Roark Aerospace
真实运营足迹,后期披露偏薄
Roark Aerospace 看起来是真在运营的 counter-UAS 公司,已有实质收入和真实融资动作;但相对于外界流传的后期估值和规模叙事,它异常不透明。
封面要素
公司概况
Roark Aerospace 是一家英国反无人机公司,2023 年 11 月 30 日注册成立。公司把自己定位为无人机探测和 C-UAS-as-a-service 平台,覆盖 Aero-Ark 传感器、Plexus 指挥软件,以及快速部署 / 监测服务。官方备案显示,这家公司不只是幻灯片故事:FY2024 账目披露营业额 £36.9 million、净资产 £79.3 million、英国平均员工 49 人,并采用覆盖英国和美国实体的集团销售模式。公开产品证据支持其确有多模态探测、成像和决策支持供给;2025 年 3 月至 2026 年 4 月的股权备案也显示重大融资活动和一套复杂演进的资本结构。核心尽调问题不是公司是否存在,而是透明度:客户、投资人名单、安全与合规状态,以及从备案到流传的独角兽级估值之间的桥梁,对一家声称后期规模的公司来说都明显偏薄。
- 成立时间
- 2023-11-30
- 总部
- London, England, UK
- 产品
- Roark 销售 Aero-Ark 探测设备和 Plexus 软件,供给覆盖直接硬件销售、Detection as a Service、Rapid Deploy 现场服务,以及更广的 WAMI/HSI 监测数据流。公开材料称,其分层技术栈在陆地、空中和海上环境中使用 RF 感知、数据包捕获、被动雷达、边缘处理和指挥控制流程。
- 客户
- 表面上的目标客户包括政府和军方用户、公共安全与安保运营方、关键基础设施业主、临时活动买家和私人监测客户,但公开来源没有点名任何生产级客户。
- 商业模式
- 混合硬件与服务模式,把设备销售、经常性 detection-as-a-service、快速部署费用和平台访问捆在一起,而不是纯软件订阅业务。
- 阶段
- Private company; Series B claimed by third parties in late 2025
- 融资情况
- 官方备案显示,2025 年 3 月有重大股权活动,2025 年 12 月有一笔 £156.96 million 的实缴配股;第三方来源流传的则是 $210 million Series B、投后估值 $1.6 billion,但公开层面仍只得到部分验证。
执行摘要
主要优势
- 官方文件支撑其真实运营质地,包括成立仅一年后 FY2024 £36.9M 营业额和 £79.3M 净资产。
- 产品证据支持 Aero-Ark 加 Plexus counter-UAS 栈真实存在,并通过硬件和服务两种包装销售。
- 2025 年 3 月至 2026 年 4 月的文件证明公司完成了实质资本形成并仍能获得融资,而不是纯宣传式融资故事。
- 2026 年,防务、安保和关键基础设施领域的 counter-UAS 需求和采购活动都在扩张。
主要风险
- 所称 $1.6B 估值和 $210M Series B 缺少清晰公开投资者名单、签署版股权结构桥和完整透明权利包。
- 公开证据仍没有指向具名量产客户或强独立部署验证。
- FY2024 账目显示毛利为负、应收账款高、库存可观,且相对于复杂资产负债表,披露现金有限。
- 在准入受管制的防务市场,买方越来越要求可验证信任信号;Roark 的治理、合规、供应商和安全披露仍偏薄。
未决问题
- 2025 年末融资的签署版股权结构表、投资者名单和完整优先权桥。
- 具名客户引用、量产部署证据,以及集中度 / 留存指标。
- 截至 2025 年 11 月 30 日的账目、当前现金跑道,以及 FY2024 后毛利质量改善的证明。
- 适用于防务和公共安全采购的出口管制、安全供应和数据安全文件。
目录
01公司概览
1.1 身份与足迹
对一家声称达到十亿美元规模的公司而言,Roark Aerospace 的公开身份异常压缩,但官方记录把法律壳体说得很清楚。Companies House 将 RO-ARK AEROSPACE LIMITED 列为活跃的私人有限公司,2023 年 11 月 30 日注册,注册地址为 Monomark House, 27 Old Gloucester Street, London, WC1N 3AX。归档的法律页面还增加了 Cambridge Science Park 地址和纽约办公室,说明其运营足迹可能比单一英国备案显示的更分散;归档首页则展示 NATO NCAGE、UEID、DUNS 和 VAT 标识。合起来看,官方和归档记录支持一个结论:Roark 很快就围绕防务相邻商业活动完成组织搭建,至少在营销材料中维持了英国—剑桥—美国足迹,并在生命周期很早就向盟国政府买家释放采购就绪信号。Cushman & Wakefield 2026 年一季度 marketbeat 提到 Cambridge 一处 13,650 平方英尺租约,这是 Cambridge 存在不只是门面地址的最强独立佐证。[CO001, CO002, CO003, CO004, CO005, CO007]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 / 备注 |
|---|---|---|---|---|
| 法律实体 | RO-ARK AEROSPACE LIMITED | 2026-06-20 | 高 | Companies House 已确认 |
| 成立时间 | 30 Nov 2023 | 2023-11-30 | 高 | 法律实体非常年轻 |
| 注册办公地址 | Monomark House, 27 Old Gloucester Street, London WC1N 3AX(伦敦注册地址) | 2026-06-20 | 高 | 归档网站还显示 Cambridge 和 New York 办公室 |
| 商业模式标签 | 无人机探测即服务 / C-UAS 即服务 | 2026-01-08 归档 | 中 | 归档营销话术,不是合同收入拆分 |
| 2024 年营业额 | £36.9M | 2024-11-30 | 中 | 已申报账目,英国和美国实体按集团销售口径处理 |
| 2024 年税前利润 / 亏损 | -£16.56M | 2024-11-30 | 中 | 尽管收入规模可观,仍处亏损 |
| 英国平均员工数 | 49 | 2024-11-30 | 中 | 不含承包商和国际员工 |
| 公开融资轮主张 | Series B $210M,估值 $1.6B | 2025-12 | 中 | 有媒体和数据库支持,但投资人身份仍验证很薄 |
财务 KPI 混合了一手申报和第三方融资轮报道;估值和全球规模指标仍需要管理层会议室确认。
[CO001, CO002, CO003, CO006, CO012, CO013]当前证据集中可站得住的高信号身份、财务和融资数据点。
[CO003, CO012, CO014, CO015, CO018, CO021]1.2 领导层、创始人控制与治理
公开治理披露集中在一个人身上。Companies House 显示,Patton Atkinson-French 是唯一在任董事,注册当日获任;他也是唯一的 significant control person,持有超过 75% 的股份和投票权,并有权任免董事。高管和任职页面还显示身份验证已完成、英国国籍、居住在英格兰、出生日期为 1985 年 2 月。Roark 因此不像一家成熟的风险资本支持防务主承包商,更像一个由创始人紧控的运营载体,在搭起广泛公开治理表面之前,就快速做大收入和融资。这本身不构成否决项,但会抬出典型尽调问题:董事会构成、投资人保护、授权机制和关键人依赖。考虑到外部来源报道的融资规模,公开可见董事会名单缺失尤其值得注意。[CO008, CO009, CO010]
| 人物 | 角色 | 公开记录 | 控制权 / 职能覆盖 | 关键人物依赖 |
|---|---|---|---|---|
| Patton Atkinson-French | 创始人 / 唯一董事 / PSC | Companies House 中唯一 active 董事;唯一 PSC,持有 75%+ 所有权和投票权 | 创始人控制、治理、融资、外部代表 | 高——所有公开治理和控制记录都指向一个人 |
| 未识别投资人 / SPV 对手方 | 可能的 2025 年末融资参与方 | Nordic9 和 Resilience 提到,但公开申报未完整具名 | 提供资本,并可能拥有董事会权利或清算优先权 | 中到高——身份和条款仍不透明 |
| 未具名的更广泛高管 / 董事会班底 | 未获公开确认 | 公开材料未浮现完整董事会或具名职能负责人 | 对销售、工程、财务和受监管采购执行都很重要 | 信息风险高,因为后期融资通常意味着更广泛的治理基础设施 |
这张表有意保持不完整,因为创始人 / 控制人之外的公开领导层披露很少。
[CO008, CO009, CO010, CO022, CO023]| 利益相关方 | 角色 | 控制权 / 经济重要性 | 尽调要求 |
|---|---|---|---|
| Patton Atkinson-French | 创始人、唯一董事、唯一 PSC | 根据 PSC 申报,持有 75%+ 股份和投票控制权,并拥有任免董事权 | 审查股东协议、保留事项,以及上次 PSC 更新以来的任何稀释 |
| 2025 年 3 月普通股和优先股认购人 | 大型早期机构或关联方融资群体 | 2025 年 3 月 13 日 SH01 在普通股和优先股中创造了 £125M 合计名义资本 | 识别认购人名单、受益所有人,以及这究竟是定价机构资本还是内部重组 |
| 2025 年 12 月 B 普通股投资人 / SPV | 后期融资轮对手方 | 2025 年 12 月 15 日 SH01 显示,131 股 B 普通股实缴现金 £156.96M;Resilience 称有一家美国投行通过 SPV 参与 | 获取认购协议、瀑布条款和董事会权利包 |
| 优先股股东 | 可能受下行保护的资本类别 | 2025 年 12 月 28 日和 2026 年 4 月 4 日申报显示,优先股带有股息 / 流动性优先权表述 | 确认清算优先权、反稀释和 pari passu 排名 |
| 债券贷款人 | 非股权资本提供方 | 2024 年账目披露 £16.42M 债券,年利率 6.5%,2029 年 3 月到期 | 理解契约、担保,以及债务是否限制未来融资或出口 |
公开网页证据支持资本工具和控制权集中,但不支持完整 cap table。
[CO009, CO016, CO017, CO018, CO019, CO022]1.3 商业模式与产品定位
Roark 自有材料和第三方材料都持续把它描述为一张一体化空域防御网络,而不是单一探测器或效应器。归档首页使用了 “Drone Detection as a Service | C-UAS as a Service” 标签;2024 年账目称,公司主要从事无人机探测系统、反无人机系统和无人载具的制造与服务部署。DroneXL 在 2025 年 1 月补充了运营细节,称 Aero-Ark 系统结合 ADS-B、Remote ID、DJI 数据包收集和解密、被动雷达以及边缘 AI 分类;同一报道还称 Roark 通过 Plexus 平台扩展 WAMI 和高光谱成像覆盖。Global Defense Tech News 也把它描述为跨陆地、空中和海上环境部署的统一传感器、反无人机工具和决策支持。共同主线是一套服务主导架构,把感知、地图、分析和指挥软件打包在一起。在一个越来越奖励一体化指挥控制和经常性监测合同的市场里,这个策略说得通。问题在于,大部分技术细节仍停留在二级报道,而不是第一方工程文档。[CO006, CO011, CO027, CO028, CO029, CO040]
