初创公司尽调
尽调报告 Defense Technology / AI Autonomy Late-stage private / pre-IPO 2026-10-01

Shield AI

防务自主平台兼具强劲融资势头与未决披露缺口

Shield AI 将差异化自主平台与防务落地应用相结合,但当前估值仍计入了大量执行风险与披露风险。

覆盖范围与披露说明

基于 2026-10-01 基准日期执行的复核研究;估值、收入、员工人数及融资数据均已结合当前章节重新核验。

封面要素

投后估值 01
12700 USD million [CI013]
FY2025 收入 02
300 USD million [CI007]
2026 收入指引 03
540 USD million [CI008]
员工人数 04
1500 employees [CO015]

公司概况

Shield AI 是一家成立于 2015 年、总部位于圣地亚哥的成长期防务科技私有企业。核心产品体系由 Hivemind 自主软件以及 V-BAT、X-BAT 飞行器项目构成,通过硬件采购、自主软件集成及配套任务支持服务,面向美军各军种与盟国防务部门交付。公司目前仍保持私有化且信息披露极少,在毛利率、客户经济模型以及最新融资方案的最终交割条款上,关键问题仍未得到解答。

官网
shield.ai
成立时间
2015-01-01
创始人
Andrew Reiter, Brandon Tseng, Ryan Tseng
创立地点
San Diego, CA
总部
San Diego, CA
产品
任务自主软件与自主防务飞行平台,以 Hivemind、V-BAT 与 X-BAT 为核心。
客户
美国各军种及盟国防务部门
商业模式
防务硬件采购结合自主软件授权与集成服务
阶段
Late-stage private / pre-IPO
融资情况
2026 年 3 月宣布 Series G 融资,包含 $1.5B 股权与 $500M 固定回报优先股,估值达 $12.7B
[CO001, CO003, CI013, CI014]

执行摘要

主要优势

  • 任务自主软件已在自研与第三方防务平台上落地,并有明确的集成案例。
  • 由 2026 年 3 月一笔大额融资方案支持,经确认估值达 $12.7B。
  • 广泛覆盖美国及盟国防务客户,降低了对单一项目的依赖。

主要风险

  • 毛利率、现金消耗与客户集中度仍未披露。
  • X-BAT 与制造规模化扩张带来了显著的执行风险。
  • 安全合规、监管审批与并购交割依赖可能推迟商业化进程。

未决问题

  • 经审计的 ARR、现金余额与活跃客户数仍未披露。
  • 完整的公司治理、董事会与股权结构表细节仍未公开。
  • 优先股条款与最终并购交割机制仍有待确认。
  • NRR 与合同续约持久度均未公开披露。

目录

Chapter 01

01公司概况

1.1 业务定位与商业模式

Shield AI 是一家成立于 2015 年、总部位于圣地亚哥的防务科技私有企业。独立留存资料均将 Brandon Tseng、Ryan Tseng 与 Andrew Reiter 列为联合创始人。公司业务横跨 Hivemind 自主软件与自研飞行器:V-BAT 是现役无人机产品线,X-BAT 则处于在研阶段。由此形成两条潜在变现路径:整机或执飞系统销售,以及将 Hivemind 授权给第三方平台;但留存证据并未披露合同层面的经济指标或经常性软件收入。此外,官方网站 shield.ai 需与 shieldai.ai 区分开来,后者是 2024 年成立的一家无关的数据保护与合规平台。[CO001, CO002, CO003, CO016]

核心指标快照表
指标数值 / 状态日期置信度证据缺口或限定说明
企业性质防务科技私有企业2026-10-01中未留存上市公司申报文件
成立时间2015历史中由公司官方口径及独立资料报道
总部地点加利福尼亚州圣地亚哥2026-10-01中多份独立资料一致;未留存具体注册地
产品体系Hivemind 软件及 V-BAT、X-BAT 飞行器2026-03-26中合同层面的软硬件经济指标未予披露
投后估值$12.7B2026-03-26高对应宣布的 Series G 轮
Series G宣布募集 $1.5B2026-03-26高公告称正在融资;保留证据无法证实资金已到位
优先股公告披露 $500 million2026-03-26高固定回报条款与偿付顺位未予披露
延迟提款信贷额度已承诺 $250 million2026-03-26高提款条件及实际提款金额未予披露
FY2026 营收预计超过 $540 million2026中第三方报道的预测值;非经审计营收或 ARR
员工人数据报道 1,500 人2026-03低单一二手汇编数据,援引自 Revelio Labs
累计融资总额2026-10-01低不同工具及公告金额无法直接累加为已到账现金
ARR2026-10-01低保留来源均未披露经常性软件收入
客户数量2026-10-01低已点名项目无法构成活跃客户基数
办公地点宣称在美国、欧洲、中东和亚太地区设有业务机构2026-03-26中公司未披露完整的网点数量

G 轮、优先股及信贷额度金额均为已公布交易额,并非可用现金;营收属于预测数据,员工人数为第三方报道,null 表示保留证据不足以支持给出精确数值。

[CO001, CO003, CO010, CO011, CO014, CO015]

1.2 领导团队与公司治理

创始团队与防务作业、工程研发的匹配度极高:Brandon Tseng 曾服役于海豹突击队,Ryan Tseng 是 WiPower 创始人,Andrew Reiter 则曾任 Draper Labs 工程师。管理层在 2025 年 3 月发生重大调整,Gary Steele 出任首席执行官,Ryan Tseng 转任战略职务。2026 年 3 月的融资公告还把治理权同新注入资金绑定:Advent 董事长 David Mussafer 预计进入董事会,JPMorganChase 高管 Todd Combs 预计出任董事会观察员。创始人现任头衔、完整董事会名单、投票权控制结构、清算优先权以及接班梯队深度,仍属于私有尽调范畴,尚非经核实的公开事实。[CO002, CO009, CO012]

领导层与创始人表
人员角色 / 状态背景职能覆盖关键人物依赖度
Brandon Tseng联合创始人;当前运营职衔未证实前海豹突击队成员防务用户背景与公司初创组建对创始人存在重大依赖;当前职权范围有待确认
Ryan Tseng联合创始人;据报道在 2025 年 3 月后担任战略职务创立 WiPower,后被 Qualcomm 收购公司战略与过往高管领导经历首席执行官交接后承担关键业务连续性职责
Andrew Reiter联合创始人;当前运营职衔未证实前 Draper Labs 工程师技术创始覆盖在保留证据中,其当前职责与留任意向均未公开
Gary Steele首席执行官前 Cisco 高管高管领导力与规模扩张外部 CEO 降低了单一创始人依赖,但任职时间相对较短
David Mussafer公告候任董事会董事Advent International 主席投资方治理与董事会监督融资公告中将该任命描述为预期安排
Todd Combs公告候任董事会观察员JPMorganChase 投资高管投资方观察与战略财务对接观察员权利及权限范围未予披露

仅为部分列举:创始人当前职衔、完整高管团队、董事会全员、各委员会席位及投票权控制安排在保留来源中均未披露。

[CO002, CO009, CO012]

1.3 资本结构与业务规模

2026 年 3 月 26 日,Shield AI 宣布按 $12.7B 投后估值募集 $1.5B 的 Series G 股权融资,以及 $500M 固定回报优先股。同一公告还称,Blackstone 旗下基金承诺提供额外 $250M 的延迟提款信贷额度。上述数字均为宣布的交易金额,并不等于已到账现金;消息源也未提供优先股回报率、清算受偿顺位、提款条件或完全稀释后持股比例。独立资料显示,公司曾在 2022 年和 2023 年分别完成 $165M 和 $200M 融资。第三方预计其 2026 财年收入超 $540M,但留存资料中没有任何经审计的财务报表能证实该预测、ARR、毛利率或客户数。[CO006, CO007, CO010, CO011, CO012, CO014, CO015]

利益相关方或投资方图谱
利益相关方角色控制权或经济重要性尽调重点追问
Advent InternationalG 轮领投方领投 $1.5 billion 股权轮;其主席获公布为候任董事获取持股比例、投票权、优先权及董事会席位条款
JPMorganChase 安全与韧性计划G 轮联合领投方战略机构资本;高管获公布为董事会观察员确认投资规模与观察员知情权
Blackstone 旗下管理基金优先股及信贷额度提供方$500 million 优先股投资外加 $250 million 延期提款承诺额度审查固定回报率、受偿顺位、限制性契约、提款测试条件及已出资额
Snowpoint Ventures现有投资方G 轮融资公告中点名的参投方确认同比例跟投权与累计持股比例
InnovationX、Riot Ventures、Disruptive 与 Apandion现有投资方财团点名参投方,各自出资金额未披露逐一梳理各持有人持股比例、权益及老股交易动态
Sagewind Capital / Aechelon出售方及拟收购标的收购 Aechelon 占用了部分融资收益核实收购价格、对价结构、审批进度及交割状态

仅为抽样列示:保留公告虽列举了参投方,但未提供完整资本结构表、个人持股比例、老股交易、融资条款或收购对价明细。

[CO010, CO011, CO012, CO013]

1.4 发展里程碑与尽调重点

公开编年史显示,公司从 2016 年获得国防创新部门(DIU)合同起步,历经 2021 年对 Heron Systems 和 Martin UAV 的收购、两轮估值上调的私募融资、2025 年首席执行官更迭,再到 2026 年最新融资与拟收购 Aechelon 的交易。公开记录中也包含一起反向运营事件:Wikipedia 援引报道指出,2024 年 4 月一架 V-BAT 着陆时导致一名美海军官兵三根手指部分截肢,并引发了阶段性禁飞限制。该事件证实了安全与部署风险的存在,但不能直接推算全局事故率。当前最高优先级的尽调事项依然是:交易交割进展、融资具体条款、经审计的收入质量、客户集中度、完整管理团队履历以及人员受伤后采纳的运营控制措施。[CO004, CO005, CO006, CO007, CO008, CO009, CO010, CO013]

FO001: 公司里程碑时间线

八个有信源依据的里程碑追溯了 Shield AI 从成立至 2026-03 融资及拟议收购的历程。

仅在收录信源明确标明月份时使用月度日期;仅标明年份的里程碑不添加虚构月份。

公司成立 [CO001, CO002]

获得首个国防创新部门合同 [CO004]

收购 Heron Systems 与 Martin UAV [CO005]

$165 million 融资轮次 [CO006]

$200 million 融资轮次 [CO007]

V-BAT 伤人事件与飞行限制 [CO008]

首席执行官更迭 [CO009]

宣布 Series G 融资与拟收购 Aechelon [CO010, CO013]

[CO004, CO005, CO006, CO007, CO008, CO009, CO010, CO013]
Chapter 02

02市场分析

2.1 市场边界与替代方案

可投资的市场边界锁定在军事任务自主系统与自主飞行器,而非宽泛的国防开支或通用人工智能。纳入的收入池包括自主软件授权与集成、自主无人机销售以及外包执飞服务;仿真属于相邻领域,可用于加速自主训练与验证。行业现状主要依靠遥控或人工操纵飞行器,传统机载软件与总包集成商则是独立自主技术栈的直接替代品。Shield AI 的自身演示覆盖了 VISTA、MQ-20 Avenger 以及双机编队 Firejet 飞行器,证明了跨平台通用性,但这并不意味着每笔防务飞行器预算都能被其撬动。此外,位于 shieldai.ai 的同名合规平台属于无关实体,已从本市场定义中剔除。[CM001, CM002, CM004, CM005]

市场定义与边界
细分赛道 / 类别纳入支出剔除支出采购方 / 付费方相关性
任务自主软件与军用平台绑定的授权、集成、测试及技术支持通用企业级 AI 与无关软件国防采办项目、军种与平台 OEMHivemind 核心市场
自主飞行器与无人机系统飞行器采购、有效载荷集成及后勤维保不含自主组件的有人机支出美军各军种、海岸警卫队及盟国防务部门V-BAT 核心市场及 X-BAT 潜在市场
ISR 即服务外包飞行器运营及任务支援无关的监视服务政府机构及作战司令部采购方不拥有飞行平台所有权的替代采购路径
仿真与合成环境自主算法训练、验证及仿真软件通用游戏与非防务仿真自主技术开发商、主承包商及防务项目拟进行的 Aechelon 收购所强化的邻近市场
现有方案与替代方案远程遥控、手动操纵及主承包商自研自主集成不相关的 shieldai.ai 合规平台现有项目办公室与既有承包商奠定转换成本与自研还是外购的基准

本边界系分析师基于本章保留的产品、客户及合同证据划定;并非出版机构定义的市场分类法。

[CM001, CM002, CM003, CM004, CM005, CM013]

2.2 受限证据下的市场测算

公开数据引用的统计分母各异,不能混为单个 TAM。二手消息源称 2023 年全球军用无人机市场规模约 $14.4B,预计 2032 年达 $47.16B,但未给出可核验的方法论或 Shield AI 可触达部分的拆分口径。另一项分析指出,2025 年防务科技风险投资交易规模达 $49.1B,其中约 $12B 流向自主系统,同时还引用了五角大楼 FY2026 基础预算申请 $961.6B 加上 $150B 的补充拨款。这些都属于资本供给与政府预算背景,并非客户实际支出的 TAM。Fortune 报道的公司 2026 年收入预测至少为 $540M,只是一项当期捕获信号,无法视为 SOM,因为消息源并未界定目标市场或一致的分母口径。[CM006, CM007, CM008, CM009, CM010, CM011]

