在北京形成有意义的配送路线规模。
- 公司曾吸引强势战略和财务投资人,包括 Great Wall、Meituan、Qualcomm Ventures 及其他后期投资者。
- 技术栈横跨乘用车 ADAS 和低速自动配送,即便陷入困境,仍留下 IP 残值和赞助方期权。
- 中国辅助驾驶和自动配送系统的终端需求仍真实存在,因此在救助情景下,Haomo 资产并不自动归零。
主要风险
- 2025 年末公开报道提到运营停摆、欠薪和可能的账户冻结压力,估值问题已经变成偿付能力问题。
- 已审阅记录中没有公开的 Haomo 2025–2026 年财务报表、现金余额、债务计划或当前估值标记。
- Great Wall 集中度仍极高:它同时是控制生态、关键历史客户,也是最可能的救助路径。
- 停摆之前,Great Wall 以外的多元化信号就已偏弱;Hyundai 证据在恶化,当前客户延续性也未解决。
- 中国 2025–2026 年 OTA、数据和自动驾驶安全规则,会抬高任何重启的成本和复杂度。
未决问题
- 当前现金、债务、应付款、欠薪和或有负债
- 哪些 OEM 和配送项目仍在运行且获得支持的签署证明
- 当前完全稀释股权结构、清算优先权,以及现有股东在救助方案中的经济性
- Great Wall 或其他赞助方是否提供了有约束力的过桥融资或资产收购支持
- Haomo 能在更新后的 OTA、事故和汽车数据合规制度下重启的证据
目录
01公司概览
1.1 身份、起点,以及 Haomo 原本要卖什么
Haomo 应被看作一家 Great Wall Motor 分拆公司:它试图用一套自动驾驶栈,同时服务两个相邻却运营逻辑不同的市场——乘用车辅助驾驶和低速物流自动化。公司发布的信息与独立画像在创立基线上基本一致。Great Wall 分离智能驾驶部门后,Haomo 于 2019 年 11 月成为独立公司,公开资料显示公司位于北京。它的产品野心不止一个 ADAS 模块。公司多次描述一套全栈方案:围绕 HPilot 和 NOH/NOA 式乘用车辅助驾驶、HPark 与记忆驾驶功能、Little Camel 或 HDeliver 等低速自动配送车,以及后来的 DriveGPT 和 MANA OASIS 计算中心,作为把这些产品连起来的技术层。尽调需要重视这种宽度,因为 Haomo 从来不只是功能供应商。它想成为自动驾驶平台供应商,模块可外供、代码与云端协作可开放,商业路径横跨 OEM 项目和终端物流。但同样的宽度也放大了执行风险:公司必须在两种截然不同的采购动作里同时证明产品市场匹配。[CO001, CO002, CO008, CO009, CO026, CO030]
| 指标 | 数值 / 状态 | 日期锚点 | 置信度 | 缺口 |
|---|---|---|---|---|
| 成立 / 剥离 | November 2019 | 2019 年公司成立报道 | 高 | 需要中国登记摘录确认法律实体细节 |
| 总部 | 中国北京 | 2021-2026 年公开资料 | 高 | 不同资料中的区级地址不一致 |
| 当前运营状态 | 据报道自 Nov. 24, 2025 起运营停摆 | 2025 年末媒体报道 | 中 | 未找到官方清算或重启通知 |
| 核心业务 | 乘用车 ADAS 和低速无人物流车 | 2025 年业务描述 | 高 | 未披露分部收入 |
| 可支撑的最近一次重大融资数据点 | RMB 100m Series B1 首批款 | Feb. 2024 TechCrunch 报道 | 中 | 后续轮次金额披露不清晰 |
| 独角兽估值标记 | 2021 年末高于 US$1b | Series A 报道 | 中 | 停摆后的当前估值未知 |
| OEM 部署证据 | 20+ 款 Great Wall 关联车型 | 2023-2025 年公开报道 | 中 | 没有可用的 2026 年活跃车型清单 |
| 辅助驾驶里程证据 | 到 Feb. 2024 为 120m km;到 May 2024 为 141m km | TechCrunch 和 Friendly Auto | 中 | 无法获得独立审计 |
| 物流证据 | 北京 200k-300k 件包裹配送;声称有九个用例 | 2023-2025 年公开报道 | 中 | 未披露各客户收入贡献 |
| 当前公开员工数标记 | 停摆前近 300 名员工 | Nov. 2025 Yicai 报道 | 中 | 没有已验证的停摆后员工数 |
本表把历史证据和当前状态分开;停摆、估值和员工数行仍有部分未解,因为未找到 2025 年 11 月之后公司官方声明。
[CO001, CO002, CO005, CO006, CO007, CO008]Haomo 的投资论点把 Great Wall 控制、双产品线和重研发技术栈连在一起;同一套逻辑也把客户风险和生存风险集中起来。
[CO008, CO009, CO011, CO026, CO030, CO032]最影响决策的概览 KPI,把历史验证与仍未解决的当前状态风险放在一起。
最后一项 KPI 刻意写成未解决状态,因为公开记录仍缺少正式重启、重组或清算声明。
[CO001, CO005, CO006, CO011, CO021, CO032]1.2 公司背后的领导层、控制权与资本结构
资本与治理一直是 Haomo 故事的核心,因为 Great Wall 同时是创始方、战略客户和控制锚。公开报道显示,Haomo 2021 年末完成近 RMB 1 billion 的 Series A 融资,投资方包括 GL Ventures、Meituan、Qualcomm Ventures、Shoucheng Holdings 和 JZ Capital,并进入独角兽行列。TechCrunch 后来报道,2024 年 2 月 Haomo 获得由 Chengdu Wufa 领投的 Series B 首笔 RMB 100 million;CB Insights 则将公司列为 Series B-II,并估算披露融资总额约 $245 million。这些数字方向上利好,但无法完全对齐,本身就是重要尽调信号。Yicai 后续报道称,Haomo 已完成七轮融资,Great Wall Motor 董事长 Wei Jianjun 仍是实际控制人,持股 37%。Great Wall 自己的 2025 年年报没有披露这家创业公司的完整股权结构,但确认 Haomo Zhixing 是联营企业和关联方,并记录了与该投资有关的大额累计亏损分担。融资动能放慢时,领导层稳定性也恶化。2025 年中报道提到董事长 Zhang Kai 辞任,以及产品和技术高管离职,说明公司在进入最艰难商业阶段时,并没有稳定的领导梯队。[CO003, CO004, CO005, CO006, CO007, CO014]
| 人物 / 角色 | 故事中的角色 | 公开支撑 | 关键人物依赖 | 尽调要求 |
|---|---|---|---|---|
| Wei Jianjun / Great Wall 董事长 | 通过 Great Wall 提供最终控制锚点和战略支持方 | Yicai 称其为持股 37% 的实际控制人;GWM 文件显示 Haomo 为关联联营公司 | 很高 | 确认直接和间接投票权、救援义务以及当前董事会控制权 |
| Gu Weihao / CEO | 商业化和 DriveGPT 叙事的公开代表 | 出现在融资和 DriveGPT 发布报道中 | 高 | 确认运营停摆后他是否仍在任 |
| Zhang Kai / 董事长 | HPilot 和 AI Day 早期产品及定价发言人 | 2023 年产品发布中被引用;2025 年报道称其已辞任 | 历史上高,目前不清楚 | 要求提供正式辞任日期和继任者状态 |
| Ai Rui / 技术副总裁 | 与路线图执行绑定的技术负责人 | 2025 年报道称其在离职高管之列 | 中 | 厘清核心工程领导层是否仍在位 |
| Cai Na / 产品副总裁 | 辅助驾驶落地的产品负责人 | 2025 年离职名单中被点名 | 中 | 确认产品管理是否已重建,或被 Great Wall 吸收 |
| Great Wall 内部智能驾驶团队 | 外部但关键的运营对手方 | GWM 2025 年报强调自研 Coffee Pilot 3/4 和 VLA 路线图 | 很高 | 判断 Haomo IP、员工或项目是否已被内部化 |
覆盖不完整,因为公开记录清楚识别了控制关系和几名高管,但没有提供停摆后稳定、最新的董事会或管理层名单。
[CO003, CO016, CO018, CO024, CO025]| 利益相关方 | 角色 | 控制或经济重要性 | 重要原因 | 尽调要求 |
|---|---|---|---|---|
| Great Wall Motor | 创始方、最大客户和控制支持方 | 战略母生态,加上关联方关系 | Haomo 的生存能力和退出选项高度取决于 Great Wall 是否愿意支持或吸收它 | 取得当前服务、许可和资金承诺 |
| Meituan | Series A 投资者和物流生态伙伴 | 战略资本加需求侧验证 | 重要证据:Haomo 曾经具备非 OEM 平台相关性 | 确认是否有任何活跃商业项目延续到 2025 年 |
| GL Ventures / Hillhouse | Series A 投资者 | 独角兽阶段的成长股权验证 | 显示早期风投资金曾有信心,但不代表当前支持 | 梳理优先权堆叠和任何下行保护 |
| Qualcomm Ventures | Series A 投资者和生态信号 | 融资时的技术市场可信度 | 关系到 Haomo 最初作为严肃 AV 软件公司的定位 | 确认任何技术或渠道关系是否延续 |
| Chengdu Wufa Fund | 2024 年 Series B1 领投方 | 最近一次清晰报道的新资金轮 | 说明地方政府背景资金取代了海外 VC 动能 | 索取轮次条款、资金用途和里程碑约束 |
| Alibaba DAMO / Wumart Dmall / Dada Nexus 物流合作方 | 物流侧商业对手方 | 需求证据,而非核心控制 | 显示 Haomo 的物流故事触达了具名运营方,但不一定形成规模 | 按对手方拆分付费部署规模 |
该图谱凸显战略重要性与可见经济性之间的不对称:Great Wall 和 Meituan 最重要,但公开披露中的持股比例、清算优先权和实时商业承诺仍很薄。
[CO003, CO004, CO005, CO006, CO007, CO011]1.3 商业验证、客户足迹,以及危机前真正做到的规模
2025 年危机前,Haomo 的确拿出过不轻的公开商业化标记。最强的历史验证来自 Great Wall 车型部署和北京物流运营,而不是广泛的第三方 OEM 名单。公司关联和独立来源称,HPilot 或 NOH 产品已覆盖 20 多款 Great Wall 相关车型;官方 2023 年 DriveGPT 发布还称,Haomo 在过去主要销售给 Great Wall 之后,已成为三家车企的供应商。物流侧,TechCrunch 称 Haomo 配送车在北京为超市配送了接近 300,000 件生鲜杂货包裹;Logistics Tech Outlook 描述,截至 2023 年 7 月,与 Meituan、Wumart Dmall 和 Dada Nexus 的合作累计超过 200,000 单。Friendly Auto 的 2025 年总结还补充,Little Camel 2.0 和 3.0 已在九个场景投入商业运营;截至 2024 年 5 月,Haomo 驱动的乘用车累计辅助驾驶里程超过 141 million 公里。这些指标说明 Haomo 不只是实验室项目。但格局也很窄。Great Wall 仍是核心 OEM 分发通道,物流验证集中在北京试点区域;即便在停摆报道出现前,能证明多客户收入规模持久化的独立证据也仍然薄。[CO010, CO011, CO012, CO013, CO019, CO020]
1.4 里程碑、承压信号与公司现状
里程碑记录清楚分成上升期和崩解期。上升期从 2019 年分拆开始,经过早期融资、2021 年独角兽状态、2023 年 4 月 DriveGPT 发布,以及 2023 年 10 月低成本 HP170、HP370、HP570 辅助驾驶产品发布。崩解期在 2025 年浮出水面。KrASIA 2025 年中翻译的 36Kr 报道称,Haomo 只剩两个 Hyundai 乘用车项目,城市 NOA 交付目标落空,Little Camel 销售目标已下调到约 50 台,并把重心放在清库存而不是推出新车型上。2025 年末,危机加深:KrASIA、CarNewsChina 和 Yicai 均报道,员工被通知不用到岗,工资已经拖欠,冻结账户和可能诉讼在内部被讨论。Great Wall 自己的年报也间接加重担忧:它把公开智能驾驶叙事转向自研 Coffee Pilot 3 和 4、VLA 与世界模型系统,同时确认更多与 Haomo 相关的投资亏损。因此,当下概览里最重要的事实,不只是 Haomo 曾经达到独角兽状态;而是截至 2026 年 8 月,公开渠道仍没有官方声明清楚确认重启、重组或清算。[CO015, CO016, CO017, CO019, CO020, CO021]
| 日期 | 事件 | 类型 | 金额 / 估值 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2019-11 | Haomo 从 Great Wall 剥离后成为独立公司 | 创立 | 公司成立 | Great Wall Motor 和 Haomo 创始团队 | 为整个投资案例建立法律和战略起点。 |
| 2021-02 | Great Wall 及随后 Pre-A 投资者在早期融资序列中支持 Haomo | 融资 | 未披露 + 数亿元人民币 | Great Wall、Shougang、Meituan、Hillhouse 等投资方 | 说明独角兽轮之前,早期资本支持已经到位。 |
| 2021-12 | Series A 融资近 RMB 1 billion,并赋予独角兽地位 | 融资 | 近 RMB 1b;>US$1b 估值标记 | GL Ventures、Meituan、Qualcomm Ventures、Shoucheng、JZ 等投资方 | 确认私募市场对平台故事的信心达到峰值。 |
| 2023-04 | Haomo 在 AI Day 发布 DriveGPT | 产品 | DriveGPT 亮相;声称进入三家车企供应链 | Haomo 和合作伙伴生态 | 标志着转向基础模型驱动的差异化和出海叙事。 |
| 2023-10 | Haomo 发布 HP170、HP370 和 HP570 低成本 ADAS 套件 | 产品 | RMB 3,000-8,999 标价 | Haomo | 展示以价格切入的大众市场 ADAS 策略。 |
| 2024-02 | Series B1 首批款被报道 | 融资 | RMB 100m / US$14m 融资额 | Chengdu Wufa 和 Haomo | 说明资本仍可获得,但规模远小于 2021 年。 |
| 2025-05 | 领导层离职和商业化疲软公开化 | 治理 | 董事长和 VP 变动 | Haomo 高管 | 标志着叙事从规模化转向重组风险。 |
| 2025-11 | 员工被通知不用到岗;工资和账户问题被报道 | 负面 | 运营停摆 | Haomo 员工和媒体 | 对连续性和股权价值提出生死问题。 |
| 2026-08 | 截至报告日,未找到官方重启、清算或重组声明 | 负面 | 状态未决 | Haomo / Great Wall 公开沉默 | 当前尽调必须把连续性视为未确认,而不是默认存在。 |
这是本报告唯一的权威时间线;2025 年后的条目把据报道的运营事件与官方确认的公司行动区分开。
[CO001, CO004, CO005, CO006, CO021, CO022]Haomo 的公开时间线从 Great Wall 支持下的扩张,走向商业化疑虑,最后到被报道的运营停摆。
[CO001, CO004, CO005, CO006, CO018, CO021]1.5 图表与证据
02市场分析
2.1 先界定 Haomo 的真实市场边界,再谈规模
Haomo 并不是卖进一个简单的“自动驾驶市场”。它位于两个相关但经济属性不同的需求池交叉点。第一个是中国乘用车智能驾驶,尤其是面向国内 OEM 的 L2+ 到 L2.9 辅助驾驶系统,车企需要在大众价格带里组合传感器、算力、软件和安全验证。第二个是低速自动末端配送,机器人、园区和社区路线、生鲜配送、零售物流会带来另一套买方、运营指标和监管约束。把二者当成一个 TAM,会夸大机会,也会遮住销售现实。乘用车 ADAS 依赖 OEM 平台定点、SOP 节点、芯片选择,以及不同车价带的功能渗透。低速物流依赖路线密度、人工替代、利用率、城市许可,以及客户把车辆视为 资本开支、服务还是混合基础设施。Haomo 自己的产品组合也反映了这种拆分:HPilot 和 NOH 瞄准 OEM 部署,HDeliver 或 Little Camel 瞄准物流运营商与零售商。因此,市场章节采用两条平行视角,而不是一个混合的标题数字。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 与 Haomo 的相关性 |
|---|---|---|---|---|
| 乘用车智能驾驶软件和域控制栈 | 量产车内的 ADAS 软件、传感器融合、域控制、验证、轻地图和 NOA 能力 | Robotaxi 车队经济性和通用半导体 TAM | 车企工程和产品团队 | HPilot、NOH 和低成本 L2+/L2.9 套件的直接目标市场 |
| 国内 OEM 的大众市场 ADAS 功能渗透 | RMB100k-RMB300k 车型中的功能搭载和供应商定点 | 仅限豪华车型的展示项目 | OEM 产品 P&L 和平台负责人 | 相关,因为 Haomo 为主流普及给 HP170/370/570 定价 |
| 低速无人末端配送 | 受控区域内的车辆、车队运营、路线规划、维护和配送服务自动化 | 长途卡车、中国以外的人行道微型机器人和通用电商 GMV | 零售商、外卖平台、校园 / 社区运营方、物流服务商 | HDeliver / Little Camel 的直接目标市场 |
| 无人物流平台服务 | 围绕配送车辆的车队管理、远程运营、维护和部署服务 | 广义仓储自动化或仅无人机基础设施 | 平台运营商和企业物流管理者 | 重要,因为物流价值来自利用率,而不只是车辆出货 |
| 智能驾驶的具身 / 物理 AI | 提升驾驶认知和交互的模型、数据和仿真层 | 与交通无关的通用基础模型收入 | OEM、AV 公司和赋能伙伴 | 相关,因为 Haomo 用 DriveGPT 和 MANA OASIS 区分路线图 |
本表把决策相关市场定义得足够窄,以避免重复计算 Robotaxi、通用 AI 和无关半导体机会。
[CM001, CM002, CM003, CM004, CM005, CM006]| 细分 | 买方 | 用户 | 付款方 | 工作流 / 采用触发点 |
|---|---|---|---|---|
| 国内 OEM 主流车型 | OEM 智能驾驶团队 | 驾驶员和安全验证团队 | 车型平台 P&L | 需要可负担、可量产的辅助驾驶功能 |
| 高端 / 旗舰 OEM 项目 | OEM 战略和 CTO 办公室 | 重度用户和首发车型团队 | 品牌投资预算 | 需要旗舰 NOA 或 VLA 差异化 |
| 零售和商超配送 | 零售运营经理 | 门店拣货打包和本地配送网络 | 零售商或平台运营商 | 需要高密度短途配送并降低人工成本 |
| 校园 / 园区 / 社区物流 | 园区或物业运营方 | 设施运营团队和终端用户 | 运营方预算或服务费 | 需要在审批更容易的受控区域跑自动化 |
| 外卖或即时零售平台 | 平台运营负责人 | 商户 + 骑手 + 调度团队 | 平台经济模型 | 需要更快的高峰期配送和更可预期的服务水平 |
Haomo 不得不面对 OEM 和物流渠道截然不同的采购动线,GTM 复杂度因此被拉高。
[CM003, CM006, CM024, CM025, CM026, CM027]Haomo 的两个核心市场,需要不同买方、用户和采用触发因素。
[CM001, CM006, CM024, CM025, CM026, CM038]2.2 中国 ADAS 仍在增长,但增长正向上层能力和全栈供应商迁移
中国智能驾驶大盘方向仍然有利。Research-and-markets 聚合内容和专业行业媒体把 2025-2026 年描述为结构升级期:非智能和低阶配置继续退场,高阶辅助驾驶功能走向主流。最有用的数字不是泛泛 TAM,而是渗透率和架构变化。独立 OEM 从 2022 年由非智能配置主导的市场,转向 2025 年高阶智能驾驶占 62.7% 的市场,城市 L2.9 NOA 渗透率从 2.1% 升至 17.2%。另一个供应商侧视角显示,到 2025 年前四个月,搭载 L2.5 和 L2.9 的新车型占比分别达到 30.2% 和 34.82%。Gasgoo 和 DVN 随后显示,2026 年规模正进入零部件与产业化:2026 年前四个月 LiDAR 装机超过 1.3 million 台,供应商栈集中到少数大玩家手里。这对品类需求是利好,但对创业公司站位并不中性。Robotics and Automation News 总结 IDTechEx 研究称,车企优先选择全栈且有成本效率的合作伙伴,纯软件 ADAS 供应商丢掉阵地。这一变化对 Haomo 战略不利,因为它的商业化问题恰好出现在市场奖励更宽生态和更紧芯片-软件集成的时候。[CM008, CM009, CM010, CM011, CM012, CM013]
| 发布方 / 视角 | 年份 | 地域 | 数值 | 方法论 / 相关性 | 局限 |
|---|---|---|---|---|---|
| Research and Markets / OEM ADAS 结构视角 | 2026 | 中国乘用车 | 没有单一 $TAM;结构升级和渗透证据 | 更适合判断渗透率和架构迁移,而不是直接拿来当收入口径 | 付费报告摘要没有拆出 Haomo SAM |
| The Automotive Data / 供应商份额视角 | 2025 | 中国乘用车 | 2025 年 1-4 月新车型中,L2.5 占 30.2%;L2.9 占 34.82% | 可较好代理供应商定点需求正在放大的位置 | 新车型占比不是供应商收入 |
| Grand View Research / 全球无人末端配送 | 2024-2030 | 全球 | US$1.6b 至 US$5.9b | 无人配送硬件和服务的全球上限视角 | 包含与 Haomo 无关的玩家和地域 |
| DataM / 中国无人末端配送 | 2026-2035 | 中国 | 2026 年 US$6.05b,到 2035 年 US$25.98b | 对 Haomo 物流可选性最相关的方向性视角 | 预测式估计,不是已实现支出 |
| Haomo / 物流可服务市场叙事 | 2023 | 中国 | RMB450b 末端配送市场 | 显示管理层对物流机会的上限判断 | 若没有利用率和路线约束,作为可投资 SAM 过宽 |
这些视角有意混合:有的是收入池,有的是渗透代理,有的是管理层给出的上限。不应相加。
[CM008, CM009, CM020, CM021, CM022, CM023]| 驱动因素 / 约束 | 方向 | 时间 | Haomo 影响 | 尽调追问 |
|---|---|---|---|---|
| 高阶 ADAS 渗透率持续上升 | 正向 | 2025-2026 | 主流辅助驾驶供应商的品类需求仍在 | Haomo 实际把多少增长转化成了仍在推进的 OEM SOP 定点? |
| 市场转向全栈、芯片 + 软件平台 | 负向 | 当前 | 更利好 Huawei、Horizon 这类生态,而不是单点型创业公司 | 没有强势芯片或 OEM 伙伴护城河时,Haomo 技术栈还有哪些环节真正差异化? |
| 激光雷达和工业验证在 2026 年放量 | 混合 | 当前 | 验证了品类,但也抬高了资本开支和供应商质量门槛 | Haomo 能否撑住所需算力、验证和零部件关系? |
| 地面车辆仍是最后一公里配送的主流形态 | 正向 | 2024-2030 | 支撑 Haomo 在物流场景选择车辆形态 | Haomo 实际跑出了怎样的路线经济性和接管率? |
| 最后一公里自动化仍取决于利用率和地方许可 | 负向 | 当前 | 小试点容易,规模化难 | 停摆前,哪些城市级许可和单元经济模型已经真实运行? |
| Great Wall 客户集中度 | 负向 | 当前 | 即使总市场增长,也会压缩实际可触达 SOM | Great Wall 分散供应商后,Haomo 需求还剩多少? |
品类增长与 Haomo 适配度出现分化;本表拆出品类顺风未能转化为公司特定优势的环节。
[CM014, CM015, CM016, CM021, CM027, CM033]Haomo 的机会横跨乘用车 ADAS 和自动驾驶配送两个增长快、但口径敏感的收入池。
各行混合了直接市场规模估算和以渗透率锚定的机会代理值;它们展示的是区间差异,不是一套经审计、完全同口径的数据。
[CM020, CM021, CM022, CM023, CM030, CM036]乘用车 ADAS 的品类增长,如今靠架构、芯片、验证和跨车型规模推进,不能只靠功能营销。
[CM010, CM011, CM012, CM013, CM014, CM015]2.3 自动末端配送是真需求,但本质仍是 ROI 和运营模型市场
Haomo 的物流业务瞄准的是有真实结构性需求、但运营上很苛刻的市场。Grand View Research 估算,全球自动末端配送市场 2024 年约 US$1.6 billion,并预计 2030 年约 US$5.9 billion;DataM 则把中国市场单独估在 2026 年约 US$6.05 billion、2035 年接近 US$26 billion。这些市场数字不能直接当作 Haomo 的近期收入池;它们是由硬件成本、城市密度、餐饮和生鲜需求、智慧城市投资共同塑造的采用上限。更有用的是读运营逻辑。地面车辆占主导,因为它们比空中系统更适配受控、重复路线;中国最密集的场景仍是园区、商圈、社区、小区、超市和社区物流。DataM 明确把市场框定为进入关键投资时点:利用率、监管批准、运营员比例、电池经济性和保险框架,决定试点部署能否变成可规模化业务。Logistics Tech Outlook 对 Haomo 的画像正好贴合这幅图:有真实包裹流转和具名客户,但仍局限在特定地理和用例环境。因此,物流机会仍在,但持久优势的门槛正在从原型新鲜感升到车队经济性和生态集成。[CM020, CM021, CM022, CM023, CM024, CM025]
物流市场会从广泛包裹需求,收窄到少数经济上跑得通的自动驾驶路线。
[CM021, CM024, CM025, CM026, CM027, CM028]2.4 市场含义:两个市场都变成规模游戏,Haomo 的时间窗口随之收窄
市场含义不是 Haomo 选错了品类,而是它选的品类后来都变得更难,尤其不利于一家中等规模、战略上依赖母公司的创业公司。在乘用车 ADAS 里,重心转向能组合芯片、软件、安全和广泛 OEM 覆盖的平台,Huawei、Horizon 及类似生态在定义标杆。在低速物流里,需求仍然真实,但经济性越来越依赖可复制的路线密度、更低车辆成本,以及与 Meituan、Alibaba、JD 或专业机器人运营商等巨型平台的运营集成。Haomo 每条故事都有一些组件——面向主流车型的低成本 HPilot 套件、具名物流试点、围绕 DriveGPT 的研究叙事——但到 2025 年,它缺少足够广泛的 OEM 多元化证据,也没有足够强的物流规模来抵消对 Great Wall 的依赖。因此,底层市场可以继续增长,Haomo 自身可实现的 SAM 和 SOM 却在收缩。这种背离对估值很关键:当公司的份额获取路径已被商业化延迟和报道中的停工削弱,投资者不能只引用品类增长来支撑乐观判断。[CM030, CM031, CM032, CM033, CM034, CM035]
2.5 图表与证据
03竞争格局
3.1 竞争版图:Haomo 同时面对创业公司、成熟厂商和替代方案
Haomo 的竞争集合远不只是一个创业公司排行榜。乘用车里,它必须在 OEM 采购周期中击败 Horizon、Momenta、Huawei 相关栈和其他第三方智能驾驶供应商。低速自动化里,它还要用配送经济性证明自己,面对 robotaxi 或 robotruck 运营商,这些公司的技术、数据和资本基础可能外溢到相邻商业场景。同样重要的是,Haomo 还面对 Great Wall 内部智能驾驶路线图、Mobileye 等全球成熟供应商,以及人工配送加传统 ADAS 现状的替代竞争。这很重要,因为 Haomo 试图同时跨越两种采购动作:整车项目采购和物流运营自动化。公开证据显示,行业向少数规模化生态集中后,多线竞争变得更难;这些生态在芯片、数据或分发上有更清晰优势。因此,最相关的比较不是“哪家创业公司的演示最好”,而是哪家公司在真实采购决策中变得最难被 OEM 和平台运营商绕开。[CP001, CP002, CP003, CP004, CP005, CP006]
| 竞争者 / 替代方案 | 类别 | 规模或融资证明 | 目标客户 | 差异化 | 可比性限制 |
|---|---|---|---|---|---|
| Haomo AI | ADAS 与低速物流双线创业公司 | 独角兽阶段融资,并在 Great Wall 商业化,但 2025 年出现停摆报道 | Great Wall 和部分物流运营方 | 低成本 HPilot 套餐,加上物流可选项 | 高度依赖 Great Wall,当前运营不明 |
| Horizon Robotics | 芯片 + ADAS 一体化平台 | 披露 10 个 OEM 品牌、20+ 款 HSD 定点车型 | OEM ADAS 与产品团队 | Journey 6 芯片叠加 HSD 技术栈,覆盖多个价格带 | 打法比 Haomo 早期软件主导模式更重资产 |
| Momenta | 端到端量产 ADAS 供应商 | 多家头部 OEM 合作,2026 年被市场视为第一梯队 | 大型 OEM 项目 | 数据飞轮和广泛 OEM 触达 | Momenta 是供应商,消费者品牌可见度间接 |
| Huawei Qiankun ADS | 生态主导的智能驾驶平台 | 2026 年报道引用 10.47b km 和 77% 国内第三方供应商份额 | Huawei 阵营及合作 OEM | 传感器冗余、本地数据、生态力量 | 硬件成本更高,可能限制低价格带适配 |
| Pony.ai | 以 Robotaxi 为主,延伸到 POV 和卡车的自动驾驶公司 | Toyota / GAC 合资公司,加上官方多产品线 | 出行运营商、卡车运输伙伴、OEM | L4 运营,加上相邻量产产品 | 核心规模证明仍然最强在车队运营 |
| WeRide | 统一 L2-L4 自动驾驶平台 | 公开定位为一套技术栈横跨出行、物流、环卫 | 城市、出行运营商、企业伙伴、OEM | 跨场景复用运营平台 | 乘用车直接定点披露少于 Huawei / Horizon |
| Mobileye | 全球老牌 ADAS / AV 平台 | 截至 2025 年,230m+ 辆车搭载 EyeQ | 全球 OEM | 装车基数数据、REM 地图、长期 OEM 信任 | 与中国 OEM 的深度关系弱于头部本土生态 |
| Great Wall 自研 / 人工配送 | 现状替代方案组合 | Great Wall Coffee Pilot 路线图,加上既有人工作业经济性 | Great Wall 和物流运营方 | 内部控制,或更低的组织变革成本 | 不是一一对应的外部供应商类别 |
本表混合直接供应商和替代方案,因为 Haomo 的真实竞争场里还包括自研和现状运营模式。
[CP001, CP002, CP003, CP004, CP006, CP014]到 2026 年,Haomo 在商业化广度和生态深度上都落后于中国领先生态。
坐标轴采用有证据支撑的 1-5 档序数评分,而不是经审计的市场份额数字。
[CP009, CP013, CP014, CP018, CP021, CP024]3.2 中国 ADAS 对手转向更强 OEM 嵌入的一体化生态
Horizon、Momenta 和 Huawei 说明了 Haomo 乘用车位置为什么变弱。Horizon 公开材料给出一个高度集成的故事:Journey 6 芯片、HSD 辅助驾驶软件、参考工具,以及针对 RMB100k+、RMB150k+、RMB200k+ 车型价格带的分层方案。Momenta 的公开合作与第三方报道呈现另一种强模式:端到端数据飞轮,通过多个 OEM 项目销售,而不是依靠一个锚定客户。Huawei 的 Qiankun ADS 展示第三条路径:把传感器冗余、本地数据密度和足以影响市场层面供应商份额的生态力量合在一起。这些对手技术路线不同,但商业优势收敛到同一点:它们要么提供芯片加全栈集成,要么拥有广泛车型覆盖,或两者都有,因此车企难以忽视。Haomo 的低成本 HPilot 信息在战略上看似合理,但到 2026 年,它面对的是更大的对手;这些对手同样能按车价层级切分产品,同时带来更多量产定点、更多资本和更厚的数据闭环。[CP007, CP008, CP009, CP010, CP011, CP012]
| 采购标准 | Haomo | Horizon | Momenta | Huawei ADS | 点评 |
|---|---|---|---|---|---|
| 芯片 + 软件一体化技术栈 | 未知 / 公开信息不清 | 强 | 中等,借助伙伴芯片 | 强 | Horizon 和 Huawei 公开主打芯片 + 技术栈优势;Haomo 未显示同等芯片控制力 |
| 大众市场价格带套餐 | 历史上强 | 强 | 中等 | 中等 | Haomo 和 Horizon 都公开过面向主流车型的分层产品 |
| OEM 定点广度 | Great Wall 之外公开证明有限 | 强 | 强 | 生态内强 | Haomo 的多元化弱于领先者 |
| 端到端 / 大模型叙事 | 通过 DriveGPT 具备 | 强 | 强 | 强 | 叙事曾经齐平,但商业化广度拉开差距 |
| 当前公开信任 / 连续性 | 2025 年停摆报道后偏弱 | 强 | 强 | 强 | 汽车买方看重路线图连续性,不亚于演示效果 |
「未知」表示公开信息不足,并非负面判断。决策相关的模式是相对广度,不是绝对技术排名。
[CP007, CP008, CP009, CP010, CP011, CP012]领先者超过 Haomo 的地方,不在功能词汇本身,而在围绕这些功能搭出的商业化支撑宽度。
[CP007, CP009, CP012, CP015, CP017, CP021]3.3 L4 运营商与全球标杆拉大了广度和可信度差距
Haomo 的物流和自动驾驶叙事,也必须与规模不在 Great Wall 生态内的公司比较。Pony.ai 官方材料显示,它横跨 robotaxi、robotruck 和个人拥有车辆的智能驾驶方案;WeRide 的平台描述强调 L2 到 L4 在出行、物流和环卫之间复用。DeepRoute 2026 年 IO 2.0 发布又增加一个压力点:即便更新的同行,现在也在讲 VLA、多传感器灵活性和已锁定的 OEM 合作。Mobileye 仍是嵌入式 ADAS 规模的全球标杆,而不是中国配送的直接对手;但这个标杆仍重要,因为它设定了地图、数据量和长周期 OEM 信任预期。累积效果是,Haomo 不是只被某一个国内对手跑赢。到 2026 年,它被拿来对照多种竞争强度模型:消费车规模、robotaxi 车队运营、垂直集成地图与 ADAS 阶梯、多产品复用;公开材料却没有清楚证明它在任何一个维度领先。[CP018, CP019, CP020, CP021, CP022, CP023]
| 公司 | 公开套餐或合同模式 | 价格信号 | 包含能力信号 | 未知项 / 含义 |
|---|---|---|---|---|
| Haomo | HP170 / HP370 / HP570 产品分层 | 2023 年发布稿显示低至约 US$411 | 面向低端到高端乘用车的 ADAS 套件 | 定价激进,但 2026 年公开合同经济性不可得 |
| Horizon | 按车型细分的 HSD 300 / 600 / 1200 | 没有标价,但 RMB100k+ / 150k+ / 200k+ 价格带明确 | 基于 Journey 6 架构的城区 / 高速 / 泊车辅助 | 说明更大对手即使不公开标价,也能切分市场 |
| Momenta | OEM 合作与供应商模式 | 未公开披露 | 伙伴部署中的高速 NOA、城区 NOA、泊车、代客泊车 | 供应商经济性大概率按项目逐一谈判 |
| Pony.ai | Robotaxi、自动驾驶卡车和 POV 产品线 | 未公开披露 | L4 车队系统,加上量产智能驾驶解决方案 | 即使没有公开价目表,也显示出比 Haomo 更宽的产品包装 |
| Mobileye | Cloud-Enhanced ADAS / EyeQ 阶梯 | 未公开披露 | 由 REM 支撑的 ADAS 到更高阶自动化阶梯 | 老牌厂商定价不透明,限制逐项可比 |
这个市场公开定价很少。套餐结构、集成范围和目标价格带,往往比实际合同 ASP 更容易观察。
[CP008, CP018, CP019, CP021, CP023, CP032]3.4 护城河耐久度取决于量产锁定、数据闭环和资本续航
最重要的竞争教训是,自动驾驶护城河如今更多是运营性的,而不是叙事性的。供应商拿下一个量产项目,就获得验证历史、软硬件调校和持续道路数据,后续替换成本随之上升。2026 年公开视野里最强的供应商——Horizon、Momenta、Huawei、Mobileye、Pony.ai 和 WeRide——各自至少有一层可持续护城河:OEM 嵌入、芯片控制、数据规模、全球车队运营或财务韧性。Haomo 的公开证据更窄:历史商业化高度绑定 Great Wall 和北京试点,产品组合强调成本意识,技术故事可能有前景,但已不在领先生态叙事之内。管理层动荡、欠薪报道和运营停摆使信任恶化后,即便技术栈仍有可信度,新车型项目也更难承销。由此得到的护城河评估偏弱:Haomo 没有一层清晰、可持续的竞争力,看起来能强过客户集中和执行风险。[CP033, CP034, CP035, CP036, CP037, CP038]
| 护城河主张或弱点 | 威胁 | 严重性 | 重要性 | 缓释措施 / 尽调追问 |
|---|---|---|---|---|
| Great Wall 作为锚定客户的关系 | Great Wall 自研或分散供应商 | 高 | 锚定客户集中可能从护城河翻成依赖 | 确认 Great Wall 之外是否仍有在跑项目 |
| 低成本产品包装 | 更大对手在补齐可负担价格的同时增加生态深度 | 高 | 如果其他玩家能按配置和芯片栈切分市场,单靠价格不耐久 | 索取按车型价格带和项目阶段拆分的胜率数据 |
| DriveGPT 和具身 AI 叙事 | 对手普遍讲端到端 / VLA 故事,叙事商品化 | 中 | 没有交付规模,叙事齐平已无法形成差异 | 要求提供专门对应 DriveGPT 时代产品的量产指标 |
| 物流试点经验 | 全球 Robotaxi 和物流玩家积累更大车队数据 | 高 | 运营数据和安全证明可能很快超过 Haomo | 索取路线经济性、接管数据和城市许可 |
| SOP 后 OEM 切换成本 | 对手一旦拿下项目,可能形成锁定 | 高 | 汽车项目里,后来挑战者很难追回 | 梳理 Haomo 实际 SOP 覆盖和续约可见性 |
| 公司连续性和信任 | 停摆和欠薪报道损害采购信心 | 高 | 汽车客户会避开不稳定供应商 | 需要重启、重组或资产转移状态的文件证据 |
本风险登记表聚焦护城河耐久性,而非纯技术排名;汽车采购中,买方信任和量产连续性才是决定因素。
[CP033, CP034, CP036, CP037, CP038, CP039]公开 KPI 更支持几类对手:已有装车规模、OEM 覆盖广,或运营数据深。
[CP009, CP014, CP018, CP019, CP037, CP040]3.5 图表与证据
04财务
4.1 收入模型:多条变现路径可见,但真实收入质量不可见
Haomo 公开材料显示,它不是一家单一纯软件公司,而是混合收入模型。乘用车侧,公司向 OEM 项目销售 ADAS 套件和更宽的智能驾驶方案,HP170、HP370、HP570 被定位为覆盖低端到高端车型的分层产品。物流侧,Haomo 推广 HDeliver 和相关低速自动驾驶车辆,意味着硬件收入、部署收入,以及可能存在的持续运营或支持费。公开媒体和公司关联报道还提到代码、模块、硬件、软件和云服务,暗示 Haomo 变现的是一组产品与集成工作,而不是单一经常性 SaaS 费用。牵引力信号确实存在——20 多款车型、到 2024 年春约 140 million 辅助驾驶公里、接近 300,000 件北京生鲜杂货配送——但这些是活动代理指标,不是经审计收入。因此,财务解读只能是混合的:Haomo 有真实商业触点,但公开证据仍没有说明,产品销售与支持服务各贡献多少收入,也没有说明这些收入流是否有吸引人的利润率或持久留存。[CI001, CI002, CI003, CI004, CI005, CI006]
| 收入流 | 机制 | 单位 | 当前公开数值 / 状态 | 质量 | 尽调追问 |
|---|---|---|---|---|---|
| 乘用车 ADAS 套件 | 按车或按项目销售 HP170/HP370/HP570 | 按车型 / 按套件 | 已披露 20+ 款车型和标价;实际收入未披露 | 中 | 按产品层级拆出出货量、ASP 和毛利率 |
| 全栈 OEM 智能驾驶解决方案 | 面向 OEM 项目的软件、模块、硬件和云端打包 | 项目合同 | 有公开描述,但未披露已签合同金额 | 低 | 提供合同结构、里程碑时点和收入确认政策 |
| 自动配送车销售 | HDeliver / Little Camel 车辆和硬件部署 | 整车销售 | 已披露低于 90k 元价位;销量未披露 | 低 | 披露销量、底盘车辆成本和贡献毛利 |
| 运营 / 技术支持服务 | 面向客户的支持、数据、云和自动驾驶服务 | 服务费 | 公司描述和品类惯例暗示存在这类收入;未披露 Haomo 收入数字 | 低 | 把经常性支持收入与一次性集成工作分开 |
| 相邻智能硬件 / 机器人应用 | 可能延伸到清洁、安防及相关机器人场景 | 项目或产品销售 | 2024 年融资用途说法提及;商业化状态不清 | 低 | 列出非核心相邻业务的在跑客户、试点和收入贡献 |
Haomo 可能靠产品 + 服务的混合模式变现。公开来源没有揭示各收入流占比。
[CI003, CI004, CI005, CI006, CI009, CI033]| 产品 | 价格 / 合同信号 | 标价与实际 | 包含内容 | 未知项 / 含义 |
|---|---|---|---|---|
| HP170 | 套件家族中披露 CNY3,000 档标价 | 类似标价的公开定价 | 面向低端乘用车的入门 ADAS 套件 | 不能证明实际 OEM ASP 或毛利 |
| HP370 | CNY3,000-CNY8,900 区间内的中端层级 | 类似标价的公开定价 | 能力更高的 ADAS 套餐 | 折扣、搭载率和项目经济性未披露 |
| HP570 | 高端约 CNY8,900 | 类似标价的公开定价 | 高端乘用车 ADAS 套餐 | 若包含较重硬件,毛利仍可能偏低 |
| HDeliver 最新一代 | 报道资料称单车低于 CNY90,000 | 产品报价,不是合同记录 | 自动驾驶最后一公里配送车 | 低价位可能帮助采用,但会压缩硬件毛利 |
| OEM 集成与支持 | 定制 / 未披露 | Unknown | 大概率包含软件、验证、云和支持服务 | 没有合同结构,无法判断收入质量 |
官方价目表是有用的需求信号,但不能替代实际 ASP、折扣和毛利披露。
[CI002, CI005, CI006, CI022]Haomo 公开业务模型看起来是把 OEM 项目和物流部署转化为产品收入与服务收入混合体。
[CI003, CI004, CI005, CI006, CI033]4.2 成本结构与单位经济:停摆前,Haomo 就已经显得资本消耗重