Roark 对外营销的服务模型如何连接传感器、地图、指挥软件、客户和资本需求。
[CO006, CO011, CO027, CO028, CO029, CO031]1.4 融资、财务画像与里程碑
Roark 的备案在一个关键点上比媒体报道更有力:到 2024 年底,公司不是空壳。其提交的账目披露,截至 2024 年 11 月 30 日的年度营业额为 £36.9 million、净资产 £79.3 million、英国平均员工 49 人、税前亏损 £16.6 million,同时披露一笔 2029 年 3 月到期的 £16.42 million 债券。与此同时,备案历史显示融资节奏很快:2025 年 3 月 13 日大额配股,2025 年 12 月 15 日 B 普通股备案显示 £156.96 million 现金实缴,随后在 2025 年 12 月、2026 年 1 月、2026 年 4 月又有资本表调整。这些官方备案本身不能证明 $1.6 billion 估值,但确实确认第三方媒体讨论的日期附近发生了实质性资本形成。Nordic9 和 Resilience 都报道了 2025 年末 $210 million Series B 及 $1.6 billion 隐含估值,但 Nordic9 明确标注该交易待验证,Resilience 则强调投资人披露不透明。最干净的结论是:Roark 几乎肯定在 2025 年末完成了一次重大融资事件,但准确投资人名单、轮次结构和估值计算在公开领域仍缺少文档支撑。[CO012, CO013, CO014, CO015, CO016, CO017]
| 日期 | 事件 | 类型 | 金额 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2023-11-30 | RO-ARK AEROSPACE LIMITED 在英国成立 | 创立 | active 状态的私人有限公司 | Patton Atkinson-French | 公司法律起点现在与后续收入和融资主张绑定 |
| 2024-11-30 | 首个已报告会计期结束 | 规模 | £36.9M 营业额;£79.3M 净资产;49 名英国平均员工 | Ro-Ark 集团 | 在公开独角兽叙事前,显示了有意义的早期商业活动 |
| 2025-01-30 | DroneXL 报道 OTC Aero-Ark 产品可用,并报道与 Plexus 相关的地图扩张 | 产品 | 单价自 £4,500 起;称 WAMI/HSI 覆盖 30 个美国城市和 12 个英国城市 | Roark Aerospace / DroneXL | 暗示从定制防务安装转向更可复制的产品 / 服务包装 |
| 2025-03-13 | 大额 SH01 配股申报 | 融资 | 100,000,000 股;£125,000,000 合计名义资本 | Ro-Ark 认购人 | 表明在 2025 年末融资轮主张之前,已有大规模早期资本化或重组 |
| 2025-06-20 | 2024 年账目提交至 Companies House | 治理 | 小公司制度下的账目 | Ro-Ark / Companies House | 为收入、亏损和债务提供第一份硬性公开证据 |
| 2025-12-15 | B 普通股配股记录 | 融资 | 131 股 B 普通股实缴 £156,963,450 | Ro-Ark 和融资轮投资人 | 与 2025 年末重大融资事件最一致的一手申报 |
| 2025-12-18 | Nordic9 报道 $210M Series B,估值 $1.6B | 融资 | $210M;“仍待验证” | Nordic9 / Roark Aerospace | 公开独角兽叙事出现,但带有验证保留 |
| 2025-12-26 | Resilience 强调围绕 Roark 新闻稿主张的验证担忧 | 反向 | 投资人身份和主张受到质疑 | Resilience Media / Roark 管理层 | 让透明度成为一阶尽调问题,而不是脚注 |
| 2026-04-06 | 2026 年 4 月 SH01 再次更新 cap table | 治理 | 配发 20,079 股 B 普通股;总股数 1,372,418 | Ro-Ark 认购人 | 显示融资轮后 cap table 管理延续到 2026 年 |
| 2026-06 | NATO 和盟军机构加强 C-UAS 采购和测试路径 | 市场 | AIRCOM 重点日;NATO 市场平台工作流 | NATO AIRCOM / U.S. Army / 盟友 | 支持 Roark 扩张时所处买方环境确实火热这一论点 |
日期和金额直接锚定在申报或检索到的文章上;客户名称大多仍未披露。
[CO003, CO012, CO014, CO017, CO018, CO020]从 2023 年底到 2026 年中,法律设立、产品信号、账目、融资申报和验证关键节点的时间顺序。
时间线混合使用一手备案和独立文章;公开网络上的投资人身份仍不完整。
[CO003, CO008, CO012, CO013, CO014, CO017]1.5 验证缺口与尽调风险
主要尽调问题不是 Roark 毫无证据,而是最强证据和最大胆主张分属记录的不同层。官方备案支撑了注册、创始人控制、运营收入、债务和多次注册后资本事件。但头部风险叙事——55,000 台已部署设备、3,000 名最终用户、244 名员工、$210 million Series B 和 $1.6 billion 估值——主要由 Tracxn、行业媒体和一篇部分付费墙的批评文章支撑,而不是由开放的投资人或客户披露支撑。Resilience 明确把 Roark 描述为一个“无人能验证的防务独角兽”,归档官网也只露出极少营销文案和标识。在 DroneShield、Shield AI 等成熟同行会定期发布投资者材料和更完整产品叙事的行业里,Roark 的披露水平明显低于后期防务科技投资人希望看到的标准。后续章节应以一个工作假设展开:Roark 是真实存在、商业活跃且资金充足的公司,但许多规模和客户主张必须在管理层会议室验证后,才应进入承销判断。[CO023, CO025, CO026, CO031, CO032, CO033]
1.6 附件
02市场分析
2.1 市场边界与分层逻辑
Roark 自己的公开表述很重要,因为该赛道的头部数字宽到容易误导。归档首页把公司定位为 "Drone Detection as a Service | C-UAS as a Service";DroneXL 和 Defense Tech News 描述的组合则包括 OTC 探测设备、广域感知、反无人机工具和 Plexus 数据层。这意味着最接近的市场不是整个自主系统技术栈,也不是所有无人系统支出,而是分层的 counter-UAS 任务链:探测、跟踪、识别、压制和作战指挥支持。NATO AIRCOM 2026 年 focus day 强化了这个边界,它强调从探测到 neutralisation 的全链条;美国陆军 / NATO 的 marketplace 工作同样先从反无人机系统切入,再扩展到相邻雷达和传感器。落到实际口径,纳入的支出包括分层战场和场站防御项目、可部署 / 机动 C-UAS 套件、机场或关键基础设施空域保护,以及经常性监测或托管服务叠加层。排除的支出包括广义自主软件、通用无人机,以及不解决同一反无人机任务的相邻防空品类。这个更窄的边界更适合承销,因为它匹配 Roark 声称服务的买家问题,也凸显现有企业已经发布了比 Roark 更强的部署证明。[CM001, CM002, CM003, CM004, CM005, CM008]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 相关性 |
|---|---|---|---|---|
| 分层军事 C-UAS 项目 | 面向战场或基地防御的探测、跟踪、压制、C2、训练和可部署系统 | 不解决反无人机任务的一般导弹防御或广义自主项目 | 国防部、军种、盟军联盟项目 | 核心目标市场,因为 Army、NATO 和供应商页面都把问题定义为分层探测-跟踪-压制 |
| 固定场地关键基础设施保护 | 围绕公用事业、港口、园区和战略场地的持续监测、缓解和响应 | 没有空域防御需求的一般物理安防或摄像头预算 | 基础设施所有者 / 运营方、安全集成商、国家监管机构 | 与 Roark 相关,因为服务主导的监测和低附带损害缓解可适配连续场地防御 |
| 机场 / 飞行场空域安全 | 保护飞行运行的无人机探测、分类和安全缓解 | 一般机场 IT 或无关周界安防支出 | 机场运营方、航空主管机关、公共安全伙伴 | 重要相邻细分,因为同行明确把机场列为核心终端市场 |
| 公共安全 / 活动 / VIP 保护 | 面向大型活动、公共场馆、监狱和 VIP 行动的临时或持续保护 | 没有无人机拦截功能的通用警务软件或人群管理工具 | 执法机构、国土安全、活动安保预算负责人 | 对快速部署系统有吸引力,但证明和法律 / 监管适配极其重要 |
| 广义自主和无人系统相邻领域 | 软件、自主工厂、边缘自主和一般无人系统预算 | 把所有自主或所有 UAS 支出都视为直接 C-UAS 收入机会 | 混合:国防实验室、OEM、平台项目、软件买方 | 可作为背景,但若不收窄,太宽,不能作为 Roark TAM |
纳入 / 排除逻辑基于反无人机待完成任务,而不是泛防务科技标签。
[CM001, CM002, CM003, CM004, CM005, CM008]相关市场从宽泛的自主系统预算,收窄到规模小得多、资格门槛更重的 C-UAS 细分;后者更贴合 Roark 对外营销的服务方案。
这是边界收窄逻辑图,不是可相加的 TAM 堆栈;各层分母不同,不应求和。
[CM001, CM006, CM007, CM008, CM009, CM034]2.2 规模口径与需求背景
需求背景很强,但每个公开规模口径用的分母都不同。DefenseScoop 报道称,Pentagon FY26 预算请求包含 $13.4 billion 的自主与自主系统支出,其中 $3.1 billion 专门用于跨军种 counter-UAS。这立即说明,如果把广义自主系统数字当作 TAM,会高估 Roark 的直接市场。国际侧,Unmanned Airspace 称 2026 年前几个月公开宣布的政府 C-UAS 合同超过 $29 billion,但这是一个滚动合同总额,而不是干净的年度市场定义。Defense News 给出另一个视角:NATO 正在九个用例中筛选 18 个经审查方案,并提供购买和租赁选项,这更应理解为近期资质瓶颈,而不是支出总额。供应侧数量也显示同样的模糊性。Unmanned Airspace 目录称其跟踪 550+ 家公司和 1,000+ 个项目,而其支出文章称拦截无人机制造商已超过 70 家。正确结论不是某一个数字为真、其他为假,而是公开证据指向一个庞大且迫切的市场;真正可买的切片更窄、经过筛选,也比单一通用 TAM 标题暗示的更拥挤。[CM006, CM007, CM008, CM009, CM012, CM018]