市场规模与背景观察维度
发布机构统计周期地域 / 市场数值复合年均增长率 (CAGR)统计口径与方法置信度局限性
TSG Invest2023 实际 / 基准全球军用无人机$14.4B未说明二手估算数据;抓取文本中未标明原始发布机构低包含了超出 Shield AI 自主软件及平台业务范畴的军用无人机
TSG Invest2032 预测全球军用无人机$47.16B未说明二手预测数据;抓取文本中未呈现底层测算模型低未按地域、类别拆分,亦无 Shield AI 可触达份额
Angel Investors Network 援引 PitchBook/Defense News2025防务科技风险投资交易$49.1B未说明报道的一级市场交易总规模中资金供应并非客户需求或 TAM
Angel Investors Network2025自主系统在防务科技风投中的占比约 $12B未说明防务科技投资中报道的细分类别投向低投资类别划分边界与交易统计方法未披露
Angel Investors NetworkFY2026 预算申请美国五角大楼基础预算加补充预算$961.6B 基础预算 + $150B 补充预算不适用报道的联邦预算申请及调解补充案中宽泛的国防预算;仅有一小部分未公开比例与自主化直接相关
Fortune2026 预测Shield AI 营收至少 $540M预计增速超 80%管理层基于 2025 营收所作预测;不计入 Aechelon中公司营收属于远期捕获值,缺乏市场分母前提下不能等同于 SOM

各项数值保留了各自来源的时间周期与统计分母。由于各来源口径不可比,本表特意未做加总或折算为 TAM/SAM/SOM。

[CM006, CM007, CM008, CM009, CM010, CM011]

2.3 买方、用户、付费方与采纳路径

核心付费方为政府军工项目办公室或盟国防务部门,实际使用者则是作战官兵、任务规划人员与装备操作员。采购可通过直接整机购买、服务合同(如海岸警卫队 $198M 的 V-BAT IDIQ 合同),或将自主系统集成进总包商与政府已有平台来推进。独立报道称,预计 2026 年收入中有超过一半来自国际客户,使得出口许可与盟国预算成为采纳的关键变量。采购触发点主要在于通信受阻、GPS 遭拒止或无可用跑道等极端环境下的实战表现;公司将高韧性机载自主作为应对方案。不过,美空军在 Hivemind 与 Anduril 的 Lattice 之间切换的测试表明,买方可能倾向于要求互操作性,而非接受单一供应商锁定。[CM002, CM003, CM012, CM013, CM014]

细分赛道与采购方图谱
细分赛道采购方使用方付费方工作流预算归属方采纳触发因素
协同作战飞机自主系统美国空军项目与飞机团队任务规划员与操纵员美国联邦国防预算跨机型集成并试飞自主系统军种采办 / 项目办公室经充分验证且具备必需互操作性的任务自主能力
海上 ISR 服务美国海岸警卫队巡逻艇官兵与 ISR 操作员美国联邦政府采购合同在 IDIQ 框架下部署 V-BAT 执行海上监视任务海岸警卫队采办 / 项目办公室摆脱跑道依赖的持久 ISR 能力
盟国自主飞行器盟国防务部门本国军事作战部队盟国国防预算采购飞行器或获取自主软件授权用于本土任务国家级采购主管机构强对抗环境作战能力与盟军协同互操作性
主承包商与 OEM 集成防务主承包商与飞机 OEM平台工程与任务团队主承包商、采办项目或政府专项集成预算将 Hivemind 嵌入第三方平台平台项目领导层相比内部自研更快速的自主化部署
X-BAT 远征空中战力美国海军及潜在盟国采购项目航空与海上任务操纵员未来战机研发与采办项目研发、测试、制造并列装无跑道依赖飞机机型项目办公室无跑道环境下的分布式空中战力部署需求

采购方角色系基于保留的客户、合同及平台证据分析归纳得出;具体项目预算归属与采办职权仍需采购方实证确认。

[CM002, CM003, CM004, CM012, CM013, CM014, CM015]
FM001: 市场证据锚点

以独立卡片呈现,避免暗示异构的预算、资本、合同、收入与产能指标构成嵌套的 TAM/SAM/SOM 层级体系。

自主系统的 $12B 数据及 150 架战机生产速率为信源所述的近似值或公司规划值;未做任何换算或汇总。

[CM007, CM008, CM009, CM011, CM013, CM015]

2.4 采纳制约因素与估值关联

需求信号十分充沛,但落地节奏与转化率仍存不确定性。防务采购容易因项目变更、预算重新划拨以及国会程序而延期,且多供应商架构削弱了软件独占的假设。国际市场扩张加大了出口管制风险敞口。此外,悬而未决的 Aechelon 收购形成了近期的监管依赖:Fortune 报道最终融资交割取决于该收购审批,若交易未获批准公司将重新评估融资。X-BAT 则拉高了资本密集度与排期风险;GeekWire 报道的公司规划要求 2028 年在德梅因产出首架原型机,并在 2030 年代初达到年产约 150 架的规模。因此,估值工作应针对合同转化、国际审批与工厂爬坡进行情景测试,而非直接把宽泛的防务预算算作自身营收机会。[CM010, CM012, CM014, CM015, CM016]

2.5 附注与论证

Chapter 03

03竞争格局

3.1 竞争格局与买方替代方案

买方可将 Shield AI 与几类不同的解决方案进行横向比对。在所分配证据中,Anduril 是最接近的风投背景全栈同行;Palantir 属于相邻的战场数据平台,而非飞行自主供应商。Skydio 专注较小型的商用及战术无人机;General Atomics、Northrop Grumman 与 Lockheed Martin 则代表传统主机厂或主承包商替代方案。Shield AI 自身业务横跨 Hivemind 自主系统、V-BAT 与 X-BAT 飞行器以及 Aechelon 仿真技术。这一格局亦具备竞合属性:Hivemind 入选 Anduril 的 Fury 战机,使得 Anduril 既是竞争对手,又是接入第三方机体的渠道。买方亦可选择自研或继续使用遥控及旧有系统,但分配证据并未量化其成本、实施周期或实战效能。[CP001, CP003, CP004, CP005, CP006]

竞品与替代方案概况
替代方案类别公开规模或融资证据目标采购方或应用场景已证实的差异化优势分配证据中的局限性
Shield AI自主软件加自研飞行平台规划占地 420,000 sq. ft. 的 X-BAT 厂区;据报道有 391 个在招岗位军用飞机自主系统、ISR 及自主战斗机Hivemind 跨多机型适配;V-BAT 与 X-BAT 产品组合可比定价及规模化量产表现未予披露
Anduril直接全栈防务科技对标企业Acquinox 披露 2024 估计营收约 $1B,2025 报道估值为 $30.5B一体化防务系统与协同作战飞机全栈硬件、传感器、软件及 Lattice OS分配证据未提供可比单机经济模型
Palantir相邻防务数据与决策软件Acquinox 披露 2024 营收为 $2.87B,公开市值约 $350B战场数据集成与企业级决策支持侧重数据集成而非战机飞行自主未见有可比飞行自主产品的证据
Skydio相邻商用与战术微小型无人机厂商Acquinox 披露估计营收约 $180M,2023 报道估值为 $2.2B商用与战术微小型无人机任务聚焦微小型无人机自主飞行分配证据未比对强对抗环境战机的性能
General Atomics既有自主飞行器竞争对手分配证据中未量化协同作战飞机与军用无人机系统 (UAS)深度参与成熟军机项目据报道坠机事故后暂停测试
Northrop Grumman传统航空航天既有巨头分配证据中未量化协同作战飞机与防务平台战机研制与主承包商地位当前自主能力与定价未量化
Lockheed Martin传统航空航天既有巨头分配证据中未量化军用战机与平台捆绑能力传统平台与采办渠道优势地位当前自主能力与定价未量化
内部自研或沿用旧有方案维持现状的替代路径分配证据中未量化采购方自主集成或继续使用远程遥控完全掌控集成流程与现有工作流成本、周期、战备就绪度及转换权衡均未获证实

仅基于分配证据池的部分竞品全景。Acquinox 将其营收标注为估计值并按时期列出报道估值;公司规划与招聘人数并不代表实际产量或营收。

[CP001, CP003, CP004, CP005, CP006, CP007, CP009, CP015, CP016]

3.2 技术能力与产业地位

Shield AI 经证据支持的最强优势,在于公司声称 Hivemind 已搭载过的机型广泛度:X-62 VISTA、MQ-20 Avenger 以及双机编队 MQM-178 Firejet。这支撑了软件可移植性的逻辑,但这并非独立对比基准,也无法证明其在所有平台上都具备同等战备状态。与此同时,公司正在构筑自研机体能力。GeekWire 报道其规划了一处 420,000 平方英尺的 X-BAT 厂房,未来可能再扩建 100 万平方英尺,目标是在 2030 年代初达到年产约 150 架的水平。行业媒体报道亦指出,复合材料设计、分析与制造是 X-BAT 的关键投入要素。上述规划有望缩小传统巨头的制造优势,但在产能、质量、成本和交付证据充分显现之前,仍属前瞻预期。[CP002, CP007, CP008, CP012, CP014]

核心能力与采购考量矩阵
采购考量标准Shield AIAndurilPalantirGeneral Atomics / 传统主承包商证据限定说明
战机飞行自主系统Hivemind;官方案例包括 VISTA、Avenger 及双机协同 Firejet据报道 Fury 已集成 Hivemind;原生技术栈细节未作比对未见提供飞行自主产品的证据据报道参与 CCA 竞标;暂无详尽技术栈比对所分配证据中缺乏独立的横向基准测试
自研战机平台V-BAT 及处于研发中的 X-BATAcquinox 报道的 Fury 及更广泛的全栈系统未见有战机产品的证据据报道参与飞机项目战备状态与已交付机队规模在此不具可比性
跨平台分发能力多项官方飞行范例及据报道的 Fury 集成可在 Fury 上运行 Hivemind;更广泛的第三方移植能力未知属企业级数据集成,非飞机平台移植内部及第三方自主化现状未知集成时间与可复制性未予披露
零跑道依赖定位V-BAT 与 X-BAT 均定位于 VTOL 飞行器保留对比证据中未知不适用于纯软件方案归类传统厂商中未知无通用任务测试支撑相对评分
制造规模规划 420,000 平方英尺 X-BAT 厂区;目标年产约 150 架飞机指定证据中未量化不适用于飞机制造指定证据中未量化Shield 数据属规划性质,非已验证产能
公开定价可比性未知未知未知未知保留资料中缺乏可比的公开标价依据

「未知」表示指定证据未确立该单元格内容,并不代表缺乏该能力。Shield AI 飞行范例均为公司官方口径;制造数据属报告规划,而非已实现产能。

[CP002, CP004, CP005, CP006, CP007, CP008, CP013, CP014, CP016]
FP001: 序数定位:自主可移植性与自有平台广度对比

分析师仅基于指定信源按 1(低或无证据支持)至 5(高)打分;横轴代表有证据支持的自主可移植性,纵轴代表有证据支持的自有平台或项目广度。

条目X (1–4)Y (1–5)说明
1. Shield AI 4 4 拥有多个机型试飞案例,外加 V-BAT 与 X-BAT
2. Anduril 3 5 全栈广度;Fury 可搭载 Hivemind
3. Palantir 1 1 相邻数据集成,非飞行自主
4. Skydio 2 2 专注小型无人机;更广泛的可移植性尚未确立
5. General Atomics 2 4 飞机项目广度;自主可移植性尚未确立
6. Northrop Grumman / Lockheed Martin 2 5 传统平台广度;当前可移植性未量化

分析师序数评分并非实测产品指标。评分仅汇总收录证据,在证据缺失处取保守值,绝不可解释为对性能、采购或市场份额的估计。

[CP003, CP004, CP005, CP014, CP015]

3.3 定价、GTM 与分销路径

留存公开资料未提供 Shield AI 或上述替代方案的可比公开牌价、单机软件费用、折扣标准或实际履约合同经济数据。因此合理的比对维度是产品打包与入市路径,而非凭空臆测价格。Shield AI 自身在波罗的海的一则招聘启事提到了区域战略规划、销售管线全流程管理、与防务部门及采购机关直接接洽、区域伙伴关系维护、合同谈判以及背负销售指标。整套配置对应的是直销主导、围绕项目捕获的大客户与政府销售打法,而非自助式定价。招聘数据显示有 391 个在招岗位(含 173 个工程岗位和 62 个运营岗位),外部产品管理指南也强调对从概念到部署的全程掌控,均印证了组织在技术交付与客户集成上的大力投入。然而,两类资料均无法印证其销售效率或竞标胜率。[CP009, CP010, CP011, CP013]

定价与打包方案证据
方案或替代品观察到或表明的合同模式包含功能公开价格 / 折扣证据尽调启示
Shield AI Hivemind潜在授权至合作伙伴飞机;具体结构未予披露飞机自主飞控软件未披露标价、单价、折扣或实际费率需索取单机、单项目及支持费用明细表
Shield AI V-BAT / X-BAT显示为政府竞标与飞机项目承包模式自研飞机结合自主系统与任务系统未披露可比整机价格或维持保障费用明细需拆分研发、生产、软件及保障各项经济指标
Anduril Fury / Lattice据报道采用全栈系统打包方案;具体条款未披露飞机、传感器及软件栈无可比标价或折扣证据需验证捆绑销售是否改变 Hivemind 搭载经济性
Palantir 防务软件显示为企业级软件方案;具体条款未披露战场数据集成与决策支持指定证据中无可比单价或合同金额视作邻近领域软件,非直接飞行自主可比标的
内部自研 / 沿用旧有方案买方出资内部自研或沿用既有作业流程买方主导集成与既有工作流无成本、工期或机会成本证据在宣称供应商降本优势前需建立总成本对比

本表专门记录打包特征与公开数据缺口;保留资料均未提供对等的标价、实际折扣或总成本分母。

[CP006, CP011, CP013, CP014, CP015]