成本图景只能间接重构,但方向清楚。Haomo 披露的资金用途集中在大模型、算力、数据和量产,而不是近期盈利。它的产品组合也意味着天然不同的利润结构:主流 ADAS 套件如果规模化销售,可能支撑不错的毛利;自动配送车、域控制器、传感器栈和车队支持,通常会把业务拖向硬件式经济性和营运资金需求。可比上市自动驾驶公司凸显了问题。WeRide 2025 年约 US$97.9 million 收入,仍低于约 US$196 million 研发开支和 US$264 million 经营亏损;Pony 的 2025 年和 2026 年 Q1 披露显示,快速收入增长与重度运营投入、车辆 BOM 优化和新增资本开支并存。这不能证明 Haomo 的精确现金消耗率,但说明周边品类经济性:即便更强、资本更充足的自动驾驶公司,也离轻松自我造血很远。Haomo 同时有 Great Wall 集中、物流硬件和前沿模型野心,除非锚定客户或新一轮融资持续吸收扩张成本,否则现金转化大概率为负。[CI011, CI012, CI013, CI014, CI015, CI016]
| 指标 | 数值 / null | 置信度 | 重要性 | 尽调追问 |
|---|---|---|---|---|
| ADAS 套件实际 ASP | Null | 低 | 建模 OEM 定点毛利需要该数据 | 按 HP170/370/570 和客户提供实际 ASP |
| 配送车毛利率 | Null | 低 | 决定 HDeliver 放量是改善还是拖累现金生成 | 披露 BOM、组装成本、服务准备金和贡献毛利 |
| 物流利用率证明 | Wumart 单个案例日均 ~300 单,峰值 ~700 单 | 中 | 使用密度是配送车队回本的路径 | 提供路线级成本、接管率和客户付款条款 |
| 客户集中度 | 很可能偏高,但准确收入占比未披露 | 中 | 集中度决定收入质量和议价能力 | 展示头部客户收入结构和应收账款账龄 |
| R&D 强度相对收入 | Haomo 公开资料无从判断;上市同业显示 R&D 往往高于收入 | 中 | 判断规模能否自我造血,还是依赖融资 | 提供 2024-2025 年收入、毛利率、R&D、SG&A 和资本开支桥表 |
这些空占位本身就是重要发现:公开收入质量不能只靠活动指标支撑。
[CI007, CI008, CI011, CI018, CI019, CI020]| 公司 | 收入结构信号 | R&D / 运营成本信号 | 流动性信号 | 对 Haomo 的启示 |
|---|---|---|---|---|
| Pony.ai | 2025 收入 US$90.0m;Q1 2026 收入 US$34.3m | 2025 非 GAAP 净亏损 US$174.0m;Gen-7 投入和资本开支仍在继续 | Dec. 31, 2025 现金 / 投资 US$1.515b;Mar. 31, 2026 为 US$1.436b | 即使规模化且已上市的头部公司,仍需要很大的资产负债表支撑增长 |
| WeRide | 2025 收入 US$97.9m,包含产品和服务收入 | 2025 R&D US$196.2m;运营亏损 US$264.1m | Dec. 31, 2025 流动性约 US$1.0b | 商业牵引没有消除高 R&D 和部署烧钱 |
| Mobileye | 2025 收入 US$1.894b | 2025 年产生 US$602m 经营现金 | 2025 年末现金 US$1.8b | ADAS 真正自我造血,只出现在远高于 Haomo 已披露规模的体量上 |
| WeRide 申报文件细节 | 前五大客户占 2025 收入 40.5% | 即便增长,经营现金仍然流出 | 公开申报文件披露收入结构和风险因素 | 说明投资者需要怎样的披露标准,而 Haomo 尚未提供 |
| Pony 单车经济性披露 | 广州和深圳 UE 盈亏平衡;Gen-7 车辆单车日净收入创纪录,达 RMB394 | 仍前置投入,并继续推进 BOM 降本 | 仍需要大规模资产负债表 | 凸显 Haomo 若要证明利润路径可信,需要跨过多高的门槛 |
这些不是直接估值可比公司;它们是自动驾驶业务资本强度的经营财务基准。
[CI015, CI016, CI017, CI018, CI019, CI020]即便 Haomo 的收入在增长,业务同时混合研发、硬件和运营支持,成本栈大概率仍然很重。
由于 Haomo 没有公开必要的毛利输入,该桥为定性判断。
[CI011, CI015, CI016, CI017, CI020, CI027]4.3 资本充足性:公开记录暗示先依赖融资,后出现流动性压力
最有支撑的财务结论是,Haomo 到 2024 年仍依赖外部融资或战略支持,而且很可能延续到 2025 年。Yicai 报道称,Haomo 2024 年初完成超过 CNY100 million 的 B1 轮和 CNY300 million 的 B2 轮,资金用于量产、机器人扩张,以及继续投入大模型、算力和数据。这种资金用途语言是成长投资语言,不是成熟自由现金流语言。2025 年末报道随后急转负面:裁员、欠薪、停工、冻结账户传闻,以及危机后对现金、负债或重启没有公开澄清。Great Wall 2025 年年报确认 Haomo 仍是联营企业和关联方,但没有给公共投资者可以承销持续偿付能力的清晰现金续航披露。相对上市同行,这种不透明很严重。Pony 和 WeRide 可以披露现金余额、收入构成、资本开支和亏损;Haomo 不能。实际含义是,Haomo 的资本充足性应被视为未解到偏弱,举证责任在管理层:它必须证明公司仍有经营流动性,而不只是历史融资标题。[CI001, CI012, CI013, CI014, CI015, CI016]
| 项目 | 公开证据 | 含义 | 置信度 | 尽调索取项 |
|---|---|---|---|---|
| 2024 新融资 | B1 >CNY100m,B2 CNY300m | Haomo 仍需要外部资本才能继续扩张 | 高 | 确认准确交割日期、实收现金和剩余现金余额 |
| 资金用途 | 量产、大模型、算力、数据、机器人业务扩张 | 资金投向增长和 R&D,不是进入收割期 | 高 | 提供董事会批准预算和各项实际支出 |
| 当前流动性 | 未公开现金、债务或现金跑道 | 无法判断现金跑道 | 低 | 提供现金、受限现金、债务、AP、欠薪和烧钱速度 |
| 2025 年末困境信号 | 裁员、欠薪、运营停摆、账户被冻结的说法 | 指向资金充足性薄弱或已经断裂 | 中 | 提供诉讼状态、债权人行动和重启资金方案 |
| 战略支持选项 | Great Wall 仍是关联锚点,但支持确定性不清楚 | 母公司纽带可能有帮助,但并非公开披露的担保 | 中 | 说明救助义务、关联方融资安排和任何资产转移 |
公开资料能拼出历史融资时间线,但这里判断资金充足性,靠的是当前偿付能力证据和披露质量。
[CI001, CI012, CI013, CI014, CI021, CI028]公开同业显示,自动驾驶项目通常需要披露透明度和资本缓冲才能撑下去;Haomo 没有给出同等可见度。
[CI014, CI015, CI016, CI017, CI020, CI028]4.4 财务结论:阻断承销的是基础数据缺失,而不是品类需求不足
Haomo 的财务结论为负面,因为缺失的数据正是承销所需的核心输入。投资者看不到公开收入、毛利率、应收账款账龄、现金余额、债务、契约状态、在手订单、续约率、按收入计的客户集中度或月度现金消耗。也没有证据证明公司把令人印象深刻的商业化叙事转化成稳定、多元且高质量的收益。公开记录反而指向自动驾驶公司常见模式:高研发强度、产品加服务的复杂度、可能的营运资金拖累,以及持续需要新融资或强大战略赞助。这种模式在公司仍处于放量路径且能显示资产负债表强度时可以存活;但 Haomo 2025 年末的困境报道让同一模式变得危险。尽调后果很直接。除非管理层提供当前现金、负债、活跃合同、单位经济和客户回款数据,否则唯一负责任的财务立场是:到 2026 年尽调日,Haomo 是一家资本密集、报告可见度低、偿付风险未解的公司。[CI031, CI032, CI033, CI034, CI035, CI036]
| 缺失指标 | 为什么重要 | 当前影响 | 精确尽调路径 |
|---|---|---|---|
| 审计后年度收入和在手订单 | 判断规模和需求韧性必须要看 | 阻断任何严肃的基准情景模型 | 获取 2024-2025 年审计后 P&L 和已签订单积压 |
| 分产品线毛利率 | 需要拆开软件经济性和硬件拖累 | 无法判断毛利率路径 | 分别索取 ADAS、配送车和服务毛利率 |
| 现金、受限现金和债务 | 现金跑道的核心输入 | 无法评估偿付能力 | 获取最新资产负债表和银行流水 |
| 应收 / 应付账款账龄 | 揭示营运资本压力和客户回款健康度 | 可能掩盖冻结的现金循环或困境 | 索取 AR/AP 账龄、争议发票和供应商账期 |
| 头部客户集中度和关联方收入 | 决定收入质量和议价权 | Great Wall 依赖仍未厘清 | 提供逐客户收入、毛利和续约状态 |
基本财务披露缺失,本身就是建模的最大阻碍。
[CI031, CI032, CI034, CI035, CI036]4.5 图表与证据
05产品与技术
5.1 产品表面:Haomo 把驾驶辅助和自动配送卖进两套完全不同的工作流
Haomo 的产品图谱横跨两个不同客户任务。乘用车里,它提供 HPilot 及相关 NOH 或 NOA、记忆驾驶、泊车和安全辅助功能,以分层 ADAS 套件或更宽的全栈方案卖给 OEM。物流里,它提供 HDeliver 和 Little Camel 系列,面向超市、快递类流程、园区、机场和其他受控环境做短路线自动配送。公司公开发布强调,Haomo 不是只卖一个盒装功能。它描述了横跨代码、模块、软件、硬件、云服务和全栈方案的“6P”合作模式,意味着愿意在多个集成层级承接 OEM 需求。这种宽度有战略意义,因为它给 Haomo 更多方式进入客户工作流;但它也迫使公司同时成熟两种截然不同的交付模型:达到汽车 SOP 等级的 ADAS,以及重运营的物流自动化。最终的产品故事比许多创业公司叙事更宽,但公开记录仍没说明其中多少已经充分部署,多少仍带有愿景成分。这个模糊点对尽调尤其重要,因为模块宽度本身无法说明 Haomo 到 2026 年是否还能服务、修补并迭代每一类工作流。[CE001, CE005, CE006, CE010, CE011, CE012]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| HPilot 乘用车 ADAS | OEM 和终端驾驶员 | 历史上已商业部署 | 面向主流车型的分层辅助驾驶产品 | 需要当前在售车型清单和支持状态 |
| HP170 | 入门到中端市场乘用车 | 2023 年公开发布 | 低成本、无高精地图,聚焦高速 / 快速路 | 需要实际搭载率和客户名单 |
| HP370 | 中端 OEM 项目 | 2023 年公开发布 | 算力和功能更强,包含记忆驾驶 / 避障 | 需要 SOP 车型和软件更新节奏 |
| HP570 | 高端城市辅助驾驶项目 | 2023 年公开发布 | 72/100 TOPS,更丰富摄像头栈,可选 LiDAR,具备城市道路能力 | 需要发布材料之外的真实世界规模证据 |
| DriveGPT | 内部产品栈、OEM 项目,并间接触达驾驶员 | 技术平台 / 路线图层 | 面向驾驶认知的可解释 GPT 式规划叙事 | 需要当前量产使用情况和算力足迹 |
| HDeliver / Little Camel | 零售商、园区运营方、物流管理者 | 商业试点和有限运营 | 低成本自动驾驶最后一公里车辆平台 | 需要实时车队规模、接管率和 2025 年后支持状态 |
表格把已产品化 SKU 和底层能力分开,因为 Haomo 既销售具体 ADAS 套件,也营销基础自动驾驶能力。
[CE001, CE002, CE006, CE011, CE012, CE032]| 用户任务 | 当前工作流 | Haomo 方案 | 可衡量收益 | 限制 |
|---|---|---|---|---|
| 高速辅助驾驶 | 驾驶员监督车道保持、车距和导航任务 | HP170 / HP370 / HP570 | 降低驾驶员负担,提供类似 NOA 的便利性 | 当前可靠性指标未公开更新 |
| 城市辅助驾驶 | 驾驶员在系统支持下处理密集城市复杂场景 | HP570 加 DriveGPT 叙事 | 可能带来城市道路自动化和更平顺行为 | 商业推出似乎落后于预期 |
| 泊车和记忆驾驶 | 手动泊车,或重复记忆路线 | HPilot 泊车和记忆驾驶功能 | 在重复的本地环境中提高便利性 | 边缘场景可靠性未公开量化 |
| 超市最后一公里配送 | 人工骑手或混合骑手流 | HDeliver / Little Camel | 在受控短线路上降低人力强度 | 规模经济仍取决于地理位置和路线 |
| 园区 / 公园 / 机场物流 | 人工或小车完成受控区域移动 | Little Camel 部署 | 适配可预测路线和封闭 ODD | 需要许可、维护和本地运营支持 |
同一家公司覆盖差异很大的工作流,既带来产品宽度,也抬高执行复杂度。
[CE001, CE010, CE011, CE012, CE013, CE021]Haomo 的产品以不同方式进入 OEM ADAS 和物流运营两类客户工作流。
[CE001, CE011, CE021, CE026, CE027]5.2 架构与技术差异化:DriveGPT、传感器分层和研究工具构成核心技术叙事
Haomo 的技术叙事集中在同时让辅助驾驶更便宜、更强、也更可解释。2023 年 AI Day 发布把 DriveGPT 描述为基于“Drive Language”的 GPT 式自回归规划模型,并用类似 RLHF 的人类反馈训练,让决策更安全、更平顺、更可解释。HP170、HP370、HP570 三个产品层级显示,这套叙事不是抽象 AI 品牌包装,而是落在具体传感器和算力组合上:三款 SKU 在算力预算、摄像头、雷达、泊车能力和可选 LiDAR 上差异明显。第三方对 DriveGPT 的解释强调,规划模型仍依赖高质量感知输入,这与 Haomo 自己的 6P 框架,以及 MANA OASIS 和上游算力伙伴等基础设施的可能角色一致。与此同时,Haomo 的 GitHub 存在表明,公司在 LiDAR-相机融合、占用预测和地点识别等方向有真实工程产出。重要技术结论是,Haomo 看起来确有研究能力和可信的架构论点——但它仍依赖传感器、芯片、数据管线、算力伙伴和车型项目执行,而不是只靠模型新鲜感。[CE002, CE003, CE006, CE007, CE008, CE009]
| 层级 / 组件 | 角色 | 依赖 | 风险 |
|---|---|---|---|
| 感知传感器栈 | 向驾驶系统输入目标、车道和场景信息 | 摄像头、雷达,以及按 SKU 可选的 LiDAR | 上游感知或标定薄弱时,性能会下降 |
| DriveGPT / 规划层 | 生成或改进行为决策和推理 | 大模型训练数据、算力和人类反馈闭环 | 可解释性叙事不保证量产延迟或安全性 |
| 车端算力平台 | 按 SKU 运行 5、32、72/100 TOPS 级算力的 HPilot 分层产品 | 车规级芯片和热管理 / 可靠性集成 | 芯片成本、伙伴准入和部署优化都很关键 |
| 云 / 数据 / 6P 栈 | 支撑代码、模块、软件、硬件、云和全栈协同 | OEM 集成、数据管道、算力中心、合作伙伴 | 架构宽度会制造集成和支持复杂度 |
| 物流车平台 | 把自动驾驶栈与车辆、货箱和路线运营结合起来 | 整车制造、维护和路线运营 | 硬件和运营负担可能压垮软件毛利的期待 |
Haomo 披露具体传感器或算力配置时,架构证据最强;只剩高层营销词时,证据最弱。
[CE002, CE006, CE007, CE008, CE009, CE014]Haomo 的架构看起来把感知、规划、算力和云式协同叠在一起,同时服务 ADAS 与物流产品。
[CE001, CE002, CE006, CE014, CE022, CE034]Haomo 的产品成熟度不只取决于模型 IP,还取决于上游传感器、算力、数据、OEM 和监管方。
[CE014, CE020, CE028, CE029, CE031]5.3 成熟度、路线图与信任:交付足以证明重要性,但不足以消除支持风险
Haomo 的公开路线图既有已交付证据,也有脆弱性证据。正面看,官方发布显示,HPilot 到 2023 年 4 月已覆盖接近 20 款车型,到 2024 年超过 20 款;DriveGPT 被连接到量产车型叙事;Little Camel 2.0 和 3.0 据称已在多个场景商业运营。公司还声称进入 EU、Israel 并规划后续区域。但成熟度不等于连续性。后续报道称,城市 NOA 目标滑坡,Hyundai 交付项目延迟,公司的活跃商业表面到 2025 年可能大幅收窄。信任与合规证据同样不均衡。Haomo 公开了若干安全或硬件主张,例如 HP170 的五星 AEB 和不依赖 HD-map;但保留来源没有显示当前公开可靠性仪表盘、支持 SLA、软件发布日志、网络安全项目或危机后的质量声明。对买方来说,这个缺口很关键:汽车和物流客户需要长期支持与合规证明,而不只是技术上可信的功能。因此,成熟度判断是混合到偏负面:历史上不轻,当前不确定。另一个实际含义是文档深度:公开材料足以勾勒 2023 年技术栈,却不足以验证买方是否仍收到活跃版本、问题修复或长期维护。在安全敏感领域,缺失运营证据会压低原本亮眼功能披露的价值。[CE004, CE017, CE019, CE020, CE023, CE024]
| 控制 / 质量信号 | 状态 | 范围 | 缺口 |
|---|---|---|---|
| HP170 AEB 五星 E-NCAP 声称 | 公开声称 | 入门 ADAS 安全营销 | 需要独立测试结果和当前认证状态 |
| 无高精地图架构声称 | 公开声称 | 乘用车 ADAS 产品设计 | 需要证明其在中国更严格的 OTA / AD 规则下表现如何 |
| RLHF / 人类反馈训练叙事 | 公开描述 | DriveGPT 模型改进流程 | 需要安全论证文件、监控和失效模式证据 |
| 监管 / OTA 合规可见度 | 只能从行业监管资料间接判断 | 中国自动驾驶软件更新 | 未找到 Haomo 专门的合规页面或政策披露 |
| 当前支持连续性 | 2025 危机后不清楚 | 所有已交付产品 | 未找到停摆后质量、补丁或 SLA 的公开声明 |
公开信任证据远比产品功能证据薄。
[CE018, CE019, CE020, CE031, CE036]| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源 |
|---|---|---|---|---|
| Apr. 2023 | DriveGPT 在第 8 届 AI Day 发布 | 发布叙事,并进入向伙伴开放阶段 | 标志 Haomo 转向大模型叙事 | 官方发布 / Gasgoo |
| Apr. 2023 | WEY Mocca DHT-PHEV 被称为首款搭载 DriveGPT 的车型 | 接近量产的声称 | 暗示研究叙事与 OEM 部署之间已有连接 | 官方发布 / Gasgoo |
| Oct. 2023 | HP170 / HP370 / HP570 发布 | 公开发布 | Haomo 乘用车技术栈的 SKU 更清晰 | 官方发布 |
| Oct. 2023 | DriveGPT 数据集和多模态感知声称扩展 | 声称取得进展 | 显示模型训练仍在推进 | 官方发布 |
| 2025 年报道 | 据报道,城市 NOA 和部分客户部署延迟,或规模小于预期 | 反向 | 加深了对当前成熟度和路线图执行的疑问 | 行业报道 |
路线图到 2023 年末仍清晰,之后弱得多;这本身就是重要尽调事实。
[CE002, CE004, CE017, CE023, CE024, CE035]Haomo 的强项是产品描述够具体,短板是当前信任基础和路线图透明度。
[CE015, CE016, CE023, CE024, CE030, CE032]5.4 图表与证据
06客户
6.1 客户分层:真正付费者通常是车企或平台运营商,而不是终端用户
Haomo 至少服务两种不同客户动作。乘用车里,直接付费者是购买 ADAS 栈的车企或车型项目团队,驾驶员只是下游用户。低速物流里,付费者是零售商、超市运营商、即时配送平台或场景业主,他们判断自动配送能否降低人力压力并维持服务水平。这个区别很重要,因为 客户名单数量可能夸大多元化。Great Wall 不是众多客户之一;它是 Haomo 的历史母公司生态、最大可见量产通道,也可能是商业化可信度的核心来源。物流名单提供了重要证据,说明 Haomo 技术进入过真实运营路线;但这些关系更像项目化路线部署,而不是宽广透明的经常性收入基础。因此,尽调问题不是 Haomo 是否曾经有客户。它确实有。更难的问题是,多少经济价值留在 Great Wall 和少数北京中心配送项目里,而不是分散在一组独立合同中。[CU001, CU002, CU003, CU004, CU005, CU006]
| 客群 | 买方 / 用户 / 付款方 | 用例 | 规模 / 可见度 | 战略价值 | 缺口 |
|---|---|---|---|---|---|
| Great Wall 乘用车项目 | 买方:GWM 车型项目团队;用户:驾驶员;付款方:GWM / OEM | HPilot ADAS 在量产车型上部署 | 可见度最高,量产证明最强 | 锚定客户关系,且可能是核心收入信号 | 未披露客户级收入拆分或合同金额 |
| 其他车企项目 | 买方:车企工程 / 采购团队;用户:驾驶员;付款方:OEM | Great Wall 之外的其他 ADAS 项目 | 公开资料中有声称,但多数未具名 | 可能降低 GWM 之外的集中度 | 具名客户、SOP 日期和规模披露不足 |
| 超市和零售运营商 | 买方:零售商 / 本地门店运营方;用户:门店和家庭收件人;付款方:零售商 / 平台 | 社区生鲜和包裹自动配送 | 主要通过 Wumart Dmall 和北京线路可见 | 最能证明 Haomo 解决了真实低速物流任务 | 当前实时车队规模和合同条款未公开 |
| 即时配送和物流平台 | 买方:平台运营方;用户:商户和终端收件人;付款方:平台 / 运营方 | 高频最后一公里配送 | Meituan、Dada、Alibaba 多次出现 | 可能提供密集订单环境和强学习闭环 | 公开证据很少把实时车队数量直接归因到 Haomo |
| 战略投资者和伙伴生态 | 买方也可能是投资方或生态赞助方;用户不一;付款方不一 | 借助资本和生态协同获取客户 | Meituan 和 Great Wall 同时出现在资本和客户叙事中 | 可以加速信任、试点和路线准入 | 也会提高集中度和关联方风险 |
表格拆开付款方、用户和路线所有者,避免把终端驾驶员或生鲜收件人的采用误当成直接客户多元化。
[CU001, CU002, CU003, CU004, CU005, CU009]Haomo 的客户旅程从 OEM 或平台采购开始,而不是直接获取终端用户。
[CU001, CU003, CU004, CU034, CU035]6.2 采用轨迹与具名客户验证:Great Wall 和北京配送项目是核心证据
最可信的采用证据集中在两处。第一,多家 2024 和 2025 年来源称,HPilot 已装配在 20 多款车型上,且主要通过 Great Wall 生态销售;截至 2024 年 2 月,辅助驾驶里程超过 120 million 公里。第二,Haomo 的自动配送项目在北京创造了真实订单流证据。独立和公司关联来源引用了截至 2024 年初接近 300,000 件生鲜杂货包裹、截至 2023 年 7 月超过 200,000 累计订单,以及 Wumart Dmall 在顺义覆盖 60 多个社区的路线基础。Logistics Tech Outlook 和 EqualOcean 也把这些部署与 Meituan、Dada Nexus、Wumart Dmall 和 Alibaba 联系起来。话虽如此,各个具名客户的证明质量不同。Great Wall 显然达到量产级。Wumart Dmall 有具体路线和订单结果。Meituan 有强战略逻辑,作为平台也有很大自动配送规模,但公开来源没有量化 Meituan 线上车队里具体有多少使用 Haomo 车辆。Dada 和 Hyundai 更弱:两者均被公开提及,但合同经济性和当前部署连续性仍缺少文档支持。[CU010, CU011, CU012, CU013, CU014, CU015]
| 指标 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 搭载 HPilot 的车型 | 20+ 款车型 | 2024-02 | China Daily / Chinapev / TechCrunch / Yicai 等来源 | 高 | Haomo 在危机前已真正落地乘用车部署 | 按 OEM 和车型配置拆分的精确数据未知 |
| 辅助驾驶里程 | 120+ 百万公里 | 2024-02 | China Daily / Chinapev / TechCrunch | 高 | 证实已部署车辆有持续终端使用 | 没有按客户或车型分组披露里程 |
| Great Wall 相关车型数 | 20+ 款 Great Wall 车型搭载 | 2025-11 | Yicai 裁员报道 / CarNewsChina | 中 | 显示业务仍集中在母公司 OEM 生态周围 | 未披露收入贡献或当前支持状态 |
| 北京生鲜杂货包裹配送量 | 近 300,000 个包裹 | 2024-02 | China Daily / TechCrunch / Yicai | 高 | 说明物流部署已走出实验室试点 | 包裹量不等于盈利,也不等于合同续约 |
| Wumart 路线扩张 | 顺义 60+ 个社区、120,000+ 笔订单 | 2024-04 | EqualOcean | 中 | 给出了具名客户和路线级细节 | 量化只来自单一来源,且没有当前状态更新 |
| 自动配送累计订单 | 截至 2023 年 7 月超过 200,000 笔 | 2023-07 | Gasgoo / Logistics Tech Outlook | 中 | 佐证零售配送项目持续产生重复订单 | 未披露标准化客户分组或单位经济性 |
大多数数字来自公司披露或合作伙伴转述,只能作为牵引力信号,不能当作经审计的收入或留存指标。
[CU010, CU011, CU013, CU015, CU016, CU017]| 客户 / 项目 | 细分领域 | 部署 / 使用场景 | 量产还是试点 | 结果 / 信号 | 局限 |
|---|---|---|---|---|---|
| Great Wall Motor | 锚定 OEM / 关联方生态 | HPilot 部署在量产乘用车型上 | 量产 | 最持续获得佐证的客户关系;20+ 款车型和长期生态绑定 | 具体收入占比、定价和续约条款仍未披露 |
| Wumart Dmall | 零售 / 超市运营商 | 顺义及北京其他路线的生鲜杂货自动配送 | 商业路线部署 | 披露了具体订单数、社区数和日峰值吞吐 | 2025 年后当前是否延续,公开层面未验证 |
| Meituan | 即时配送平台和投资方 | 低速自动配送合作与路线运营 | 显示为项目式 / 客户关系 | 多次被列为合作伙伴或客户;以 Meituan 车队规模看,战略价值重要 | 公开来源未把 Meituan 实时机器人数量直接对应到 Haomo 车辆 |
| Dada Nexus / Dada Kuai Song 配送平台 | 配送平台 | 场景化无人超市配送和路线运营 | 项目式 / 可能从试点转运营 | EqualOcean 和 Logistics Tech Outlook 将其列为合作伙伴 | 缺少合同范围、收入贡献和当前规模 |
| Beijing Hyundai / Hyundai China 线索 | 其他车企多元化候选 | 2024 年末报道的中国 EV 项目,搭载 DriveGPT | 不明确 / 信息冲突 | 若获证实,将是 Great Wall 之外的重要多元化 | Hyundai 2026 年官方材料点名 Momenta,而非 Haomo,可信度下降 |
各行把量产级证据和合作伙伴式提及分开,避免把标识出现过度解读为可持续收入。
[CU012, CU014, CU016, CU019, CU021, CU024]Haomo 的采用路径靠车型项目集成和路线运营推进,而不是客户自助上手。
[CU010, CU015, CU017, CU031, CU034]Great Wall 的量产证明最强,物流场景里 Wumart Dmall 最强;连续性可见度几乎处处偏弱。
[CU012, CU016, CU021, CU025, CU026, CU037]6.3 留存、扩张与集中度:扩张逻辑可见,但耐久性大多只能推断
Haomo 的公开客户记录支持一个显而易见的先落地再扩张逻辑,却几乎没有真正留存披露。ADAS 里,自然扩张路径是一项 OEM 项目扩到更多车型、更多配置和更长生命周期销量。配送里,扩张路径是一片区域或一条路线扩到更多社区、更多用例和更密订单吞吐。这些路径都出现在公开记录中,但投资者仍拿不到最关键的承销指标:NRR、GRR、流失率、合同期限、续约时点、客户级收入或按客户的附加率。耐久性最好的代理指标是 Great Wall 连续性,因为 Haomo 在 Great Wall 2025 年年报中仍是联营企业和关联方生态名称。但同一事实也制造集中风险。如果 Great Wall 承载大部分 ADAS 经济价值,而公开物流名单仍然狭窄,那么客户黏性与客户集中度会一起上升。2025 年末裁员和停工报道让问题更尖锐:即便历史客户验证是真的,如今的连续性和支持质量也不能再默认成立。[CU028, CU029, CU030, CU031, CU032, CU033]
| 指标 | 数值 / 状态 | 细分领域 | 置信度 | 尽调要求 |
|---|---|---|---|---|
| 净收入留存 | 未公开 | OEM 和物流客户 | 低 | 按主要客户分组和产品线提供 NRR |
| 总收入留存 / 流失 | 未公开 | OEM 和物流客户 | 低 | 列出已流失项目、暂停部署及原因 |
| 合同期限 / 续约周期 | 未公开 | OEM 项目 | 低 | 提供标准 ADAS 项目期限和续约里程碑 |
| 路线重复活动 | 可从 200,000+ / 300,000 笔配送订单里程碑看到 | 零售配送 | 中 | 展示路线级毛利率、人工介入率和续约历史 |
| 客户满意度 | 仅有间接信号;Dmall 在 2026 年称 AI 零售运营满意度 >95%,但并非专指 Haomo | 零售合作伙伴环境 | 低 | 提供客户归因的 Haomo 部署服务或配送满意度指标 |
| 2025 年末停摆后的当前支持连续性 | 不明确 | 全部细分领域 | 低 | 提供活跃客户名册、服务 SLA 和当前软件支持状态 |
这些 null 有意保留,因为截至尽调日期,公开渠道没有找到 Haomo 的 NRR、GRR、流失或续约时间表。
[CU028, CU030, CU031, CU032, CU038]| 扩张驱动 | 集中风险 | 影响 | 尽调路径 |
|---|---|---|---|
| 单一 OEM 项目扩展到更多车型 | Great Wall 可能主导战略价值和经济性 | 高上行空间可与极端客户集中并存 | 按 OEM 和车型映射收入、装车量和积压订单 |
| 单一超市区域扩展到更多社区 | 路线成功可能仍是地域性,而非全国性 | 运营证明增强,但收入广度可能仍有限 | 索取按城市、按客户拆分的部署数据 |
| 与 Meituan 和 Dada 的平台合作 | 大平台可施加价格压力和服务质量要求 | 战略客户标识未必转化为有吸引力的利润率 | 审查商业条款、排他性和硬件所有权 |
| 向 Hyundai 或其他 OEM 多元化 | 合作伙伴技术栈一变,公开多元化叙事就可能瓦解 | 多元化薄弱会让 Haomo 继续绑定母公司生态需求 | 获取已签署 SOP 时间表和当前集成状态 |
| 历史客户赢单延续到 2026 年 | 运营停摆可能同时冻结扩张和续约 | 客户集中风险转化为连续性风险 | 索取当前活跃客户名单和未履行支持义务 |
集中度在这里被视为商业模式的结构性问题,而不是单纯负面标题。
[CU034, CU035, CU036, CU037, CU039]6.4 客户结论:危机前验证真实,但多元化弱,且 2026 年新鲜度缺口很大
客户结论不是 Haomo 缺少市场采用,而是采用证据从未升级成清晰、多元、持续验证的客户基础。Great Wall 是唯一明确达到量产深度的关系。Wumart Dmall 及相邻配送伙伴提供了有说服力的证据,证明 Haomo 在北京路线中解决过真实末端问题。但公开记录在客户经济性上仍太薄,关系集中在少数客户上,2025 年运营危机后也太陈旧,无法支撑强耐久性结论。Hyundai 线索尤其说明问题:2024 年末一篇文章称 Haomo 已接入 Beijing Hyundai EV,但 Hyundai 自己 2026 年发布材料点名的是 Momenta。同样,当前 Dmall 仍被认可,只能证明这个零售交易对手仍在扩大 AI 运营,并不能证明 Haomo 仍驱动这些工作流。尽调中,客户牵引力应评为历史上已验证;当前连续性未解;多元化弱于 Haomo 标题客户名单给人的第一印象。[CU036, CU037, CU038, CU039]
| 未决事项 | 最新公开信号 | 重要性 | 严重性 | 下一步尽调 |
|---|---|---|---|---|
| 当前 Great Wall 项目清单 | 2025 年报道仍提到 20+ 款 Great Wall 车型搭载 | 需要知道哪些车型今天仍在线并获得支持 | 高 | 索取当前 SOP 车型名册和 OTA 支持矩阵 |
| 当前 Wumart / Dmall 车队状态 | 2024 年路线指标和 2026 年 Dmall 扩张存在,但没有当前 Haomo 提及 | 历史成功不能证明部署仍在持续 | 高 | 索取实时路线清单、车队数量和人工介入指标 |
| 当前 Meituan / Dada 部署份额 | 合作伙伴被点名,但缺少当前在线 Haomo 车辆数 | 没有在线份额,这些标识的战略价值难以判断 | 高 | 按合作伙伴获取已签署协议和活跃车辆数量 |
| Hyundai 多元化状态 | 2024 年末 Haomo 文章与 Hyundai 2026 年官方合作伙伴点名冲突 | 这是检验 Haomo 是否摆脱 Great Wall 依赖的重要测试 | 中 | 直接确认项目取消、替换或上线状态 |
| 客户级经济性 | 没有公开收入集中度、合同期限或利润率数据 | 缺少这些数据,无法对客户质量形成投资判断 | 高 | 索取客户级收入、毛利率和续约数据 |
这张表单独列出截至 2026 年 8 月阻碍高信心耐久性判断的具体缺失事实。
[CU032, CU033, CU036, CU038, CU039]6.5 图表与证据
07风险
7.1 监管与法律风险:中国现在要求 ADAS 和自动驾驶公司像安全关键制造商一样运作
对 Haomo 来说,最重要的背景变化是中国智能驾驶监管环境明显收紧,而这种收紧恰好发生在 Haomo 自身经营状况恶化之时。2025 年 2 月,MIIT 和 SAMR 通知收紧了搭载组合驾驶辅助系统的智能网联汽车在产品准入、召回、事故报告和 OTA 治理方面的要求。这些规则直接影响 Haomo,因为它的历史叙事依赖频繁软件迭代、不断演进的 NOA 能力和安全敏感的自动化主张。新框架要求更完整的测试验证、更清晰的系统边界、备案,以及当 OTA 变更修复安全缺陷时采取召回式处理。北京自动驾驶法规和 2026 年跨境汽车数据指南进一步说明,公路和数据治理预期在上升,不是在下降。随后 Xiaomi SU7 事故又加速了行业内对营销语言和功能推出谨慎度的执法压力。对一家已经陷入困境的公司,这很危险:重启不会发生在宽松政策窗口里,而是在比 Haomo 2023 和 2024 年扩张期更严格的安全、披露和合规制度下进行。[CR001, CR002, CR003, CR004, CR005, CR006]
| 规则 / 问题 | 管辖区 | 状态 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| MIIT/SAMR 对配备 CDAS 的 ICV 产品准入、召回和 OTA 通知要求 | 中国国家层面 | 2025 年 2 月起生效 | 高 | 高 | 低-中 | 高 —— 直接约束 Haomo 的 ADAS 和 OTA 更新 | 获取 Haomo 的准入申报、OTA 记录和缺陷响应文件 |
| 影响主要技术参数或自动驾驶功能的 OTA 变更需要许可 | 中国国家层面 | 已生效 | 高 | 高 | 低 | 高 —— 放慢快速推出,并抬高失败成本 | 索取公司的许可历史和待批产品变更申请 |
| 配备 CDAS 车辆的事故和事件报告义务 | 中国国家层面 | 已生效 | 中-高 | 高 | 低-中 | 高 —— 报告纪律薄弱会迅速放大监管风险 | 索取全部 2024–2026 年事件日志和监管往来函件 |
| 北京面向 L3+ 和配送场景的自动驾驶汽车条例 | 北京市 | 2025 年 4 月 1 日生效 | 中 | 中-高 | 中 | 中-高 —— 本地运营仍有条件且受监督 | 映射每一项北京测试、许可和安全员安排 |
| 2026 年汽车数据跨境传输指南和重要数据控制 | 中国国家层面 | 2026 年 2 月发布 | 中 | 高 | 低-中 | 中-高 —— 增加海外数据使用和跨境模型迭代复杂度 | 索取汽车数据目录、出境路径和安全评估状态 |
| 强制性 GB 44721-2026 L3/L4 安全标准 | 中国国家层面 | 2026 年 8 月发布,2027 年 7 月 1 日生效 | 中 | 高 | 低 | 高,适用于任何未来更高阶自动驾驶路线图 | 展示对照 GB 44721-2026 的产品差距分析及相关 DSSAD / 验证要求 |
| 裁员后的劳动、工资和竞业限制纠纷 | 中国劳动法 / 争议框架 | 2025 年扰动后仍在发生的风险 | 中-高 | 中-高 | 低 | 中-高 —— 会损害重启能力,并增加知识产权泄露风险 | 获取欠薪状态、社保状态和已签署竞业限制清单 |
各行按直接阻止部署、触发责任或挡住重启的能力排序,依据是当前公开事实。
[CR001, CR002, CR003, CR004, CR005, CR006]监管、偿付能力和客户支持义务交叠处,剩余风险最高。
[CR001, CR005, CR010, CR017, CR025, CR036]7.2 运营与客户连续性风险:2025 年末停摆把所有历史牵引力信号都变成支持问题
Haomo 的运营风险最好理解为连续性问题。危机前,公司有真实采用信号:20 多款车型、超过 120 million 辅助驾驶公里,以及北京零售路线中的大量配送订单。但公开报道一旦称停工、欠薪、数百名员工前景不明,每个历史验证点都会变成新的风险问题。谁还在修补软件?谁在响应事故升级?哪些客户仍获得支持?如果运营团队被削减,质保、路线运营或 ADAS 问题修复会怎样?客户章节已经说明,停摆前 Haomo 的基础就显得集中;集中会让中断更严重,因为太多已装基础可能落在少数交易对手手里。当前 Dmall 被认可、Meituan 车队规模大,只能证明交易对手仍有战略相关性,不能证明 Haomo 仍是活跃供应商。在安全敏感品类里,缺少当前支持证明本身就是重大风险事件。[CR015, CR016, CR017, CR018, CR019, CR020]
| 失效模式 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 未解缺口 |
|---|---|---|---|---|---|
| 运营停摆后,客户支持和 OTA 补丁失效 | 高 | 高 | 低 | 高 | 当前公开层面没有活跃支持团队、SLA 或更新节奏证明 |
| 人手减少削弱安全事件响应或召回执行 | 中-高 | 高 | 低 | 高 | 没有当前事件管理准备度的公开证据 |
| 历史配送路线不再运营,或缺少强支持仍在运营 | 中 | 中-高 | 低 | 中-高 | 当前 Wumart/Dmall/Meituan 实时部署状态未经验证 |
| 若 Great Wall 转向替代供应商,ADAS 存量装车信心受损 | 中-高 | 高 | 低-中 | 高 | 公开证据已显示 Great Wall 正用其他供应商补充 Haomo |
| 算力或工程资源不足,大模型路线图停滞 | 中 | 高 | 低 | 中-高 | 当前公开层面没有算力入口、训练节奏或留任模型团队证明 |
这里把支持失效和安全流程退化放在前面,因为即便代码库本身很强,这些风险也会落地。
[CR016, CR017, CR018, CR019, CR020, CR021]Haomo 的主要风险会相互叠加,而不是各自孤立。
[CR013, CR017, CR018, CR021, CR041, CR043]7.3 合作方、算力与人员风险:Haomo 依赖的关键对手方并不完全可控
Haomo 的依赖栈横跨客户、监管、芯片和人员。Great Wall 既是锚定客户,也是控制生态,给了 Haomo 历史可信度;但 Great Wall 的采购策略一旦变化,收入、数据获取和战略相关性可能同时受损。公开记录已经显示,Great Wall 正与其他智能驾驶供应商合作,Haomo 不能假设自己拥有排他地位。技术侧,Haomo 围绕 DriveGPT 的大模型叙事抬高了它对算力可得性和成本的暴露。Nvidia 2025 年针对中国的出口限制不是抽象的宏观噪音;它说明先进 AI 算力可能变得受限、延后,或明显变贵。对一家主打前沿模型自动驾驶的公司,这比只卖成熟低成本 L2 套件的公司更关键。人员层面同样严峻。裁员和运营停摆会加大高级工程师和产品负责人流失风险;中国 2025 年更明确的竞业限制指引也意味着,靠限制性契约留住技术员工并不顺滑。团队削弱、IP 防线可能被争议、算力路径不确定,合在一起就是一组高执行风险。[CR025, CR026, CR027, CR028, CR029, CR030]
| 依赖项 | 交易对手 | 角色 | 集中度 | 失效场景 | 严重性 | 缓释 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| 锚定 OEM 与控制方 | Great Wall Motor / Wei Jianjun 生态 | 客户、战略支持方、治理节点 | 极高 | Great Wall 重新分配采购或支持,Haomo 同时失去收入和可信度 | 高 | 除历史关系外,公开层面看不到其他缓释 | 高 |
| 配送平台入口 | Meituan、Wumart Dmall、Dada、Alibaba 相关路线 | 物流部署和路线需求 | 中-高 | 平台更换供应商,或不再续约试点式项目 | 中-高 | 只有历史路线证据;当前合同耐久性未体现 | 中-高 |
| 其他 OEM 多元化 | Hyundai / 未披露车企 | Great Wall 之外的潜在多元化 | 中 | 项目延期、被替换,或始终没跑到可持续 SOP 规模 | 中高 | 未看到有力的公开缓释措施 | 中高 |
| 高端 AI 算力获取 | NVIDIA / 更广泛的 AI 芯片供应链 | 支撑大模型训练和推理路线图 | 中 | 出口管制收紧或成本上升,会拖慢 DriveGPT 级路线图 | 中高 | 未披露公开的替代算力方案 | 中高 |
| 监管机构和数据主管部门 | MIIT、SAMR、北京主管部门、数据监管机构 | 准入、召回和数据流的闸门 | 高 | 文件缺失或不合规会拖慢重启或扩张 | 高 | 只能靠补齐文档和执行合规来缓释 | 高 |
表格把监管机构列为运营依赖,因为审批和报告义务会直接卡住产品部署。
[CR015, CR021, CR022, CR023, CR024, CR025]| 角色 / 职能 | 依赖或缺口 | 发生概率 | 严重程度 | 缓释成熟度 | 尽调路径 |
|---|---|---|---|---|---|
| 核心自动驾驶和 ADAS 工程团队 | 裁员和运营停摆可能削弱了维护安全关键产品所需的团队 | 高 | 高 | 低 | 索取当前组织架构、在岗工程人数和值守责任归属 |
| 高级技术人才留存 | 竞业限制执行成本更高;若补偿不到位,执行可能失效 | 中高 | 中高 | 中低 | 审阅竞业协议、付款状态和现有争议 |