| 发布方 | 年份 | 地域 | 数值 | 方法 | 置信度 | 局限 |
|---|---|---|---|---|---|---|
| DoD / DefenseScoop | FY2026 | 美国 | $3.1B counter-UAS 预算申请 | 高级防务官员披露的 FY26 申请预算线 | 中 | 单一买方预算线,不是全球 TAM,也不是 Roark SAM |
| DoD / DefenseScoop | FY2026 | 美国 | $13.4B 自主与自主系统 | 同一简报披露的更广义自主能力总额 | 中 | 仅属相邻领域;包含非 C-UAS 软件和平台 |
| Defense News / NATO | 2026 | NATO 成员国 | 9 个用例覆盖 18 套系统 | 带购买和租赁选项的精选市场平台资质设计 | 中 | 统计的是解决方案,不是支出;更适合看作资质漏斗 |
| Unmanned Airspace | 2026 | 全球 | 2026 年最初数月公开宣布的政府 C-UAS 合同超过 $29B | 基于已宣布奖项和项目汇总的公开合同统计 | 中 | 滚动合同总额,不是标准化年度市场规模估算 |
| Unmanned Airspace Directory | 2026 | 全球 | 550+ 家公司 / 1,000+ 条项目描述 / 70+ 家拦截器制造商 | 行业目录和编辑市场地图 | 中 | 供应方广度视角,不是需求侧支出 |
| Fortem / Lockheed and DroneShield IR(竞品 / 投资者关系) | 2026 / 2025 | 全球 | >$10B 到 >$12B 市场叙事 | 供应商 / 投资者关系口径的板块增长叙事 | 低 | 推广性或公司自撰市场口径,而不是中立的自下而上测算 |
这些视角有意混合预算线、精选项目数量和供应商数量,因为没有公开来源能干净隔离 Roark 特定的 SAM 或 SOM。
[CM006, CM007, CM009, CM018, CM019, CM020]公开数量信号显示:买方侧正在筛选,供应侧则相当拥挤。
数值是计数或保守下限,不是支出总额;70 代表引用来源中的「超过 70」。
[CM009, CM019, CM020]2.3 买家、预算与采用路径
2026 年的采购路径不像缓慢定制的防务采购周期,更像筛选型 marketplace 加验证引擎。美国陆军和 DVIDS 对 Eurosatory 意向书的报道显示,盟国买家正在明确试图缩短采购和训练周期,把供应商放到士兵反馈面前,并放大有效方案。Defense News 还称,NATO 希望合同能让各国在共同资质流程之后购买或租赁预审系统。AIRCOM 和 Counter-UAS Europe 显示了资质流程强调的内容:互操作性、机动与可部署架构、低附带损害的压制手段,以及来自乌克兰实战和公共空间安保的经验。公开供应商材料也把买家地图勾得更清楚。D-Fend、Dedrone/Axon、SRC、Fortem、Leonardo DRS 和 DZYNE 都指向反复出现的细分场景:机动力量和军事基地、机场和飞行场、关键基础设施、公共安全组织,以及高关注活动或 VIP 保护。因此,预算归属按细分场景移动,而不是落在一个泛化防务买家身上:部委和军种掌握战场支出,基础设施运营方掌握场站安保,机场和航空主管机构掌握机场连续性,公共安全组织掌握活动和城市空域风险。Roark 的服务叙事可能会打动需要持续监测或灵活部署的客户,但它仍必须跨过这些渠道正在制度化的互操作、测试和证明门槛。[CM009, CM010, CM011, CM012, CM013, CM014]
| 细分 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 联盟战场 / 基地防御 | Army、防空指挥部、盟军采购办公室 | 操作员、防空人员、部队防护团队 | 政府国防预算 | 资质、靶场测试、部署、现场反馈、规模化订单 | 军种采购或国防部项目办公室 | 紧迫无人机威胁和互操作性要求 |
| 机场 / 飞行场保护 | 机场主管机构或航空安保负责人 | 飞行场运营和安保团队 | 机场运营预算或公共安全支持 | 探测侵入、保护运行,只在缓解安全时升级处置 | 机场高管 / 监管背书的安保负责人 | 避免停摆,并证明低附带损害缓解 |
| 关键基础设施场地防御 | 公用事业、港口、园区或工业场地安保负责人 | 本地安全运营中心或合同操作员 | 场地所有者 / 运营方 | 围绕固定范围的持续监测加快速响应 | 首席安全官或设施运营负责人 | 保护运营连续性和高价值资产 |
| 公共安全 / 活动 / VIP 安保 | 执法机构或活动安保协调方 | 战术团队、公共安全操作员、场馆安保 | 机构预算或活动安保支出 | 高关注活动前和期间的便携或临时部署 | 公共安全指挥部或场馆所有者 | 需要以最小附带风险快速部署 |
| 服务主导 C-UAS / 托管监测利基 | 想获得覆盖、但不想自建空域技术栈的组织 | 安保人员或托管服务客户 | 运营预算,而不只是一次性 capex | 订阅、部署传感器、监测、升级处置、刷新 | 安保 / 运营高管 | 偏好低摩擦上线和持续支持 |
买方和预算负责人标签综合自 Army、NATO 和主要供应商引用明确点名的终端市场。
[CM010, CM011, CM013, CM014, CM024, CM025]各细分市场采购路径不同,但所有可信渠道都在转向更强证明、互操作性和部署证据。
[CM011, CM012, CM013, CM014, CM024, CM025]2026 年采购改革把采用过程变成分阶段证明引擎:先有紧迫需求,再资格筛选、反馈、签约,最后才规模列装。
这是工作流地图,不是已披露的转化漏斗;公开来源描述了阶段和资格逻辑,但没有通过率统计。
[CM010, CM011, CM012, CM013, CM014, CM033]2.4 增长驱动、约束与 Roark 匹配度
三股力量同时把市场往前拉:战场迫切性、采购改革和成本交换压力。Defense News 对拦截无人机的报道说明了买家为什么如此关注价格不对称;Fortem 和 Lockheed 也从供应商侧提出同一点,认为低成本、软件中心的拦截能显著降低交战成本。与此同时,这个市场并不空。Shield AI、Epirus、D-Fend、Dedrone、Fortem、DroneShield、Leonardo DRS、DZYNE 和 SRC 都拿出了证据,表明严肃资本、具名部署或战略收购方已经进入该品类。这给 Roark 带来采用约束。第一,市场拥挤且越来越整合,分销伙伴、客户背书和渠道信任的门槛更高。第二,NATO 和陆军采购改革让互操作性和测试验证更重要,而不是更不重要。第三,Roark 的公司级披露相对那些会点名客户、用例或地域的同行仍然偏薄。因此,Resilience 的批评即便放在市场章节也相关:赛道热度不会自动转化为公司可信度。最有证据支撑的工作判断是,Roark 瞄准的确是一个真实且扩张中的细分市场,尤其是服务主导或固定场站项目;但投资人仍需要客户名称、资质证据和管线转化数据,才能把通用市场增长转成公司级份额假设。[CM015, CM016, CM017, CM021, CM022, CM023]
| 驱动因素 / 约束 | 方向 | 时点 | 含义 | 尽调要求 |
|---|---|---|---|---|
| 盟军采购紧迫性和战场反馈循环 | 驱动因素 | 当前 | 买方想要快速证明、并由操作员反馈塑形的系统,而不是缓慢的需求文件 | Roark 是否已经进入任何经验证市场平台、试验或盟军反馈循环? |
| 互操作性、测试和资质门槛 | 约束 | 当前且在加剧 | 市场与项目改革让通用接口和标准化测试更难绕开 | 哪些 Roark 系统通过了第三方测试、开放架构集成或操作方评估? |
| 低成本拦截与成本交换压力 | 驱动因素 + 约束 | 当前 | 可负担的压制或监测能力可以拉动需求,但买家会把 Roark 同更便宜或已验证的替代方案比较 | Roark 每个受保护站点、部署或交付任务的成本,相比同业方案是多少? |
| 供应商拥挤与行业整合 | 约束 | 当前 | 名录中 550+ 家供应商叠加战略并购,抬高了差异化、分销和信任门槛 | Roark 靠什么渠道、客户或数据优势,在同业和主承包商面前脱颖而出? |
| 服务主导 / 类租赁采购模式 | 驱动因素 | 正在出现 | 购买叠加租赁选项和持续监测,可能更有利于托管服务定位 | Roark 能否以经常性服务条款签约,并守住可接受的毛利和留存? |
| Roark 特有披露与客户证明缺口 | 约束 | 当前 | 同业客户案例已经点名而 Roark 没有,行业需求本身不能证明公司牵引力 | 分配份额前,要求 Roark 提供具名客户、试点转生产数据和资质证明 |
时间标签是定性判断,用来说明该因素已经在影响采购,还是仍处于萌芽阶段。
[CM011, CM012, CM015, CM016, CM017, CM023]2.5 附件
03竞争对手
3.1 格局与竞争类别
Roark 面对的不是一组干净的同类公司。它自己的归档首页和早期行业报道把公司定位为 “Drone Detection as a Service | C-UAS as a Service”,配有便携或固定 Aero-Ark 硬件和服务层,降低客户前期支出。这使最接近的直接竞争来自 counter-UAS 中服务主导、软件中心的部分,而不是整个防务自主市场。围绕这条直接赛道,还有更广的专家公司和既有厂商。Shield AI 销售任务自主基础设施,Epirus 开发用于分层短程防空的高功率微波效应器,D-Fend 围绕 RF 网络接管,Dedrone 和 Fortem 强调一体化 detect-track-identify-mitigate 技术栈,DroneShield 是公开市场中的模块化 counter-UxS 专家,Leonardo DRS、DZYNE 和 SRC 则代表成熟集成商或既有厂商替代方案。Unmanned Airspace 2026 年目录跟踪 550 多家公司和 1,000 多个项目或系统,强调 Roark 进入的是一个拥挤品类,买家可以混合专家传感器、效应器和指挥层,而不必押注一个新创业故事。美国陆军 2026 年 marketplace 语言在这里也重要,因为它暗示买家越来越偏好能嵌入盟国资质和互操作体系的已验证供应商。在这种环境下,Roark 的故事最容易被理解为一枚低摩擦探测和托管服务楔子,切入一个高端越来越由资本更雄厚的一体化供应商掌控的市场。[CP001, CP002, CP003, CP004, CP005, CP006]
| 竞争对手 | 类别 | 规模 / 融资标志 | 目标客群 | 差异化 | 相对 Roark 论点的局限 |
|---|---|---|---|---|---|
| Roark Aerospace | 直接同业 / 服务主导挑战者 | OTC 硬件和服务公开定价;私域规模说法仍有争议 | 寻求更低门槛切入的国防、公共安全和商业监测买家 | 以 DDaaS / C-UASaaS 包装,给出明确入门价和托管服务选项 | 具名客户证明偏薄,公开部署验证弱于同业 |
| Shield AI | 相邻自主平台 | 2025 年 3 月融资 $240M,估值 $5.3B | OEM、政府、机器人和国防集成商 | 争议环境中的任务自主软件和开发者工具 | 更接近自主基础设施,而非低成本探测服务 |