3.4 护城河持久度与被替代风险

护城河的持久度取决于 Hivemind 能否成为可跨机型复用的通用自主层,抑或始终局限于少数演示与试验项目。官方飞行范例以及据报道的 Fury 集成均为软件可移植性的积极信号,但分配证据未披露实际集成周期、经常性软件收入、客户切换行为或相较于竞品技术栈的表现。老牌总包商亦可将自研自主系统同现有战机、采购人脉及维保基础设施打包捆绑。技术风险在整个赛道依然突出:ePlane AI 报道称 General Atomics 在发生坠机后暂停了飞行测试。这虽不代表 Shield AI 自身失败,但印证了自主战机研发可能因测试挫折而拖延进度。因此,真正稳固的护城河必须建立在经检验的跨平台性能、量产执行力与商业转化上,而非仅凭产品宣传口径。[CP006, CP014, CP015, CP016]

3.5 附注与论证

Chapter 04

04财务情况

4.1 收入来源与变现机制

Shield AI 通过兼具硬件系统、自研软件授权以及综合战术仿真的双轨模型实现防务自主变现。其主要交付的硬件产品是 V-BAT 垂直起降无人机系统,采购方涵盖美国海军陆战队与海岸警卫队等军种,以及日本、印度和亚美尼亚等外国军方。除整机销售外,Shield AI 还将核心自主技术栈 Hivemind 作为模块化独立软件授权推向主承包商与政府重大项目(包括 Collaborative Combat Aircraft 协同作战飞机计划)。公司在 2026 年 3 月收购老牌防务仿真供应商 Aechelon Technology,进一步强化了仿真与自主管线,后者的合成环境工具支持着 1,000 余套军方实训飞行系统。然而,实际价格机制依然不透明:除最新发布的 X-BAT 第六代自主攻击机预估单机造价约 $27M 之外,官方标准牌价普遍未公开。包括成本加成研发合同与固定总价批量生产在内的典型防务订购惯例主导着现金转化节奏,导致收入确认呈现块状不均衡。[CI001, CI002, CI003, CI004, CI005]

Shield AI 收入来源与商业结构
收入来源商业机制单价 / 合同模式当前状态收入质量尽调评估所需资料
V-BAT UAS 硬件与备件政府直接采购与对外军售(FMS)确定固定总价(FFP)平台批次与备件包处于量产阶段;已列装 USMC、美国海岸警卫队及盟国军力中等;硬件交付积压订单较高,受制于国防供应链核实物料清单(BOM)成本、制造毛利率与交付节奏
Hivemind 自主飞控软件平台授权、任务套件集成及企业级软件软件许可与开发性集成服务已部署于 V-BAT、VISTA X-62 及第三方 CCA 测试机高;具备软件级高毛利特征与深度工作流绑定索取 Hivemind 独立 ARR、单席位/单平台定价及续约率
X-BAT 自主战机重大国防采办项目与国际伙伴销售成本加成研发逐步转向固定价格量产(单机目标 $27M)处于工程原型阶段;获美国海军 $50M 合同;预计 2028 年测试投机性较高;存在多年研发风险且转化周期漫长获取里程碑 B 时间表、DoD 成本分摊比例及客户采购承诺
Aechelon 战术模拟模拟软件许可、合成环境生成与维保企业级软件许可与多年期维保合同2026-03 宣布收购;已装机 1,000+ 台飞行模拟器高;高粘性经常性收入,已嵌入五角大楼联合仿真环境确认 Aechelon 独立 EBITDA 贡献、年化收入运行率(run rate)及交易估值

收入来源覆盖截至 2026 年已披露业务线;硬件与软件之间的财务拆分仍属非公开专有数据,管理层未作分部披露。

[CI001, CI002, CI003, CI004]
Shield AI 定价、商业化与合同条款
产品 / 功能价格与合同结构标价与实际价格动态折扣、未知因素及或有事项来源与证据基准
V-BAT 平台与备件定制化任务采购合同(政府定价)实际价格通过 DoD 主承包商与对外军售谈判确定适用批量折扣;维保与有效载荷套件各异公司与防务报道(Tectonic、Wikipedia)
Hivemind 软件许可分级自主授权与工程集成服务费实际合同金额挂钩机队规模与自主等级提供捆绑集成折扣;独立定价未予披露防务报道与公司通告(Tectonic、Built In)
X-BAT 自主飞机$27 million 目标单机标价;$50M 初期海军研发量产前目标标价;远期实际成本取决于生产批次规模LRV 移动拖车及地面支援设备成本未知;含一次性工程(NRE)费用防务新闻报道与高管表态(Tectonic、GeekWire)
Aechelon 仿真工具企业级仿真许可、传感器建模工具套件及技术支持采用政府标准 GSA 采购目录及防务直接分包合同维持保障费与硬件模拟器集成成本未单独列项公司收购公告(Tectonic、Built In、Fortune)

官方防务目录标价属于非公开信息;合同金额仅反映公开披露的单机价格目标与初期研发合同授予额。

[CI001, CI002, CI003, CI004, CI005]
FI001: Shield AI 商业合同至毛利率转化流程

直观呈现从防务采购需求到软硬件收入、再到利润兑现的结构性转化链条。

节点 / 连接节点 / 起点终点说明
节点 1 防务采购与盟国需求
节点 2 合同载体:FMS、直接商业销售与研发
节点 3 硬件交付:V-BAT UAS 与备件
节点 4 自主许可:Hivemind 平台与任务套件
节点 5 战术仿真:Aechelon 虚拟环境
节点 6 综合毛利率与营运资金再投资

展示从获取合同到产品交付的业务流程;因分部利润率未予披露,省略转化百分比。

[CI001, CI002, CI003, CI004, CI015]

4.2 公开业务进展、预测准确度与估值倍数现实

公开报道展现了显著的营收扩张,同时也暴露出收入轨迹与市场定价维度的重大投资评估风险。历史营收从 2022 年大致翻倍,于 2023 年达到约 $163M,根据公司口径,截至 2025 年 3 月的财年进一步扩张至约 $300M。尽管规模持续攀升,Shield AI 在截至 2025 年 3 月的周期内较其最初设立的 $400M 目标低了 $100M,交付缺口达 25%,充分说明了防务采购周期固有的波动性与执行风险。管理层前瞻指引预计 2026 日历年收入增速将超 80%,在不含 Aechelon Technology 并购带来的外延收入贡献下争取达到至少 $540M,并设想在截至 2028 年 3 月的财年冲击 $1B 收入。独立二级市场评估强调了依赖未经核实数据平台的危险:虽然自动数据抓取工具给出了 2025 年收入 $1.6B 的线性预测模型,暗示仅 3.8x 的偏低市销率,但结合其 $12.7B 的一级市场估值基准与 $13.9B 的隐含二级市值,Shield AI 真实的静态市销率已接近 2025 财年实际营收的 46x。[CI006, CI007, CI008, CI009, CI010, CI011]

单位经济效益与财务评估指标
指标类别报告数值置信度与出处对投资评估的重要性尽调问询 / 缺口诉求
FY2023 收入$163M中(第三方尽调转引公司口径)确立自 2022 年以来的收入翻倍增长基准轨迹对账 GAAP 审计报表与内部未审计管理层汇报材料
FY2025 收入(截至 3 月)$300M高(多方信息相互印证的公司披露)构成历史估值倍数的核心分母调阅过去 12 个月按客户与合同渠道细分的收入明细
FY2025 收入偏离计划缺口-25%(低于 $400M 目标达 $100M)高(二手分析引用的公司披露)提示国防采购波动性与管理层预测风险审计收入里程碑推迟与合同延期的根本原因
CY2026 预测收入$540M+(YoY 增长 >80%)中(高管接受 Fortune 专访时的预测)评估支撑 $12.7B 估值的远期收入轨迹核查已签约在手订单、确定性任务令与未授予期权份额
毛利率 %null低(未披露的私有数据)研判软件高毛利能否对冲重资产机体制造业务要求提供按软件与硬件拆分的 GAAP 毛利率
经营性现金消耗 / EBITDAnull低(未披露的私有数据)揭示净现金消耗率与自我造血能力索取月度现金消耗表、EBITDA 调节表与资金跑道测算模型
净收入留存率(NRR)null低(未披露的私有数据)衡量各机队营运方对 Hivemind 授权的增购与持续性获取美国及国际客户的群组留存与增购指标

Null 值代表 Shield AI 未公开的专有数据;尽调诉求代表机构投资评估前所需的关键底稿文件。

[CI006, CI007, CI008, CI009, CI015]

4.3 资本结构、资产负债表依赖与投资判断

自 2015 年成立以来,Shield AI 历经 17 轮机构融资累计募集超 $3.5B 资金,并在 2026 年 3 月敲定了巨额融资方案:包括由 Advent International 和 JPMorgan Chase 领投、投后估值 $12.7B 的 $1.5B Series G 股权融资,以及由 Blackstone 提供的 $500M 非稀释性固定回报优先股。这笔注资的大部分已指定用于现金收购 Aechelon Technology,并扩建华盛顿州德梅因 420,000 平方英尺的 X-Forge 3 先进制造厂房,该设施需在 2028 年目标交付日前配置数千名航空技工并投入巨额资本支出。从投资评估角度看,Shield AI 存在严重的财务披露缺口:公司未公开披露毛利率、EBITDA、经营性现金烧钱速度、净留存率(NRR)或 Hivemind 的合同续约率。在二级市场上,投资者需求趋于谨慎,买单相较股票供给处于 0.9:1 的偏淡比率,股票交易价格在 $168.17 附近徘徊。鉴于 ~46x 的高静态估值倍数、历史预算目标的未达标以及整机量产的高额资本支出需求,投资人对未来增长与利润率扩张必须持审慎态度。[CI012, CI013, CI014, CI015, CI016]

4.4 附注与论证

Chapter 05

05产品与技术

5.1 Hivemind 自主架构与边缘任务执行

Shield AI 依靠平台通用的软件技术栈 Hivemind 提供军事任务自主能力,旨在卫星定位与无线电通信遭拒止的环境下驾驶无人机及协同作战战机。该平台直接在嵌入式机载计算硬件上运行机器学习感知、动态路径规划、战术决策及多机协同,无需依赖地面控制站。通过将飞行自主性与特定飞行器机体解耦,Shield AI 把 Hivemind 定位为通用数字飞行员,能够适配战术无人机系统、喷气战斗机及盟国防务资产等异构装备。公司在多种平台上验证了实际系统集成,包括在美空军 X-62A VISTA 战机上完成自主战术飞行测试,通过政府自主参考架构(A-GRA)接口在 General Atomics MQ-20 Avenger 平台上进行试飞,以及在 Kratos Firejet 无人靶机上执行编队协同演示。[CE001, CE002, CE004, CE011]

产品模块与防务装备资产矩阵
产品模块 / 资产目标操作方 / 客户作战成熟度核心技术壁垒关键尽调阻碍
Hivemind 自主核心系统DoD 及盟国防务军力已在 30 类载具上进入量产应用摆脱 GPS 与通信链路的跨平台边缘自主飞控强电磁对抗环境下多机蜂群可扩展性验证
V-BAT 无人机(Group 3)美国海军、海军陆战队及国际盟友全面实战部署复杂电磁环境下长航时、零跑道依赖的 VTOL硬件与软件维保毛利的完整细分拆解
X-BAT 自主战机美国空军、海军及第一梯队防务盟友处于研发 / 原型阶段具备 2,000 海里航程且摆脱跑道依赖的 VTOL 战机单机飞离成本与首飞节点的独立佐证
Aechelon 仿真平台联合作战训练司令部及无人机研发机构处于成熟商业 / 政府列装阶段用于自主系统训练的多光谱合成现实引擎收购后与 Hivemind 核心软件融合的时间表

装备现状反映截至 2026 年的公开飞行验证记录与公司披露;量产数据未包含涉密的无人机测试构型。

[CE001, CE002, CE003, CE004, CE006, CE008]

5.2 飞行器平台、X-BAT 研制与先进制造

虽然 Hivemind 属于跨平台的通用自主软件,Shield AI 仍将其自主引擎与专用实体无人机机体结合,以满足军方客户在战术侦察与打击任务中的实战需求。公司的量产机型以 V-BAT 为代表,这是一款专为电子战战场设计、已在全球部署执行情报、监视与侦察(ISR)任务的长航时 Group 3 垂直起降无人机。为应对高速空战任务,Shield AI 推出了 X-BAT,这是一款翼展 26 英尺、设计航程达 2,000 海里的无人垂直起降自主战机,摆脱了对传统硬化跑道或易受攻击的前沿空军基地的依赖。为支撑产能爬坡,公司正在华盛顿州德梅因建设 420,000 平方英尺的 X-Forge 3 先进制造厂房,并设定了未来再扩建 100 万平方英尺的长期区域目标,以在 2030 年代初支持多机型量产。[CE003, CE006, CE007, CE014]

FE001: Shield AI 全栈自主架构

六层技术架构栈,阐明从云端模型编排到物理边缘机体的演进递进。

机队指挥与任务编排
战术作战中心接口、机队 UI/CLI 与任务目标指派
模块
  • 任务指派控制台
  • 机队状态仪表板
  • 空地协同网关
云端自主 DevOps 与分发
集中式开发、模型打包及边缘安全分发基础设施
模块
  • Hivemind 企业级 SDK
  • AWS IoT Core 流水线
  • Amazon ECR 容器镜像仓库
合成现实与仿真引擎
高保真数字孪生训练与多智能体虚拟验证环境
模块
  • Aechelon 合成现实
  • A-GRA 合规试验台
  • 多智能体态势仿真器
Hivemind 边缘自主核心
执行实时感知与战术决策的嵌入式飞行智能
模块
  • 无 GPS 状态估计器
  • 动态航迹规划器
  • 近距格斗战术机动逻辑
飞行器管理与飞控系统
子系统控制回路、安全攸关看门狗与操纵舵面接口
模块
  • 飞行控制计算机
  • 航电总线接口
  • 故障安全返航协议
机体与物理边缘硬件
部署于实战战区的实体无人机体与传感器载荷
模块
  • V-BAT 3 类垂直起降无人机(Group 3 VTOL UAS)
  • X-BAT 自主战斗机
  • 第三方机型(VISTA/Avenger)