| 客户支持和现场运营 | 配送线路和已装车 ADAS 车队需要训练有素的支持人员 | 高 | 高 | 低 | 索取工单日志、支持团队名单,以及按城市划分的现场服务覆盖 |
| 管理层重启可信度 | 2025 年底陷入困境后,未看到公开的重启计划、资本计划或治理动作 | 高 | 高 | 低 | 获取董事会决议、应急融资步骤和重启里程碑 |
| 战略聚焦 | 同时跑 ADAS、配送车和大模型研发,会挤压稀缺的管理层注意力 | 高 | 中高 | 低 | 要求管理层给业务排序,并明确关闭 / 延续选择 |
人员风险不只是留才问题;它会直接影响安全响应、客户信任和重启可行性。
[CR031, CR032, CR033, CR034, CR035, CR038]Great Wall、监管方、算力通路和关键员工是四个最重要的外部控制点。
[CR015, CR025, CR027, CR028, CR034, CR044]7.4 财务 / 模型风险与否决标准:缺少重启证据现在与战略同样重要
这里的财务风险不只是 Haomo 曾经资本密集。更关键的是,公开证据已经看不到一座从历史进展通向当前生存能力的财务桥。公开记录里没有 Haomo 2025 或 2026 年经审计财报,没有披露现金余额、债务清单、客户应收账款账龄,也没有可执行的 Great Wall 资金兜底。与此同时,Haomo 的模式仍然结构性吃资源:低成本 ADAS、无人配送车、前沿模型研发,各自以不同方式消耗现金和管理层注意力。ML Vehicle 报道 Little Camel 2025 年销售目标偏低、Hyundai 相关交付时间后移,说明停摆之前,公司已经很难把产品宽度转成规模化、稳定的商业执行。IC 结论因此很清楚。管理层证明运营已重启、现金仍在、客户活跃且合规准备到位之前,举证责任全部在公司。实际尽调中,这些缺失事实不是普通信息缺口,而是足以打断投资逻辑的触发器。[CR036, CR037, CR038, CR039, CR040, CR041]
| 模型 / 财务风险 | 当前信号 | 发生概率 | 严重程度 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|
| 未公开现金和资金续航 | 当前 Haomo 现金、债务或应收账款明细均未公开 | 高 | 高 | 高 | 获取最新银行余额、负债和 13 周现金流 |
| Great Wall 没有可执行兜底 | Great Wall 文件显示关联关系,但没有支持担保 | 中高 | 高 | 高 | 索取股东贷款协议、担保或支持函 |
| 双业务模式烧钱 | ADAS、配送车和大模型研发的资本消耗方式各不相同 | 高 | 高 | 高 | 提供分业务线损益表和资本分配计划 |
| 物流商业化动能弱 | ML Vehicle 报道,2025 年 Little Camel 目标仅 50 台,且没有新车型 | 中高 | 中高 | 中高 | 提供已签订单、管线和产品路线图 |
| OEM 多元化未必抵消集中度 | Hyundai 和未具名 OEM 的叙事仍薄弱或相互冲突 | 中 | 中高 | 中高 | 按 OEM 提供已签客户合同和 SOP 时间表 |
这张表把产品和客户事实转成承做变量,而不是重复泛泛的烧钱警告。
[CR022, CR036, CR037, CR038, CR039, CR040]| 风险 | 可监测触发项 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 运营重启风险 | 公司在明确领导层和支持覆盖下恢复正常运营 | 尽调结束前未验证重启 | 视为打破投资论点,不把复苏纳入承做假设 |
| 流动性风险 | 披露现金和负债 | 没有当前现金证明或负债明细 | 假设偿付能力仍未解决,并要求看到救助融资证据 |
| Great Wall 依赖 | Great Wall 确认仍提供商业支持或继续承担采购方角色 | 没有当前 Great Wall 支持声明,也看不到项目延续 | 假设集中度风险仍是关键问题 |
| 客户连续性风险 | 具名客户确认活跃部署和 SLA | Great Wall 或物流伙伴没有客户证明 | 大幅折价历史牵引力 |
| 监管准备度 | 提供准入、OTA 和事故报告材料 | 合规文件缺失或监管问题未结 | 假设重启或推出会被拖延 |
| 人员 / IP 风险 | 核心工程人员留任,或在有效保护下合法转移 | 关键团队离职,加上未付义务或有争议的竞业限制 | 假设复苏更慢,IP 外泄风险更高 |
否决标准刻意设得具体,让投委会能区分可修复的不透明和真正打破论点的失败。
[CR041, CR042, CR043, CR044]7.5 要点展示
08估值
8.1 建议:除非入场价格明确按困境资产定价,并有最新尽调支撑,否则回避
当前正确判断是回避。不是因为 Haomo 从未做出真实产品,而是因为公开证据已经无法支撑把它按运营中的成长公司承销。Haomo 确实拿过有意义的里程碑:2021 年跻身独角兽,2024 年继续获得 B1 和 B2 轮融资,落地到 20 多款乘用车车型,末端配送量也达到数十万件包裹或订单。换一个背景,这些事实足以支持观察建议。但到 2025 年末,故事已经从商业化转向连续性。多家媒体报道称,公司停工、工资欠发,且账户可能被冻结。Great Wall 2025 年年报仍把 Haomo 列为联营公司并确认累计亏损,但公开披露没有把这一点连接到重启、再资本化,或干净的剩余股权叙事。这个缺口比市场规模或产品野心更重要。投资人要定价的不只是自动驾驶软件的可选性;还要判断底层运营实体是否仍有足够团队、流动性、合规准备和客户支持能力,值得继续讨论。没有新尽调证明这些点,唯一站得住的立场就是:相对旧独角兽叙事,假设大幅减值。[CV001, CV002, CV004, CV005, CV018, CV019]
| 决策字段 | 当前观点 | 决策含义 |
|---|---|---|
| 投资建议 | 避免 | 公司现在更像困境承做问题,不是常规成长期自动驾驶投资。 |
| 置信度 | 中等 | 减值方向明确,但当前资本结构和负债没有公开披露。 |
| 风险评级 | 极高 | 股权价值主要取决于偿付能力、重启执行和出资方行为。 |
| 估值立场 | 偏贵 | 任何贴近 2021-2024 年独角兽叙事的价格,都明显跑在当前公开证据之前。 |
| 入场纪律 | 只有低于困境水平且完成新一轮尽调后,才重新打开 | 投资人承做上行前,需要看到最新资产负债表、客户、监管和救助结构证据。 |
这并不是说 Haomo 没有创造过价值,而是说历史价值证明和当前可投资价值已经是两个问题。
[CV026, CV034, CV035, CV036, CV037, CV048]| 论点 | 方向 | 什么会改变观点 |
|---|---|---|
| Haomo 在乘用车 ADAS 和低速配送上确有历史商业化证明,也拿到过可信机构投资人支持。 | 支持论点 | 如果能证明这些历史资产仍有完整团队、活跃客户和支持义务,投资论点会更强。 |
| Great Wall 仍是最可能的出资方或救助路径,因为它站在控制权、客户历史和战略背景的中心。 | 支持论点 | 若披露资金兜底、资产收购或重组框架,可承做性会明显改善。 |
| Meituan、Wumart/Dmall 和更广泛的末端自动化需求说明,即便 Haomo 变弱,终端市场并未消失。 | 支持论点 | 具名 2026 年在运项目若能直接绑定 Haomo,将抬高资产价值底线。 |
| 2025 年底停摆后,每个历史证明点都变成连续性问题,不再是估值溢价。 | 反论点 | 经审计或由管理层认证的重启更新会降低这项折价。 |
| 已审阅材料中没有公开的 2025-2026 年 Haomo 财务报表、现金余额、债务明细或股本桥。 | 反论点 | 当前财务报表和股权结构披露,会把公司从不透明变成可定价。 |
| 中国针对 OTA、CDAS、数据和 L3/L4 安全的监管制度,比 Haomo 最初放量时更严。 | 反论点 | 监管就绪流程和许可的证据,会提高可行重启的概率。 |
投资论点仍指向非零战略可选性;反论点解释了为什么今天不能按正常运营的软件公司给这份可选性定价。
[CV018, CV019, CV021, CV032, CV033, CV038]历史证明仍在,但放到当下,运营困境、不透明和对支持方的依赖把结论推向回避。
[CV001, CV004, CV005, CV018, CV033, CV034]8.2 关键估值事实,是历史独角兽标记与当前困境信号之间的断裂
Haomo 的历史价格锚很容易描述,但已经不能安全复用。2021 年 Series A 以超过 $1 billion 估值塑造了独角兽叙事,公司还在 2024 年 B1 和 B2 轮合计融到超过 CNY400 million。若公司随后证明 2025 年收入稳定增长、客户连续性仍在,且流动性有出资方支撑,这些旧估值仍有意义。公开记录却指向相反方向。到 2025 年中,Hyundai 多元化不足、低速车目标下调和组织动荡,已经暴露商业化问题。到 2025 年 11 月,多篇报道描述了停摆、欠薪和运营瘫痪。Great Wall 仍是实际控制生态,年报仍记着 Haomo 相关亏损;但同一份文件也强调 Great Wall 自己的 Coffee Pilot、VLA 和世界模型路径,而不是简单的 Haomo 救援叙事。合在一起,历史定价已经过期。历史投资人投的是一家正在扩张的自动驾驶供应商;当前投资人投的会是重组、救援或资产保全故事。这是两种根本不同的估值问题,二者之间应有很大的折价。[CV001, CV002, CV004, CV006, CV007, CV008]
8.3 公开可比公司给的是上限,不是估值捷径
Pony.ai、WeRide 和 Mobileye 有参考价值,但主要证明 Haomo 缺了什么。2026 年 8 月,Pony.ai 和 WeRide 即便收入不高,仍有数十亿美元市值,因为市场至少看得到持续运营、经审计或申报文件支撑的披露、充足现金余额和仍在推进的商业化动能。Mobileye 的绝对估值高得多,因为它同时具备规模、现金生成能力和成熟的 ADAS 装车基础。Haomo 没有可比的披露栈。没有公开 2025 或 2026 年收入报表,没有披露现金余额,没有可见的股权结构更新,没有清晰支持兜底,也没有已验证的重启公告。这意味着公开可比公司测算只能用来压住乐观假设。若有运营、有资本支持且有真实财务披露的 AV 同行在公开市场约值 $1.9 billion 到 $6.8 billion,那么连续性未解决的困境私营公司应该相对它们极大折价,而不是享受叙事溢价。CompaniesMarketCap 和 SEC 关联申报目录能让投资人直接找到同行年报,反而凸显了对比:Haomo 的问题不是缺少行业可比公司,而是缺少当前公司层面的可观测性。[CV009, CV010, CV011, CV012, CV013, CV014]
| 可比对象 | 指标 | 倍数 / 估值 / 状态 | 相关性 | 局限 |
|---|---|---|---|---|
| Pony.ai | 中国公开市场自动驾驶可比公司 | ~$3.39B 市值;FY2025 收入 $90.0M;毛利率 15.7%;现金 / 投资 $1.515B | 最接近的证据:资本市场仍会给有正常披露、仍在运营的中国自动驾驶平台提供资金。 | 仍是一家运营中、在扩张的上市公司,有近似审计口径披露,融资通道远好于 Haomo。 |
| WeRide | 中国公开市场自动驾驶可比公司 | ~$1.92B 市值;FY2025 收入 $97.9M;毛利率 30.2%;现金 / 投资约 $1.0B | 对仍亏损但完整运营的自动驾驶公司,这是有用的低端公开可比。 | Robotaxi 占比较高,且已上市;它仍比 Haomo 健康、透明得多。 |
| Mobileye | 成熟上市 ADAS 基准 | ~$6.77B 市值;FY2025 收入 $1.894B;现金 $1.8B;经营现金流 $602M | 当收入、现金创造和装车成熟度可见时,真正规模化的 ADAS 基准应是什么样。 | 规模大得多、成熟得多,结构上也不同于 Haomo。 |
| Haomo 历史私募估值 | 2021 年底独角兽 / 2024 年融资叙事 | 历史估值信号 >$1B;2024 年 B1 和 B2 合计融资 >CNY400M | 只能作为过时的上限锚,用来衡量故事已经偏移多远。 | 反映困境前的融资条件,不能当作当前公允价值。 |
| Haomo 推断的 2026 年基准情形 | 困境可选性 / 救助情形 | $80M-$200M 推断当前公允区间 | 比正常收入倍数法更贴合证据组合,因为连续性、偿付能力和救助经济条款才是主导因素。 | 这是估算值,不是观察到的价格;对当前未公开的负债和出资方条款高度敏感。 |
可比公司组合刻意偏向天花板:用途是防止买贵,而不是为直接套用正常运营公司倍数到 Haomo 找理由。
[CV009, CV010, CV011, CV012, CV013, CV014]Haomo 在历史证明上的得分高于当下可投性。
分数是基于公开证据集综合出的 0-10 序数判断,仅供投资委员会讨论。
[CV018, CV020, CV021, CV033, CV035, CV043]8.4 情景估值应按困境可选性框定,而不是精确定价
当前财务缺失,估值只能按情景做。悲观情景不是正常的倍数压缩,而是清算或资产出售:工资债权、负债、优先权和流失的客户信心会让普通股几乎没有或完全没有剩余。基准情景是假设一次受控救援,最可能由 Great Wall 或另一家战略买方推动,部分工程资产、客户关系和支持义务保留下来。这个情景可以支撑非零股权价值,但相对历史独角兽标记仍要大幅折价,因为公司是在更严监管下从困境中重启。乐观情景不只是自动驾驶情绪改善。它需要有文件证明的重启、新资本、核心客户仍在且得到支持的证据,以及 OTA、数据和安全合规栈已经准备好重启。只有这样,Haomo 才可能从困境可选性重新走向后期成长估值。即使在这个更好情景下,今天的公开证据也只支持远低于旧的标志性估值的价值,因为连续性断裂已经发生。[CV022, CV023, CV027, CV028, CV029, CV030]
| 情景 | 概率信号 | 假设 | 估值 / 回报逻辑 | 关键风险 |
|---|---|---|---|---|
| 悲观 | 45% | 停工最终证明不可逆,负债和优先权吸收价值,只能零散出售 IP 或合同资产。 | $0M-$50M 股权价值;实际是清算或残值回收结果,普通股几乎没有剩余。 | 未知负债、客户流失、团队流失,以及没有出资方救助。 |
| 基准 | 40% | Great Wall 或其他战略方设计受控重启或资产保全交易,保住部分团队、代码和支持义务。 | $80M-$200M 股权价值;相对独角兽估值大幅折价,反映战略可选性,但不再是干净的成长型公司。 | 稀释、隐藏负债、客户连续性有限,以及监管重新合格进展慢。 |
| 乐观 | 15% | Haomo 证明已实际重启、融到新资金、客户支持连续,并在更严的 2025-2026 年规则下具备合规准备度。 | $250M-$450M 股权价值;仍远低于旧独角兽估值,因为业务是从困境中重建。 | 重启仍失败、客户不回归,或救助经济条款让存量股权居于劣后。 |
区间是供委员会讨论的美元股权情景结果,不是有文件证明的交易价格。
[CV027, CV028, CV029, CV030, CV031, CV043]估值对重启证明和负债更敏感,行业 TAM 叙事的影响小得多。
数值是基于情景的十亿美元估值,用来展示对证据里程碑的敏感性,不是管理层指引。
[CV022, CV023, CV029, CV030, CV031, CV043]当前最站得住脚的结果大多远低于历史独角兽叙事。
区间是基于情景的十亿美元估值,仅用于讨论当下股权价值。
[CV027, CV028, CV029, CV030, CV031]8.5 什么会改变建议:偿付能力、支持连续性和出资方经济性的证据
建议可以改变,但只有具体证据补上最大的承销缺口。第一,需要当前资产负债表包:现金、债务、应付款、欠薪、或有负债,以及任何诉讼或账户冻结历史。第二,是客户连续性:Great Wall、Wumart/Dmall、Meituan、Dada 或其他项目哪些仍在运行,今天由谁服务。第三,是治理和资本结构:Great Wall 或其他出资方是否有可执行的救援义务,普通股前面压着哪些优先权,少数股东在经济上是否已经没有价值。第四,是更严格 2025-2026 年制度下的监管准备:OTA 控制、事件响应、产品准入状态和汽车数据合规的证据。若这些材料证明公司有资金支持的重启、活跃客户和干净救援结构,公允区间可从困境水平大幅上移。在此之前,本章纪律很简单:不要把历史商业化证据和当前可投资性混为一谈。[CV021, CV030, CV034, CV035, CV036, CV037]
| 触发项 | 阈值 | 对投资论点的传导 | 行动含义 |
|---|---|---|---|
| 没有重启证据 | 到下一轮尽调周期,仍没有经验证的运营重启、融资事件或客户支持声明 | 打破救助可选性框架,把案例推向清算逻辑。 | 估值转向悲观情形,避免任何历史估值锚定。 |
| 负债悬顶高于预期 | 出现实质性的欠薪、诉讼、债务或账户冻结影响 | 即使 IP 或客户合同保留部分企业价值,剩余股权也会缩水。 | 假设普通股深度劣后,或实际已受损。 |
| Great Wall 不提供救助经济条款 | 控制生态没有绑定支持、资产收购或重组框架 | 移除最可能通向存续的战略桥梁。 | 把基准情形视为过于乐观,并放大下行情景概率。 |
| 客户连续性失败 | Great Wall、Dmall/Wumart、Meituan 或其他具名项目不活跃,或没有支持 | 削弱“存量部署义务能形成可防守资产底线”的论点。 | 大幅下调基准和乐观区间。 |
| 合规栈不具备重启准备 | 新规下没有 OTA、事故或汽车数据合规证据 | 再次商业化所需成本和时间都会上升。 | 下调乐观情形概率,并维持避免立场。 |
这些是决策触发项,不是泛泛的观察清单。
[CV021, CV030, CV033, CV040, CV041, CV042]| 主题 | 缺失证据 | 重要性 | 负责人 / 尽调路径 |
|---|---|---|---|
| 当前偿付能力 | 现金、债务、应付款、欠薪、账户冻结状态和或有负债 | 决定投资人承做的是救助、重组,还是剩余股权接近归零的案例。 | 索取当前管理层资料包和法律 / 财务尽调备忘录。 |
| 客户连续性 | 仍在运的 OEM 和配送项目签字确认,以及当前支持责任 | 区分历史牵引力和当前可变现关系。 | 索取客户函、SLA 时间表和当前部署地图。 |
| 股权结构和优先权 | 当前完全摊薄所有权、清算优先权和出资方权利 | 困境救助中,剩余股权可能与资产价值大幅背离。 | 审阅融资文件和股东协议。 |
| Great Wall 经济条款 | Great Wall 是否提供过桥资金、资产购买意向或其他可执行支持 | Great Wall 是最可能的救助路径,也是主要集中度节点。 | 索取董事会材料、关联方协议和出资方往来函件。 |
| 合规准备度 | OTA 许可历史、事故响应流程、产品准入记录和数据传输合规状态 | 没有安全和数据证据,2025-2026 年制度下的重启不可信。 | 索取监管工作底稿和内部合规负责人。 |
| 人员留存 | 停摆后关键工程、产品和支持人员状态 | 如果执行核心已经散掉,IP 价值会很快下滑。 | 索取组织架构、留任计划和已签署的雇佣状态清单。 |
这些问题按推进速度排序:哪些证据最能快速把 Haomo 从无法定价的困境,拉回边界清晰的战略机会案例。
[CV040, CV041, CV042, CV043, CV044, CV048]8.6 要点展示
免责声明
本报告仅供信息参考,不构成投资建议。它完全基于截至 2026-08-05 已审阅的公开材料。若干决策关键事项——包括当前偿付能力、 客户连续性、股权结构经济账和重启状态——在公开记录中仍未披露;一旦这些事项得到澄清,评估可能发生重大变化。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Haomo became an independent company in November 2019 after Great Wall Motor separated out its intelligent-driving unit. | 中 | SO003, SO007 |
| CO002 | Public profiles and media coverage consistently place Haomo’s headquarters in Beijing, China. | 中 | SO008, SO010 |
| CO003 | Yicai reported in November 2025 that Great Wall Motor chairman Wei Jianjun was Haomo’s actual controller and held a 37 percent stake. | 中 | SO008 |
| CO004 | Haomo’s December 2021 Series A raised nearly RMB 1 billion from GL Ventures, Meituan, Shoucheng Holdings, Qualcomm Ventures, and JZ Capital. | 中 | SO003 |
| CO005 | TechCrunch reported that Haomo raised RMB 100 million, or about US$14 million, in the first tranche of its Series B funding in February 2024, led by Chengdu Wufa. | 中 | SO004 |
| CO006 | The late-2021 financing established Haomo as a unicorn, implying a valuation above US$1 billion. | 中 | SO003, SO008 |
| CO007 | Public database coverage of Haomo’s total disclosed funding is inconsistent, with CB Insights showing Series B-II status and about US$245.3 million total raised while TechCrunch cited more than US$200 million of equity funding. | 中 | SO004, SO010 |
| CO008 | Haomo’s two core business lines were passenger-vehicle assisted driving systems and low-speed autonomous last-mile logistics vehicles. | 中 | SO005, SO009 |
| CO009 | Haomo publicly framed HPilot, NOH, memory-driving functions, and lower-cost ADS kits as modular passenger-car products that could be offered to external automakers under an open collaboration model. | 中 | SO001, SO002 |
| CO010 | At the end of 2021, Haomo expected its assisted-driving system to be available on 34 Great Wall passenger vehicle models by the end of 2022. | 中 | SO003 |
| CO011 | By February 2024, TechCrunch reported that Haomo’s ADAS solutions had powered more than 20 vehicle models and accumulated about 120 million kilometers of driving. | 中 | SO004 |
| CO012 | Friendly Auto reported that Haomo-powered vehicles had logged over 141 million kilometers of assisted driving by May 2024 and that Little Camel products had entered nine use cases. | 中 | SO009 |
| CO013 | Haomo’s April 2023 official DriveGPT release said the company had become a supplier for three automakers and that HPilot-equipped vehicles had already entered the EU and Israel. | 中 | SO001 |
| CO014 | Yicai reported that Haomo had nearly 300 employees across Beijing, Shanghai, Shenzhen, and Baoding before the November 2025 halt. | 中 | SO008 |
| CO015 | KrASIA’s June 2025 reporting said Haomo’s passenger-vehicle segment had only secured projects on two Hyundai models and was struggling to deliver them on time. | 中 | SO005 |
| CO016 | Great Wall Motor’s 2025 annual report identified Haomo Zhixing as an associate of the group and also as a related-party entity controlled by the ultimate controlling shareholder. | 中 | SO011 |
| CO017 | Great Wall’s 2025 annual report recorded RMB61.7 million of unrecognized current-year loss share for Haomo and RMB413.7 million of accumulated loss share. | 中 | SO011 |
| CO018 | Mid-2025 public reporting said Haomo chairman Zhang Kai and senior product and technology executives left the company during the year. | 中 | SO005, SO009 |
| CO019 | KrASIA reported that Haomo had made minimal commercialization progress in both its passenger-car and logistics business lines by mid-2025. | 中 | SO005 |
| CO020 | KrASIA reported that Haomo’s 2025 sales target for low-speed delivery vehicles was only about 50 units and was mainly focused on clearing inventory rather than scaling new models. | 中 | SO005 |
| CO021 | Multiple late-2025 outlets reported that Haomo told employees not to report to work starting November 24, 2025, indicating an abrupt halt of operations. | 高 | SO006, SO007, SO008 |
| CO022 | Late-2025 reporting said Haomo employees faced salary arrears and that internal communication channels were muted or disabled after the work-stop notice. | 高 | SO006, SO007, SO008 |
| CO023 | Yicai reported that employees believed litigation tied to unmet performance targets had frozen Haomo’s accounts, although the company had not publicly confirmed the cause. | 中 | SO008 |
| CO024 | KrASIA reported that Great Wall shifted advanced driving programs to Momenta in 2024 and used DJI solutions for midtier and entry-level brands while Haomo’s remaining passenger-vehicle business narrowed. | 中 | SO006 |
| CO025 | Great Wall’s 2025 annual report emphasized in-house Coffee Pilot 3 and Coffee Pilot 4 plus VLA and world-model development as the group’s main intelligent-driving path. | 中 | SO011 |
| CO026 | Haomo publicly launched DriveGPT at its eighth AI Day in April 2023 and described it as an autonomous-driving generative foundation model. | 中 | SO001, SO013 |
| CO027 | Haomo’s April 2023 official release said the new WEY Mocca DHT-PHEV would be the first mass-produced vehicle powered by DriveGPT. | 中 | SO001, SO013 |
| CO028 | Haomo’s October 2023 AI Day release priced HP170 at RMB3,000-3,999, HP370 at RMB5,000-5,999, and HP570 at RMB8,000-8,999. | 中 | SO002, SO014 |
| CO029 | Haomo’s October 2023 release said DriveGPT had screened 10 billion internet-image frames and 4.8 million autonomous-driving 4D clips 200 days after launch. | 中 | SO002 |
| CO030 | Logistics Tech Outlook said MANA OASIS provided 670,000 trillion floating-point operations per second as Haomo’s compute center for autonomous-driving model training. | 中 | SO012 |
| CO031 | Haomo’s HDeliver or Little Camel line was positioned as a medium-sized autonomous terminal-logistics vehicle priced below RMB90,000 in 2023 coverage. | 中 | SO012 |
| CO032 | Logistics Tech Outlook said Haomo’s HDeliver vehicles had exceeded 200,000 cumulative distribution orders by July 2023 and were used in collaborations with Meituan, Wumart Dmall, and Dada Nexus. | 中 | SO012 |
| CO033 | Across public reporting, Great Wall appears as both Haomo’s largest strategic backer and its main early passenger-vehicle commercialization channel. | 中 | SO003, SO004, SO011 |
| CO034 | By the August 2026 run date, the retained public source pack did not contain a formal Haomo or Great Wall statement clearly confirming restart, restructuring, or liquidation after the November 2025 halt. | 中 | SO006, SO007, SO008 |
| CO035 | Because continuity is unresolved and Great Wall has visibly developed alternative intelligent-driving paths, any current investment view on Haomo depends more on rescue or absorption mechanics than on stand-alone growth metrics. | 中 | SO006, SO011 |
| CO036 | Research and Markets characterized 2026 as a qualitative inflection point for China intelligent driving rather than a period of stalled demand. | 中 | SO020 |
| CO037 | Yahoo Finance syndication of the 2026 OEM ADAS report said high-level intelligent driving among independent OEMs doubled to 62.7% by 2025 and urban L2.9 penetration reached 17.2%. | 中 | SO021 |
| CO038 | Grand View Research estimated the global autonomous last-mile-delivery market at about US$1.6 billion in 2024 and projected it to reach about US$5.9 billion by 2030. | 中 | SO022 |
| CO039 | DataM estimated China’s autonomous last-mile-delivery market at about US$6.05 billion in 2026, underscoring that Haomo’s logistics opportunity remained structurally large even as the company faltered. | 中 | SO023 |
| CO040 | The Automotive Data reported that L2.5 and L2.9-equipped new vehicles in China reached 30.2% and 34.82% respectively by the first four months of 2025, showing why low-cost ADAS scale mattered strategically. | 中 | SO025 |
| CM001 | Haomo’s opportunity was split between passenger-vehicle intelligent driving and low-speed autonomous last-mile delivery rather than a single unified AV market. | 中 | SM010, SM025 |
| CM002 | Passenger-vehicle intelligent driving economics are driven by OEM platform wins, feature penetration, and mass-production validation rather than by robotaxi-style utilization. | 中 | SM001, SM003 |
| CM003 | Low-speed autonomous logistics depends on route density, operator economics, and city permissions more than on headline autonomous-driving capability alone. | 中 | SM009, SM024 |
| CM004 | Haomo’s own product mix mapped to those two markets through HPilot for passenger cars and HDeliver or Little Camel for logistics. | 中 | SM011, SM010 |
| CM005 | Embodied or physical-AI driving models were relevant to Haomo only insofar as they improved intelligent-driving products, not as a separate addressable AI market. | 中 | SM001, SM015 |
| CM006 | OEM buyers and logistics buyers use different budget owners and adoption triggers, increasing Haomo’s go-to-market complexity. | 中 | SM001, SM009 |
| CM007 | Mixing robotaxi TAM with Haomo’s assisted-driving and delivery-vehicle markets would overstate the company’s practical addressable opportunity. | 中 | SM001, SM008 |
| CM008 | Research and Markets framed 2026 as a key inflection point for China intelligent driving, moving from quantitative growth to qualitative capability change. | 中 | SM001 |
| CM009 | The 2026 OEM ADAS market summaries say non-intelligent and low-level configurations kept retreating while higher-level assisted-driving functions became the growth center. | 中 | SM001, SM002 |
| CM010 | Independent OEMs in China reached 62.7% high-level intelligent-driving mix by 2025 according to the Research in China report syndication. | 中 | SM002 |
| CM011 | Urban L2.9 NOA penetration among independent OEMs rose from 2.1% in 2022 to 17.2% in 2025. | 中 | SM002 |
| CM012 | The Automotive Data reported that L2.5-equipped new models reached 30.2% and L2.9-equipped models reached 34.82% by the first four months of 2025. | 中 | SM003 |
| CM013 | Gasgoo said cost optimization and large-scale production capabilities had become critical as ADAS moved into the mass market. | 中 | SM004 |
| CM014 | DVN reported that LiDAR installations in China’s passenger-vehicle market exceeded 1.3 million units in the first four months of 2026. | 中 | SM005 |
| CM015 | DVN reported that Hesai and Huawei together accounted for 65.5% of automotive LiDAR installations in China in the first four months of 2026. | 中 | SM005 |
| CM016 | Robotics and Automation News summarized IDTechEx research saying software-only ADAS startups were under increasing pressure as automakers prioritized cost-effective full-stack partnerships. | 中 | SM006 |
| CM017 | The same IDTechEx-linked summary said new contracts signed by software-only vendors dropped 37 percent year over year. | 中 | SM006 |
| CM018 | Research and Markets said multiple technical routes including VLA and world models were developing in parallel in China’s 2026 intelligent-driving market. | 中 | SM001 |