| Epirus | 定向能专家 | 2025 年 3 月 $250M Series D;风险融资总额超过 $550M | 分层近程防空、关键资产、军事编队 | 高功率微波效应器和空域感知 | 控制点不同;对 Roark 探测服务的直接替代性较弱 |
| D-Fend Solutions | 直接专业厂商 | 2024 年 12 月 $31M 成长融资;Motorola 于 2026 年同意以 $1.5B 收购 | 政府、公共安全、机场、关键基础设施 | RF 网络接管并安全降落,强调开放 API 和连续性 | 企业销售以报价为主,而非透明低端入门定价 |
| Dedrone / Axon | 有生态背书的直接专业厂商 | 数百家客户;现已进入 Axon 公共安全平台 | 公共安全、机场、公用事业、军方、惩教、关键基础设施 | AI 驱动的空域 C2、30+ 项集成、硬件无关技术栈 | 公开定价不够明确;取决于平台和集成范围 |
| Fortem Technologies | 集成型专业厂商 | Lockheed 2026 年投资 $25M;拥有盟友订单和关键基础设施合同验证 | 军方、基础设施、机场、世界级场馆、政府 | 集成雷达、拦截器和软件套件,并有实战捕获证明 | 可能比 Roark 入门级方案更重,也更偏项目制 |
| DroneShield | 公开市场专业厂商 | 上市纯玩家;官网称获得 34+ 个政府机构信任 | 政府、军方、固定站点和移动防护买家 | Counter-UxS 产品覆盖软件、固定站点和移动套件 | 公开市场 IR 有助于可信度,但官方目录价仍不透明 |
| Leonardo DRS | 既有厂商 / 主承包商级集成商 | 收入 $3.6B,积压订单 $8.7B,员工 7,300 人 | DoD、盟军、海事、机动和基地防御买家 | 项目级规模、分层压制能力和更广的国防关系 | 业务范围远大于 Roark;并非为超低摩擦试用优化 |
| DZYNE / SRC | 既有厂商式替代方案 | DZYNE 称已在 50+ 个国家部署;SRC 称自 2005 年以来开发 C-UAS | 军方、机场、VIP、基础设施、固定 / 移动部署 | 分层、任务就绪系统,已建立国防相关性 | Roark 式托管服务或城市覆盖打包的证据较少 |
各行比较公开规模证明、买家聚焦和差异化;没有目录价或具名客户,反映的是引用来源披露和未披露的边界。
[CP001, CP002, CP004, CP005, CP006, CP007]| 购买标准 | Roark | D-Fend | Dedrone | Fortem | DroneShield | Leonardo DRS |
|---|---|---|---|---|---|---|
| 仅探测的入门产品 | 是 — Aero-Ark 硬件和 DDaaS 已公开 | 部分 — 集成在 EnforceAir 技术栈内 | 部分 — 探测叠加更广的空域技术栈 | 否 — 作为集成系统销售 | 部分 — 模块化传感包 | 否 — 分层项目解决方案 |
| 托管或服务主导覆盖模式 | 是 — DDaaS 和快速部署已公开 | 公开信息未知 | 公开信息未知 | 公开信息未知 | 公开信息未知 | 没有公开的服务主导入门证明 |
| AI / 软件指挥层 | 是 — Plexus 和边缘 AI 说法 | 是 — RF 网络软件定义控制 | 是 — AI/ML 自主 C2 | 是 — SkyDome 管理器和 AI 雷达 | 是 — AI 驱动软件栈 | 是 — 集成 C2 和传感器融合 |
| 集成动能或压制效应器 | 声称可按授权客户缓解威胁,但公开细节有限 | 是 — 接管 / 驱离 / 干扰选项 | 是 — 通过集成技术栈提供缓解选项 | 是 — DroneHunter 拦截器 | 是 — 跨产品线的压制工具 | 是 — 动能、非动能和 DE 选项 |
| 开放架构 / 集成姿态 | 部分 — 声称具备互操作性 | 是 — 突出开放 API | 是 — 30+ 项集成且硬件无关 | 是 — 集成套件被包装为兼容性解决方案 | 部分 — 方案家族广,但官网表层 API 细节有限 | 是 — 平台无关和全域集成 |
| 具名部署或采购证明 | 公开记录偏薄 | 强 | 强 | 强 | 中等至强 | 强 |
| 机场 / 关键基础设施相关性 | 间接声称 | 明确 | 明确 | 明确 | 明确 | 明确 |
| 主承包商或生态渠道杠杆 | 公开证明低 | Motorola 之后高 | Axon 之后高 | 通过 Lockheed 和政府合同达到高 | 借公开市场和机构触达达到中等 | 通过国防主承包商关系达到很高 |
单元格刻意区分公开证据和未知项;比较的是有文件支持的能力信号,不是实验室基准测试。
[CP001, CP002, CP003, CP010, CP013, CP014]竞争重心落在两类能力兼具的供应商身上:部署证明足够广,渠道杠杆也更强。Roark 的差异化目前主要在入门价格和服务打包。
坐标轴是序数估计:x 近似表示部署证明 / 信任,y 近似表示渠道杠杆 / 集成广度。分数是有证据支撑的排序,不是实验室测量。
[CP003, CP005, CP007, CP009, CP011, CP012]3.2 能力、定价与买家匹配
同行能力分化出清晰的控制点。Roark 的公开材料强调易接入探测、模块化部署和服务化覆盖;2023 年 DRONELIFE 与 sUAS News 文章对入门价格披露得异常明确,包括 £2,150 到 £4,450 的硬件型号、每用户每月 £95 的 DDaaS,以及每天 £185 的快速部署。相比之下,大多数披露更充分的同行不会在官方表面发布标价,而是销售一体化或报价制系统。D-Fend 的 EnforceAir 页面强调 RF-cyber 探测、跟踪、识别、安全降落和开放 API 集成。Dedrone 营销一套硬件无关、AI 驱动的空域安全平台,拥有超过 30 个已验证集成;Fortem 则认为,多供应商部署常因组件无法干净集成而失败,因此销售端到端 SkyDome 套件。DroneShield 为政府客户打包固定场站、单兵、移动中和软件选项。Leonardo DRS、DZYNE 和 SRC 把问题描述为跨军事、关键基础设施、机场、海上或远征场景的分层防护。商业含义是,Roark 的公开价格透明度和托管服务叙事,可能会吸引需要快速探测覆盖、活动部署或低资本开支试用路径的客户;但主承包商级采购买家会把这套供给与那些已经发布更深集成、部署和任务验证性能证据的供应商相比较。[CP001, CP002, CP003, CP009, CP010, CP013]
| 供应商 | 公开价格或合同模式 | 包含能力 | 未知项 / 折扣注意事项 | 含义 |
|---|---|---|---|---|
| Roark Aero-Ark WiRED | £3,200(不含 VAT)起 | 前向散射无人机探测 | 实际成交价、支持和部署条款未知 | 面向探测主导客户的低摩擦切入点 |
| Roark Aero-Ark CRI | £2,150(不含 VAT)起 | 传统 RF 扫描叠加识别功能 | 没有公开折扣或批量价目表 | 让 Roark 对标低端探测硬件和试点 |
| Roark WiRED + CRI 组合 | £4,450(不含 VAT)起 | 前向散射和 CRI 组合包 | 没有企业 ASP 或经常性附加率证明 | 暗示上行销售路径,但不是完整主承包商级系统价格 |
| Roark DDaaS | 每用户每月 £95(不含 VAT) | 无需客户部署硬件即可获得覆盖 | 覆盖足迹和 SLA 条款未经独立验证 | 如果覆盖真实,服务模式可缩短试点采用周期 |
| Roark Rapid Deploy | 每日 £185(不含 VAT)起 | 临时硬件部署和服务支持 | 时长、人员配置和支持假设不清楚 | 可作为活动或紧急站点覆盖的切入口 |
| D-Fend, Dedrone, Fortem, DroneShield | 引用的官方页面没有公开目录价 | 集成探测、跟踪、C2 和缓解技术栈 | 可能是基于报价的企业或政府采购 | 定价不透明会拖慢自助采用,但也指向更高价值的项目销售 |
| Leonardo DRS, DZYNE, SRC | 项目或解决方案销售;引用页面没有目录价 | 分层系统、集成和任务包 | 定价绑定项目范围、平台和客户 | 竞争点是信任和集成,而非易获得的入门价 |
在引用集合中,只有 Roark 给出了清晰的公开入门价;对于同业,可辩护的比较对象是合同模式和打包方式,而不是臆造的美元估算。
[CP001, CP002, CP003, CP034, CP035]| 供应商 | 分销渠道 | 信任信号 | 部署或规模证明 | 对 Roark 的含义 |
|---|---|---|---|---|
| Roark | 创始人主导的直接叙事;本章证据中没有公开锚定大型渠道伙伴 | 公开信任表层低至中等 | 规模和部署说法仍只得到部分验证 | 必须逐个客户赢得信任 |
| D-Fend | Motorola 公共安全和企业网络 | 签署收购并披露收入后达到高 | 在 30+ 个国家完成数千次部署 | 渠道杠杆提高客户信心,也降低从小厂商切换出去的摩擦 |
| Dedrone | Axon 公共安全生态和伙伴网络 | Axon 集成和客户历史之后达到高 | 数百家客户和广泛垂直行业证明 | 公共安全买家可以采购更大的工作流平台 |
| Fortem | Lockheed 生态叠加直接国防和 DHS 关系 | 凭运营证明和主承包商背书达到高 | 盟友订单、关键基础设施合同、CBP、Ukraine、就职典礼安保 | 强证明让 Fortem 在集成部署中难以被替代 |
| DroneShield | 公开市场可见度和机构关系 | 中等至高 | 34+ 个政府机构和持续投资者更新 | 透明度有助于偏好上市公司披露的买家 |
| Leonardo DRS | 主承包商和 DoD 既有地位 | 很高 | 收入、积压订单、员工规模和已部署 C-UAS 项目 | 在大型项目中,既有厂商信任会挤出创业公司挑战者 |
| DZYNE / SRC | 已建立国防和任务驱动部署历史 | 中等至高 | DZYNE 覆盖 50+ 个国家;SRC 有陆军 / 空军经历和长期历史 | 替代集成商会减少 Roark 唯一适配的场景数量 |
本表比较的是信任姿态,而非原始产品性能;在国防和公共安全中,渠道和采购证明往往与传感器质量同样重要。
[CP011, CP012, CP015, CP018, CP019, CP020]Roark 手里最清楚的是低摩擦定价楔子;同行则控制更难复制的层,包括 RF 接管、定向能效应器、集成拦截器或主承包商级部署信任。
[CP001, CP002, CP010, CP013, CP014, CP016]3.3 分销、切换成本与信任