架构图综合了公开的技术披露、职位说明书与演练报告;专有子系统互联关系经概括处理。

[CE001, CE002, CE003, CE005, CE008, CE010]

5.3 云端自主工具链、机队部署与合成仿真

面向不断增加的机型部署任务自主软件,需要持续集成、敏捷测试与安全的边缘分发管道。针对复杂的企业级部署需求,Shield AI 与 Amazon Web Services 建立战略合作,展示了打通 Hivemind Enterprise SDK、AWS IoT Core 与 Amazon Elastic Container Registry 的自动化工作流。这一从云到端的纽带允许自主软件开发者在云端训练并仿真自主模型,随后通过命令行或图形界面将就绪的容器化飞行员程序直接推送到边缘平台硬件。自主能力的验证高度依赖数字孪生仿真,Shield AI 对 Aechelon Technology 的收购增强了这一环节,后者提供高保真合成现实引擎,可在获取实飞许可前对复杂模拟战场环境下的多机协同战术进行压力测试。[CE005, CE008]

作战工作流与自主系统部署矩阵
作战任务职能传统作业流程Shield AI 解决方案展现的实战效益技术与部署局限
强对抗环境战术侦察依赖活跃卫星数据链路的人工远程遥控配备 Hivemind 无 GPS 自主导航的 V-BAT面对敌方强电磁与 GPS 干扰仍能确保持续 ISR 搜集载荷更换受限以及恶劣天气下的传感器探测距离衰减
协同蜂群编队预设航路点规划或单机单人操作基于 Firejet/V-BAT 的多机 Hivemind 自主蜂群摆脱机间射频依赖的战术协同机动复杂非常规环境构型下的算法协同边界
战术空战(近距格斗)战机飞行员在座舱生理极限内的手动操纵在 X-62A VISTA 上得到验证的 Hivemind 自主驾驶亚秒级战术决策执行,机动表现超越基准人类飞行员实弹交战场景受制于适航认证与监管限制
全机队自主算法更新在中央维修基地进行的人工本地航电固件刷新集成 AWS IoT Core 与 ECR 的 Hivemind 企业级 SDK容器化虚拟飞行员算法直接敏捷分发至分布式飞机远程下发数 GB 级大模型时受制于边缘节点带宽瓶颈

工作流效益源自公开试飞演示;边缘部署工作流假设初期软件推送具备安全的国防网络连接。

[CE001, CE002, CE005, CE008]

5.4 高安全性系统架构、接口治理与技术风险

在杀伤性防务装备上运行自主飞行算法带来了严峻的技术与实战故障风险,算法偏差或边缘感知失效可能直接损毁价值数百万美元的战机或导致战术任务失败。Shield AI 建立高安全性系统架构,在任务管理、飞行控制与自主软件边界设立专门的工程团队,以确保形式化可追溯性、网络韧性及故障隔离标准。早期实战部署暴露了实体集成层面的挑战,其中一次人工辅助 V-BAT 着陆事故造成一名美海军士兵受伤,并引发作战禁飞限制。为减少在盟国防务项目中的集成阻力,Shield AI 使 Hivemind 契合政府参考架构(如 A-GRA),并积极与新加坡 ST Engineering 等军工主承包商探索集成合作,同时角逐美空军 Collaborative Combat Aircraft 等长期重大项目。[CE009, CE010, CE012, CE013]

技术架构与运营风险分析
组件 / 子系统主要架构职责关键技术依赖核心工程 / 运营风险
边缘自主推理核心机载感知、状态估计与航迹规划执行高性能边缘计算模块(如 NVIDIA / FPGA)大过载机动下的热衰减与耐冲击抗震性能
Hivemind 企业云端 SDK模型编排、容器化及 AWS 边缘端推送AWS IoT Core、ECR 镜像仓库及安全隔离飞地网络部署延迟及物理隔离(air-gapped)前线环境的分发瓶颈
Aechelon 合成仿真系统数字孪生产生与多智能体训练环境精准物理动力学模型与高保真传感器仿真仿真到现实(sim-to-real)迁移落差导致真实边缘行为产生偏差
X-Forge 3 先进制造X-BAT 战斗机机体制造与总装特种复合材料与航空发动机供应链厂房建设延期及重资产工装夹具费用超支

技术依赖与风险综合自工程架构公开信息、招聘岗位说明及国防航空基准标准。

[CE005, CE007, CE008, CE009, CE010]

5.5 附注与论证

Chapter 06

06客户情况

6.1 客户分层与防务采购机制

Shield AI 将客户群划分为三大机构层级:美国本土一线军种、国际盟国防务部门以及商业科技生态伙伴。本土防务客户是公司收入基石与实战验证的历史依托,涵盖美空军、美海军、美海军陆战队以及美特种作战司令部。该层级的采购通过多年期国防拨款、小企业创新研究(SBIR)成果转化以及前沿项目招标展开,包括五角大楼的 Collaborative Combat Aircraft 项目与美海军 $50M 的首期 X-BAT 研制合同。盟国防务力量是增速最快的客户群体,主要体现为政府对外军售(FMS)以及与新加坡、印度、波兰和日本等国签署的双边联合开发协议。商业技术合作(如与 Amazon Web Services 联合搭建的云边部署框架)则充当技术分发渠道以简化软件容器交付,并非终端采购渠道。[CU001, CU005, CU006, CU008, CU011, CU013]

客户细分表
客户群体买方 / 使用方 / 付款方主要应用场景观察到的规模与采纳度收入与战略价值尽调与数据缺口
美国国防部美国海军、空军、海军陆战队、特种作战司令部(SOCOM)战术 ISR、空战自主飞控、协同作战飞机(CCA)实战战区部署;获美国海军 $50M X-BAT 研发合同核心营收支柱;多年期国防预算拨款的关键锚点涉密采购资金总额及具体批次交付量未予披露
盟国军方新加坡(RSAF/DSTA)、印度、波兰、日本自主任务主权、边境监控、海上安保获多个国家采购选型:印度陆军、波兰海军、日本海上自卫队(JMSDF)地域多元化与软件开发套件商业化交付节点与经常性软件续约条款尚未证实
国土安全与海事机构美国海岸警卫队、意大利海岸警卫队、欧洲边境管理局(Frontex)海上拦截、海岸巡逻、边境监控V-BAT 海试,含 RIMPAC 2026 及意大利海上实战部署承包商自建自营(COCO)服务收入来源机构试用向多年期服务合同转化的转化率未知
云与技术合作伙伴Amazon Web Services、国防主系统集成商云边协同容器部署、机队自主协同编排Hivemind 与 AWS IoT Core 联合架构演示生态系统分发与企业级集成标准制定除联合营销外的直接商业营收贡献尚未明确

客户细分基于已核实的公开防务订单、盟军官方发布及官方企业合作伙伴公告。

[CU001, CU005, CU006, CU007, CU008, CU011, CU013]

6.2 采纳轨迹、商业规模与行业巨头对比

Shield AI 的商业采纳轨迹展现出强劲的营收势头,但与现代防务科技头部同行对比时暴露出悬殊的体量差距。高管披露 2025 年营收规模约 $300M,同时预计 2026 年增速超 80%,年化营收将突破 $540M。这一扩张受到 V-BAT 飞行器批量交付、X-BAT 首批工程里程碑以及 Aechelon Technology 仿真平台纳入五角大楼联合仿真环境的驱动。然而,与 Anduril Industries 的竞品比对揭示了显著的竞争劣势。Anduril 在 2025 年录得 $2.2B 营收,对 Shield AI 形成了 7.3 倍的体量优势,并在防空、水下系统及指挥软件等更多军工项目中斩获批量订单。Anduril 已将多款自主平台转化为大批量、经常性列装项目,而 Shield AI 的营收结构仍高度依附于战术无人机采购和预期中的软件授权,其商业毛利率亦尚未经过审计检验。[CU003, CU004, CU010, CU012]

客户增长 / 采纳轨迹表
采纳指标报告数值日期出处置信度战略启示缺失分母 / 披露缺口
FY2025 收入基数约 $300M2025-12-31Signals Inbox / Fortune 高管披露中印证自早期风投原型阶段以来的规模化扩张硬件采购与软件许可间的分拆数据未予披露
FY2026 预测收入目标>$540M(YoY 增长 >80%)2026-07-21Signals Inbox 报道的公司指引中反映预期的产能爬坡与并表 Aechelon 收购未拆分已签约在手订单与投机性竞标项目储备
Hivemind 操纵的载具类别26 类载具2026-03-26Aechelon 收购公告高展现跨空、陆、海多平台的自主系统移植能力缺乏各平台类别的活跃经常性软件授权收入数据
相对于 Anduril 的收入规模比例0.137x(约 7.3 倍规模差距)2025-12-31Signals Inbox 对比财务分析中凸显与顶级国防自主系统主承包商之间的巨大商业规模差距未记录经审计的 EBITDA、毛利率及现金消耗对比数据

财务与运营采纳指标综合自公司公开目标、媒体报道及技术发布资料。

[CU003, CU004, CU010, CU012]
FU001: 客户旅程图

国防机构客户从实验室仿真和飞行试验演进至主权自主机队列装的六阶段进程。

直观示意机构客户经历标准防务采办与自主软件部署各阶段的推进历程。

[CU001, CU002, CU007, CU008, CU010]

6.3 知名客户实证、留存情况与客户集中度风险

机构证据证实,Shield AI 的硬件与自主平台在关键主权防务部门的国际采购上持续扩大。2026 年 2 月,新加坡空军与国防科技局(DSTA)深化战略合作,将 Hivemind 软件开发工具包部署于多种无人机型,使新加坡防务工程师能够自主编写并部署符合主权需求的 AI 飞行逻辑。在此之前,印度陆军于 2026 年 1 月选定 V-BAT 与 Hivemind 并配套本土工业制造路径;波兰军备局亦在 2025 年日本海上自卫队选型的基础上,于 2026 年 6 月签约采购 V-BAT 系统以增强波兰海军海上安全能力。尽管斩获众多知名订单,客户集中度依然是核心薄弱环节。Shield AI 的营收极度依赖主权国防预算与单一项目拨款节奏。至关重要的一点是,公开文件未披露基于队列的净留存率、软件流失率数据及单机毛利率,这使得外界难以判断国际订单究竟意味着持久的经常性软件现金流,抑或仅是一次性硬件采购合同。[CU001, CU002, CU005, CU006, CU007, CU009, CU014]

知名客户实证表
客户 / 军兵种客户细分部署与应用场景量产对比试点状态作战实效局限与尽调缺口
新加坡空军与国防科技局(DSTA)盟国主权军方使用 Hivemind 软件开发套件开展无人机自主能力研发活跃联合研发与部署工程师独立编写、测试并部署 AI 飞行员任务行为软件许可费总额与主权平台部署规模未公开
印度陆军盟国主权军方战术侦察与本土自主工程采购入选选中 V-BAT 飞行器及 Hivemind,并配套本土工业工装交付时间表、批次数量及本土生产里程碑未披露
波兰军备局 / 波兰海军盟国主权军方波罗的海海上监视与海岸侦察采购合同就舰载 V-BAT 海上无人作业签署正式合同综合保障包细节与合同财务金额未列出
日本海上自卫队盟国主权军方舰载与海上监视作业采购入选选中多套 V-BAT 系统用于首批海军部署后续采购订单时间表与维保合同条款未获确认

亚太与欧洲战区防务机构主权客户入选情况的一手信源核实。

[CU001, CU002, CU005, CU006, CU011, CU013]

6.4 附注与论证

Chapter 07

07风险

7.1 安全、质量与工业执行风险

最明确的已显现风险在于 V-BAT 的安全性与产品成熟度。2024 年一次演示致伤一名美军官兵,军方客户为此下达了数月禁飞令;Forbes 亦披露过更早的螺旋桨警告、机身开裂、燃油系统故障以及演示期间坠机等事故。Shield AI 声称改用起落架后无需人工协助并已通过五角大楼安全审计,但相关报道未能独立证实该审计结果。X-BAT 则构成了第二重执行风险:管理层目标是在 2028 年自新建的 420,000 平方英尺厂房交付首批整机,并在 2030 年代初达到年产约 150 架的水平。因此投资逻辑应将经审计的安全记录、试飞验收及工厂里程碑视为准入门槛,而不可径直假定改版设计已彻底消除故障隐患。[CR001, CR002, CR003, CR004, CR008]

运营 / 质量 / 安全风险登记表
失效模式可能性严重程度缓解成熟度残留风险敞口未解决缺口 / 监控项
人工辅助降落期间人员触碰 V-BAT 螺旋桨已于 2024 年发生;再次发生率无数据极严重部分具备——据报道已引入自主起落架、警示标识及操作规程若控制措施失效将导致重伤并招致客户限制机队飞行小时、事故率、审计结论与复飞/中止降落数据
V-BAT 可靠性缺陷与演练损失历史上多起指控;当前机队比率无数据高未经验证——公司称已对飞行器完成大修与重新设计停飞、飞机坠毁、决策推迟与保障负担失效模式历史、MTBF、任务完好率与根本原因闭环
X-BAT 厂房建设与产能爬坡远期规划;进度概率无数据高早期——设施在建且人员招聘同步展开成本超支、鉴定延误及错过 2028/2030 年代初期里程碑施工、工装、供应商、测试与首件里程碑
Hivemind 跨异构飞行器集成已有多项演示;量产可靠性未披露针对项目采纳为高已完成演示,但在所列全部平台上均未获量产验证适航认证、延迟、接口及操作员信心缺失导致的失效特定平台测试报告、软件适航保障与验收标准