| CM019 | These market shifts increased the value of integrated chip, software, data, and validation stacks versus narrower feature vendors. | 中 | SM003, SM006 |
| CM020 | Grand View Research estimated the global autonomous last-mile-delivery market at about US$1.615 billion in 2024 and about US$5.93 billion by 2030. | 中 | SM008 |
| CM021 | Grand View Research said ground delivery vehicles accounted for more than 82% of autonomous last-mile-delivery market share in 2024. | 中 | SM008 |
| CM022 | DataM estimated China’s autonomous last-mile-delivery market at about US$6.05 billion in 2026 and about US$25.98 billion by 2035. | 中 | SM009 |
| CM023 | DataM described the China autonomous last-mile-delivery market as entering a critical investment-timing phase rather than having already reached scaled profitability. | 中 | SM009 |
| CM024 | DataM said leading adoption zones for autonomous last-mile delivery in China include tier-1 cities, smart-city pilot zones, campuses, industrial parks, and high-density residential communities. | 中 | SM009 |
| CM025 | DataM identified vehicle utilization, remote-operator ratio, battery economics, and regulatory approval as key variables for scaling autonomous-delivery fleets. | 中 | SM009 |
| CM026 | Logistics Tech Outlook described Haomo’s HDeliver deployments across nine use-case environments such as shopping malls, smart communities, campuses, retail, and courier services. | 中 | SM010 |
| CM027 | Logistics Tech Outlook and DataM together imply that Haomo’s logistics strength was most relevant in dense, controlled, short-route environments rather than in unconstrained urban delivery at national scale. | 中 | SM009, SM010 |
| CM028 | The Haomo logistics story therefore fit the right vehicle format but still depended on utilization and operations proof that public sources do not fully disclose. | 中 | SM009, SM010 |
| CM029 | Grand View and DataM both show that autonomous-delivery demand remains structurally real even though near-term monetization is still operationally constrained. | 中 | SM008, SM009 |
| CM030 | The underlying China ADAS market remained attractive in 2026 even though that did not guarantee that any specific supplier would capture share. | 中 | SM001, SM006 |
| CM031 | Haomo’s pricing strategy for HP170, HP370, and HP570 shows it was aiming at the mass-market assisted-driving opportunity rather than only premium flagships. | 中 | SM011, SM012 |
| CM032 | The rise of full-stack ecosystems made Haomo less well positioned than players such as Horizon, WeRide, Pony.ai, and Mobileye that publicly present broader platform scope or larger-scale product families. | 中 | SM014, SM015, SM016, SM017 |
| CM033 | KrASIA and Friendly Auto both suggest Haomo’s commercial position worsened just as the market became more demanding on delivery, scale, and ecosystem breadth. | 中 | SM019, SM020 |
| CM034 | Public evidence does not provide a clean current Haomo market-share figure in either passenger ADAS or autonomous delivery. | 低 | |
| CM035 | Haomo’s feasible passenger-vehicle SAM was materially smaller than total China ADAS demand because Great Wall remained its central visible OEM channel. | 中 | SM013, SM019 |
| CM036 | Autonomous-delivery market ceilings are much larger than Haomo’s practical SAM because only a subset of routes, cities, and customers can support profitable autonomous ground-vehicle operations today. | 中 | SM008, SM009 |
| CM037 | Category growth alone did not protect vendors from failure in 2025-2026 because commercialization discipline and ecosystem scale mattered more than narrative. | 中 | SM006, SM019, SM020 |
| CM038 | Haomo’s addressable share depended heavily on Great Wall staying committed as both a customer and ecosystem sponsor. | 中 | SM013, SM019, SM025 |
| CM039 | Once Haomo’s commercialization delays surfaced, the market’s shift to scale and full-stack integration likely reduced, rather than expanded, its realistic path to winning new OEM share. | 中 | SM006, SM019, SM020 |
| CM040 | A positive Haomo market-share thesis would have required broader OEM diversification, lower-cost deployment economics, and live route-level logistics scale that the retained public record does not demonstrate. | 中 | SM009, SM019, SM020 |
| CP001 | Haomo competed across passenger-vehicle ADAS and low-speed autonomous logistics rather than in one single product category. | 高 | SP001, SP002, SP025 |
| CP002 | Haomo’s most direct passenger-vehicle rivals by 2026 were scaled Chinese third-party intelligent-driving suppliers such as Horizon, Momenta, Huawei-linked ecosystems, and emerging peers like DeepRoute. | 中 | SP018, SP021, SP024 |
| CP003 | Pony.ai and WeRide were relevant autonomy benchmarks for Haomo because their fleet and logistics-adjacent operations gave them broader data and commercialization proof, while Mobileye was the global incumbent benchmark for embedded ADAS scale. | 中 | SP012, SP017, SP018 |
| CP004 | Great Wall’s public shift toward Coffee Pilot and related in-house smart-driving programs made internal build a more dangerous substitute for Haomo than it had been during Haomo’s earlier growth phase. | 中 | SP003, SP004, SP005 |
| CP005 | Haomo’s dual-market strategy forced it to compete in both OEM sourcing cycles and logistics-operations economics at the same time. | 中 | SP001, SP002, SP025 |
| CP006 | A realistic competitive map for Haomo must include substitutes such as OEM self-build, incumbent ADAS platforms, and human-led delivery operations rather than only AI startups. | 中 | SP018, SP021, SP024 |
| CP007 | Horizon’s HSD is a software-hardware integrated full-scenario assisted-driving system built on the Journey 6 platform. | 高 | SP007, SP008 |
| CP008 | Horizon publicly segments HSD around RMB100k+, RMB150k+, and RMB200k+ vehicle bands, showing broad mass-market packaging discipline. | 高 | SP007, SP008 |
| CP009 | Horizon says HSD has secured design wins with 10 OEM brands across more than 20 vehicle models. | 高 | SP007, SP008 |
| CP010 | Journey 6 emphasizes integrated compute, tooling, and faster mass-production deployment rather than stand-alone algorithm licensing. | 高 | SP007, SP008 |
| CP011 | Honda and Momenta announced a partnership to build mass-production-ready end-to-end assisted-driving systems for China. | 高 | SP009, SP011 |
| CP012 | Momenta positions itself as a mass-production flywheel company that unifies perception and planning in an end-to-end assisted-driving solution. | 中 | SP009, SP010 |
| CP013 | Third-party 2026 coverage places Momenta among the first tier of Chinese intelligent-driving suppliers rather than among fringe challengers. | 中 | SP010, SP011, SP024 |
| CP014 | According to 36Kr coverage citing Huawei data, Huawei Qiankun ADS had accumulated 10.47 billion assisted-driving kilometers and 77% share among domestic third-party intelligent-driving suppliers by May 2026. | 中 | SP010 |
| CP015 | Huawei’s competitive strength comes from multi-sensor redundancy and local-road data accumulation, but its hardware-heavy configuration can be less natural for the lowest mass-market price tiers. | 中 | SP010, SP021 |
| CP016 | Horizon and Huawei benefit from chip-plus-software integration that Haomo did not publicly demonstrate with the same clarity. | 中 | SP007, SP008, SP021, SP024 |
| CP017 | The China ADAS market has been tilting away from software-only vendors and toward full-stack providers with integrated hardware, software, and mass-production support. | 高 | SP008, SP021, SP022, SP024 |
| CP018 | Mobileye remained the global incumbent benchmark for embedded ADAS scale, with more than 230 million vehicles built with EyeQ through 2025. | 中 | SP012 |
| CP019 | Mobileye says its REM crowdsourcing stream covers more than 95% of public roads in the United States and Europe and used data from over 8 million vehicles covering 34 billion miles last year. | 中 | SP012 |
| CP020 | Mobileye’s main competitive advantage relative to Haomo is global installed-base data, mapping, and a vertically integrated ADAS ladder rather than China-specific low-speed-delivery experience. | 中 | SP012, SP021 |
| CP021 | Pony.ai publicly spans robotaxi, robotruck, and personally owned vehicle smart-driving product lines on one business surface. | 中 | SP013 |
| CP022 | Pony.ai says its Toyota China and GAC Toyota joint venture involved combined investment exceeding RMB1 billion to scale Level 4 robotaxis. | 中 | SP013 |
| CP023 | Pony.ai says its three smart-driving product lines for personally owned vehicles had received orders and entered mass production. | 中 | SP013 |
| CP024 | WeRide One is described as a unified platform spanning L2 to L4 across mobility, logistics, and sanitation. | 中 | SP014 |
| CP025 | WeRide says its shared operations platform improves deployment speed, vehicle utilization, and day-to-day operating control across new verticals. | 中 | SP014 |
| CP026 | DeepRoute IO 2.0 uses a VLA model and supports both LiDAR-equipped and pure-vision configurations across automakers. | 中 | SP015, SP016 |
| CP027 | DeepRoute said it had secured five confirmed OEM partnerships for IO 2.0 with first production vehicles scheduled to reach market later in 2026. | 中 | SP015 |
| CP028 | Research and Markets frames 2026-2030 as the large-scale commercialization phase for robotaxi and L3/L4 personal vehicles after an earlier consolidation period. | 中 | SP018 |
| CP029 | EEWorld’s 2026 market summary says tech-driven operators such as Pony.ai and WeRide lead commercialization speed, operating coverage, and deployment scale in robotaxi. | 中 | SP019 |
| CP030 | Chinese robotaxi leaders are expanding globally in 2026, broadening their operating-data and regulatory experience relative to Haomo’s geographically narrow logistics proof. | 中 | SP017, SP018, SP019 |
| CP031 | Haomo’s historical commercialization proof remained concentrated in Great Wall vehicle programs and Beijing-area logistics pilots rather than in a broad multi-OEM installed base. | 高 | SP002, SP003, SP004, SP025 |
| CP032 | Haomo’s low-cost HP170/HP370/HP570 packaging was strategically sensible, but similar mass-market segmentation is now offered by larger rivals such as Horizon. | 中 | SP007, SP008, SP025 |
| CP033 | Once an OEM validates a smart-driving supplier into production, switching costs rise because hardware integration, safety validation, and road-data loops all become sticky. | 高 | SP008, SP009, SP012, SP014 |
| CP034 | Ecosystem-level multi-homing exists, but individual vehicle programs still tend to favor a few deeply embedded suppliers rather than frequent stack swapping. | 中 | SP011, SP018, SP024 |
| CP035 | By 2026 buyers increasingly evaluate chip alliances, production tooling, and operations platforms as part of the product—not as optional extras. | 中 | SP007, SP008, SP014, SP021 |
| CP036 | Haomo lacked public evidence of a proprietary chip, dominant open platform, or global data asset comparable with Horizon, Mobileye, Pony.ai, WeRide, or DeepRoute. | 高 | SP007, SP012, SP013, SP014, SP015, SP025 |
| CP037 | Reported salary arrears, management churn, and the operations halt materially weakened Haomo’s competitive trust position versus still-shipping rivals. | 中 | SP003, SP005, SP006 |
| CP038 | In low-speed logistics, human delivery and platform-run operations remained viable substitutes, so Haomo also competed against unit-economics discipline rather than only against autonomy startups. | 中 | SP002, SP017, SP018 |
| CP039 | The strongest rivals by 2026 combined either broad OEM embed, global data, or fleet-operations breadth, while Haomo had no public evidence of leading any one of those dimensions. | 中 | SP010, SP011, SP012, SP013, SP014, SP017 |
| CP040 | Competitive moat in 2026 is best understood as a three-part stack of deployment data, production integration, and capital resilience. | 高 | SP012, SP017, SP018, SP021 |
| CI001 | Haomo raised more than CNY100 million in Series B1 and CNY300 million in Series B2 during 2024, meaning it brought in over CNY400 million of fresh capital that year. | 中 | SI001, SI002, SI003 |
| CI002 | Management said the 2024 financing would be used for mass production, large-model research, computing power, big data, and robotics expansion. | 中 | SI001, SI003 |
| CI003 | Haomo’s public business model included ADAS kits, full-stack smart-driving solutions, and autonomous delivery vehicles rather than one pure software stream. | 中 | SI003, SI005, SI006, SI007 |
| CI004 | Haomo likely monetized a blended product-and-service mix, but the percentage split between hardware, software, cloud, and support work is not publicly disclosed. | 中 | SI003, SI005, SI015 |
| CI005 | Haomo publicly disclosed HP170, HP370, and HP570 list pricing in a roughly CNY3,000 to CNY8,900 range, signaling a mass-market ADAS attachment thesis. | 高 | SI001, SI006 |
| CI006 | Partner-style coverage said the latest HDeliver generation was priced below CNY90,000, implying a vehicle-hardware monetization model with more hardware-like economics than pure software. | 中 | SI007 |
| CI007 | Haomo disclosed traction markers such as more than 20 equipped models and nearly 140 million assisted-driving kilometers by spring 2024, but those are activity metrics rather than audited revenue. | 中 | SI001, SI005 |
| CI008 | Public logistics proof included nearly 300,000 grocery deliveries plus one case study with about 300 daily orders on average and a 700-order peak, but no unit-profit disclosure accompanied those figures. | 中 | SI005, SI007 |
| CI009 | Haomo described software, hardware, modules, code, and cloud services for automakers, which implies contract structures more complex than a simple subscription or pure per-vehicle fee. | 中 | SI003, SI005 |
| CI010 | Haomo’s real revenue quality likely remained concentrated because Great Wall stayed the actual controller and the strongest public customer anchor even after claims of other OEM deals. | 中 | SI002, SI004, SI008 |
| CI011 | Public sources do not disclose Haomo’s annual revenue, ARR, or backlog, making it impossible to convert traction metrics into a clean top-line base case. | 高 | SI005, SI008, SI016 |
| CI012 | Layoffs, salary-arrears reporting, and the operations halt indicate that Haomo’s capital needs likely outran available liquidity or investor support by late 2025. | 中 | SI002, SI004, SI023, SI024 |
| CI013 | The Yicai halt report’s reference to possible frozen accounts and litigation is consistent with working-capital or financing distress rather than orderly self-funded scaling. | 中 | SI004 |
| CI014 | Haomo has no public cash-balance, debt, or runway disclosure available in retained sources as of August 2026. | 高 | SI008, SI016 |
| CI015 | WeRide’s 2025 public disclosures show the scale of autonomy-company burn: about US$97.9 million of revenue, about US$196.2 million of R&D expense, about US$264.1 million of operating loss, and roughly US$1.0 billion of liquidity at year-end 2025. | 高 | SI014, SI015 |
| CI016 | Pony.ai’s 2025 and Q1 2026 official disclosures show a still capital-hungry model despite growth: US$90.0 million of 2025 revenue, US$34.3 million of Q1 2026 revenue, and liquidity above US$1.4 billion. | 高 | SI011, SI012, SI013 |
| CI017 | Mobileye’s 2025 results show that meaningfully self-funded ADAS economics appear only at a far larger scale than anything Haomo publicly disclosed: US$1.894 billion of revenue, US$602 million of operating cash generation, and US$1.8 billion of cash. | 中 | SI017 |
| CI018 | Peer disclosures suggest autonomy companies often monetize a mix of product revenue, service revenue, and support or licensing work rather than one clean recurring-software line item. | 高 | SI011, SI012, SI014, SI015 |
| CI019 | WeRide’s Form 20-F shows product revenue, service revenue, top-customer concentration, and continuing operating cash outflow, illustrating the level of detail investors normally need to judge an AV business. | 高 | SI015, SI016 |
| CI020 | Because listed peers with higher revenue still post large losses and R&D bills, Haomo was likely financing-dependent absent exceptional parent support or unusually strong undisclosed margins. | 高 | SI011, SI014, SI015, SI017 |
| CI021 | Haomo’s use-of-funds language emphasized growth investment and technical capability building rather than margin optimization or near-term free cash flow. | 中 | SI001, SI003 |
| CI022 | Official and public list-pricing signals should be treated as list pricing, not as realized ASP, discount, payment timing, or gross margin evidence. | 中 | SI006, SI007, SI015 |
| CI023 | OEM validation cycles and milestone-based contracts likely lengthened Haomo’s cash conversion because automotive revenue recognition typically follows testing, acceptance, and SOP milestones rather than instant subscription billing. | 中 | SI009, SI015, SI020 |
| CI024 | Autonomous delivery vehicles likely absorbed working capital through BOM, inventory, maintenance, and deployment support before they could become recurring cash generators. | 中 | SI007, SI011, SI012 |
| CI025 | Public peer disclosures show that top-line growth does not remove capital intensity: WeRide and Pony both continued to invest heavily in R&D, BOM optimization, and fleet expansion while losses persisted or free-cash demands remained meaningful. | 高 | SI011, SI012, SI014, SI015 |
| CI026 | Pony’s disclosed unit-economics breakeven milestones in Guangzhou and Shenzhen highlight the bar Haomo would have needed to clear publicly to defend a credible profitability path, yet no analogous Haomo unit-economics proof was found. | 高 | SI012, SI011 |
| CI027 | Haomo’s published emphasis on a 70% cost reduction for HDeliver shows cost discipline mattered, but it also confirms that delivery-autonomy economics remained burdened by hardware cost. | 中 | SI007 |
| CI028 | Great Wall’s associate relationship likely helped Haomo fund expensive R&D longer than a purely standalone startup could have, even if public investors cannot quantify the exact support. | 中 | SI005, SI008, SI009 |
| CI029 | That same Great Wall linkage did not guarantee durable support, because Great Wall’s own internal smart-driving roadmap reduced the certainty that Haomo would remain the preferred capital sink or technical vehicle. | 中 | SI008, SI020, SI021, SI022 |
| CI030 | A unicorn valuation and multiple fundraising rounds did not eliminate financing dependency; by late 2025 the public record still pointed to solvency stress. | 中 | SI004, SI010, SI024 |
| CI031 | Haomo’s public financial proof is dominated by operational activity metrics rather than by audited or management-level financial disclosure. | 中 | SI005, SI007, SI016 |
| CI032 | The supportable financial verdict is that Haomo was a capital-intensive autonomy company with low public reporting visibility and unresolved solvency risk by August 2026. | 高 | SI004, SI014, SI017, SI024 |
| CI033 | The most supportable best-case revenue architecture for Haomo was a combination of ADAS product sales, OEM integration or support services, delivery-vehicle sales, and possible cloud or data services. | 中 | SI003, SI005, SI006, SI015 |
| CI034 | If Haomo had hoped to raise or justify more capital, the natural proof points would have been diversified OEM SOP wins, disclosed margins, or profitable fleet economics; retained public sources do not show those proof points clearly materializing. | 中 | SI002, SI004, SI022, SI023 |