counter-UAS 的竞争耐久性越来越取决于渠道和证据,而不只是传感器。Motorola 计划以 $1.5 billion 收购 D-Fend,Axon 收购 Dedrone,把两家专家公司拉入更大的公共安全生态。Lockheed Martin 投资 Fortem,以及 Fortem 相关的关键基础设施和盟国订单公告,显示主承包关系如何加快部署、验证和制造规模。Leonardo DRS 起点大得多——收入 $3.6 billion、积压订单 $8.7 billion、员工 7,300 人——再把 counter-UAS 叠进更广的防务项目。这些事实重要,因为这个市场的切换成本真实存在,但并非绝对。D-Fend 和 Dedrone 都强调开放 API 或硬件无关集成,意味着买家可以组装多供应商技术栈;与此同时,Fortem 声称部署失败常源于兼容性缺口,也说明一旦端到端系统经过验证并落地,就会变得黏性很强。对 Roark 来说,现实问题是公开证据中具名客户、重复部署或采购胜利仍远少于这些竞争对手披露的内容。因此,Resilience 的怀疑正好落在竞争章节内部:即便 Roark 有聪明的服务模式,防务、公共安全、机场和关键基础设施买家也更可能信任有部署文档、交易验证和更广渠道支持的供应商。[CP010, CP011, CP012, CP013, CP014, CP015]
| Roark 护城河主张 | 威胁 | 严重性 | 证据 | 缓解措施 / 尽调问题 |
|---|---|---|---|---|
| 低摩擦公开定价 | 专业厂商和既有厂商可以压低低端探测经济性,或把探测打包进更大的技术栈 | 高 | Roark 对入门价的透明度异常高,而同业销售集成系统 | 要求提供赢单 / 输单分析,证明客户选择 Roark 不只是因为低前期成本 |
| DDaaS / 托管覆盖模式 | 如果客户可以多归属或之后采购报价型方案,覆盖主张未必能转化为可防守的经常性收入 | 高 | Roark 官方和行业来源描述了服务打包,但缺少公开续约数据 | 要求提供留存、被监测站点数量,以及按城市 / 客户类型划分的利用率 |
| 模块化、可互操作架构 | 开放 API 和硬件无关竞争技术栈会削弱硬锁定 | 中 | D-Fend 和 Dedrone 公开强调集成,Fortem 则主张集成质量才是真正护城河 | 验证 Roark API 深度、伙伴生态,以及客户更换传感器时会坏掉什么 |
| 可负担的便携硬件 | 相比已验证的压制或指挥层,便携探测更容易商品化 | 高 | 专业厂商和既有厂商已经覆盖便携、固定、移动和软件主导方案 | 展示从探测设备走向更高价值数据、C2 或缓解收入的路线图 |
| 进入试点和活动的速度 | 有主承包商背书的供应商如今也有自己的快速部署路径,并在关键基础设施或高曝光活动中拿到证明 | 中高 | Fortem、D-Fend、Dedrone 和 DroneShield 发布的部署或渠道证明都强于 Roark | 提供具名试点转合同案例,以及快速部署的毛利数据 |
| 叙事差异化 | 即使产品可用,围绕 Roark 规模和融资的负面审视也会削弱买家信任 | 高 | Resilience 质疑验证情况,而同业交易和采购证明更强 | 为尽调室准备客户推荐、部署文件和融资 / 估值证据 |
严重性反映的是可承保性,而非技术优雅度;最大风险是信任、商品化和渠道不对称,而不是缺少市场需求。
[CP003, CP010, CP011, CP012, CP015, CP016]这些核心指标强调:Roark 的竞争对手目前公开披露的资本、采购证明和生态杠杆,远多于 Roark。
[CP002, CP005, CP007, CP011, CP016, CP017]3.4 护城河耐久性与替代风险
证据最多指向一条混合型、偏弱的护城河。Roark 最有支撑的优势是进入门槛低:公开价位、便携硬件和服务打包比许多同行更清晰。这有助于赢得试点、活动覆盖,或不想购买全栈的客户。但同一组证据也显示,低端探测和托管监测是技术栈中最容易商品化的部分。资本更雄厚的供应商已经拥有更难复制的护城河:Shield AI 在自主基础设施,Epirus 在定向能效应器,D-Fend 在 RF 接管,Fortem 在已验证拦截运营和一体化雷达,Dedrone 在 AI 指挥控制及生态杠杆,DroneShield 在公开市场融资渠道和政府专业化,Leonardo DRS 或 SRC 则在项目级信任。Roark 还面临内部或主承包商主导集成的替代风险:买家可以组合专家传感器、干扰器和指挥软件,而不是标准化采用一家年轻托管服务供应商。因此,本章承销结论偏谨慎:Roark 有可能切出一个探测主导的服务细分,但除非管理层能证明具名客户、续约行为、集成防御性,以及从低成本探测走向更高价值指挥或效应层的路径,否则公司仍会同时暴露在来自下方的商品化和来自上方的一体化既有厂商替代之下。[CP003, CP004, CP006, CP008, CP010, CP011]
3.5 附件
04财务
4.1 收入模式、定价与公开牵引力
公开记录支持一套混合变现模式,但不能给出干净收入结构。Roark 自己的归档首页使用 "Drone Detection as a Service | C-UAS as a Service" 标签,行业报道则同时描述硬件销售和经常性服务供给。sUAS News 发布了标价:Aero-Ark CRI 从 £2,150 ex VAT 起,Aero-Ark WiRED 为 £3,200 ex VAT,组合设备 £4,450 ex VAT,Detection as a Service 从每用户每月 £95 ex VAT 起,Rapid Deploy 每天 £185。DroneXL 另称 OTC 系统税前起价 £4,500,并将其与 Plexus 平台访问绑定;DroneLife 称 Roark 也提供已安装覆盖作为服务。这足以说明,到 2025 年初 Roark 已在市场销售硬件加服务的混合包,但不足以推断实际 ASP、折扣行为、订阅渗透率、续约画像或合同期限。唯一官方收入披露仍是 2024 年账目,披露营业额 £36.9 million,并明确把 turnover 定义为英国和美国实体的集团总销售额。这项会计政策很关键:它指向真实销售活动,但也意味着外部投资人无法通过备案拆分英国法律实体收入与更广集团活动,也无法区分硬件、订阅、部署和数据服务各自的贡献。[CI001, CI011, CI022, CI024, CI025, CI026]
| 收入流 | 机制 | 单位 | 当前价值 / 状态 | 质量 | 尽调问题 |
|---|---|---|---|---|---|
| Aero-Ark CRI 硬件 | 直接设备销售 | 每台 | £2,150(不含 VAT)起 | 中 — 仅有公开目录价 | 要求提供实际 ASP、销量和标准折扣区间 |
| Aero-Ark WiRED 硬件 | 直接设备销售 | 每台 | £3,200(不含 VAT)起 | 中 — 仅有公开目录价 | 要求提供 BOM、安装成本和单台毛利 |
| WiRED + CRI 组合包 | 捆绑硬件销售 | 每台 | £4,450(不含 VAT)起;DroneXL 引述系统价格从 £4,500(不含税)起 | 中 — 目录价获得行业报道佐证 | 要求提供附加率、组合包结构和渠道折扣 |
| 探测即服务 | 经常性监测覆盖服务 | 每用户 / 月 | 起价为每月 £95(不含 VAT) | 中——公司公布的标价,范围限制不明 | 要求提供活跃付费用户数、合同期限、流失率和支持成本 |
| Rapid Deploy | 短期部署服务 | 每日 | 起价为每日 £185(不含 VAT) | 中——仅有标价 | 要求提供利用率、机组 / 人工假设,以及转化为经常性合同的情况 |
| WAMI / HSI / Plexus 关联服务 | 数据和监控服务包 | Unknown | 已披露能力;没有公开独立价格 | 低——变现规模未公开量化 | 要求提供价格卡、贡献毛利率和营业额占比 |
表格把公开标价和实际经济性分开;这些数字都不能证明已签收入结构或毛利率。
[CI022, CI024, CI025, CI026, CI027]| 产品 | 价格 / 合同模式 | 标价 vs 实现价格 | 确认 / 质量问题 | 来源 |
|---|---|---|---|---|
| CRI 硬件 | 标价 £2,150,不含 VAT | 仅标价 | 没有批量折扣或渠道毛利证据 | sUAS News |
| WiRED 硬件 | 标价 £3,200,不含 VAT | 仅标价 | 没有服务附加或安装费处理证据 | sUAS News |
| 组合硬件包 | 标价 £4,450,不含 VAT / DroneXL 称约 £4,500(不含税) | 仅标价 | 两家行业媒体大体一致,但都没有展示实际合同价值 | sUAS News;DroneXL |
| Detection as a Service(探测服务) | 每用户 / 月 £95,不含 VAT | 仅标价 | 未知按席位、按站点,还是按企业最低消费计费 | sUAS News |
| Rapid Deploy | 每日 £185,不含 VAT | 仅标价 | 未知动员、操作员或差旅转嫁收费 | sUAS News |
| 备案营业额 | FY2024 集团总销售额 £36.9M | 已实现收入数字 | 账目汇总英国和美国集团销售额,遮蔽产品线结构和实体边界 | 2024 年账目 |
定价证据大多停留在市场宣传层面;公开记录里唯一的实际收入数字是 FY2024 总营业额。
[CI001, CI011, CI024, CI025]公开证据指向一个混合模型:先从硬件或服务部署切入,再尝试把监测和响应覆盖变现。
[CI001, CI011, CI022, CI024, CI025, CI027]4.2 成本结构、营运资本与单位经济
2024 年备案异常有价值,因为它显示 Roark 报告的收入规模尚未转化为可公开验证的优质单位经济。公司披露营业额 £36.9 million、销售成本 £49.3 million,因此毛利为负 £12.4 million,随后还有 £3.1 million 管理费用和 £1.07 million 利息成本。资产负债表上,年末现金只有 £6.1 million,存货为 £13.6 million,应收款达到 £87.1 million,其中 £36.1 million 在一年以上到期。这些数字指向一家承担实质履约、部署和回款负担的企业,而不是干净的软件单一模式。它们也意味着,公开可见的营运资本消耗大幅超过流动现金。因此,按传统 SaaS 口径讨论 CAC/payback 不可能,甚至可能误导:Roark 看起来把硬件、现场部署和持续监测服务组合在一起。证据能支持的最佳判断更窄。Roark 确有真实商业活动,但唯一披露年度显示负毛利、显著应收款堆积、存货敞口,以及很可能超出普通经营现金生成能力的融资需求。任何假设其具备有吸引力的软件式利润率的投资备忘录,都在远远超出证据投射。[CI001, CI002, CI003, CI004, CI005, CI006]
| 指标 | 数值 / null | 置信度 | 重要性 | 尽调要求 |
|---|---|---|---|---|
| FY2024 营业额 | £36,906,898 | 中 | 说明 2025 年末融资叙事之前,公司已有真实商业活动 | 按产品线、地区和客户细分进行核对 |
| FY2024 毛利润 | -£12,418,074 | 中 | 唯一披露年度里,报告收入未覆盖直接交付成本 | 提供产品级毛利率和服务交付成本桥表 |
| FY2024 税前结果 | -£16,561,770 | 中 | 披露账目确认公司仍在亏损 | 提供 EBITDA 和调整后烧钱桥表 |
| 银行现金(2024 年 11 月 30 日) | £6,109,227 | 中 | 后续融资事件前的流动性基础 | 提供当前现金、受限现金和最低运营现金需求 |