评级严格区分已发生事件与未来执行风险。指定证据中尚无公开的事故率、机队飞行小时、MTBF、网络安全测试或建厂成本分母数据。

[CR001, CR002, CR003, CR004, CR008, CR014]
FR001: 风险热力图:基于证据的残留风险敞口

六项风险的定性排序;标签如实保留发生可能性是已发生、待决、可能抑或公开数据无法量化。

此处未推算具体数字概率。影响程度、缓解措施成熟度与残余严重度均为分析师定性判断,基于所引述事件、交易或有条件、扩张规划、政策分析及采购结构得出。

[CR001, CR002, CR003, CR004, CR005, CR008, CR009, CR010]

7.2 监管、法律与地缘政治风险敞口

监管风险既有针对特定交易的,也有结构性的。Fortune 报道 2026 年融资的最终交割以 Aechelon 收购案获批为先决条件,若审批受阻 Shield AI 将重新评估融资。另外,分析师指出潜在的“人在回路”规则、针对致命自主武器的任务限制以及各国互异的标准均构成了实战部署与海外销售的掣肘。中国于 2025 年将 Shield AI 列入不可靠实体清单,禁止其从中国采购、在华运营、投资或入境;管理层声称公司刻意规避了对中国的依赖,但分配资料中没有公开的物料清单证据可证实该表述。一票否决指标包括:Aechelon 审批遭拒或严重延误、无法取得必需的出口或作战许可,或发现受禁的底层供应链依赖。[CR005, CR009, CR012, CR013]

监管与法律风险登记表
风险 / 司法辖区证据与现状可能性严重程度缓解成熟度残留风险敞口尽调路径 / 否决条件
Aechelon 收购审批 / 美国据报道融资最终交割以收购获批为前提;在已核阅证据中仍处于待决状态。待决;概率未公开高——若未获通过将需重新评估融资有限——管理层仅表示将重新评估,未给出确定的备用方案交易、融资与产品路线图不确定性获取审批文件及已签署的融资先决条件;若审批被拒或延期导致运营计划所需承诺资本落空,则触发否决。
自主武器与出口管制规则 / 美国及盟国市场独立分析指出潜在的人在回路限制、各国标准差异以及出口审批构成约束。可能;规则演变路径与市场范围未量化对受限任务剖面或延期的国际项目影响严重有限——公开证据未按产品与司法辖区逐项列出获批情况特定区域的产品需求、销售延期或禁止使用建立按司法辖区与产品划分的审批矩阵;若核心 Hivemind 任务无法在目标项目中获得许可,则触发否决。
V-BAT 伤人事件、安全监管与潜在法律责任 / 美国2024 年伤害事件及客户施加的飞行限制由 Forbes 报道并由 Wikipedia 汇总。已发生事件;再次发生概率未公开对人员安全属于极严重;若再次发生将造成重大商业冲击部分具备——据报道已更改设计与规程,但审计证据未经独立确认再次停飞、索赔主张、错失采购机会以及声誉受损核查调查报告、索赔主张历史、保险、审计结论与机队安全数据;若再次发生严重伤人或系统性隐患未消除,则触发否决。

严重程度系分析师依据各行所列记录后果作出的评估;指定信源未提供分母时,发生概率不做量化。样本仅限于截至 2026-10-01 指定信源库中有证据支持的重大风险。

[CR001, CR002, CR004, CR005, CR009]

7.3 客户、伙伴与竞争依赖

Shield AI 面向政府采购体系销售,后者的多年期预算、政治意图更迭、持续决议案与冗长审批流程,并不一定契合风险投资的成长节奏。现有证据表明其拥有广泛的美军及盟国客户群,但未量化客户集中度或在手订单转化率。与此同时,老牌主机厂可将自主系统同飞行器、弹药及维保捆绑销售,资金充裕的防务科技新贵亦在争夺订单与紧缺的 AI 人才。Shield AI 在 VISTA、Avenger 与 Firejet 上的跨平台官方演示是集成广度的证据,但并非经常性量产收入或全机型高可靠性能的明证。后续需严密监测期权行权率、已获拨款订单、续约与量产合同落地、国际审批进展,以及第一大客户和单一项目占营收的比重。[CR010, CR011, CR014]

合作伙伴 / 依赖风险登记表
依赖项对手方角色集中度失效情景严重程度缓解措施残留风险敞口
政府采购与预算美国国防部(DoD)及盟国防务买家客户、审批方与项目出资方未知;指定信源库中未公开客户与项目份额预算转移、持续决议案(CR)、定标延误或未行使期权高服务于多个美国军种及国际买家所有买家均与政府挂钩且受审批驱动
Aechelon 交易与仿真集成Aechelon 及融资投资方收购标的、仿真能力及交易资本重大但未量化审批受阻将改变融资结构并推迟既定集成高管理层称届时将重新评估融资无公开承诺的备选方案或集成里程碑
X-BAT 工业化爬坡劳动力、供应商与测试基础设施厂房建设、生产制造、资格鉴定及后期维保供应商集中度未披露人才、零部件、工装或测试瓶颈导致无法达成生产目标高西雅图地区航空航天人才以及工程/生产同址布局无公开供应商图谱、固定总价风险敞口或爬坡经济模型
自主平台竞争防务主承包商及获资金支持的防务科技同行竞争性自主技术、集成平台、人才及采购渠道项目层面的竞争份额未披露买家选择传统捆绑系统或竞争对手的自主层高跨平台演示及差异化拒止环境定位技术演示无法证明可持续的量产合同或利润率

集中度字段保持未知状态,避免以客户名单、管线声称或融资额替代经审计的收入、在手订单或可用现金。

[CR005, CR006, CR008, CR010, CR011, CR014, CR015]

7.4 财务模型敞口与逻辑被证伪触发点

据报道,2026 年 3 月的融资方案包含 $1.5B 的 Series G 融资及 $500M 的固定回报优先股,同时管理层预计 2026 年收入将超 $540M、增速超 80%。募集到的资金并不等同于无限制的可用现金:该方案旨在支持并购与产品扩产,内含优先受偿工具,并设有交割前提条件。EquityZen 亦将该业务定性为具备长研发周期的资本密集型资产,并提示持续的盈利风险。分配资料中的公开来源并未披露烧钱速度、毛利率、EBITDA、营运资金需求、优先股具体条款、厂房资本承诺或单项目经济模型。逻辑被证伪的触发因素包括:再度遭遇安全停飞禁令、未如期完成 X-BAT 试飞或厂房建设里程碑、并购审批失败导致融资生变,或有证据表明在手订单与利润率无法在不依赖反复稀释性股权或优先股注资的前提下支撑硬件扩产。[CR005, CR006, CR007, CR015, CR016]

7.5 附注与论证

Chapter 08

08估值

8.1 投资逻辑与反向逻辑

Shield AI 的核心投资逻辑立足于其在现代军事冲突中的战略站位——在电子战和 GPS 遭拒止的严苛战场中担当一流的自主系统供应商。公司自研的自主软件平台 Hivemind 扮演 AI 飞行员的角色,支持无人机和有人机在无卫星导航信号或易受干扰的射频链路下机动、导航并执行战术动作。美空军部在 Collaborative Combat Aircraft(CCA)项目中对 Hivemind 的入选,证实了其在防务项目中的架构领先地位。对 Aechelon Technology 的收购提供了合成仿真环境与传感器建模,可为持续的飞行算法训练提供数据喂养,构筑抵御传统航空航天防务巨头的护城河。此外,向盟国防务采购的渗透(如波兰海军选定其海上无人机,以及北约伙伴国的采购意向)拓宽了长期的合同储备。 反向逻辑则揭示了值得警惕的重大结构性、财务性与运营性风险。首先,Shield AI 身处重资产、高资本密集的军工领域,需要长期的制造扩张投入,在华盛顿州投建 420,000 平方英尺的 X-Forge 3 厂房以量产 X-BAT 自主战机即是明证。其次,公司仍处非上市且未盈利状态,毛利率数据尚未明晰,并在发生安全事故后推迟了盈利目标的兑现节点。第三,联邦防务采购集中度极高;与美国国防部及空军订立的合同工具造成收入确认的不均衡,且全额软件授权款项与持续实战表现考核挂钩,带来奖励费扣减风险。第四,虽然私募二级市场估价与媒体报道流传着冲击 $20B 估值融资的愿景讨论,但在当前 $12.7B 估值下入局,一旦遭遇政府防务拨款调整或战机研制进度滞后,将面临巨大的下行风险。 [CV001, CV002, CV003, CV004, CV005, CV006, CV007, CV008, CV014, CV015]

投资论点与反论点矩阵
核心支柱投资论点反论点风险验证触发事件
自主技术护城河Hivemind 平台在无 GPS 或通信受限条件下跨多种飞行器提供任务自主能力。自主算法面临持续的电子战适应与验证挑战。在争夺战区部署中验证实战生存能力。
核心项目支柱入选美国空军(USAF)协同作战飞机(CCA)及低成本 LUCAS 攻击无人机核心项目。奖励费用风险结构将合同收入与用户实战评级及主观里程碑直接挂钩。DoD 合同变更备案与正式行使期权通知。
战略并购收购 Aechelon Technology,将高保真合成现实仿真直接融入 AI 飞行员训练。整合阻力以及吸收传统仿真业务部门带来的运营开销。整合后的软件功能发布及联合仿真基准测试结果。
工业布局在西雅图地区建立 420,000 sq ft 的 X-Forge 3 先进制造中心,用于生产 X-BAT 战斗机。沉重的厂房资本开支与漫长的培育周期(目标到 2030 年代初年产 150 架战机)加剧资金消耗。Des Moines 厂房入驻时间表与 2028 年首架原型机下线。
盟国市场拓展凭借波兰海军波罗的海海上态势感知合同维持国际销售势头。严格的出口管制与 NATO 技术共享限制可能制约国际市场渗透。对外军售(FMS)与直接商业销售的监管审批。

对照一手及二手信源核实的操作与市场风险,评估关键投资驱动因素。

[CV002, CV003, CV004, CV006, CV007, CV008]
FV001: 建议逻辑推演序列

本分析链条将核心自主能力验证、资本结构现实与制造风险逐步推演,得出“观察”的投资建议。

节点 / 连接节点 / 起点终点说明
节点 1 1. 自主能力与项目验证 [autonomy-validation]Hivemind 自主系统入选美空军 CCA 项目,并获波兰海军部署 V-BAT。
节点 2 2. Series G 资本结构现实 [capital-structure]经确认估值达 $12.7B,且附带 $500M 固定回报优先股的高位受偿压制。
节点 3 3. 重工业资本支出与现金消耗 [manufacturing-burn]为 X-BAT 建造 420,000 平方英尺的 X-Forge 3 厂房,需要投入巨额资本。
节点 4 4. 未经审计的毛利率与奖励费风险 [margin-opacity]私有财务不透明、软硬件毛利率拆分未经验证,且面临合同奖励费考核风险。
节点 5 5. 情景差异评估 [scenarios-evaluation]$24B 乐观情景与 $6B 悲观情景差异巨大,要求采取严格的回报门槛。
节点 6 6. 投资建议:观察 / 估值偏高 [track-decision]在经审计财务披露与 X-BAT 首批生产里程碑达成之前,维持观察立场。
连接 1 1. 自主能力与项目验证 [autonomy-validation]2. Series G 资本结构现实 [capital-structure]导向
连接 2 2. Series G 资本结构现实 [capital-structure]3. 重工业资本支出与现金消耗 [manufacturing-burn]叠加
连接 3 3. 重工业资本支出与现金消耗 [manufacturing-burn]4. 未经审计的毛利率与奖励费风险 [margin-opacity]推高
连接 4 4. 未经审计的毛利率与奖励费风险 [margin-opacity]5. 情景差异评估 [scenarios-evaluation]支撑
连接 5 5. 情景差异评估 [scenarios-evaluation]6. 投资建议:观察 / 估值偏高 [track-decision]得出
[CV001, CV004, CV007, CV008, CV011, CV014]

8.2 投资建议与融资背景

我们的正式尽调建议为“观察”(TRACK),估值立场为“偏高”(STRETCHED)。公司最新经确认的融资轮次是 2026 年 3 月公布的总规模 $2.0B 的组合方案,包含按 $12.7B 投后估值完成的 $1.5B Series G 定价股权融资,以及由 Blackstone 旗下基金提供的 $500M 固定回报优先股。该估值较 2025 年 3 月录得的 $5.3B Series F-1 估值增长了 140%。自 2015 年成立以来,历经 17 笔融资事件,累计实缴股本与债务融资总额已达约 $2.9B 至 $3.16B。 严苛的建仓纪律要求对二级市场撮合交易与传闻保持警惕。虽然 Forge Data 追踪的二级交易迹象显示出较高的活跃度,每股成交价达到 $172.04,且媒体推测性报道提及潜在的 $20B 目标估值,但未经证实的价格标记无法构成具有实操意义的一级市场入场定价。 此外,Blackstone 持有的 $500M 固定回报优先股在资本结构中享有优先受偿主张,对普通股及初级股权持有方构成了高位受偿压制。叠加未经验证的软件毛利率以及缺少独立审计的现金流数据,以超过 2026 年预期收入($540M)23x 的倍数入场,需要计入深度的风险折价。 [CV001, CV009, CV010, CV011, CV012, CV013, CV014]