| CI035 | Underwriting remains blocked by missing cash, debt, gross margin, burn, receivables, and customer-concentration disclosures. | 高 | SI008, SI016 |
| CI036 | The absence of a current cash and liabilities package matters more than category growth because investors need proof of survival before they can underwrite scale. | 中 | SI014, SI017, SI025 |
| CE001 | Haomo’s public product surface included passenger-vehicle ADAS and low-speed autonomous delivery vehicles rather than only an autonomy research platform. | 高 | SE001, SE006, SE017 |
| CE002 | Haomo introduced DriveGPT in 2023 as a GPT-style autoregressive model for autonomous driving built around a “Drive Language.” | 高 | SE001, SE003, SE031, SE010 |
| CE003 | Haomo framed DriveGPT as a way to make autonomous driving decisions safer, smoother, more human-like, and more explainable to drivers. | 高 | SE001, SE003, SE031 |
| CE004 | Haomo linked DriveGPT to a mass-production vehicle narrative, naming the WEY Mocca DHT-PHEV as the first vehicle to be powered by the model. | 高 | SE001, SE003, SE031 |
| CE005 | At its 2023 DriveGPT launch, Haomo said it had become a supplier for three automakers after previously supplying Great Wall. | 高 | SE001, SE003, SE017 |
| CE006 | The HP170, HP370, and HP570 releases show that Haomo’s productization extended beyond one generic ADAS claim into explicit SKUs with differentiated hardware and capability tiers. | 高 | SE002, SE004, SE032 |
| CE007 | HP170 was positioned as a low-cost highway and expressway product with about 5 TOPS compute, a forward camera-led sensor stack, and optional front radars. | 中 | SE002, SE032 |
| CE008 | HP370 raised the stack to about 32 TOPS and added memory-driving and intelligent barrier-avoidance features. | 中 | SE002, SE032 |
| CE009 | HP570 was positioned for urban-road assisted driving with 72 or 100 TOPS compute and optional LiDAR. | 中 | SE002, SE032 |
| CE010 | Haomo’s 2023 product messaging emphasized no-HD-map or map-free assisted driving, lowering reliance on high-definition map infrastructure in the pitch. | 高 | SE002, SE003 |
| CE011 | HDeliver and Little Camel targeted controlled last-mile environments such as supermarkets, courier scenarios, campuses, and airports. | 中 | SE005, SE006 |
| CE012 | By 2025 reporting, Little Camel 2.0 and 3.0 had reportedly entered commercial operation across nine use cases. | 中 | SE005, SE006 |
| CE013 | Public partner-style coverage said the latest HDeliver version cut cost about 70% versus the market average and was priced below CNY90,000. | 中 | SE006 |
| CE014 | Independent technical explanation of DriveGPT suggests the planning model still depends on high-quality perception inputs and tokenization, meaning model ambition does not remove upstream sensing dependencies. | 中 | SE010, SE012, SE013 |
| CE015 | Haomo’s public GitHub footprint shows practitioner output in LiDAR-camera fusion, occupancy forecasting, and place recognition rather than a dead or purely empty research surface. | 中 | SE007, SE008, SE026, SE027, SE028, SE029, SE030 |
| CE016 | GitHub activity and stars are useful developer signal, but they do not prove product reliability, OEM support quality, or production-scale autonomy performance. | 中 | SE007, SE008, SE026, SE028 |
| CE017 | Haomo publicly described export or deployment ambition into the EU, Israel, the Middle East, South Africa, Australia, Mexico, and Russia. | 中 | SE001, SE003, SE031 |
| CE018 | Haomo’s trust-and-safety messaging included claims such as five-star AEB for HP170 and RLHF-like human-feedback training for DriveGPT. | 中 | SE002, SE003 |
| CE019 | No retained source provided a current public Haomo reliability dashboard, support SLA, cybersecurity page, or post-2025 quality statement. | 高 | SE014, SE015 |
| CE020 | China’s tighter OTA and autonomous-driving compliance rules raise the bar for vendors that need to update safety-impacting software. | 中 | SE015 |
| CE021 | Haomo’s two core product families addressed fundamentally different workflows: consumer or OEM-assisted driving versus controlled-route logistics automation. | 高 | SE001, SE006, SE017 |
| CE022 | 2025 reporting said Haomo co-developed the MANA OASIS autonomous-driving compute center with Volcano Engine, implying a material compute and infrastructure layer beneath the product story. | 中 | SE005 |
| CE023 | By 2025, public reporting suggested city-NOA and some customer programs had slipped or narrowed, which weakens the claim that Haomo’s frontier features were fully mature at commercial scale. | 中 | SE005 |
| CE024 | Haomo’s public roadmap is detailed through 2023 but much less transparent after 2024, making present-tense product maturity harder to judge. | 中 | SE005, SE014, SE016 |
| CE025 | Relative to 2026 peer product surfaces from Horizon, WeRide, Pony.ai, DeepRoute, and Mobileye, Haomo’s retained public architecture looks conceptually competitive but less current and less supported by present-tense deployment evidence. | 中 | SE018, SE020, SE021, SE022, SE023, SE024, SE025 |
| CE026 | The passenger-vehicle workflow for Haomo centered on OEM integration, validation, and then driver-facing use of assisted-driving, parking, and memory-driving features. | 中 | SE002, SE017 |
| CE027 | The logistics workflow centered on route-bounded deployments where delivery volume, supervision, and controlled environments mattered more than open-road generality. | 中 | SE005, SE006 |
| CE028 | Haomo’s critical dependencies included sensors, chips, training data, compute partners, OEMs, logistics partners, and regulators rather than solely its own model IP. | 中 | SE001, SE005, SE010, SE015 |
| CE029 | A no-HD-map architecture can reduce cost and deployment friction, but it shifts more of the burden onto perception quality, model robustness, and local-road learning. | 中 | SE003, SE010, SE019 |
| CE030 | Haomo’s developer signal looked stronger in public research artifacts than in buyer-facing public docs, support surfaces, or compliance disclosures. | 中 | SE007, SE008, SE014, SE015 |
| CE031 | The late-2025 operating crisis materially raises trust risk for shipped automotive and logistics software because customers need patching, maintenance, and continuity as much as features. | 中 | SE005, SE014 |
| CE032 | The best supportable maturity verdict is that Haomo had historically shipped enough to matter but not enough to remove current support and roadmap uncertainty. | 高 | SE005, SE016, SE017, SE025 |
| CE033 | Haomo’s pinned GitHub repositories carried hundreds of stars and updates as late as 2025, indicating some external practitioner interest in its research output. | 中 | SE007, SE008, SE026, SE028 |
| CE034 | Haomo’s 6P framing suggests the company wanted to meet OEMs across code, module, software, hardware, cloud, and full-scheme integration layers, but no public API or SDK documentation was found in retained sources. | 中 | SE001, SE014, SE015 |
| CE035 | By October 2023 Haomo claimed DriveGPT had screened 10 billion frames of internet images and 4.8 million 4D autonomous-driving clips. | 中 | SE002, SE032 |
| CE036 | No retained public source provided MTBF, intervention-rate, recall, or support-SLA data for Haomo’s active products. | 高 | SE014, SE015 |
| CU001 | In Haomo’s ADAS business, the buyer and payer were usually the OEM program team while the driver was the downstream user. | 高 | SU001, SU004, SU022 |
| CU002 | Great Wall was Haomo’s most important visible customer relationship because it combined production deployment, strategic control, and ecosystem dependence. | 高 | SU001, SU004, SU010 |
| CU003 | In Haomo’s low-speed logistics business, the payer was typically a retailer, supermarket operator, or delivery platform rather than the end grocery recipient. | 中 | SU005, SU008, SU020 |
| CU004 | Great Wall mattered more than raw logo count because Haomo’s strongest passenger-vehicle proof came from the parent OEM ecosystem rather than from a broad set of named automakers. | 高 | SU001, SU004, SU021 |
| CU005 | The logistics customer roster provided real operating proof but not the same level of disclosed revenue quality as the Great Wall relationship. | 中 | SU005, SU006, SU008 |
| CU006 | Haomo publicly implied other automaker relationships beyond Great Wall, but the names and commercial scale of those customers were mostly withheld. | 高 | SU004, SU010, SU022 |
| CU007 | Customer evidence should be separated into production-grade Great Wall deployment, named Beijing retail-route operations, and weaker partner-style diversification claims. | 高 | SU001, SU005, SU012 |
| CU008 | Wumart Dmall is the clearest named retail customer because public sources describe specific routes, communities, and daily-order behavior. | 中 | SU005, SU006, SU008 |
| CU009 | Meituan, Dada, and Alibaba references imply some breadth in logistics relationships, but they do not by themselves prove a diversified revenue base. | 高 | SU004, SU005, SU008 |
| CU010 | By February 2024, public sources said HPilot had been equipped in more than 20 vehicle models. | 高 | SU002, SU003, SU004, SU024 |
| CU011 | By February 2024, public sources said Haomo had accumulated more than 120 million kilometers of assisted-driving mileage. | 高 | SU002, SU003, SU004 |
| CU012 | Great Wall still identified Haomo as an associate and related-party ecosystem entity in its 2025 annual reporting. | 高 | SU001, SU018 |
| CU013 | By late 2025, Yicai and CarNewsChina still described more than 20 Great Wall models as equipped with Haomo HPilot. | 中 | SU010, SU021 |
| CU014 | By early 2024, TechCrunch described Meituan, Alibaba, and JD.com as logistics clients of Haomo. | 中 | SU004 |
| CU015 | By February 2024, multiple sources said Haomo delivery vans had ferried nearly 300,000 grocery parcels for supermarkets in Beijing. | 高 | SU002, SU004, SU009, SU024 |
| CU016 | EqualOcean said Haomo's cooperation with Wumart had delivered more than 120,000 orders across more than 60 communities in Shunyi Mapo, Beijing. | 中 | SU005 |
| CU017 | Gasgoo said Xiaomotuo had completed more than 200,000 delivery orders by July 2023 and had been operating since November 2020 in Shunyi before expanding to Shanghai and Yizhuang. | 中 | SU006, SU007 |
| CU018 | Public delivery-program sources tied Haomo to nine active or target scenarios and regular operation in Beijing's policy pilot zone. | 中 | SU006, SU007, SU008 |
| CU019 | EqualOcean and Logistics Tech Outlook both linked Haomo’s logistics business to Wumart Dmall, Meituan, and Dada-style partnerships. | 中 | SU005, SU008 |
| CU020 | Logistics Tech Outlook said HDeliver averaged about 300 daily orders and recently peaked near 700 in the Wumart Dmall case. | 中 | SU008 |
| CU021 | By June 2026, Meituan operated 8,400 autonomous delivery vehicles across 47 Chinese cities, showing that Haomo’s platform-side counterparties were operating at very large scale. | 中 | SU015 |
| CU022 | Because Meituan’s autonomous-delivery footprint was already large in 2026, any successful Haomo relationship there would likely be strategically valuable but operationally demanding. | 中 | SU015, SU020 |
| CU023 | Haomo’s public narrative suggested some OEM diversification beyond Great Wall, including a claim in 2023 that it had become a supplier to three automakers. | 中 | SU004, SU022 |
| CU024 | A late-2024 BusinessKorea article said Beijing Hyundai was developing a China-market EV that would incorporate Haomo DriveGPT technology. | 中 | SU012 |
| CU025 | Hyundai’s official Auto China 2026 materials instead named Momenta as the ADAS technology partner for IONIQ V. | 高 | SU013, SU014 |
| CU026 | The conflict between the late-2024 Haomo reporting and Hyundai’s 2026 Momenta announcement makes Haomo’s Hyundai diversification case unproven as of August 2026. | 高 | SU012, SU013, SU014, SU019 |
| CU027 | Public evidence for Dada is partner-style and scenario-specific rather than a clean proof of large, durable customer economics. | 中 | SU005, SU008 |
| CU028 | No public NRR, GRR, churn, or renewal metrics for Haomo customers were identified in retained sources. | 高 | SU001, SU004, SU015 |
| CU029 | No retained public source disclosed customer-level revenue concentration or contract values for Haomo. | 高 | SU001, SU004, SU010 |
| CU030 | The best public proxy for customer durability is the continuity of Great Wall sponsorship and related-party status rather than any published retention dataset. | 中 | SU001, SU012, SU018 |
| CU031 | Supermarket-delivery milestones prove repeat order activity, but they do not prove contract renewal quality or positive unit economics. | 中 | SU006, SU008, SU020 |
| CU032 | Current live deployment status for Wumart, Meituan, and Dada after the late-2025 halt was not verified in retained public sources. | 高 | SU011, SU016, SU021 |
| CU033 | Late-2025 layoffs and work-stoppage reporting materially weakened confidence in Haomo’s ability to keep supporting customers without interruption. | 中 | SU010, SU011, SU021 |
| CU034 | Haomo’s customer expansion logic followed a land-and-expand pattern: one OEM to more models and one delivery route to more communities or scenarios. | 中 | SU003, SU005, SU006, SU023 |
| CU035 | Top-customer concentration risk was likely high because the public record points to Great Wall as the core ADAS channel and a narrow set of logistics counterparties as the main named users. | 高 | SU001, SU004, SU010, SU019 |
| CU036 | By August 2026, the most supportable customer verdict is that Haomo had real pre-crisis adoption proof but an under-disclosed and concentrated customer base. | 高 | SU001, SU004, SU011, SU021 |
| CU037 | Current Dmall and Wumart 2026 recognition shows the retail counterparty remained capable of scaled AI operations even though Haomo was no longer named in that current coverage. | 中 | SU016, SU017 |
| CU038 | Because current 2026 Dmall coverage omits Haomo, it cannot be used as proof that the historical autonomous-delivery relationship remained active. | 高 | SU016, SU017 |
| CU039 | ML Vehicle’s 2025 reporting that Haomo supported only two Hyundai models with delays further weakens the case for broad OEM diversification. | 中 | SU019 |
| CR001 | China’s February 2025 MIIT/SAMR notice materially tightened product admission, recall, and OTA governance for intelligent connected vehicles with CDAS. | 高 | SR001, SR002 |
| CR002 | The 2025 notice requires manufacturers to conduct sufficient testing and verification, define system boundaries, and strengthen quality and safety management. | 高 | SR001, SR002 |
| CR003 | CDAS and OTA-related technical parameters must be filed, and already-admitted products had to supplement those materials by June 1, 2025. | 高 | SR001, SR002 |
| CR004 | OTA changes affecting main technical parameters or autonomous-driving functions require additional permits before implementation. | 高 | SR001, SR002 |
| CR005 | If an OTA upgrade is used to remedy a defect, the company must handle it under recall rules and stop producing or selling defective products until compliant. | 高 | SR001, SR002 |
| CR006 | The 2025 framework also creates accident and event reporting obligations, including 24-to-48-hour deadlines for serious CDAS accidents and periodic event reporting. | 高 | SR001, SR002 |
| CR007 | Beijing’s autonomous-vehicle regulation took effect in April 2025 and explicitly covers L3+ autonomous-vehicle activity including freight and delivery scenarios. | 高 | SR003, SR004 |
| CR008 | The Beijing framework remains conditional and supervision-based rather than a blanket permission for unconstrained commercialization. | 中 | SR003, SR004, SR021 |
| CR009 | China’s 2026 automotive-data cross-border transfer guide explicitly covers driving-automation and software-upgrade scenarios, raising compliance burdens for automotive data flows. | 高 | SR005, SR006 |
| CR010 | Legal analysis of the 2026 guide indicates automotive data controllers must identify important data and use the correct transfer mechanism, adding process overhead to any overseas or cloud-linked operation. | 高 | SR005, SR006 |
| CR011 | Chinese legal commentary indicates accident liability at L3+ increasingly shifts toward manufacturers or operators rather than remaining solely with drivers. | 中 | SR003, SR007 |
| CR012 | GB 44721-2026, China’s first mandatory L3/L4 autonomous-driving safety standard, takes effect on July 1, 2027. | 中 | SR008 |
| CR013 | After the fatal Xiaomi SU7 crash, China moved to rein in “smart driving” hype and tighten rollout scrutiny for assisted-driving systems. | 中 | SR009, SR010 |
| CR014 | A distressed company like Haomo would face a harder restart because the current Chinese policy environment is stricter on testing, reporting, OTA, and marketing than it was in 2023-2024. | 高 | SR001, SR008, SR009, SR010 |
| CR015 | Great Wall’s continued related-party and associate linkage means customer concentration is also a governance and dependency risk. | 高 | SR013, SR027 |
| CR016 | Late-2025 reporting tied Haomo’s staff cuts and operational stress to weak car sales, high R&D costs, and broader commercialization pressure. | 中 | SR014, SR015, SR028 |
| CR017 | Public reports that Haomo halted operations mean software support, OTA patching, and customer incident response cannot be assumed to be intact. | 中 | SR015, SR016 |
| CR018 | In a safety-sensitive product category, missing proof of a current support team is itself a major operational risk. | 高 | SR001, SR015, SR016 |
| CR019 | Haomo’s historical deployment metrics prove pre-crisis use but do not prove current continuity or customer satisfaction after the halt. | 高 | SR017, SR018, SR019, SR020 |
| CR020 | Dmall’s 2026 scaling evidence shows the retail counterparty remained active, but because Haomo is not mentioned it cannot be treated as proof of current Haomo route activity. | 中 | SR025 |
| CR021 | Meituan’s 2026 robot fleet scale makes any platform-side relationship strategically valuable but also demanding on support quality, economics, and reliability. | 中 | SR024 |
| CR022 | Haomo’s Hyundai diversification story is fragile because late-2024 reporting naming Haomo conflicts with Hyundai’s 2026 official Momenta-based ADAS announcement. | 高 | SR022, SR023, SR031 |
| CR023 | Great Wall had already supplemented Haomo with other intelligent-driving suppliers such as DeepRoute and Hesai, showing Haomo did not hold an untouchable position. | 中 | SR014 |
| CR024 | The narrow public logistics roster around Wumart Dmall, Meituan, and Dada implies partner dependence rather than a broadly diversified customer base. | 高 | SR018, SR019, SR032 |
| CR025 | NVIDIA’s April 2025 Form 8-K said exports of H20-class chips to China would require a license for the indefinite future. | 高 | SR011, SR012 |
| CR026 | NVIDIA’s Q1 FY2026 results showed a $4.5 billion H20-related charge and $2.5 billion of unshipped H20 revenue, proving China AI compute access had become materially volatile. | 高 | SR011, SR012 |
| CR027 | Because Haomo pitched DriveGPT and large-model autonomy, advanced-compute volatility is a more relevant risk than it would be for a simpler L2-only supplier. | 高 | SR011, SR012, SR022 |
| CR028 | Great Wall is Haomo’s most important dependency because it concentrates customer demand, ecosystem trust, and governance influence in one node. | 高 | SR013, SR018 |
| CR029 | Haomo’s public record does not disclose a robust multi-supplier or multi-cloud mitigation strategy for its higher-end autonomy ambitions. | 中 | SR011, SR012, SR013 |