| 库存 / 营业额比率 | 0.3679 | 中 | 对一家有时被描述成软件的公司来说,库存负担很重 | 提供库存账龄和减记政策 |
| 应收账款 / 营业额比率 | 2.3591 | 中 | 相比年销售额,应收款回收看起来非常沉重 | 提供 AR 账龄、前几大债务人名单和付款条款 |
| 债券义务 | £16.42M,6.5% APR,2029 年 3 月到期 | 中 | 债务服务影响跑道和资金用途优先级 | 提供债券持有人身份、契约和担保包 |
| 披露平均员工数 | 仅英国 49 人 | 中 | 公开员工数没有覆盖全部用工基础或成本结构 | 按职能提供总 FTE、承包商和国际薪酬 |
| CAC / 回本期 | 低 | 无法评估 go-to-market 效率或渠道经济性 | 按产品提供渠道组合、CAC、转化率和回本期 | |
| 积压订单 / 利用率 / 续约 | 低 | 需要用来判断收入韧性和服务产能利用率 | 提供已签积压订单、部署利用率和续约 cohort |
数字比率是作者基于 2024 年备案账目计算,应作为诊断信号,而非独立承保指标。
[CI001, CI002, CI004, CI005, CI006, CI007]公开记录暗示,在出现任何剩余毛贡献之前,收入会先被硬件、部署、营运资本和融资负担拉低。
这座桥是定性分析,因为 Roark 没有披露产品线毛利率、人工负担、托管成本或续约 cohort。
[CI002, CI003, CI005, CI006, CI007, CI009]公开来源显示,这家公司比纯软件国防公司资本强度更高,现金转化风险也更大。
矩阵采用定性评分,因为产品线成本会计和现金转化数据并不公开。
[CI005, CI006, CI007, CI009, CI034, CI035]4.3 资本形成与资本充足性
2025 年末和 2026 年初备案确认,Roark 在首份账目之后没有停留在一个小型、资本单薄的壳里。Companies House 显示,2025 年 3 月 SH01 配发 81 million 股普通股和 19 million 股优先股,每股 £1.25,现金认购;优先股带有 full-ratchet 反稀释和优先分红权。随后,2025 年 12 月 15 日备案显示 131 股 B 普通股,现金实缴 £156,963,450;接着 12 月 23 日和 12 月 28 日的资本重组备案大幅改变股份数量和面值;2026 年 4 月备案又以极低的 USD 1,390 增发 20,079 股 B 普通股。这些都是真实融资活动加资本表工程的强信号,但不等同于完全透明的轮次公告。Nordic9 报道了 $210 million Series B、$1.6 billion 估值和 $260 million 累计投资,但标注交易待验证。Tracxn 将 Roark 列为 Series B 公司,已融资 $287 million。Resilience 调查称,公司流传了同样的 $210 million 和 $1.6 billion 标题,但唯一披露的具名交易对手是一个 “US-based investment bank via an SPV”。承销结论是:资本形成显然真实,但当前在手现金、月度 burn、资金用途计划,以及精确投资人权利包仍明显披露不足。[CI009, CI013, CI014, CI015, CI016, CI017]
| 项目 | 公开证据 | 当前数值 / 状态 | 含义 | 尽调要求 |
|---|---|---|---|---|
| 手头现金 | FY2024 备案账目 | 2024 年 11 月 30 日为 £6.1M | 仅为历史数据;不能对 Series B 后流动性给出确定结论 | 提供当前现金和月度净烧钱 |
| 月度烧钱 | 无公开披露 | 没有该数据就无法计算跑道 | 提供月度净烧钱和季度现金桥表 | |
| 跑道月数 | 无公开披露 | 无法测试回款变慢或订单推迟时的生存能力 | 提供基准 / 下行 / 无新增资本情景下的跑道 | |
| 计划资金用途 | 无公开开源披露 | 无法区分增长 capex、运营支持或债务服务 | 提供董事会批准的资金用途计划 | |
| 下一轮触发条件 | 无公开开源披露 | 未知下一轮融资取决于收入规模、出口,还是采购胜利 | 提供现金触发、契约触发和运营里程碑 | |
| 债务 / 项目融资义务 | 账目披露 £16.42M 债券,6.5% APR,2029 年 3 月到期,另有融资租赁负债 | 已在公开记录中出现 | 对现金流的非股权索偿已经存在 | 提供完整债务明细和贷款人权利 |
| 2025 年 3 月一级资本信号 | 日期为 2025 年 3 月 13 日的 SH01 | 配发 100M 股换取现金;名义资本合计 £125M | 指向重大的早期资本化和结构化优先权条款 | 提供认购方名单和定价轮文件 |
| 2025 年末融资信号 | 日期为 2025 年 12 月 15 日的 SH01,加上 Nordic9 / Resilience / Tracxn 报道 | 131B 普通股对应备案现金 £156.96345M;外部来源称一轮 $210M | 公司筹得了重大资本,但确切轮次边界和投资者组合仍不清楚 | 提供已签 cap table、认购文件,以及 FX / 对账桥表 |
本表区分已备案事项和仍依赖管理层披露的事项;公开证据证明有融资活动,不证明当前跑道。
[CI005, CI009, CI014, CI015, CI016, CI017]最干净的公开区间围绕已披露价格点、历史流动性和外部报道融资总额,而不是当前 burn 或 runway。
金额保留来源原生单位;公开记录给出点估计而非真实区间时,low=high。
[CI001, CI005, CI024, CI025, CI029, CI030]4.4 披露缺口与财务结论
Roark 的财务证据集在方向上强于其公开叙事,但仍低于后期防务科技承销所需标准。多家数据库跟踪该公司,但 PitchBook 基本被门禁挡住,Crunchbase 在抓取时被阻断,两条路径都没有在开放网络暴露可用投资人名单或当前运营指标。roark-aerospace.com 根域也不再提供可工作的页面,让归档快照和第三方报道承担了过多尽调负担。这很重要,因为披露更充分的同行展示了更强披露应是什么样子。Shield AI 公开宣布一轮 $240 million 融资、估值 $5.3 billion;Motorola 称 D-Fend 将以 $1.5 billion 被收购,预计 2026 年收入 $185 million;Fortem 点名了一笔由 Lockheed 领投的 $25 million tranche,以及具体后续订单和合同;Axon 和 Dedrone 描述了具体部署和最终用户类别。相比之下,Roark 仍缺少具名客户、披露积压订单、按产品线拆分的毛利、CAC/payback、当前现金、当前 burn,以及从备案配股到独角兽标题的透明桥梁。因此,财务结论是混合的,但只有拿到管理层会议室证据才具备可投性:收入质量已足以支持继续工作,利润率路径仍属推测,2025 年末融资后的资本充足性无法仅凭公开证据评估。[CI028, CI040, CI041, CI042, CI043, CI044]
| 缺失的私有指标 | 对承保的影响 | 精确尽调路径 |
|---|---|---|
| 硬件 vs 经常性服务的收入结构 | 无法判断收入韧性或利润率画像 | 要求提供经审计的分部收入桥表,以及 2024 年 12 月以来的月度结构 |
| 实现价格和折扣 | 标价可能高估实际变现和毛利率 | 要求提供前 20 大合同,包含标价、折扣、期限和支持义务 |
| 客户集中度和积压订单 | 无法衡量订单可见性或少数交易对手暴露 | 要求按阶段提供已签积压订单和头部客户收入计划 |
| 当前现金和月度烧钱 | 即使有大额融资说法,跑道仍无法计算 | 要求提供财资仪表盘和 12 个月现金桥表 |
| 投资者名单和权利包 | 无法建模清算瀑布、治理或稀释风险 | 要求提供认购协议、股东协议和按类别 cap table |
| 按产品线划分的毛利率 | 汇总毛利为负可能掩盖不同收入流的经济性差异 | 要求提供硬件、服务和数据毛利率桥表 |
| CAC / 回本期 / 渠道经济性 | 无法测试销售效率或分销商经济性 | 要求按销售动作提供 pipeline 转化、CAC 和回本期 |
| AR 账龄和库存周转 | 应收款和库存水平可能意味着营运资本承压或里程碑计费 | 要求提供账龄桶、坏账历史和库存账龄表 |
这些缺口是关键阻塞点,公开网络证据还不能据此组成完整财务承保包。
[CI038, CI039, CI045, CI048]4.5 附件
05产品与技术
5.1 产品表面与客户工作流
最强证据显示,Roark 销售的是服务主导的 counter-UAS 套餐,而不是单一探测器 SKU。归档首页把公司定位为 "Drone Detection as a Service | C-UAS as a Service";两家独立无人机行业媒体描述了 Aero-Ark 硬件、DDaaS 席位和 Rapid Deploy 项目,并给出明确公开价格。跨这些来源看,核心商业动作一致:客户要么购买或临时部署 Aero-Ark 硬件,要么购买监测覆盖服务,然后通过 Plexus 平台接收空域警报。同一批来源还把 WAMI 和高光谱成像放在空域产品旁边,说明 Roark 想销售的是更广的监测技术栈,而不是纯 RF 探测。薄弱处仍是部署证明。证据能显示产品包装、形态和商业意图,但还没有显示具名运营场站、实时 DDaaS 城市列表,或足以确认网络中多少处于生产、多少处于市场化路线图的客户结果。[CE001, CE004, CE010, CE014, CE016, CE019]
| 模块 / 产品线 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| Aero-Ark WiRED | 基地 / 场站安保操作员 | 已公开营销并标价 | 前向散射感知、卫星辅助信号输入、FPGA 支撑的边缘 AI 主张 | 需要独立现场测试、误报数据和 BOM / 供应商细节 |
| Aero-Ark CRI | 安保 / 空域操作员 | 已公开营销并标价 | RF 扫描加 2.4/5.8 GHz 频段的数据包分析,并声称覆盖 SATCOM / LTE / LoRa | 需要验证协议覆盖、解密广度和真实场景距离 |
| WiRED + CRI 组合 | 需要更广覆盖的一体化买家 | 已公开标价的捆绑包 | 在一个包里组合多种感知模式,支持便携或固定部署 | 需要安装模式、附加率和支持负担 |
| Detection as a Service(探测服务) | 想要覆盖但不自有硬件的客户 | 已商业化提供,范围仍不透明 | 低摩擦月度入口和基于网络的覆盖叙事 | 需要具名覆盖城市、SLA 条款和活跃客户数 |
| Rapid Deploy | 活动、事件或临时场站用户 | 已商业化提供的服务 | 日费部署降低前期 capex,并支持短通知任务 | 需要机组模式、利用率证明和动员经济性 |
| Plexus + WAMI / HSI | 需要更广监视语境的操作员 | 已声称能力;定价和证明偏薄 | 把空域警报与地面影像、合作飞手网络主张组合起来 | 需要产品文档、数据权利模式、留存政策和具名用例 |
各行把可见产品表面与仍未验证的部署和性能主张分开。