投资建议与治理摘要
维度评估结论决策含义
投资建议跟踪保持观察式尽调;在经审计的毛利率与现金消耗率披露前,暂缓直接股权出资。
置信水平中获证实的 Series G 估值与入选第一梯队 CCA 提供了坚实支撑,但作为非上市公司的财务不透明限制了分析确定性。
风险评级高软硬件双轨运营带来高额资金消耗,叠加华盛顿州飞机制造爬坡风险。
估值倾向偏高以投后估值 $12.7B(约 23.5 倍 FY2026E 预测收入)计,此定价对防务项目延期或预算削减几乎未留安全边际。
融资结构优先股高位受偿压力Series G 包含来自 Blackstone 的 $500M 固定回报优先股,确立了优于普通股的清算优先权。

投资建议与风险评级综合反映了 Series G 资本结构、Forge 二级市场数据及联邦采购敞口。

[CV001, CV009, CV011, CV014]

8.3 情景分析、可比标的与尽调问询

未来 24 至 36 个月的估值情景因不同的运营落地路径呈现出巨大差异。在乐观情景下,Shield AI 达成其预计的 2026 年 $540M 营收目标,顺利完成 X-BAT 机体试飞,全面交付 CCA 自主软件增量,并以 $20B 至 $24B 估值(对应预期收入的 37x 至 44x)完成 IPO 或高估值倍数的战略退出。在基准情景下,营收达到 $450M 至 $540M,国际部署平稳推进,但硬件资本支出与仿真集成成本压制了营业利润率,使得合理的私募估值区间维持在 $12B 至 $15B(对应收入的 22x 至 28x)。在悲观情景下,德梅因制造厂的瓶颈拖累了 X-BAT 交付进度,防务采购预算缩减,技术评审引发 CCA 奖励费扣减,导致估值倍数回落并以 $6B 至 $8B(对应收入的 11x 至 15x)的折价轮重新定价。 同行可比对标突显了这种分化。Lockheed Martin、RTX 及 L3Harris Technologies 等上市军工主承包商市销率在 1.0x 至 2.5x 之间,反映出成熟的资产负债表与较低的增速。纯防务软件与数据平台 Palantir Technologies 则凭借经检验的 GAAP 盈利能力与高软件毛利率,享受 35x 至 50x 收入的高溢价倍数。以硬件为主的无人机制造商 Kratos Defense 市销率则在 3x 至 4x 之间。Shield AI 当前 $12.7B 估值对应预测的 $540M 收入隐含约 23.5x 的远期倍数,已计入了软件平台转型的显著溢价。在出资前最终的尽调诉求包括:获取经 GAAP 审计的财务报表、核实 Blackstone 优先股的清算受偿顺位与股息条款、审查 CCA 项目的奖励费考核标准,以及核查国际出口管制的合规状况。 [CV005, CV007, CV008, CV010, CV012, CV015, CV016]

乐观、基准与悲观估值情景
情景FY2026E 收入隐含估值估值倍数概率关键运营驱动因素与下行情景触发事件
乐观情景$540M+$20.0B - $24.0B37.0x - 44.4x20%DoD 预算无约束扩张、X-BAT 试飞毫无瑕疵,且国际软件许可采纳带动高溢价 IPO。
基准情景$450M - $540M$12.0B - $15.0B22.2x - 27.8x55%按期稳步交付 USAF CCA 基础期权,V-BAT 逐步向盟国出货,制造端资本开支持续压缩净利润率。
悲观情景<$400M$6.0B - $8.0B11.1x - 14.8x25%CCA 奖励费用被削减、试飞安全问题引发延期、工厂投产成本超支,以及防务科技估值倍数严重受压收缩。

远期估值情景基于 2026 年预测收入里程碑、执行纪律以及估值倍数扩张/收缩,评估 24 至 36 个月周期的表现。

[CV005, CV007, CV010, CV012, CV014, CV016]

8.4 附注与论证

免责声明

本报告为基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术及合同事实仍未公开,在做出任何投资决策前,应直接向管理层及原始文件核实。