| CR030 | Platform relationships with Meituan, Wumart, and Dada are strategically useful but do not by themselves prove durable economics or low churn risk. | 中 | SR019, SR024, SR032 |
| CR031 | Yicai’s halt reporting said employees were already two months in salary arrears by August 2025, creating labor and reputational exposure. | 中 | SR015 |
| CR032 | Bird & Bird’s 2025 China guidance makes post-termination non-competes more evidence-intensive and requires meaningful monthly compensation. | 中 | SR026 |
| CR033 | If non-compete compensation is not maintained, departing employees may cease complying, weakening Haomo’s ability to prevent leakage of technical know-how during distress. | 中 | SR015, SR026 |
| CR034 | Layoffs in a safety-critical autonomy company create execution risk even if financing returns, because rebuilding product, support, and validation capacity is slow. | 高 | SR015, SR026, SR028 |
| CR035 | The absence of a public restart plan, board action, or operating update keeps governance and execution risk elevated. | 高 | SR013, SR015 |
| CR036 | No public audited Haomo financial statements, current cash balance, debt schedule, or customer receivables aging were identified by the diligence date. | 高 | SR013, SR027 |
| CR037 | Great Wall’s public filings do not provide an enforceable shareholder backstop that would let outside investors underwrite Haomo’s solvency. | 高 | SR013, SR027 |
| CR038 | Haomo’s model combined ADAS, autonomous-delivery vehicles, and large-model R&D, increasing capital intensity and strategic focus risk. | 中 | SR015, SR018, SR028 |
| CR039 | ML Vehicle reported that Little Camel’s 2025 sales target was only 50 units and that no new models were in the pipeline, signaling weak logistics momentum. | 中 | SR028 |
| CR040 | The same ML Vehicle report said Haomo was then supporting only two Hyundai models with delivery planned for August and facing delays, reinforcing execution fragility. | 中 | SR028 |
| CR041 | Haomo’s major risks transmit into valuation through customer loss, slower approvals, staff exits, and reduced confidence in restart financeability. | 高 | SR001, SR013, SR015, SR024 |
| CR042 | Without current proof of restarted operations, cash, and active-customer support, the investment thesis should be treated as broken rather than merely uncertain. | 高 | SR015, SR017, SR018 |
| CR043 | Haomo’s residual risk rating is critical because regulatory, solvency, concentration, and labor risks reinforce one another rather than remaining isolated. | 高 | SR001, SR013, SR015, SR026 |
| CR044 | The only credible mitigation path would combine fresh capital, explicit Great Wall or customer support confirmation, current compliance artifacts, and retention of core engineering staff. | 高 | SR001, SR013, SR015, SR026 |
| CV001 | Haomo’s late-2021 Series A created a unicorn valuation marker above US$1 billion. | 中 | SV007 |
| CV002 | Haomo raised more than CNY400 million across B1 and B2 in 2024. | 高 | SV008, SV009, SV016, SV018, SV031 |
| CV003 | Public database coverage of Haomo’s cumulative disclosed funding remains inconsistent, with CB Insights listing roughly US$245.3 million total raised and Series B-II status. | 中 | SV008, SV015 |
| CV004 | Late-2025 public reporting said Haomo halted operations and faced salary-arrears or account-freeze concerns. | 中 | SV010, SV012, SV013 |
| CV005 | The reviewed public materials do not provide audited 2025 or 2026 Haomo revenue, cash, debt, or a clean current equity mark. | 中 | SV010, SV014, SV015 |
| CV006 | Yicai reported that Great Wall Motor chairman Wei Jianjun remained Haomo’s actual controller and held a 37 percent stake. | 中 | SV010 |
| CV007 | Great Wall’s 2025 annual report still treated Haomo as an associate-like exposure and disclosed share of accumulated losses of RMB413.7 million. | 中 | SV014 |
| CV008 | Great Wall’s own public intelligent-driving roadmap now emphasizes Coffee Pilot, VLA, and world-model systems rather than an obvious Haomo-led recovery. | 中 | SV014 |
| CV009 | Pony.ai’s public market capitalization was about US$3.39 billion in August 2026. | 中 | SV001 |
| CV010 | Pony.ai disclosed FY2025 revenue of US$90.0 million. | 中 | SV004 |
| CV011 | Pony.ai disclosed FY2025 gross margin of 15.7% and cash/investments/restricted cash/wealth instruments of about US$1.515 billion. | 中 | SV004 |
| CV012 | WeRide’s public market capitalization was about US$1.92 billion in August 2026. | 中 | SV002 |
| CV013 | WeRide disclosed FY2025 revenue of US$97.9 million. | 高 | SV005, SV027 |
| CV014 | WeRide disclosed FY2025 gross margin of 30.2% and roughly US$1.0 billion of cash, deposits, wealth-management products, and restricted cash. | 高 | SV005, SV027 |
| CV015 | Mobileye’s public market capitalization was about US$6.77 billion in August 2026. | 中 | SV003 |
| CV016 | Mobileye disclosed FY2025 revenue of US$1.894 billion, operating cash flow of US$602 million, and year-end cash of US$1.8 billion. | 中 | SV006 |
| CV017 | Live public autonomy and ADAS peers provide current revenue and annual-report disclosure stacks that Haomo lacks entirely in the reviewed 2026 record. | 中 | SV004, SV005, SV006, SV027, SV028, SV029, SV030 |
| CV018 | Haomo’s historical commercialization proof was real, including 20-plus passenger-vehicle models, assisted-driving mileage above 120 million kilometers, and low-speed delivery volume measured in hundreds of thousands of orders or parcels. | 高 | SV008, SV016, SV017, SV018, SV031, SV032 |
| CV019 | Those historical proof points do not establish current support continuity after the late-2025 halt. | 中 | SV010, SV012, SV013, SV021, SV022 |
| CV020 | Diversification beyond Great Wall stayed weak before the halt, with Hyundai evidence limited and later displaced by a 2026 Hyundai-Momenta disclosure. | 中 | SV019, SV020, SV026 |
| CV021 | China’s regulatory bar tightened in 2025-2026 around OTA, product admission, data governance, and higher-level autonomy safety, increasing restart cost for a distressed player like Haomo. | 高 | SV023, SV024, SV025 |
| CV022 | With no current Haomo revenue, cash, debt, or share-count disclosure, DCF or direct peer-multiple valuation would create false precision. | 中 | SV010, SV014, SV015, SV023, SV024, SV025 |
| CV023 | A distressed-option or recovery-framework is more appropriate for Haomo than a normal software multiple framework. | 中 | SV004, SV005, SV006, SV010, SV014 |
| CV024 | Haomo cannot rationally be valued alongside Pony.ai or WeRide on a direct revenue-multiple basis because its current revenue base, liquidity, and operating continuity are not publicly evidenced. | 中 | SV001, SV002, SV004, SV005, SV010, SV014, SV015 |
| CV025 | Haomo’s unicorn-era valuation narrative is now a stale historical marker rather than a safe current price anchor. | 中 | SV007, SV009, SV010, SV013, SV014 |
| CV026 | Any current transaction near the historical US$1 billion-plus mark would be unsupported by the reviewed public evidence set. | 中 | SV007, SV009, SV010, SV014, SV015 |
| CV027 | The bear case is liquidation or piecemeal asset sale in which common-equity value could approach zero after liabilities, wage claims, or preference overhang. | 中 | SV010, SV012, SV013, SV014 |
| CV028 | The base case assumes a controlled strategic rescue that preserves part of Haomo’s engineering team, codebase, and customer-support obligations. | 中 | SV014, SV021, SV022, SV023 |
| CV029 | Under that base case, a rough US$80 million to US$200 million equity range is more defensible than any valuation close to the old unicorn mark. | 中 | SV007, SV010, SV014, SV021 |
| CV030 | The bull case requires documented restart, fresh capital, active customer support, and regulatory compliance evidence; without those, it should be treated as optionality rather than base expectation. | 中 | SV010, SV021, SV023, SV024, SV025 |
| CV031 | Even a bull case of roughly US$250 million to US$450 million would still sit far below Haomo’s historical unicorn label because the company would be rebuilding from distress. | 中 | SV007, SV010, SV014, SV021 |
| CV032 | Meituan, Wumart, and related last-mile demand mean Haomo’s IP and operating know-how are likely worth more than a pure zero in a sponsor-led rescue. | 中 | SV017, SV018, SV021, SV022 |
| CV033 | Great Wall is the most plausible rescue vector, but the public record shows no enforceable recapitalization or support backstop. | 中 | SV010, SV014 |
| CV034 | The current recommendation should be avoid rather than track for any price anchored to historical marks or narrative-only recovery. | 中 | SV001, SV002, SV003, SV010, SV014 |
| CV035 | Confidence should be rated medium because impairment direction is clear but the exact capitalization and liability stack are not public. | 中 | SV010, SV014, SV015 |
| CV036 | Risk rating should be treated as critical because equity value now depends more on solvency and restart execution than on market size. | 中 | SV010, SV012, SV013, SV023, SV024 |
| CV037 | Valuation stance should be treated as expensive relative to current public evidence if counterparties still use a 2021-2024 style private-market anchor. | 中 | SV007, SV009, SV010, SV014 |
| CV038 | Positive thesis elements remain: Haomo previously achieved real OEM integration, delivery deployments, and blue-chip funding support. | 高 | SV007, SV008, SV009, SV016, SV017, SV018 |
| CV039 | Anti-thesis elements dominate: operations halt, customer concentration, weak diversification, regulatory tightening, and missing financial disclosure. | 高 | SV010, SV011, SV012, SV013, SV014, SV019, SV020, SV023, SV024, SV025, SV026 |
| CV040 | The highest-priority diligence ask is a current schedule of cash, debt, payables, payroll arrears, and contingent liabilities. | 中 | SV010, SV014 |
| CV041 | The second-priority diligence ask is signed proof of which OEM and delivery programs remain live and supported in 2026. | 中 | SV010, SV012, SV021, SV022 |
| CV042 | The third-priority diligence ask is cap-table and preference-stack clarity, including any Great Wall support rights or rescue economics. | 中 | SV014, SV015 |
| CV043 | The fourth-priority diligence ask is evidence of regulator-ready OTA, incident-response, and automotive-data compliance for any restart. | 中 | SV023, SV024, SV025 |
| CV044 | Enterprise value and residual common-equity value may diverge sharply in a rescue because strategic buyers can pay for IP or contracts while legacy equity is diluted or subordinated. | 中 | SV010, SV014, SV015 |
| CV045 | CompaniesMarketCap’s annual-report directories show Pony.ai, WeRide, and Mobileye each maintain standardized public pathways to annual-report materials. | 中 | SV028, SV029, SV030 |
| CV046 | Mobileye’s scale, 2025 cash generation, and mature ADAS installed-base economics show what a true multi-billion ADAS benchmark looks like; Haomo is nowhere near that maturity profile in public evidence. | 中 | SV003, SV006 |
| CV047 | Pony.ai and WeRide still spend heavily on R&D despite better scale and funding, implying Haomo likely remained financing-dependent even before the halt. | 高 | SV004, SV005, SV027 |
| CV048 | Without fresh management disclosure, upside beyond speculative distressed optionality is not supportable from the public record alone. | 中 | SV010, SV014, SV015 |
| CV049 | Other China intelligent-driving leaders such as Horizon Robotics now maintain formal investor-relations reporting pages, underscoring how far Haomo has fallen behind the sector’s disclosure curve. | 中 | SV033 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Korea Newswire / Haomo.AI | Haomo unveils DriveGPT, speeds up commercialization and global expansion after landing in Europe and Israel | Haomo announced at the event that the company has become a supplier for three automakers. |
| SO002 | Korea Newswire / Haomo.AI | Haomo Launches Three ADS Products for as Low as $411, Pushes DriveGPT to Know Everything | These products, namely HP170, HP370 and HP570, are designed to be installed on low- to high-end passenger vehicles. |
| SO003 | CnEVPost | Chinese startup Haomo raises nearly $150 million in new funding, becoming latest self-driving unicorn | Investors are GL Ventures, Meituan, Hong Kong-listed Shoucheng Holdings Limited, Qualcomm Ventures, and JZ Capital. |
| SO004 | TechCrunch | Backed by Chinese carmaker Great Wall, Haomo raises $14M for autonomous driving tech | Its self-driving delivery vans have ferried close to 300,000 grocery parcels for supermarkets in Beijing. |
| SO005 | KrASIA | Haomo.AI hits crossroads amid leadership shakeup and tech setbacks | The sales target for our low-speed delivery vehicle this year is only about 50 units. |
| SO006 | KrASIA | Haomo.ai halts operations as employees told not to report to work | Haomo.ai has abruptly halted operations after instructing all employees to stop reporting to work. |
| SO007 | CarNewsChina | Great Wall’s self-driving arm Haomo.AI collapses as company reports abrupt staff layoff, report says | Employees in Beijing spoke of delayed salaries and months-long arrears, frozen company bank accounts. |
| SO008 | Yicai Global | Great Wall Motor-Backed Self-Driving Unicorn Haomo.AI Stops Work, Report Says | The prevailing theory among employees about the sudden shutdown is that litigation has resulted in the company's accounts being frozen. |
| SO009 | Friendly Auto | Haomo.AI Faces Internal Shake-Up Amid Mounting Business Challenges | As of May 2024, Haomo-powered vehicles had logged over 141 million kilometers of assisted driving. |
| SO010 | CB Insights | HAOMO.AI - Products, Competitors, Financials, Employees, Headquarters Locations | Total Raised: $245.3M. |
| SO011 | Hong Kong Exchanges and Clearing Limited | Great Wall Motor Company Limited Annual Report 2025 | The share of accumulated losses at the end of the period is RMB413,655,614.14. |
| SO012 | Logistics Tech Outlook | HAOMO.AI | Top Logistics Tech Startup in China 2023 | As of July 2023, the cumulative distribution order volume exceeded 200,000. |
| SO013 | Gasgoo | Haomo.ai launches autonomous driving large model, DriveGPT | Up to now, Haomo.ai's HPilot system has already been installed onto nearly 20 models. |
| SO014 | ArenaEV | China's Haomo delivers affordable ADAS kits to speed up adoption | The company unveiled the HP170, HP370, and HP570 ADAS kits. |
| SO015 | arXiv | DriveGPT: Scaling Autoregressive Behavior Models for Driving | We present DriveGPT, a scalable behavior model for autonomous driving. |
| SO016 | arXiv | DriveGPT4: Interpretable End-to-end Autonomous Driving via Large Language Model | DriveGPT4 represents the pioneering effort to leverage LLMs for the development of an interpretable end-to-end autonomous driving solution. |
| SO017 | GitHub | HAOMO.AI organization on GitHub | [ICRA 2024] SuperFusion: Multilevel LiDAR-Camera Fusion for Long-Range HD Map Generation. |
| SO018 | GitHub | HAOMO.AI repositories | The official implementation of [Quality over Quantity: Boosting Data Efficiency Through Ensembled Multimodal Data Curation] in AAAI2025. |
| SO019 | Law.asia | China's Autonomous Vehicle Regulations | The Beijing Autonomous Vehicle Regulations issued in 2024 provide a clear legal basis and operational rules for high-level autonomous vehicle road testing. |
| SO020 | Research and Markets | Chinese Independent OEMs' ADAS and Autonomous Driving Report, 2026 | 2026 will become a key inflection point for China's automotive industry to move from quantitative change to qualitative change in intelligent driving capabilities. |
| SO021 | Yahoo Finance / Research in China syndication | China Independent OEMs' ADAS and Autonomous Driving Research Report 2026 | The proportion of high-level intelligent driving such as L2 and above doubled to 62.7%. |
| SO022 | Grand View Research | Autonomous Last Mile Delivery Market | Industry Report, 2030 | The global autonomous last mile delivery market size was estimated at USD 1,615.4 million in 2024. |
| SO023 | DataM Intelligence | China Autonomous Last Mile Delivery Market Size & Forecast 2035 | The China Autonomous Last-Mile Delivery Market size in 2026 is estimated at USD 6.05 billion. |
| SO024 | Driving Vision News | Key ADAS/AD Markets: DVN 2026 - China Market Report Summary | LiDAR installations in China's passenger-vehicle market exceeded 1.3 million units in the first four months of 2026. |
| SO025 | The Automotive Data | ADAS and Autonomous Driving Tier 1 Suppliers in China: Seven Development Trends in the Era of Assisted Driving 2.0 | By the first four months of 2025, L2.5-equipped vehicles reached 30.2%, and L2.9-equipped vehicles surged to 34.82%. |
| SM001 | Research and Markets | Chinese Independent OEMs' ADAS and Autonomous Driving Report, 2026 | 2026 will become a key inflection point for China's automotive industry to move from quantitative change to qualitative change. |
| SM002 | Yahoo Finance / Research in China syndication | China Independent OEMs' ADAS and Autonomous Driving Research Report 2026 | The proportion of high-level intelligent driving such as L2 and above doubled to 62.7%. |
| SM003 | The Automotive Data | ADAS and Autonomous Driving Tier 1 Suppliers in China: Seven Development Trends in the Era of Assisted Driving 2.0 | By the first four months of 2025, L2.5-equipped vehicles reached 30.2%, and L2.9-equipped vehicles surged to 34.82%. |
| SM004 | Gasgoo | The next chapter for ADAS and smart cockpits丨Gasgoo Awards 2026 & China Automotive Industry Innovation Casebook | As ADAS move into the mass market, cost optimization and large-scale production capabilities have become critical competitive factors. |
| SM005 | Driving Vision News | Key ADAS/AD Markets: DVN 2026 - China Market Report Summary | LiDAR installations in China's passenger-vehicle market exceeded 1.3 million units in the first four months of 2026. |
| SM006 | Robotics and Automation News | Research report: China’s ADAS market surges as full-stack solutions dominate | Software-only ADAS startups are under increasing pressure, as automakers now prioritize cost-effective, full-stack partnerships. |
| SM007 | APA Engineering | China Implements Stricter Regulations on OTA Updates and Autonomous Driving Features | Pre-approval of safety-impacting OTA software updates is required. |
| SM008 | Grand View Research | Autonomous Last Mile Delivery Market | Industry Report, 2030 | The global autonomous last mile delivery market size was estimated at USD 1,615.4 million in 2024. |
| SM009 | DataM Intelligence | China Autonomous Last Mile Delivery Market Size & Forecast 2035 | The China Autonomous Last-Mile Delivery Market size in 2026 is estimated at USD 6.05 billion. |
| SM010 | Logistics Tech Outlook | HAOMO.AI | Top Logistics Tech Startup in China 2023 | Currently, HDeliver vehicles operate in nine major scenarios. |
| SM011 | Korea Newswire / Haomo.AI | Haomo Launches Three ADS Products for as Low as $411, Pushes DriveGPT to Know Everything | These products, namely HP170, HP370 and HP570, are designed to be installed on low- to high-end passenger vehicles. |
| SM012 | ArenaEV | China's Haomo delivers affordable ADAS kits to speed up adoption | Experts predict a whopping 70% penetration rate for smart driving features by 2025. |
| SM013 | Gasgoo | Haomo.ai launches autonomous driving large model, DriveGPT | Up to now, Haomo.ai's HPilot system has already been installed onto nearly 20 models. |