[CE001, CE004, CE010, CE014, CE017, CE019]| 用户任务 | 当前工作流 | 公司解决方案 | 可衡量收益 | 限制 |
|---|---|---|---|---|
| 持续场站监控 | 安装自有探测器,或签约分层空域监控 | 绑定 Plexus 的 DDaaS 席位或 Aero-Ark 硬件 | 与全定制系统相比,公开标价降低进入成本 | 没有 uptime、误报或具名场站的公开证明 |
| 城市空域探测 | 绕开雷达阴影和密集 RF 杂波 | 带可编程超表面和喇叭收发器的城市 Aero-Ark 变体 | 主张做了城市特定适配,而非一套通用雷达配置 | 只有行业报道描述技术路径 |
| 活动或 VIP 保护 | 短通知调入临时 C-UAS 套件和操作员 | Rapid Deploy 日费服务 | 避免大额前期采购,支持临时运营 | 没有公开部署案例研究或机组 / 服务指标 |
| 综合监视图 | 将空域警报与更广地面语境融合 | Plexus 平台加 WAMI / HSI feed | 可能比纯 RF 系统提供更丰富的态势感知 | 没有公开产品手册、隐私文档或操作员截图 |
| 第三方飞手网络扩张 | 招募外部无人机飞手,扩大成像或覆盖范围 | Part 107 赞助、免费相机和 Plexus 访问 | 数据采集扩张速度可能快于完全自有机队 | 覆盖质量控制和数据权利结构未公开 |
收益来自公开信息或基于定价与工作流主张推断;没有一项经过客户 ROI 指标验证。
[CE014, CE018, CE020, CE021, CE022, CE024]Roark 公开讲述的架构把 RF 传感、边缘处理、Plexus 编排、服务交付,以及可选成像或处置扩展层层叠起来。
[CE001, CE005, CE011, CE014, CE017, CE019]对外销售的运营流程从场站或事件需求开始,经过 Aero-Ark 感知和边缘处理,最后落到 Plexus 驱动的监测,并可选叠加成像或缓解动作。
[CE004, CE008, CE014, CE019, CE020, CE021]5.2 技术架构与差异化主张
Roark 的架构故事是可信的,但仍主要停留在营销和行业媒体层面,而不是文档级技术证明。Dronelife、sUAS News 和 DroneXL 独立指向同一个宽泛模式:WiRED 被描述为前向散射或卫星辅助感知层;CRI 被描述为 RF 扫描和数据包分析层;DroneXL 又加入 ADS-B、Remote ID、DJI 数据包解密和被动雷达;多家媒体称分类或协调在设备端或网络边缘完成,再通过 Plexus 暴露数据。这套多模态技术栈与成熟 C-UAS 供应商描述分层探测的方式方向一致,但 Roark 的公开证据明显薄于同行文档。SRC、Leonardo DRS、D-Fend、DZYNE、Dedrone 和 Fortem 都给出更清晰的传感器融合、部署套件、开放集成或完整 detect-track-defeat 链条说明。Roark 仍可能有差异化的低成本楔子,尤其如果其服务价格真实;但开放记录没有用测试数据、专利或客户验证支撑前向散射性能、数据包解密覆盖、可编程超表面有效性,或任何专有大语言模型优势。[CE005, CE006, CE007, CE008, CE009, CE011]
| 层级 / 组件 | 作用 | 依赖 | 风险 |
|---|---|---|---|
| 前向散射 / 卫星辅助感知 | 声称为 WiRED 提供远距离或抗拓扑变化探测 | 需要稳定 RF 环境、卫星信号输入和调校模型 | 没有距离、杂波处理或可重复性的公开验证 |
| RF 扫描 + 数据包分析 | 探测并识别无人机控制信号和数据包 | 协议库、解密能力和持续更新的无人机固件知识 | 若缺少持续更新,覆盖和解密主张可能很快老化 |
| ADS-B / Remote ID / 被动雷达输入 | 在单一 RF 方法之外补充更广态势图 | 第三方广播、Remote ID 采用度和被动感知质量 | 表现依赖环境,且主要由行业媒体描述 |
| 设备端计算 / FPGA 边缘层 | 在网络协同之前或同时运行分类和本地处理 | 专用硬件、模型维护和软件更新纪律 | 没有延迟、吞吐或模型准确率的公开基准 |
| Plexus 平台 | 暴露警报、数据融合和操作员工作流 | 可靠网络、UI/UX、存储和集成模式 | 检索来源中没有公开安全、API 或韧性文档 |
| 可选干扰 / 反制路径 | 支持获许可的击败动作和手动或自动响应 | 法律授权、安全运营原则,以及合作伙伴或硬件效应器 | 没有公开证据显示 Roark 在哪里获许可或经测试可执行缓解 |
架构行混合了公司声称的层级和推断的运营依赖;没有支撑的技术证明被标为风险,而不是当作事实。
[CE005, CE006, CE007, CE008, CE009, CE011]Roark 公开的产品论点靠信号接入、协议维护、边缘硬件、集成、合法缓解权利和支持组织撑着,但公开材料对这些环节的记录仍不充分。
[CE008, CE009, CE020, CE026, CE028, CE029]5.3 部署成熟度、支持与路线图
Roark 有足够公开信号证明产品已越过概念阶段,但不足以证明成熟现场执行。带日期的里程碑从 2023 年末发布文章开始,随后是 2023 年 12 月价格发布,再到 2025 年 1 月 OTC 扩展文章,后者拓宽了技术故事,并把所有探测系统接入 Plexus。DroneXL 还描述了 30 个美国和 12 个英国 WAMI/HSI 城市,以及一个招募具备认证、相机和平台访问权限的第三方无人机操作员的试点项目。这些都是有意义的商业化信号,因为它们暗示运营、支持和某种伙伴或承包商网络。即便如此,公开成熟度证明远弱于路线图宽度。已检索到的 Roark 来源没有发布 uptime、误报率、探测成功率、具名部署引用,或客户运营的集成故事。已检索证据集中唯一劳动力数据点是 2024 年账目备案显示英国平均员工 49 人;这不能排除承包商或美国团队,但会让全球服务覆盖、WAMI 扩张和复杂技术支持的组合更难仅凭开放来源承销。[CE010, CE014, CE016, CE019, CE020, CE021]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源 |
|---|---|---|---|---|
| 2023-11 发布界面 | Aero-Ark 以 DDaaS 形态亮相,叙事围绕前向散射和数据包嗅探 | 已宣布 | 说明公司一开始就走服务带动的定位,而不是后期转向 | DroneLife |
| 2023-12 商业包装 | WiRED、CRI、DDaaS 和 Rapid Deploy 公开标价 | 已宣布 / 已营销 | Roark 在该品类入门价格上异常透明 | sUAS News |
| 2025-01 OTC 扩张 | 标准版和城市版 Aero-Ark 以便携和哨兵配置对外营销 | 已宣布 / 已营销 | 暗示更大的上市野心,SKU 也多于 2023 年发布文章披露的范围 | DroneXL |
| 2025-01 成像网络扩张 | WAMI / HSI 覆盖声称扩至美国 30 个城市和英国 12 个城市,并有试点计划支持 | 声称扩张 | 如果属实,Roark 搭建的是传感器与服务混合网络,而不只是卖探测器 | DroneXL |
| 2026 尽调状态 | 公开证据仍高度依赖归档的一手网页和行业媒体报道,而不是技术文档或具名用户 | 当前现实 | 商业叙事跑在可独立验证的运营证据前面 | Roark 归档页面及独立报道 |
这些里程碑区分了已发布或已营销的界面,以及仍缺少具名客户证明的更大规模主张。
[CE004, CE010, CE014, CE016, CE019, CE021]Roark 在产品打包和定价上最成熟;独立验证性能、信任控制和支持披露的公开成熟度则低得多。
[CE002, CE014, CE017, CE020, CE021, CE029]5.4 信任控制与产品技术结论
Roark 已跨过真实、可识别、运营中且面向防务的公司最低门槛,也有连贯产品叙事,但尚未跨过许多防务和关键基础设施买家现在期待的公开信任门槛。归档官方表面展示了 NCAGE、UEID、DUNS、VAT 和法律地址;行业报道声称支持 Remote ID、授权反制选项,以及便携、哨兵和服务部署组合。同样重要的是缺失项:已检索到的 Roark 来源没有发布产品安全控制、Plexus 或 WAMI/HSI 数据的隐私处理、现场可靠性指标、正式认证,或严谨的公开集成指南。NATO 和更广 C-UAS 采购在 2026 年正转向互操作、机动部署、标准化测试,以及租赁或服务化购买模式,因此 Roark 的包装可能契合买家需求。但在管理层提供客户引用、测试报告、信任文档和硬可靠性数据之前,基于证据的结论仍是:Roark 有可信概念和产品外壳,但在运营成熟度和技术防御性上缺少足够公开证明。[CE002, CE003, CE009, CE015, CE017, CE020]
| 控制 / 信任面 | 状态 | 范围 | 缺口 |
|---|---|---|---|
| 公司 / 防务标识符 | NCAGE、UEID、DUNS、VAT 和法定地址出现在归档官方页面 | 有助于基础供应商身份和采购准入 | 不能证明产品质量、网络安全状态或涉密项目状态 |
| Remote ID / ADS-B 支持 | 行业报道中声称支持 | 扩展探测输入,并更贴合现代空域的合规话术 | 没有公开测试方法或认证显示覆盖广度 |
| 授权反制部署 | 仅在行业报道中声称面向获许可用户 | 承认缓解在法律上有条件,而非普遍可用 | 没有公开说明司法辖区、批准或安全原则 |
| 公开安全 / 隐私披露 | 除基础官方页面外,未检索到产品安全或隐私控制文档 | 这会影响 Plexus 遥测以及 WAMI / HSI 影像处理 | 需要信任中心、数据留存政策和操作员访问控制 |
| 可靠性 / 质量指标 | 未检索到 uptime、MTBF、误报或现场测试数据集 | 这些指标会支撑防务和关键基础设施买家的就绪度 | 需要正式测试报告、事件日志和支持 KPI |
本表把可见身份标记与产品安全、隐私和可靠性方面薄得多的公开记录分开。
[CE002, CE003, CE009, CE017, CE029, CE030]5.5 附件
06客户
6.1 买家类型可见,具名账户不可见
最强公开证据显示,Roark 销售给的是买家问题,而不是可公开引用的账户。归档首页和法律页确立了英国和美国运营表面,并把公司定位在 Drone Detection as a Service 和 C-UAS as a Service 周围。行业报道随后补上商业动作:Aero-Ark 硬件可直接购买,Detection as a Service 提供月度覆盖,Rapid Deploy 为临时需求提供短期使用。这种组合支持几个可能付费群体,包括政府或军方用户、公共安全和安保运营方、关键基础设施业主、临时活动买家,以及更广的监测客户。它也显示公司有意通过定价和服务打包降低接入摩擦。记录仍没有做到的是,点名一个今天明确向 Roark 付费的部委、机场、警局、场馆或企业账户。因此,分层图景在类型层面可信,但在投资人通常需要用来承销耐久性的具名客户层面偏薄。[CU001, CU002, CU003, CU006, CU008, CU010]
| 细分 | 买方 / 用户 / 付款方 | 使用场景 | 规模 | 收入 / 战略价值 | 缺口 |