证据索引

结论
编号陈述可信度来源
CO001 Shield AI is a private defense technology company founded in 2015 and headquartered in San Diego, and the company says it operates across the U.S., Europe, the Middle East, and Asia-Pacific. 中 SO001, SO003, SO007
CO002 Brandon Tseng, Ryan Tseng, and Andrew Reiter founded Shield AI. 中 SO003, SO007
CO003 Shield AI's model combines Hivemind autonomy software with V-BAT and X-BAT aircraft and includes licensing the autonomy software for integration into third-party platforms. 高 SO001, SO002, SO003, SO006
CO004 Shield AI received its first major Department of Defense contract through the Defense Innovation Unit autonomy program in 2016. 中 SO003, SO007
CO005 Shield AI acquired Heron Systems and Martin UAV in 2021. 中 SO003, SO007
CO006 An independent compilation reports that Shield AI raised $165 million at a $2.3 billion valuation in June 2022. 低 SO007
CO007 An independent compilation reports that Shield AI raised $200 million at a $2.7 billion valuation in October 2023. 低 SO007
CO008 A V-BAT landing incident in April 2024 partially amputated three fingers of a U.S. Navy service member and was followed by temporary flight restrictions. 中 SO007
CO009 Gary Steele became Shield AI's chief executive in March 2025, and co-founder Ryan Tseng subsequently moved to a strategy role. 中 SO007
CO010 On March 26, 2026, Shield AI announced that it was raising a $1.5 billion Series G at a $12.7 billion post-money valuation. 高 SO001, SO002, SO003, SO005, SO007
CO011 The March 2026 announcement described $500 million of fixed-return preferred equity from Blackstone-managed funds and a separate $250 million delayed-draw facility commitment. 高 SO001, SO002, SO003
CO012 Advent led the announced Series G with David Mussafer expected to join the board, while JPMorganChase co-led it with Todd Combs expected to become a board observer. 高 SO001, SO003
CO013 Shield AI said part of the announced financing would support its planned acquisition of Aechelon Technology, subject to customary regulatory approvals. 高 SO001, SO002, SO003, SO007
CO014 Third-party sources project more than $540 million of Shield AI revenue for fiscal 2026 and approximately 80% year-over-year growth. 中 SO002, SO003, SO005
CO015 Wikipedia reports 1,500 Shield AI employees as of March 2026, citing Revelio Labs. 低 SO007
CO016 The shieldai.ai website belongs to a separate data-protection and compliance startup founded in 2024, not the San Diego defense technology company at shield.ai. 中 SO008
CM001 Shield AI's relevant market is military mission-autonomy software plus autonomous aircraft and unmanned-aircraft products and services, not broad defense or generic AI spending. 中 SM002, SM003, SM006
CM002 The central demand case is onboard autonomy that can continue operating when GPS, communications or manual control are degraded. 中 SM002, SM003
CM003 Reported customers include U.S. military services and international defense forces. 中 SM002, SM006
CM004 Shield AI reports Hivemind demonstrations on X-62 VISTA, MQ-20 Avenger and dual-ship Firejet aircraft. 中 SM003
CM005 The shieldai.ai website describes an unrelated generative-AI data-protection and compliance platform and is excluded from the target market. 中 SM007
CM006 TSG Invest reports a global military-drone market of approximately $14.4 billion in 2023 and forecasts $47.16 billion by 2032. 低 SM001
CM007 Angel Investors Network reports that defense-tech VC deals reached $49.1 billion in 2025, up from $27.2 billion in 2024. 中 SM004
CM008 Angel Investors Network reports that autonomous systems captured approximately $12 billion of 2025 defense-tech VC and roughly 57% of defense-tech VC over 2022–2025. 低 SM004
CM009 Angel Investors Network reports a $961.6 billion FY2026 Pentagon base-budget request plus $150 billion in supplemental funding. 中 SM004
CM010 The retained drone-market, venture-funding, defense-budget and company-revenue figures use different denominators and cannot support a single defensible TAM, SAM or SOM. 中 SM001, SM004, SM009
CM011 Fortune reports that Shield AI projected at least $540 million of 2026 revenue, representing more than 80% growth based on 2025 revenue. 中 SM009
CM012 Angel Investors Network reports that more than half of Shield AI's projected 2026 revenue comes from international customers. 中 SM004
CM013 Secondary sources report a $198 million U.S. Coast Guard IDIQ for V-BAT maritime ISR services. 中 SM001, SM002
CM014 A February 2026 CCA flight test switched between Hivemind and Anduril's Lattice, indicating an Air Force requirement for interoperability rather than guaranteed single-vendor lock-in. 中 SM004
CM015 GeekWire reports Shield AI expects the first Des Moines-built X-BAT in 2028 and production of about 150 aircraft annually by the early 2030s. 中 SM008
CM016 Fortune reports that final financing close is contingent on approval of the Aechelon acquisition and that Shield AI would re-evaluate financing if the deal failed to clear regulatory hurdles. 中 SM009
CP001 Shield AI presents a portfolio spanning Hivemind autonomy, V-BAT and X-BAT aircraft, and Aechelon simulation technology. 中 SP001, SP010
CP002 Shield AI says Hivemind has flown the X-62 VISTA, MQ-20 Avenger, and two MQM-178 Firejet drones. 中 SP001
CP003 The allocated reporting identifies Anduril and Palantir as competitors and Lockheed Martin and Northrop Grumman as legacy-prime alternatives. 中 SP002, SP004
CP004 Acquinox characterizes Shield AI as software-first and platform-agnostic, Anduril as full-stack, Palantir as battlefield-data integration, and Skydio as commercial and tactical small drones. 中 SP003
CP005 ePlane AI identifies General Atomics, Anduril, and Northrop Grumman as competitors for leadership in the U.S. Air Force collaborative combat aircraft program. 中 SP004
CP006 Hivemind's reported selection for Anduril's Fury makes Anduril both a competitive peer and a third-party-airframe distribution path. 中 SP003
CP007 GeekWire reported that Shield AI planned a 420,000-square-foot X-BAT manufacturing facility and was considering another one million square feet. 中 SP007
CP008 Shield AI told GeekWire it expected X-BAT production to reach about 150 aircraft per year by the early 2030s. 中 SP007
CP009 Recruiting from Scratch reported 391 open Shield AI roles, including 173 in engineering, 62 in operations, and 131 roles added during the prior 29 days. 中 SP005
CP010 Dataford describes Shield AI product management as owning mission-autonomy products from concept through deployment while coordinating engineering and government stakeholders. 中 SP009
CP011 Shield AI's Baltics job posting describes a direct defense-market motion covering pipeline development, partner development, contract negotiation, and senior government engagement. 中 SP010
CP012 CompositesWorld reports that composite design, analysis, and manufacturing techniques are important inputs for modern VTOL aircraft such as X-BAT. 中 SP008
CP013 None of the ten retained allocated source pages discloses like-for-like list prices, discount schedules, or realized contract economics across Shield AI and its named alternatives. 中 SP001, SP002, SP003, SP004, SP005, SP006, SP007, SP008, SP009, SP010
CP014 The available multi-aircraft examples support Hivemind portability, but the allocated evidence does not quantify integration time, repeatability, or switching behavior. 中 SP001, SP003
CP015 Acquinox identifies internal autonomy development by Lockheed Martin, Northrop Grumman, and other primes as a bundling threat to third-party autonomy vendors. 中 SP003
CP016 ePlane AI reported that General Atomics suspended autonomous-fighter flight tests after a crash. 中 SP004
CI001 Shield AI deploys its flagship V-BAT vertical takeoff and landing drone across the US Marine Corps, Coast Guard, and international militaries. 中 SI004, SI006
CI002 Shield AI sells its Hivemind autonomy software both embedded on its own platforms and as a standalone product to other defense developers. 高 SI004, SI005
CI003 Shield AI completed the acquisition of tactical simulation company Aechelon Technology to advance simulation capabilities for its Hivemind Foundation Model. 高 SI001, SI007
CI004 Aechelon Technology's simulation software is utilized across more than 1,000 operational military flight training systems, including the Pentagon's Joint Simulation Environment. 中 SI004
CI005 The X-BAT autonomous vertical takeoff and landing fighter jet carries an announced target price tag of approximately $27 million per unit. 中 SI004
CI006 Shield AI's revenue roughly doubled from 2022 to reach approximately $163 million in 2023. 中 SI002
CI007 Shield AI generated approximately $300 million in revenue for the fiscal year ending March 2025, falling $100 million short of its $400 million management target. 中 SI002
CI008 Shield AI projects more than 80% revenue growth by year-end 2026 to reach at least $540 million, excluding inorganic revenue from the Aechelon acquisition. 中 SI007
CI009 Third-party linear projection reports modeled Shield AI's 2025 revenue at $1.6 billion, creating a misleading headline impression of a 3.8x sales multiple. 中 SI002
CI010 Shield AI trades near 46x actual trailing revenue on secondary markets based on its $13.9 billion implied market cap and $300 million in FY2025 revenue. 中 SI002
CI011 Secondary investor demand for Shield AI shares sat slightly below available supply at approximately 0.9 to 1 with units trading near $168.17. 中 SI002
CI012 Shield AI has raised approximately $3.52 billion in total funding across 17 financing rounds since 2015. 中 SI003
CI013 Shield AI secured a $1.5 billion Series G equity round at a $12.7 billion post-money valuation co-led by Advent International and JPMorgan Chase in March 2026. 高 SI004, SI006, SI007
CI014 Shield AI complemented its Series G with $500 million in non-dilutive fixed-return preferred equity financing provided by Blackstone. 高 SI006, SI007
CI015 Shield AI does not publicly disclose gross margins, EBITDA, operating cash burn, or net revenue retention metrics. 中 SI002
CI016 The U.S. Navy awarded Shield AI an initial $50 million development contract for the X-BAT autonomous vertical takeoff aircraft to operate from maritime vessels. 中 SI008
CE001 Hivemind is Shield AI proprietary platform-agnostic autonomy stack that enables drones and aircraft to maneuver autonomously in GPS- and communication-degraded environments. 高 SE003, SE007
CE002 Shield AI demonstrated Hivemind autonomous tactical flight on the X-62A VISTA aircraft and MQ-20 Avenger utilizing Autonomy Government Reference Architecture compliant interfaces. 高 SE001, SE007
CE003 Shield AI hardware systems portfolio includes the operational V-BAT vertical takeoff and landing drone and the planned X-BAT autonomous fighter jet. 高 SE001, SE005, SE007
CE004 Shield AI reported in fiscal 2026 that Hivemind was piloting 30 classes of vehicles across expanding defense platforms and military missions. 中 SE002
CE005 Shield AI and AWS demonstrated deploying containerized Hivemind Pilot algorithms from AWS Cloud via IoT Core and ECR directly to edge hardware platforms. 中 SE008
CE006 The X-BAT is designed as an AI-piloted vertical takeoff and landing fighter jet with a 26-foot airframe and an intended 2,000 nautical mile operational range. 高 SE007, SE009
CE007 Shield AI is constructing a 420,000-square-foot advanced manufacturing facility named X-Forge 3 in Des Moines, Washington to manufacture X-BAT aircraft. 中 SE009
CE008 Shield AI planned the acquisition of Aechelon Technology to integrate high-fidelity synthetic reality and simulation capabilities into its multi-agent autonomy development pipeline. 高 SE001, SE002, SE007
CE009 Operating AI autonomy systems controlling aircraft in physical aviation environments creates critical safety risks where errors can ground aircraft, cause crashes, or compromise missions. 中 SE003
CE010 Shield AI recruits specialized software architecture engineers to establish safety-critical workflows and cyber-relevant architecture standards across mission, vehicle, and autonomy software domains. 中 SE005, SE006
CE011 Shield AI and ST Engineering initiated exploratory collaborations to evaluate Hivemind autonomy software integration on allied uncrewed defense systems. 中 SE004
CE012 Early V-BAT vertical landing operations under manual assistance resulted in a severe injury to a US Navy service member that prompted temporary operational restrictions. 中 SE007
CE013 Shield AI positioned Hivemind autonomy software for participation in the U.S. Air Force Collaborative Combat Aircraft program. 中 SE002
CE014 Shield AI leadership stated long-term goals to expand their Seattle-area manufacturing footprint by up to 1 million additional square feet to support autonomous aviation production. 中 SE009
CU001 Singapore's DSTA and RSAF expanded their partnership with Shield AI to field autonomous drone applications using the Hivemind SDK. 高 SU001, SU002
CU002 The Republic of Singapore Air Force utilizes Shield AI's Hivemind SDK to independently design, test, and deploy sovereign mission autonomy for its operational concepts. 中 SU001
CU003 Shield AI executives disclosed a 2025 revenue base of approximately $300 million with projected 2026 growth above 80 percent to exceed $540 million. 中 SU003
CU004 Anduril Industries generated $2.2 billion in 2025 revenue, operating at roughly 7.3 times the revenue scale of Shield AI. 中 SU003
CU005 The Indian Army selected Shield AI's V-BAT aircraft and Hivemind autonomy software in January 2026 alongside development tools for Indian industrial partners. 中 SU002
CU006 The Polish Armament Agency contracted for V-BAT aircraft for Polish Navy use in June 2026, and the Japan Maritime Self-Defense Force selected multiple V-BAT systems in January 2025. 中 SU002
CU007 A maritime surveillance pilot involving Frontex and the Italian Coast Guard was conducted using V-BAT with Global Sat Tech serving as operator of record. 中 SU002
CU008 Shield AI and Amazon Web Services demonstrated containerized Hivemind Pilot deployment from AWS Cloud to edge hardware via AWS IoT Core and Amazon ECR. 中 SU004
CU009 Shield AI relies heavily on government defense appropriations and military programs, creating significant customer concentration and multi-year procurement exposure. 中 SU003, SU007
CU010 Aechelon Technology provides high-fidelity simulation and synthetic reality supporting U.S. military pilot training and the Pentagon's Joint Simulation Environment. 高 SU005, SU009
CU011 The U.S. Navy awarded Shield AI an initial $50 million development contract for the vertical takeoff and landing X-BAT autonomous strike fighter. 中 SU006
CU012 Historical company reports indicated that product sales generated 60 percent of revenue in 2022, while service contracts and consulting contributed 25 percent and 15 percent respectively. 中 SU008
CU013 Shield AI's historical defense clients include the U.S. Special Operations Command, U.S. Air Force, U.S. Marine Corps, and U.S. Navy. 中 SU007
CU014 Net revenue retention, multi-year cohort churn, and exact renewal rates for Hivemind Enterprise remain undisclosed in public records. 中 SU002, SU003
CR001 A V-BAT landing demonstration in April 2024 resulted in a U.S. service member having three fingers partially amputated, followed by a full recovery. 高 SR007, SR004
CR002 U.S. military customers imposed V-BAT flight restrictions for several months after the 2024 injury. 高 SR007, SR004
CR003 Forbes reported earlier propeller warnings, cracked fuselages, faulty fuel systems, and aircraft losses during V-BAT demonstrations. 中 SR007
CR004 Shield AI said upgraded landing gear removed the need for human assistance and that it passed two Pentagon safety audits, while Forbes reported that the units conducting the audits did not respond to its request for comment. 中 SR007, SR004
CR005 Final closing of Shield AI's 2026 financing was reported as contingent on regulatory approval of the Aechelon acquisition, with financing to be re-evaluated if clearance failed. 高 SR006, SR001
CR006 The reported 2026 financing package comprised $1.5 billion of Series G funding and $500 million of fixed-return preferred equity. 高 SR006, SR001, SR004
CR007 Shield AI projected more than 80% revenue growth and at least $540 million of revenue for 2026, excluding Aechelon. 高 SR006, SR001
CR008 Shield AI plans first X-BAT output from a 420,000-square-foot Washington facility in 2028 and a ramp to approximately 150 aircraft per year by the early 2030s. 中 SR005
CR009 Independent analyses report that human-in-the-loop requirements, restricted mission profiles, and divergent national standards could delay or limit autonomous-system deployments and international sales. 中 SR002, SR001
CR010 Independent analyses identify multiyear defense budgets, political reprioritization, and slow procurement as structural risks for Shield AI's government-dependent sales cycle. 中 SR002, SR001
CR011 Independent analyses identify established defense primes, well-funded autonomy peers, and competition for AI talent as material competitive pressures. 中 SR002, SR001
CR012 China placed Shield AI on its Unreliable Entities list in April 2025, barring the company from sourcing parts, operating, investing, or visiting the country. 中 SR004
CR013 Shield AI's president said the China restriction would not affect the company because it was intentionally not reliant on China. 中 SR004
CR014 Shield AI's website reports Hivemind demonstrations on the X-62 VISTA, MQ-20 Avenger, and two Kratos Firejet drones. 中 SR003
CR015 EquityZen characterizes Shield AI as operating in a capital-intensive industry with long development cycles and cites persistent risk of not achieving large-scale profitability. 中 SR008, SR007
CR016 Public evidence in the assigned pool does not quantify Shield AI's cash burn, gross margin, EBITDA, unrestricted runway, supplier concentration, or customer concentration. 低
CV001 Shield AI announced a $1.5 billion Series G equity round at a $12.7 billion post-money valuation alongside $500 million in fixed-return preferred equity financing. 高 SV001, SV002, SV008
CV002 The Series G equity round was co-led by Advent International and JPMorganChase's Security and Resiliency Initiative. 中 SV001, SV002
CV003 Blackstone provided $500 million in fixed-return preferred equity financing as part of the overall $2.0 billion financing package. 高 SV001, SV002, SV008
CV004 Proceeds from the March 2026 financing support the acquisition of flight-simulation and synthetic environment software provider Aechelon Technology. 中 SV002, SV004
CV005 Shield AI's valuation increased by approximately 140% to $12.7 billion from its prior $5.3 billion Series F-1 valuation twelve months earlier. 高 SV001, SV002, SV008
CV006 The Department of the Air Force selected Shield AI as a mission autonomy provider for the Collaborative Combat Aircraft (CCA) program. 高 SV002, SV007
CV007 The CCA production contract introduces financial risk via an award fee exposure structure where full licensing fees depend on demonstrated combat performance. 中 SV007
CV008 The Polish Ministry of National Defence awarded Shield AI a contract to procure the V-BAT drone system for the Polish Navy for Baltic Sea surveillance. 中 SV007
CV009 Shield AI operates in a capital-intensive defense sector with long development cycles and faces persistent risk of not achieving profitability on a large scale. 中 SV007
CV010 Shield AI projected more than 80% revenue growth in 2026 to exceed $540 million in revenue, up from 2025 levels. 高 SV008, SV002