| SM014 | Horizon Robotics | Smart Driving Technology_Smart Driving Solutions_Smart Driving Vehicles | Journey™ is China's first and largest mass-produced automotive computing solution. |
| SM015 | Mobileye | Mobileye | Driver Assist and Autonomous Driving Technologies | Driving the physical AI evolution. |
| SM016 | Pony.ai | Pony.ai | Using the power of our core technology “virtual driver”, we have positioned ourselves at the forefront of autonomous driving across three business units. |
| SM017 | WeRide | To Transform Urban Living with Autonomous Driving | WeRide offers a portfolio of five core products, including Robotaxi. |
| SM018 | DeepRoute.ai | DeepRoute.Ai | DeepRoute.Ai. |
| SM019 | KrASIA | Haomo.AI hits crossroads amid leadership shakeup and tech setbacks | Haomo.AI might follow the same path as Chery subsidiary ZDrive.ai. |
| SM020 | Friendly Auto | Haomo.AI Faces Internal Shake-Up Amid Mounting Business Challenges | Current project delivery, such as the Hyundai contracts, remains unproven, and competitors like DJI and Momenta now dominate the mid- to low-end market. |
| SM021 | Korea Newswire / Haomo.AI | Haomo unveils DriveGPT, speeds up commercialization and global expansion after landing in Europe and Israel | Haomo announced at the event that the company has become a supplier for three automakers. |
| SM022 | TechCrunch | Backed by Chinese carmaker Great Wall, Haomo raises $14M for autonomous driving tech | It also supplies Level 4 self-driving solutions to logistics vehicles, counting Meituan, Alibaba and JD.com among its clients. |
| SM023 | DVN / EAC 2026 | Key ADAS/AD Markets: DVN 2026 - China Market Report Summary | Beyond automotive, robotics and other non-vehicle applications are becoming an important growth driver. |
| SM024 | Law.asia | China's Autonomous Vehicle Regulations | Access by L3 or L4 vehicles to public roads is largely enabled through local pilot rules. |
| SM025 | Yicai Global | Great Wall Motor-Backed Self-Driving Unicorn Haomo.AI Stops Work, Report Says | Haomo.AI has three business areas -- passenger vehicle smart assisted driving, last-mile self-driving delivery vehicles, and smart hardware. |
| SP001 | CnEVPost | Great Wall-backed Haomo raises nearly $150 million in Series A funding | Haomo was founded in 2019 and mainly engages in autonomous driving for passenger vehicles and last-mile logistics. |
| SP002 | TechCrunch | Great Wall-backed autonomous driving startup Haomo is raising funding again | Haomo’s driverless delivery vans have ferried close to 300,000 grocery parcels for supermarkets in Beijing. |
| SP003 | Great Wall Motor / HKEX | 2025 Annual Report | The filing identifies Haomo Zhixing as an associate while Great Wall separately advances Coffee Pilot 3 and 4 smart-driving roadmaps. |
| SP004 | KrASIA / 36Kr translation | Haomo AI hits a crossroads as commercialization challenges intensify | Haomo was still searching for non-Great Wall growth as commercialization pressures rose. |
| SP005 | KrASIA / 36Kr translation | Haomo AI halts operations after five years on the road | Employees were told not to report to work after salary arrears and account freezes surfaced. |
| SP006 | CarNewsChina | Haomo AI collapses: China’s $1 billion self-driving startup shuts down | The report describes Haomo as shutting down after commercial struggles and management departures. |
| SP007 | Horizon Robotics | Horizon SuperDrive | HSD has secured design wins with 10 OEM brands in China and overseas across more than 20 vehicle models. |
| SP008 | Horizon Robotics | Journey 6 Series | Journey 6 offers an integrated computing solution and a complete mass-production development platform. |
| SP009 | Momenta | Honda and Momenta Partner to Advance Mass Production of End-to-End Assisted Driving Systems | Honda and Momenta will develop mass-production-ready assisted driving solutions based on end-to-end models. |
| SP010 | 36Kr / BT Finance | After FSD Enters Chinese Market: Which is More Reliable - Tesla, Huawei or Momenta? | Huawei Qiankun ADS mileage had reached 10.47 billion kilometers and its market share among domestic third-party suppliers was 77%. |
| SP011 | Zacarmate | Chinese ADAS in European EVs: Mercedes & VW Bet on Momenta, Horizon | Momenta had surpassed 800,000 production vehicles across more than 60 models from over 20 brands. |
| SP012 | Mobileye | Mobileye to supply Cloud-Enhanced ADAS for select future Stellantis vehicles | More than 230 million vehicles worldwide have been built with Mobileye’s EyeQ technology inside through 2025. |
| SP013 | Pony.ai | Pony.ai Business Overview | Pony.ai presents Robotaxi, Robotruck, and Personally Owned Vehicles product lines on one official business surface. |
| SP014 | WeRide | WeRide Technology | WeRide says its WeRide One platform supports autonomous products from L2 to L4 across mobility, logistics, and sanitation. |
| SP015 | PR Newswire / DeepRoute.ai | DeepRoute.ai Launches DeepRoute IO 2.0, Smart Driving Platform Powered by VLA Technology | DeepRoute IO 2.0 supports both LiDAR-equipped and pure vision configurations and already has five confirmed OEM partnerships. |
| SP016 | DeepRoute.ai | DeepRoute Solutions | The official solutions page presents DeepRoute as a smart-driving solutions supplier for OEM collaboration. |
| SP017 | ChinaBizInsider | China Robotaxi Global Expansion 2026: WeRide, Pony.ai vs Waymo | Chinese robotaxi players are expanding across three continents in 2026 while cutting hardware costs and increasing fleet scale. |
| SP018 | Research and Markets | L3/L4 Autonomous Driving and Startups Research Report, 2025 | 2026-2030 is framed as the large-scale commercial development phase for robotaxi and L3 autonomous private vehicles. |
| SP019 | EEWorld | Competition in Level 4 Autonomous Vehicles is intensifying | Tech-driven autonomous-driving companies such as Pony.ai and WeRide lead commercialization speed, coverage, and deployment scale. |
| SP020 | Driving Vision News | Key ADAS/AD Markets: DVN 2026 - China Market Report Summary | China’s passenger-vehicle LiDAR market exceeded 1.3 million installations in the first four months of 2026. |
| SP021 | Robotics and Automation News / IDTechEx | Research report: China’s ADAS market surges as full-stack solutions dominate | Software-only ADAS startups are under increasing pressure as automakers prioritize cost-effective, full-stack partnerships. |
| SP022 | Gasgoo | The next chapter for ADAS and smart cockpits | As ADAS move into the mass market, cost optimization and large-scale production capabilities have become critical competitive factors. |
| SP023 | Yahoo Finance / Research In China | China Independent OEMs' ADAS and Autonomous Driving Research Report 2026 | High-level intelligent driving among independent OEMs doubled to 62.7% by 2025. |
| SP024 | The Automotive Data | ADAS and Autonomous Driving Tier 1 Suppliers in China: Seven Development Trends in the Era of Assisted Driving 2.0 | Domestic suppliers gained ground through chip optimization, end-to-end algorithm-hardware coordination, and rapid localization. |
| SP025 | Korea Newswire / Haomo.AI | Haomo Launches Three ADS Products for as Low as $411, Pushes DriveGPT to Know Everything | Haomo launched HP170, HP370, and HP570 for low- to high-end passenger vehicles. |
| SI001 | Yicai Global | Chinese L4 Autopilot Firm Haomo.AI Scores USD41.4 million in Series B2 | Haomo secured CNY300 million in Series B2 after raising more than CNY100 million in B1 earlier in the year. |
| SI002 | Yicai Global | Chinese Self-Driving Startup Haomo.AI Confirms Layoffs | Haomo confirmed layoffs and said it had raised CNY100 million and CNY300 million in B1 and B2 earlier in the year. |
| SI003 | Yicai Global | Great Wall Motor-Backed Self-Driving Startup Haomo Bags Over USD14 Million in Fundraiser | The Series B1 proceeds would be used to enhance R&D and promote robotics-product industrial development. |
| SI004 | Yicai Global | Great Wall Motor-Backed Self-Driving Unicorn Haomo.AI Stops Work, Report Says | Haomo reportedly halted operations, had salary arrears, and may have had frozen accounts after missing performance targets. |
| SI005 | TechCrunch | Great Wall-backed autonomous driving startup Haomo is raising funding again | Haomo’s ADAS had powered more than 20 models and its delivery vans had ferried close to 300,000 grocery parcels in Beijing. |
| SI006 | Korea Newswire / Haomo.AI | Haomo Launches Three ADS Products for as Low as $411, Pushes DriveGPT to Know Everything | Haomo launched HP170, HP370 and HP570 priced from CNY3,000 to CNY8,900. |
| SI007 | Logistics Tech Outlook | HAOMO.AI | Top Logistics Tech Startup in China 2023 | The latest HDeliver version was priced below CNY90,000 and averaged about 300 orders a day, peaking near 700 in one Wumart case. |
| SI008 | Great Wall Motor / HKEX | 2025 Annual Report | Great Wall identifies Haomo Zhixing as an associate and related party in its 2025 annual report. |
| SI009 | Great Wall Motor / CNInfo | 2024 Annual Results Announcement | Great Wall published audited 2024 annual results, underscoring the scale of the parent ecosystem supporting Haomo historically. |
| SI010 | CnEVPost | Great Wall-backed Haomo raises nearly $150 million in Series A funding | Haomo reached unicorn status after its Series A financing. |
| SI011 | Pony.ai | PONY AI Inc. Reports First Quarter 2026 Financial Results with Continued Growth in Robotaxi Revenues and Expanding Fleet Deployment | Q1 2026 revenue was US$34.3 million and cash plus investments were US$1.4355 billion as of March 31, 2026. |
| SI012 | Pony.ai | PONY AI Inc. Scales with 160% Robotaxi Revenues Growth YoY and 500% Fare-Charging Revenues Surge as 2026 Accelerates | Pony reported US$90.0 million of 2025 revenue, US$1.5148 billion of cash/investments at year-end, and UE breakeven in multiple cities. |
| SI013 | Pony.ai | PONY AI Inc. Files Its Annual Report on Form 20-F and Publishes Inaugural ESG Report | Pony filed its annual report on Form 20-F for the fiscal year ended December 31, 2025. |
| SI014 | WeRide | WeRide Reports Record Full Year 2025 Revenue of RMB684.6 Million, Up 90% Year over Year | WeRide reported 2025 revenue of US$97.9 million, R&D expense of US$196.2 million, and about US$1.0 billion of liquidity. |
| SI015 | WeRide | Form 20-F for WeRide Inc. filed 04/23/2026 | WeRide disclosed 2025 product revenue, service revenue, top-customer concentration, and continuing operating cash outflow in its Form 20-F. |
| SI016 | WeRide | Quarterly Results | WeRide maintains public quarterly-results access, unlike Haomo. |
| SI017 | Mobileye | Mobileye Releases Fourth-Quarter and Full-Year 2025 Results and Provides Business Overview | Mobileye reported US$1.894 billion of 2025 revenue, US$602 million of operating cash flow, and US$1.8 billion of cash. |
| SI018 | Research and Markets | L3/L4 Autonomous Driving and Startups Research Report, 2025 | The sector is entering a commercialization phase in 2026-2030 rather than a low-capex software phase. |
| SI019 | EEWorld | Competition in Level 4 Autonomous Vehicles is intensifying | Robotaxi players are moving from demonstration to commercial operations but remain constrained by cost, policy, and scale. |
| SI020 | Horizon Robotics | Journey 6 Series | Journey 6 emphasizes an integrated compute platform and mass-production toolchain. |
| SI021 | Horizon Robotics | Horizon SuperDrive | Horizon explicitly targets multiple mainstream price bands with its HSD offerings. |
| SI022 | KrASIA / 36Kr translation | Haomo AI hits a crossroads as commercialization challenges intensify | Commercialization challenges were intensifying before the final halt reports. |
| SI023 | KrASIA / 36Kr translation | Haomo AI halts operations after five years on the road | Haomo halted operations after a period of rising commercial and organizational strain. |
| SI024 | CarNewsChina | Haomo AI collapses: China’s $1 billion self-driving startup shuts down | The collapse reporting reinforces that funding history did not solve Haomo’s commercialization or cash needs. |
| SI025 | ChinaBizInsider | China Robotaxi Global Expansion 2026: WeRide, Pony.ai vs Waymo | Chinese robotaxi leaders are still sequencing markets around regulation, cost, and scale rather than easy profitability. |
| SE001 | Korea Newswire / Haomo.AI | Haomo Launches DriveGPT at 8th AI Day | DriveGPT launched as a GPT-style large model and Haomo said it had become a supplier for three automakers. |
| SE002 | Korea Newswire / Haomo.AI | Haomo Launches Three ADS Products for as Low as $411 | HP170, HP370 and HP570 were launched with explicit compute, sensor, and pricing disclosures. |
| SE003 | Gasgoo | Haomo.ai launches autonomous driving large model, DriveGPT | Gasgoo says DriveGPT used RLHF-style optimization and was tied to Haomo’s first mass-produced map-free NOH product. |
| SE004 | ArenaEV | China's Haomo delivers affordable ADAS kits to speed up adoption | Haomo’s ADAS kits were positioned as low-cost accelerants of adoption. |
| SE005 | ML Vehicle | Haomo AI faces internal shake-up amid mounting technology and commercialization hurdles | The report says Haomo co-developed MANA OASIS and faced delayed city-NOA and commercial-delivery milestones. |
| SE006 | Logistics Tech Outlook | HAOMO.AI | Top Logistics Tech Startup in China 2023 | HDeliver was framed as a low-cost autonomous-delivery system and the latest version was priced below CNY90,000. |
| SE007 | GitHub | haomo-ai organization page | The org page shows pinned public repositories with hundreds of stars around AV perception and mapping topics. |
| SE008 | GitHub | haomo-ai repositories list | Public repositories were updated as recently as 2025, indicating continued visible research output. |
| SE009 | StartupIntros | Haomo AI organization profile | External profile coverage preserves a community-facing snapshot of Haomo’s open technical footprint. |
| SE010 | Lei Mao blog | DriveGPT | The blog explains DriveGPT as a planning model over tokenized scenes and notes that it still depends on strong perception inputs. |
| SE011 | Tony Xu project page | DriveGPT4 project | DriveGPT4 is presented as an interpretable end-to-end autonomous-driving system built around large language models. |
| SE012 | arXiv | DriveGPT: A Scalable Behavior Model for Autonomous Driving | The paper frames DriveGPT as a scalable transformer behavior model for autonomous driving, useful as external technical context rather than direct proof of Haomo product maturity. |
| SE013 | arXiv | DriveGPT4: Interpretable End-to-End Autonomous Driving via Large Language Model | The paper shows the broader technical direction of interpretable LLM-based driving systems, but is not itself proof of Haomo deployment quality. |
| SE014 | Great Wall Motor / HKEX | 2025 Annual Report | Great Wall’s filing confirms Haomo remained strategically adjacent to its smart-driving ecosystem. |
| SE015 | APA Engineering | China Implements Stricter Regulations on OTA Updates and Autonomous Driving Features | China now requires pre-approval of safety-impacting OTA software updates. |
| SE016 | Yicai Global | Chinese L4 Autopilot Firm Haomo.AI Scores USD41.4 million in Series B2 | Yicai said Haomo’s ADAS kits had reached more than 20 models and nearly 140 million kilometers by April 2024. |
| SE017 | TechCrunch | Great Wall-backed autonomous driving startup Haomo is raising funding again | TechCrunch reported Haomo’s delivery vans had ferried close to 300,000 grocery parcels in Beijing and that Haomo said it supplied three other OEMs. |
| SE018 | Research and Markets | Chinese Independent OEMs' ADAS and Autonomous Driving Report, 2026 | 2026 is framed as a key inflection point in China intelligent driving, raising the bar for product maturity. |
| SE019 | 36Kr / BT Finance | After FSD Enters Chinese Market: Which is More Reliable - Tesla, Huawei or Momenta? | The article compares different ADAS failure modes and reinforces why local-road reliability and sensor tradeoffs matter. |
| SE020 | Horizon Robotics | Horizon SuperDrive | Horizon markets software-hardware integration, end-to-end models, and mass-production breadth on one product surface. |
| SE021 | Horizon Robotics | Journey 6 Series | Journey 6 highlights integrated compute and a mass-production development platform. |
| SE022 | Pony.ai | Pony.ai Business Overview | Pony shows how a rival exposes multiple autonomous product lines across operational contexts. |
| SE023 | WeRide | WeRide Technology | WeRide presents a hybrid end-to-end stack and a unified operations platform, clarifying the modern product benchmark. |
| SE024 | PR Newswire / DeepRoute.ai | DeepRoute.ai Launches DeepRoute IO 2.0, Smart Driving Platform Powered by VLA Technology | DeepRoute’s launch shows how peers now market multi-chip, multi-sensor, VLA-enabled stacks with current OEM proof. |
| SE025 | Mobileye | Cloud-Enhanced ADAS for select future Stellantis vehicles | Mobileye’s REM-backed ADAS illustrates the trust, data, and integration expectations Haomo ultimately had to match. |
| SE026 | GitHub | SuperFusion repository | The repository documents Haomo research on multilevel LiDAR-camera fusion for long-range HD map generation. |
| SE027 | GitHub | OverlapTransformer repository | The repository documents Haomo research on LiDAR-based place recognition. |
| SE028 | GitHub | Cam4DOcc repository | The repository documents Haomo research on camera-only 4D occupancy forecasting. |
| SE029 | GitHub | ModaLink repository | The repository documents Haomo research on multimodal image-to-point-cloud place recognition. |
| SE030 | GitHub | EcoDatum repository | The repository documents Haomo work on multimodal data curation for training efficiency. |
| SE031 | BusinessWire / Haomo.AI | Haomo unveils DriveGPT, speeds up commercialization and global expansion after landing in Europe and Israel | The original release ties DriveGPT to commercialization, exports, and the first three-automaker supplier narrative. |
| SE032 | BusinessWire / Haomo.AI | Haomo Launches Three ADS Products for as Low as $411, Pushes DriveGPT to Know Everything | The original release preserves Haomo’s detailed public SKU, sensor, compute, and data-scale claims. |
| SU001 | Great Wall Motor / HKEX | 2025 Annual Report | Great Wall identifies Haomo Zhixing as an associate and related-party ecosystem entity. |
| SU002 | China Daily | Haomo raises 100 million yuan in Series B1 round | China Daily says HPilot was equipped in more than 20 vehicle models and Haomo vans had delivered nearly 300,000 grocery parcels in Beijing. |
| SU003 | ChinaPEV | HAOMO.AI announces the completion of over 100 million yuan in B1 round financing | ChinaPEV says HPilot was equipped on more than 20 models and user-assisted mileage exceeded 120 million kilometers. |
| SU004 | TechCrunch | Great Wall-backed autonomous driving startup Haomo is raising funding again | TechCrunch says Haomo mostly sold to Great Wall, counted Meituan, Alibaba and JD.com among logistics clients, and had ferried close to 300,000 grocery parcels. |
| SU005 | EqualOcean | Autonomous Driving Company Haomo Secures CNY 300 Million in Series B2 Financing | EqualOcean says Haomo had collaborations with Alibaba, Wumart, Meituan and Dada, and that Wumart deliveries exceeded 120,000 orders across more than 60 communities. |