CV011 Forge Price model reflects an indicative secondary private price of $172.04 per share with high market trading activity as of late September 2026. 中 SV007
CV012 Speculative investor commentary and news reports mention talks targeting a $20 billion forward valuation round. 中 SV006
CV013 Cumulative equity and debt capital raised by Shield AI reached approximately $2.9 billion to $3.16 billion across 17 financing events. 中 SV006
CV014 Shield AI is establishing a 420,000-square-foot X-Forge 3 advanced manufacturing center in Des Moines, Washington, to produce the X-BAT fighter jet. 中 SV005
CV015 First X-BAT aircraft roll-off from the Des Moines facility is scheduled for 2028, with full production targeting roughly 150 aircraft annually by the early 2030s. 中 SV005
CV016 Shield AI's implied forward multiple at a $12.7 billion valuation against $540 million projected revenue is approximately 23.5x. 高 SV001, SV008
来源
编号出版方标题引文
SO001 Shield AI Shield AI to acquire software simulation company Aechelon and raise $2B at $12.7B valuation SAN DIEGO – March 26, 2026 – Shield AI today announced it is raising $1.5 billion in Series G funding at a $12.7 billion post-money valuation and $500 million in fixed-return preferred equity financing.
SO002 FAQ Defense AI Startup Shield AI Raises $2B at $12.7B Valuation, Acquires Flight Simulator Firm Shield AI is projecting more than $540 million in revenue for fiscal year 2026, representing approximately 80% year-over-year growth. The company has publicly stated an intention to pursue an IPO in 2027, contingent on market conditions.
SO003 Sacra Shield AI revenue, valuation & funding Shield AI was founded in San Diego in 2015 by Ryan Tseng (founder of WiPower, acquired by Qualcomm), Brandon Tseng (former NAVY Seal), and Andrew Reiter (ex-Draper Labs) to build technologies like artificial intelligence-powered fighter pilots and drones for defense operations.
SO004 IPO CLUB Shield AI’s Path to a $24 Billion Valuation To justify a $24 billion public market valuation, Shield AI would need to demonstrate financial metrics aligned with both defense industry norms and high-growth tech company expectations.
SO005 Value Add VC $12.7B Shield AI IPO — Defense Tech Listing (2026) That said, a $12.7 billion valuation on ~$540 million of projected 2026 revenue is not cheap, and Shield AI will need to keep converting Air Force pilot programs into durable, repeatable contracts to justify it on public markets.
SO006 Shield AI Shield AI Our mission is to protect service members and civilians with intelligent systems.
SO007 Wikipedia Shield AI Shield AI established in 2015 is a defense and artificial intelligence technology startup founded by former-Navy Seal Officer Brandon Tseng, his brother Ryan Tseng, and Andrew Reiter in San Diego, California.
SO008 Shield AI Data Protection Shield Ai Shield AI is a real-time data protection and compliance platform that uses Generative AI to safeguard sensitive information.
SM001 TSG Invest Shield AI Stock: $12.7B Valuation — Is It a Buy? The global military drone market, valued at approximately $14.4 billion in 2023, is projected to grow to $47.16 billion by 2032.
SM002 IPO CLUB Shield AI Stock: Pre-IPO Price, Valuation & How to Buy (2026) Shield AI serves U.S. and allied defense forces, positioning itself to benefit from rising defense budgets and long-term investment in autonomous and AI-enabled warfare systems.
SM003 Shield AI Shield AI Shield AI enabled the X-62 VISTA to autonomously fly and perform tactical maneuvers against human pilots.
SM004 Angel Investors Network Shield AI $1.5B Series G: Defense Tech Deal Analysis Defense tech VC deals reached $49.1 billion in 2025, up from $27.2 billion in 2024, according to PitchBook data cited by Defense News.
SM005 Contrary Research Report: Shield AI Business Breakdown & Founding Story
SM006 Wikipedia Shield AI On April 9, 2025, Shield AI was added to China's Unreliable Entities list which barred the company "from sourcing parts, operating, investing or visiting the country."
SM007 Shield AI (unrelated namesake) Shield Ai Shield AI is a real-time data protection and compliance platform that uses Generative AI to safeguard sensitive information.
SM008 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet, plans to add 4,000 jobs by early 2030s Harris said he expects the first X-BAT aircraft to roll off the line in Des Moines in 2028, with full production ramping up to about 150 aircraft per year by the early 2030s.
SM009 Fortune Shield AI projecting more than $540M in revenue this year as valuation more than doubles to $12.7 billion Shield AI is projecting more than 80% revenue growth by the end of 2026, Shield AI cofounder Brandon Tseng and CFO Kingsley Afemikhe told Fortune.
SP001 Shield AI Defense-Tech and Autonomous Aircraft | Shield AI Shield AI enabled the X-62 VISTA to autonomously fly and perform tactical maneuvers against human pilots.
SP002 LAVX Defense Startup Shield AI in Talks for Valuation of at Least $20 Billion Competitors include Anduril Industries, valued at $14 billion as of 2024, and Palantir Technologies. Shield AI's prime contractor model—developing and owning its intellectual property rather than working as a subcontractor—has differentiated it in a market historically dominated by legacy primes like Lockheed Martin and Northrop Grumman.
SP003 Acquinox Capital Shield AI: Defence Autonomy Investment Outlook First, Shield AI's Hivemind autonomy software was selected for the Collaborative Combat Aircraft programme, integrating it onto Anduril Industries' YFQ-44A "Fury" drone.
SP004 ePlane AI Shield AI Demonstrates Autonomous Fighter Jet Capabilities Established defense contractors such as General Atomics, Anduril, and Northrop Grumman are actively competing for leadership in the U.S. Air Force’s collaborative combat aircraft program. The sector has recently faced setbacks, including General Atomics’ decision to suspend flight tests following a crash, highlighting the inherent risks associated with developing autonomous fighter jets.
SP005 Recruiting from Scratch Shield AI Is Hiring 391 People: What That Signals (2026) In total, Shield AI has 391 open roles, with engineering leading the way at 173 positions. The 131 new roles added in the last 29 days signal rapid growth and a strong demand for talent.
SP006 Wikipedia Shield AI The company has a drone swarming capability called V-BAT Teams, which enables a single human operator to command a minimum of four V-BAT drones.
SP007 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet, plans to add 4,000 jobs by early 2030s The San Diego-based company, valued at $12.7 billion following a $2 billion venture capital round earlier this year, is looking to complete construction on a 420,000-square-foot facility in Des Moines, Wash., just south of Seattle, next spring.
SP008 CompositesWorld SPE ACCE 2026 announces Shield AI speaker as first keynote Bosworth’s presentation, “Fly by: Aerospace composites for an automotive audience” will be a high-level discussion of how composites are essential in making modern aircraft, especially vertical takeoff and landing (VTOL) platforms.
SP009 Dataford Shield AI Product Manager Interview Questions & Guide 2026 As a **Product Manager** at **Shield AI**, you will sit at the intersection of cutting-edge artificial intelligence, aerospace engineering, and national security.
SP010 Shield AI via Built In Business Development Lead, Baltics (R5227) - Shield AI Lead territory strategy and pipeline development across the Baltics for Shield AI's autonomy and aircraft portfolio. Drive end-to-end deal capture, account growth, partner development, and contract negotiation while engaging senior military and government stakeholders.
SI001 Built In Shield AI Company Growth, Stability & Outlook 2026 Shield AI completed its acquisition of Aechelon Technology, bringing high-fidelity visual simulation, physics-based sensor modeling and synthetic-environment technology into the company.
SI002 Hustle Fund Shield AI Pre-IPO Shares: What Accredited Investors Should Know in 2026 | Hustle Fund The company's actual revenue for the fiscal year ending March 2025 was about $300 million, per its own disclosures. It had been targeting $400 million and came up $100 million short. So the real trailing multiple is not 3.8x. Against a $13.9 billion market cap and $300 million in revenue, you are paying closer to 46x.
SI003 Tracxn Shield AI Shield AI has raised a total funding of$3.52B over 17 round s.Its first funding round was on 2015.Its latest funding round was a Series G round on Mar 26, 2026 for $2B.7 investor s participated in its latest round.
SI004 Tectonic Shield AI Raises $2B Series G at $12.7B Valuation In October, they announced the X-BAT autonomous VTOL fighter jet (or mega-CCA), measuring 26 feet long with a 39-foot wingspan (and a $27M price tag). The company says it’ll have a 2,000-nautical-mile range and a service ceiling of 50,000 feet, just shy of the F-15’s roughly 60,000-foot ceiling. Full mission capability testing is expected in 2028.
SI005 Shield AI Shield AI Our mission is to protect service members and civilians with intelligent systems.
SI006 Wikipedia Shield AI In March 2026, Shield AI announced it is raising $1.5 billion in Series G funding at a $12.7 billion post-money valuation and $500 million in fixed-return preferred equity financing. A portion of the proceeds will support the acquisition of Aechelon Technology, Inc.
SI007 Fortune Shield AI projecting more than $540M in revenue this year as valuation more than doubles to $12.7 billion | Fortune Shield AI is projecting more than 80% revenue growth by the end of 2026, Shield AI cofounder Brandon Tseng and CFO Kingsley Afemikhe told Fortune. That would equate to at least $540 million in revenue this year, based on Shield AI’s 2025 revenue figures.
SI008 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet, plans to add 4,000 jobs by early 2030s Last month, the U.S. Navy backed the program through an initial $50 million development contract, envisioning the autonomous fighter jets operating directly from destroyers, amphibious assault ships and converted cargo vessels to project long-range airpower without relying on traditional aircraft carriers.
SE001 Shield AI Defense-Tech and Autonomous Aircraft Shield AI enabled the X-62 VISTA to autonomously fly and perform tactical maneuvers against human pilots.
SE002 Shield AI Scaling What Works, Building What’s Next Hivemind is now piloting 30 classes of vehicles and expanding across new platforms and missions. We advanced Hivemind’s role in efforts like the U.S. Air Force’s Collaborative Combat Aircraft program. We unveiled X-BAT.
SE003 TechCrunch Shield AI, Waabi, and General Motors talk AI at Disrupt 2026 At Shield AI, Nathan Michael leads the development and deployment of Hivemind, the company’s platform-agnostic mission autonomy software.
SE004 The Defense Post Shield AI, ST Engineering Explore Hivemind Autonomy Software Integration Shield AI, ST Engineering Explore Hivemind Autonomy Software Integration
SE005 Built In Boston Staff Engineer, AI Platform & Architecture (R5449) - Shield AI The Staff Engineer, AI Engineering is a senior individual contributor responsible for translating the enterprise AI engineering roadmap into scalable platform architecture, reusable technical patterns, and production-grade shared services.
SE006 DroneJobs Shield AI - Principal Engineer, Software Architect (R4562) Lead the software architecture for Shield AI’s XBAT program across Mission, Vehicle, and Autonomy software domains. Build a safe, secure, and scalable architecture that enables high‑assurance, cyber‑relevant software development for an advanced airborne system and its ground stations
SE007 Wikipedia Shield AI Hivemind, its autonomy and artificial intelligence stack, in 2015. This software helps drones and aircraft maneuver autonomously in GPS- and communication-degraded environments.
SE008 Amazon Web Services Shield AI and AWS collaborate to deliver mission autonomy at fleet scale By integrating Shield AI’s platform agnostic autonomy software development kit, Hivemind Enterprise, with AWS IoT Core and Amazon Elastic Container Registry (ECR), autonomy developers can develop a Hivemind Pilot and—after simulation and tests are complete—deploy it directly to their chosen hardware platform from the AWS Cloud using either the Hivemind command line interface (CLI) or user interface.
SE009 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet Named X-Forge 3, the initial facility will serve as the primary manufacturing and engineering center for the X-BAT.
SU001 Shield AI Shield AI, Republic of Singapore Air Force, and Defence Science and Technology Agency expand partnership to progressively field autonomy capabilities SINGAPORE (February 5, 2026) — Shield AI today announced the expansion of their partnership with Singapore’s Defence Science and Technology Agency (DSTA) and the Republic of Singapore Air Force (RSAF) to co-develop and proliferate Artificial Intelligence (AI) across a wider range of autonomous drone applications using Shield AI’s Hivemind software development kit (SDK).
SU002 Baltic Defense Initiative Shield AI: aircraft, Hivemind and autonomy software On 28 January 2026, Shield AI announced that the Indian Army had selected V-BAT aircraft and Hivemind autonomy software. It also described development tools for selected Indian partners.
SU003 Signals Inbox Anduril vs Shield AI: who will win? Anduril is winning against Shield AI today. It has roughly seven times more annual revenue, deeper access to major procurement programs, more production capacity and far more control over the systems surrounding military autonomy.
SU004 Amazon Web Services Shield AI and AWS collaborate to deliver mission autonomy at fleet scale | Amazon Web Services By integrating Shield AI’s platform agnostic autonomy software development kit, Hivemind Enterprise, with AWS IoT Core and Amazon Elastic Container Registry (ECR), autonomy developers can develop a Hivemind Pilot and—after simulation and tests are complete—deploy it directly to their chosen hardware platform from the AWS Cloud using either the Hivemind command line interface (CLI) or user interface.
SU005 Aechelon Technology Shield AI to acquire software simulation company Aechelon and raise $2B at $12.7B valuation - Aechelon Aechelon’s technology is used by the U.S. military and allies to train pilots and test advanced aircraft and autonomous systems before live flight, and supports a range of critical national security platforms, including the Pentagon’s Joint Simulation Environment (JSE).
SU006 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet, plans to add 4,000 jobs by early 2030s Last month, the U.S. Navy backed the program through an initial $50 million development contract, envisioning the autonomous fighter jets operating directly from destroyers, amphibious assault ships and converted cargo vessels to project long-range airpower without relying on traditional aircraft carriers.
SU007 Wikipedia Shield AI Its clients include the United States Special Operations Command... and several international militaries.
SU008 Vizologi Business Strategy Canvas - Vizologi In 2022, product sales accounted for 60% of our revenue, while service contracts contributed 25%. Specialized solutions and consulting fees, driven by specific client requirements, made up the remaining 15%
SU009 Shield AI Shield AI Our mission is to protect service members and civilians with intelligent systems.
SR001 Tech Insider Shield AI Raises $2B at $12.7B Valuation [2026] Despite the optimistic trajectory, Shield AI faces significant challenges. The Aechelon Technology acquisition remains contingent on regulatory approval, and the Series G’s final close is tied to the deal’s completion.
SR002 BUVCG Research Shield AI - Deep Dive Autonomous weapons remain one of the most scrutinized categories in defense.
SR003 Shield AI Defense-Tech and Autonomous Aircraft Shield AI enabled the X-62 VISTA to autonomously fly and perform tactical maneuvers against human pilots.
SR004 Wikipedia Shield AI
SR005 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet, plans to add 4,000 jobs by early 2030s Harris said he expects the first X-BAT aircraft to roll off the line in Des Moines in 2028, with full production ramping up to about 150 aircraft per year by the early 2030s.
SR006 Fortune Shield AI projecting more than $540M in revenue this year as valuation more than doubles to $12.7 billion The final close of Shield AI’s funding round will also be contingent on approval of the Aechelon acquisition. If the deal fails to clear regulatory hurdles, Tseng said the company would “re-evaluate” the financing with investors.
SR007 Forbes How A Grisly Injury Threw $5 Billion Drone Startup Shield AI Off Course CEO Ryan Tseng, confirming the U.S. Navy incident to Forbes, said that flight restrictions were imposed on the drone by U.S. military customers for several months.
SR008 EquityZen Invest In Shield AI Stock | Buy Pre-IPO Shares Shield AI operates in a capital-intensive industry with long development cycles.
SV001 Venture Capital Tracker Shield AI Valuation 2026: $12.7B Series G and Stock Status Shield AI's latest disclosed valuation is $12.7 billion post-money, announced on March 26, 2026 with a $1.5 billion Series G. The company also announced $500 million of fixed-return preferred equity, so the headline financing package is $2 billion but the priced equity round is $1.5 billion.
SV002 RevenueMemo Who owns Shield AI? Ownership structure explained (2026) Advent International and JPMorgan's Security and Resiliency Initiative co-led the $1.5 billion Series G in March 2026, with Blackstone providing an additional $500 million in preferred equity. Total capital raised in that round reached $2 billion.
SV003 Shield AI Defense-Tech and Autonomous Aircraft | Shield AI Shield AI enabled the X-62 VISTA to autonomously fly and perform tactical maneuvers against human pilots.
SV004 Wikipedia Shield AI In March 2026, Shield AI announced it is raising $1.5 billion in Series G funding at a $12.7 billion post-money valuation and $500 million in fixed-return preferred equity financing. A portion of the proceeds will support the acquisition of Aechelon Technology, Inc.
SV005 GeekWire Shield AI selects Seattle area to build X-BAT autonomous fighter jet, plans to add 4,000 jobs by early 2030s Shield AI is establishing a massive advanced manufacturing hub in Washington state, tapping the Seattle region’s deep aerospace and tech roots to build a next-generation autonomous fighter jet called the X-BAT.
SV006 Silicon Valley Investment Club Shield AI Company Profile & Tear Sheet Raising $2B ($1.5B Series G + $500M preferred equity from Blackstone) at $12.7B valuation; acquiring Aechelon Technology (Mar 2026).
SV007 Yahoo Finance Shield AI (SHAI.PVT) company profile and facts Despite a $5.3 billion valuation at the company's Series F funding round, and a company press release on Series G funding at a $12.7 billion valuation, Shield AI operates in a capital-intensive industry with long development cycles. The company still faces the persistent risk of not achieving profitability on a large scale.
SV008 Fortune Shield AI projecting more than $540M in revenue this year as valuation more than doubles to $12.7 billion Shield AI is projecting more than 80% revenue growth by the end of 2026, Shield AI cofounder Brandon Tseng and CFO Kingsley Afemikhe told Fortune. That would equate to at least $540 million in revenue this year, based on Shield AI’s 2025 revenue figures.