| SU006 | Gasgoo | Haomo.ai's Xiaomotuo completes 200,000 orders in autonomous delivery milestone | Gasgoo says Xiaomotuo completed more than 200,000 delivery orders and was serving supermarket and community scenarios. |
| SU007 | Gasgoo | Haomo.ai’s self-driving last-mile delivery vehicle greenlighted for operation in Beijing | Gasgoo says Haomo obtained vehicle coding for regular testing and operation in Beijing’s high-level automated-driving demonstration area. |
| SU008 | Logistics Tech Outlook | HAOMO.AI | Top Logistics Tech Startup in China 2023 | The article ties Haomo to Meituan, Wumart Dmall, and Dada Nexus and says Wumart Dmall route volume peaked near 700 daily orders. |
| SU009 | Yicai Global | Great Wall Motor-Backed Self-Driving Startup Haomo Bags Over USD14 Million in Fundraiser | Yicai Global says HPilot had been equipped in more than 20 vehicle models and the L4 van fleet had delivered nearly 300,000 grocery parcels. |
| SU010 | Yicai | Haomo.AI Lays Off Staff Amid Weak Car Sales, High R&D Costs | Yicai says more than 20 Great Wall models had been equipped with Haomo HPilot and other carmaker deals were signed but undisclosed. |
| SU011 | Yicai Global | Great Wall Motor-Backed Self-Driving Unicorn Haomo.AI Stops Work, Report Says | Yicai Global reported that Haomo had halted operations until further notice. |
| SU012 | BusinessKorea | Hyundai's Strategic Entry into Chinese EV Market with AI-enhanced Model | BusinessKorea said Beijing Hyundai planned a China EV using Haomo DriveGPT for launch in the second half of the following year. |
| SU013 | Hyundai Motor Group | Hyundai Motor Reaffirms China Commitment at Auto China 2026, Unveils IONIQ V as First Step in New Product Offensive | Hyundai’s official 2026 China launch materials say IONIQ V uses ADAS technologies powered by Momenta. |
| SU014 | PR Newswire / Hyundai Motor | Hyundai Motor Reaffirms China Commitment at Auto China 2026, Unveils IONIQ V as First Step in New Product Offensive | The release states that IONIQ V incorporates ADAS technologies powered by Momenta. |
| SU015 | Robotics International | Meituan, JD.com Deploy 12,000 Autonomous Delivery Units in Q2 2026 | The article says Meituan operated 8,400 autonomous delivery vehicles across 47 Chinese cities by June 2026. |
| SU016 | Dmall | From Digital Operations to AI-Native Retail: How Dmall and Wumart Won the Smartest Store in the World 2026 | Dmall says the AI-powered store model was replicated across 77 stores and customer satisfaction exceeded 95%. |
| SU017 | The Straits Times / sponsored content | From Innovation to Impact: DMALL Earns World Economic Forum Recognition for Scaled AI Deployment in Retail | The sponsored article highlights scaled AI retail deployment at Dmall and Wumart but does not mention Haomo. |
| SU018 | Great Wall Motor / CNInfo | 2024 Annual Results Announcement | Great Wall’s annual-results filing underscores the scale of the parent ecosystem around Haomo even though it does not provide Haomo customer-level disclosures. |
| SU019 | ML Vehicle | Haomo.AI Faces Commercialization Headwinds and Operational Crisis | ML Vehicle says Haomo was then supporting only two Hyundai models with planned delivery in August, but that development delays cast doubt on timing. |
| SU020 | DataM Intelligence | China Autonomous Last-Mile Delivery Market | The market report describes Meituan, Alibaba, and other platforms as major demand-side participants in China autonomous last-mile delivery. |
| SU021 | CarNewsChina | Great Wall-backed self-driving startup Haomo.AI reportedly collapsed and suspended operations | CarNewsChina says Haomo had been installed in more than 20 GWM-brand models and that the shutdown left support continuity uncertain. |
| SU022 | Korea Newswire / Haomo.AI | Haomo Launches DriveGPT at 8th AI Day | At its 2023 AI Day, Haomo said it had become a supplier to three automakers after Great Wall. |
| SU023 | ArenaEV | China's Haomo delivers affordable ADAS kits to speed up adoption | ArenaEV describes Haomo’s low-cost ADAS kits as aimed at speeding mainstream model adoption. |
| SU024 | Yicai Global | Great Wall Motor-Backed Self-Driving Startup Haomo Bags Over USD14 Million in Fundraiser | Yicai Global states that HPilot was equipped in more than 20 vehicle models and that Haomo’s L4 vans had delivered nearly 300,000 grocery parcels. |
| SU025 | Gasgoo | Haomo.ai launches autonomous driving large model, DriveGPT | Gasgoo said Haomo’s HPilot system had already been installed onto nearly 20 models and tied the roadmap to the Great Wall WEY ecosystem. |
| SR001 | State Council of the PRC / gov.cn | Notice on Further Strengthening the Administration of Automobile Product Admission, Recall, and OTA Software Upgrade for Intelligent and Connected Vehicles | The notice requires testing, system-boundary definition, accident reporting, OTA filing, and permits for autonomous-driving-function OTA changes. |
| SR002 | CMS | MIIT and SAMR Further Strengthening the Administration of Automobile Product Admission, Recall, and OTA Software Upgrade for Intelligent and Connected Vehicles | CMS explains that OTA changes related to autonomous-driving functions require permits and that certain accidents must be reported within 24 or 48 hours. |
| SR003 | Beijing Municipal Government | Beijing Autonomous Vehicle Regulation | Beijing established a legal framework for autonomous vehicles including freight and delivery applications. |
| SR004 | Beijing Government English | Beijing Passes Regulations to Support, Govern Autonomous Driving Technologies | The English summary says the rules cover infrastructure, traffic management, and safety assurance for L3 and higher systems. |
| SR005 | State Council Information Office | China releases security guide on cross-border automobile data transfers | The guide covers important-data identification and security requirements across driving automation and software-upgrade scenarios. |
| SR006 | DLA Piper Privacy Matters | China: New guidance on data transfer and identification of important data in the automotive sector | DLA Piper says automotive data controllers must identify important data and comply with the applicable cross-border transfer pathways. |
| SR007 | Law.asia | China’s Autonomous Vehicle Regulations | The article notes that Chinese AV regulation increasingly assigns operational and accident liability to manufacturers or operators under L3+ scenarios. |
| SR008 | CnEVPost | China sets unified safety baseline for L3, L4 autonomous driving | CnEVPost says GB 44721-2026 takes effect July 1, 2027 and requires autonomous systems to be at least as safe as a competent and attentive driver. |
| SR009 | KrASIA | After a fatal SU7 crash, China reins in smart driving hype | KrASIA says product language was tightened, driver-control-free parking features were curtailed, and public-road testing was restricted after the fatal SU7 crash. |
| SR010 | Mint | China eyes Level 3 autonomy rollout by 2026, tightening grip on assisted-driving tech after fatal Xiaomi crash | Mint reports stricter hardware, software, and marketing oversight after the Xiaomi crash. |
| SR011 | SEC | NVIDIA Corporation Form 8-K dated April 9, 2025 | NVIDIA disclosed that a license would be required for H20 exports to China and that the requirement would remain in effect indefinitely. |
| SR012 | NVIDIA | NVIDIA Announces Financial Results for First Quarter Fiscal 2026 | NVIDIA reported a $4.5 billion H20-related charge and said it was unable to ship an additional $2.5 billion of H20 revenue in the quarter. |
| SR013 | Great Wall Motor / HKEX | 2025 Annual Report | Great Wall still identifies Haomo Zhixing as an associate and related-party ecosystem entity. |
| SR014 | Yicai Global | Haomo.AI Lays Off Staff Amid Weak Car Sales, High R&D Costs | Yicai says Haomo let go employees, Great Wall remained actual controller, and Great Wall also tapped DeepRoute and Hesai. |
| SR015 | Yicai Global | Great Wall Motor-Backed Self-Driving Unicorn Haomo.AI Stops Work, Report Says | Yicai reported halted operations, around 300 staff across cities, and salary arrears beginning in August 2025. |
| SR016 | CarNewsChina | Haomo.AI reportedly collapsed and suspended operations | CarNewsChina framed the shutdown as a collapse that left future support and company status uncertain. |
| SR017 | China Daily | Haomo raises 100 million yuan in Series B1 round | China Daily described more than 20 vehicle models and nearly 300,000 grocery parcels, confirming meaningful historical deployment. |
| SR018 | TechCrunch | Great Wall-backed autonomous driving startup Haomo is raising funding again | TechCrunch said Haomo mostly sold ADAS to Great Wall and counted Meituan, Alibaba and JD.com among logistics clients. |
| SR019 | EqualOcean | Autonomous Driving Company Haomo Secures CNY 300 Million in Series B2 Financing | EqualOcean linked Haomo to Wumart, Meituan, and Dada and described route-level delivery volume in Beijing. |
| SR020 | Gasgoo | Haomo.ai's Xiaomotuo completes 200,000 orders in autonomous delivery milestone | Gasgoo says Xiaomotuo completed more than 200,000 orders and expanded across multiple scenarios. |
| SR021 | Gasgoo | Haomo.ai’s self-driving last-mile delivery vehicle greenlighted for operation in Beijing | Gasgoo says Haomo received the coding needed for regular testing and operation in Beijing’s high-level demonstration area. |
| SR022 | BusinessKorea | Hyundai's Strategic Entry into Chinese EV Market with AI-enhanced Model | BusinessKorea said Beijing Hyundai planned a China EV incorporating Haomo DriveGPT. |
| SR023 | Hyundai Motor Group | Hyundai Motor Reaffirms China Commitment at Auto China 2026, Unveils IONIQ V as First Step in New Product Offensive | Hyundai’s 2026 China launch materials say IONIQ V uses ADAS technologies powered by Momenta. |
| SR024 | Robotics International | Meituan, JD.com Deploy 12,000 Autonomous Delivery Units in Q2 2026 | The article says Meituan operated 8,400 autonomous delivery vehicles across 47 cities and faced strong cost and permit dynamics. |
| SR025 | Dmall | From Digital Operations to AI-Native Retail: How Dmall and Wumart Won the Smartest Store in the World 2026 | Dmall said the Wumart AI-retail model scaled to 77 stores with satisfaction above 95%, but it did not mention Haomo. |
| SR026 | Bird & Bird | APAC Employment Law – Year In Review - China | Bird & Bird says enforceable non-competes require specific trade-secret linkage and monthly post-termination compensation, typically at least 30% of prior average salary. |
| SR027 | Great Wall Motor / CNInfo | 2024 Annual Results Announcement | Great Wall’s annual results underscore the scale of the parent ecosystem but do not provide Haomo rescue-financing commitments. |
| SR028 | ML Vehicle | Haomo.AI Faces Internal Shake-Up Amid Mounting Business Challenges | The article says Haomo supported only two Hyundai models with delays and that Little Camel’s 2025 sales target was just 50 units. |
| SR029 | ChinaPEV | HAOMO.AI announces the completion of over 100 million yuan in B1 round financing | ChinaPEV says HPilot was on more than 20 models and the assisted-driving mileage exceeded 120 million kilometers. |
| SR030 | Yicai Global | Great Wall Motor-Backed Self-Driving Startup Haomo Bags Over USD14 Million in Fundraiser | Yicai said HPilot had been equipped in more than 20 vehicle models and Haomo’s L4 vans had delivered nearly 300,000 grocery parcels. |
| SR031 | PR Newswire / Hyundai Motor | Hyundai Motor Reaffirms China Commitment at Auto China 2026, Unveils IONIQ V as First Step in New Product Offensive | The PR Newswire release also says Hyundai’s new China EV incorporates ADAS technologies powered by Momenta. |
| SR032 | Logistics Tech Outlook | HAOMO.AI | Top Logistics Tech Startup in China 2023 | Logistics Tech Outlook tied Haomo to Meituan, Wumart Dmall, and Dada Nexus, with average daily route volume around 300 and peaks around 700. |
| SV001 | CompaniesMarketCap | Pony AI (PONY) - Market capitalization | As of August 2026 Pony AI has a market cap of $3.39 Billion USD. |
| SV002 | CompaniesMarketCap | WeRide (WRD) - Market capitalization | As of August 2026 WeRide has a market cap of $1.92 Billion USD. |
| SV003 | CompaniesMarketCap | Mobileye (MBLY) - Market capitalization | As of August 2026 Mobileye has a market cap of $6.77 Billion USD. |
| SV004 | Pony.ai Investor Relations | Pony AI Inc. scales 160% Robotaxi revenues growth YoY and over 500% fare-charging revenues growth in Q4, while achieving consecutive UE breakeven in Guangzhou and Shenzhen | Total revenues were US$90.0 million in 2025 and cash and cash equivalents, short-term investments, restricted cash and long-term debt instruments for wealth management were US$1,514.8 million as of December 31, 2025. |
| SV005 | WeRide Investor Relations | WeRide Reports Record Full Year 2025 Revenue of RMB684.6 Million, Up 90% Year-over-Year | Total revenue was RMB684.6 million (US$97.9 million) in 2025 and gross margin was 30.2%, while cash and equivalents, time deposits, wealth-management products, and restricted cash totaled about US$1.0 billion. |
| SV006 | Mobileye Investor Relations | Mobileye releases fourth quarter and full year 2025 results and 2026 financial outlook | Full year 2025 revenue increased 15% year-over-year to $1,894 million and the company ended the year with $1.8 billion in cash and cash equivalents. |
| SV007 | CnEVPost | Chinese startup Haomo raises nearly $150 million in new funding, becoming latest self-driving unicorn | Haomo raised nearly RMB1 billion in a Series A round and became the latest self-driving unicorn in China. |
| SV008 | TechCrunch | Backed by Chinese carmaker Great Wall, Haomo raises $14M for autonomous driving tech | Haomo raised about $14 million in the first tranche of its Series B and had HPilot in more than 20 models while its delivery vans had ferried close to 300,000 grocery parcels. |
| SV009 | Yicai Global | Chinese L4 Autopilot Firm Haomo.AI Scores USD41.4 million in Series B2 | Haomo secured CNY300 million in Series B2 after raising more than CNY100 million in B1 earlier in the year. |
| SV010 | Yicai Global | Great Wall Motor-Backed Self-Driving Unicorn Haomo.AI Stops Work, Report Says | Haomo reportedly halted operations, had salary arrears, and may have had frozen accounts after missing performance targets. |
| SV011 | KrASIA | Haomo.AI hits crossroads amid leadership shakeup and tech setbacks | Commercialization challenges intensified before the final halt reports, including weak project progress and reduced low-speed vehicle targets. |
| SV012 | KrASIA | Haomo.ai halts operations as employees told not to report to work | Employees were told not to report to work as Haomo’s crisis reached a breaking point. |
| SV013 | CarNewsChina | Haomo AI collapses: China’s $1 billion self-driving startup shuts down | The collapse reporting reinforced that Haomo’s unicorn-era financing did not solve commercialization or cash needs. |
| SV014 | Hong Kong Exchanges and Clearing Limited | Great Wall Motor Company Limited Annual Report 2025 | The share of accumulated losses at the end of the period is RMB413,655,614.14. |
| SV015 | CB Insights | HAOMO.AI - Products, Competitors, Financials, Employees, Headquarters Locations | Total Raised: $245.3M. |
| SV016 | China Daily | Haomo.AI raises over 100 million yuan | Haomo said it raised over 100 million yuan in Series B1 and had HPilot in more than 20 models with over 120 million kilometers of assisted-driving mileage. |
| SV017 | Gasgoo | Haomo.ai's Xiaomotuo completes 200,000 orders in autonomous delivery milestone | Gasgoo says Xiaomotuo completed more than 200,000 delivery orders and was serving supermarket and community scenarios. |
| SV018 | EqualOcean | Autonomous Driving Company Haomo Secures CNY 300 Million in Series B2 Financing | EqualOcean says Haomo had collaborations with Alibaba, Wumart, Meituan and Dada, and that Wumart deliveries exceeded 120,000 orders across more than 60 communities. |
| SV019 | BusinessKorea | Hyundai's Strategic Entry into Chinese EV Market with AI-enhanced Model | BusinessKorea said Beijing Hyundai planned a China EV incorporating Haomo DriveGPT. |
| SV020 | Hyundai Motor Group | Hyundai Motor Reaffirms China Commitment at Auto China 2026, Unveils IONIQ V as First Step in New Product Offensive | Hyundai’s 2026 China launch materials say IONIQ V uses ADAS technologies powered by Momenta. |
| SV021 | Robotics International | Meituan, JD.com Deploy 12,000 Autonomous Delivery Units in Q2 2026 | The article says Meituan operated 8,400 autonomous delivery vehicles across 47 cities in Q2 2026. |
| SV022 | Dmall | From Digital Operations to AI-Native Retail: How Dmall and Wumart Won the Smartest Store in the World 2026 | Dmall said the Wumart AI-retail model scaled to 77 stores with satisfaction above 95%, but it did not mention Haomo. |
| SV023 | State Council of the PRC / gov.cn | Notice on Further Strengthening the Administration of Automobile Product Admission, Recall, and OTA Software Upgrade for Intelligent and Connected Vehicles | The notice requires testing, system-boundary definition, accident reporting, OTA filing, and permits for autonomous-driving-function OTA changes. |
| SV024 | State Council Information Office | China releases security guide on cross-border automobile data transfers | The guide covers important-data identification and security requirements across driving automation and software-upgrade scenarios. |
| SV025 | Law.asia | China’s Autonomous Vehicle Regulations | The legal analysis says accident liability for higher-level autonomy increasingly shifts toward manufacturers or operators. |
| SV026 | ML Vehicle | Haomo.AI Faces Internal Shake-Up Amid Mounting Business Challenges | The article says Haomo supported only two Hyundai models with delays and that Little Camel’s 2025 sales target was just 50 units. |
| SV027 | SEC EDGAR | WeRide Inc. Annual Report on Form 20-F for the fiscal year ended December 31, 2025 | As of the date of this annual report, no subsidiary has paid dividends or made other distributions to the Cayman Islands holding company, and the filing details 2025 capital contributions, loans, and restricted net assets. |
| SV028 | CompaniesMarketCap | Pony AI - Annual Reports (20-F) | Latest form 20-F annual reports for Pony AI CIK: 0001969302. |
| SV029 | CompaniesMarketCap | WeRide - Annual Reports (20-F) | Latest form 20-F annual reports for WeRide CIK: 0001867729. |
| SV030 | CompaniesMarketCap | Mobileye - Annual Reports | Annual Report 2025. |
| SV031 | ChinaPEV | HAOMO.AI announces the completion of over 100 million yuan in B1 round financing | Haomo said it obtained over 100 million yuan in B1 financing and had HPilot on more than 20 models with user-assisted driving mileage above 120 million kilometers. |
| SV032 | Gasgoo | Haomo.ai’s self-driving last-mile delivery vehicle greenlighted for operation in Beijing | Gasgoo says Haomo obtained vehicle coding for regular testing and operation in Beijing’s high-level automated-driving demonstration area. |
| SV033 | Horizon Robotics | Horizon Robotics investor-relations reports page | The investor-relations navigation includes Financial Reports, Announcements & Notices, and Corporate Governance sections. |