|---|---|---|---|---|---|
| 政府 / 军方 | 设施或部队操作员 / 安保人员 / 国防预算负责人 | 基地或作业区域的探测与响应 | 定位和 2026 年市场环境支持这一判断,但没有具名 Roark 账户 | 采购渠道一旦验证并放量,可能成为价值最高的细分 | 没有具名买方或正式项目证据 |
| 公共安全 / 国土安全 | 警方或一线响应人员 / 操作员 / 拨款或机构预算负责人 | 围绕市政资产、事件或活动探测未经授权无人机 | DHS/FAA 买方材料和 Roark 服务包装可以支撑 | 可能受益于拨款和低摩擦服务采纳 | 没有具名警方、机场或公共安全部署 |
| 关键基础设施 | 安保或运营负责人 / 监控人员 / 设施预算负责人 | 持续监控港口、公用事业、工业资产或园区 | 产品叙事可以支撑,但缺少具名账户 | 如果合同续约,持续监控具备战略价值 | 没有公开基础设施标杆客户或 SLA 证据 |
| 临时活动安保 | 场馆或事件安保负责人 / 外包操作员 / 活动预算负责人 | 活动或临时威胁期间的短期快速部署 | Rapid Deploy 模式直接支撑 | 可能缩短销售周期,并形成可背书的短期胜单 | 没有具名活动部署或重复预订证据 |
| 城市监控 / 更广域监视 | 覆盖网络管理者 / 飞手和分析师 / 服务或数据预算负责人 | DDaaS 加 WAMI/HSI 城市覆盖,并有第三方试点支持 | 城市数量主张和试点计划报道可以支撑 | 如果真实,可能为托管服务追加销售铺出宽阔入口 | 没有城市清单、具名付费客户名单或使用率指标 |
各行描述的是可由证据支撑的买方画像,而不是具名客户;公开记录尚未识别出已确认的 Roark 账户。
[CU001, CU002, CU006, CU007, CU010, CU012]勾勒 Roark 似乎瞄准的买方问题,从事件驱动部署到持续联网覆盖。
这张旅程图综合了公开来源可支撑的买方问题;它不是 Roark 披露的内部销售流程。
[CU001, CU003, CU006, CU007, CU012, CU013]6.2 采用主张存在,但具名客户层的证明质量下降
Roark 在公开场合提出了广泛采用主张。Dronelife 把 Aero-Ark 描述为一项服务,设备安装在全球 70 个城市。sUAS News 称该服务范围有限,但目标是在 2024 年底前覆盖 100 个城市。DroneXL 描述了覆盖 30 个美国城市和 12 个英国城市的 WAMI 与高光谱覆盖,以及一个招募具备认证、相机和平台访问权限的第三方无人机飞手的项目。Tracxn 又加入更大的主张,称已部署超过 55,000 台设备、拥有超过 3,000 名最终用户。这些数据点暗示商业野心和某种运营足迹,但不等同于具名生产证明。保留下来的 Roark 来源没有一个点名客户,也没有区分试点、试用和生产部署。竞争对手页面凸显了差异:Dedrone 披露具名案例研究和广泛受保护场站数量,Fortem 则点名具体政府和活动部署。因此,基于证据的结论很窄但重要:Roark 可能有真实用户,但开放网络没有提供足够强的公开客户引用集,无法像对照来源那样验证账户质量、成熟度或可重复性。[CU004, CU005, CU006, CU007, CU008, CU009]
| 指标 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| DDaaS 已安装城市主张 | 全球 70 个城市 | 2023 | DRONELIFE | 中 | 暗示早期托管服务足迹已超出一次性硬件销售 | 没有城市清单、付费账户或活跃使用情况 |
| DDaaS 覆盖目标 | 2024 年底前全球 100 个城市 | 2023 | sUAS News | 低到中 | 说明公司有广域服务覆盖野心 | 没有确认目标已经达成 |
| WAMI / HSI 已绘制城市主张 | 美国 30 个城市、英国 12 个城市 | 2025 | DroneXL | 中 | 说明另一个监视网络可能扩展使用场景 | 没有把已绘制城市与 Roark 付费客户挂钩 |
| 第三方飞手招募 | RPAS/Part 107 赞助,加摄像头和 Plexus 访问权限 | 2025 | DroneXL | 中 | 暗示供给增长靠伙伴协助,而不完全靠自有运营 | 没有活跃飞手或付费项目数量 |
| 第三方终端用户主张 | 政府、军方和私营部门 3,000+ 终端用户 | 2026 年抓取 | Tracxn | 低 | 如果属实,意味着多个细分已有广泛采纳 | 没有具名用户、队列或独立佐证 |
| 第三方已部署设备主张 | 全球 55,000+ 台设备 | 2026 年抓取 | Tracxn | 低 | 如果属实,相对已披露公开证据而言,装机基数很大 | 没有资产清单、客户名单或与收入的公开勾稽 |
这些只是主张层面的表面信号,不是已勾稽的活跃账户数据;本章在其绑定具名客户或审计运营指标之前,只把它们视为方向性证据。
[CU004, CU005, CU006, CU007, CU008, CU009]| 客户 | 细分 | 部署 / 使用场景 | 生产 vs 试点 | 结果 | 局限 |
|---|---|---|---|---|---|
| 未具名 DDaaS 覆盖用户 | 政府、安保或私人监控用户(声称) | 70 城 / 100 城 DDaaS 叙事中的月度托管探测覆盖 | Unknown | 公开来源显示服务包装存在 | 没有具名客户、城市清单或合同披露 |
| 未具名 OTC 探测买方 | 购买 Aero-Ark 硬件并获得 Plexus 访问权限的便携或哨兵买方 | 场外硬件,加边缘 AI 监控和可选 Rapid Deploy | Unknown | 公开来源显示存在可购买的产品界面和可信的用户画像 | 没有具名买方、站点或重复订单 |
| 第三方数据库中的未具名终端用户 | 政府、军方和私营部门用户(声称) | Tracxn 记录的广泛终端用户和已部署设备画像 | Unknown | 如果属实,可能说明已有规模 | 数据库主张没有具名参考客户或官方指标独立佐证 |
本表记录的是 Roark 缺少具名证明,而不是编造客户身份;竞争对手基准另见图表和正文。
[CU003, CU004, CU005, CU008, CU009, CU010]展示从宽泛采用主张一路筛到具名客户证据后,还剩多少公开客户证据。
计数描述已审阅的公开证据面,不代表 Roark 内部管线阶段。
[CU004, CU005, CU006, CU008, CU009, CU010]对比 Roark 公开客户证明质量与可比公司披露面;后者给出了更强的具名参考。
矩阵标签是分析师基于检索页面和案例研究作出的判断,不是标准化供应商评分卡。
[CU034, CU035, CU036, CU037, CU038, CU042]6.3 耐久性、留存与集中度大多未披露
公开记录在决定 Roark 客户基础是否耐久、而不只是面向市场展示的指标上很弱。保留来源没有披露净收入留存、总收入留存、流失、续约率、合同期限、SLA 或客户满意度指标。头部客户集中度和渠道依赖也一样。投资人可以推断服务包装可能适配经常性或重复部署,但无法量化收入是分散在许多账户,还是由少数不透明项目驱动。即便最强采用主张也有分母问题,因为城市数、最终用户数和设备数没有与活跃付费账户、留存合同或按细分市场划分的收入贡献对齐。这让本章处在一种常见但重要的尽调状态:Roark 看起来拥有一种可同时支持临时和持续监测用例的产品配置,但公开耐久性证据不足,无法在没有管理层会议室数据的情况下承销扩张经济、生命周期价值或集中度风险。[CU008, CU009, CU013, CU014, CU039, CU040]
| 指标 | 数值 / null | 细分 | 置信度 | 尽调要求 |
|---|---|---|---|---|
| 净收入留存 | 全部细分 | 低 | 要求按细分和前 10 大账户提供过去 12 个月 NRR | |
| 毛收入留存 | 全部细分 | 低 | 要求提供 GRR、取消和降购历史 | |
| 流失 / 不续约率 | 全部细分 | 低 | 要求提供 logo 流失和收入流失及原因 | |
| 合同期限中位数 / 续约期限 | 全部细分 | 低 | 要求提供标准合同格式、期限和续约条款 | |
| 客户满意度 / 推荐质量 | 全部细分 | 低 | 要求按细分提供 NPS、案例研究和现场参考客户访谈 |
null 表示保留下来的公开来源没有披露该指标,并不代表底层业务价值为零。
[CU039, CU040]展示 2026 年 C-UAS 采购环境中,买方需求如何公开走向合格部署。
这条路径综合了 Army、NATO、DHS、FAA 和国土政策来源;它是市场结构视角,不是 Roark 专属流程文档。
[CU015, CU017, CU018, CU019, CU020, CU024]6.4 采购路径正在打开,但更偏向已验证、可互操作的供应商
对任何可信 C-UAS 供应商来说,2026 年需求侧环境是建设性的。美国陆军和 JIATF-401 正在打造 marketplace,让用户比较已验证系统并放大表现良好的方案;NATO 正在建立一个既可购买也可租赁的审查池;官方政策文件强调可互操作、开放、能快速测试的解决方案。DHS 给一线响应者和安全机构提供采购工具,并指向世界杯相关拨款支持;FAA 和国土政策材料显示,counter-UAS 购买和使用仍被法律授权和运营模式严格塑形。对 Roark 而言,这意味着可服务市场可能真实且扩张,但采购成功不应与营销可见度混为一谈。因此,本章结论是平衡的:Roark 的低摩擦 DDaaS 和 Rapid Deploy 包装契合新兴买家需求形态,尤其是客户想要可试用或类似租赁覆盖时;但公司缺少具名公开引用,相较那些可能赢得最高要求项目、已验证、可互操作且文档更充分的供应商,仍处于证明不足状态。[CU015, CU016, CU017, CU018, CU019, CU020]
| 扩张驱动因素 | 集中度风险 | 影响 | 尽调路径 |
|---|---|---|---|
| DDaaS 和 Rapid Deploy 降低入门摩擦 | 低摩擦入门可能带来许多小试点,却没有稳定续约 | 可能放大表层采纳主张,却无法证明长期收入 | 按阶段审查管线、续约转化和试点转生产率 |
| 已验证的美国和 NATO 市场可能扩大买方触达 | 已验证市场也可能把需求集中到证据更强、互操作资质更好的供应商 | 即使真实需求增长,Roark 仍可能遭遇资质摩擦 | 要求提供 Roark 测试结果、互操作映射和市场状态 |
| 公共安全拨款和国土安全工具扩大买方集合 | 权限和法律约束会限制谁能真正使用反制能力 | 部分表面客户可能只购买探测,或需要承包商执行 | 将每个目标细分映射到法律权限和获批运营模式 |
| 城市监控和飞手网络扩张可能扩大足迹 | 没有公开数据说明哪些城市或用户会转化为经常性合同 | 覆盖主张可能夸大已货币化采纳 | 要求提供付费账户数、活跃站点和按使用场景划分的收入 |
| 客户组合不透明,掩盖大客户依赖 | 没有公开披露量化前几大客户或渠道集中度 | 少数账户可能贡献过高收入或政策风险 | 要求提供前 10 大客户收入占比、经销商组合和集中度约束条款 |
本章可以识别出可信的扩张路径,但公开证据不足以量化客户集中度或重复经济性。
[CU015, CU017, CU018, CU020, CU024, CU025]6.5 附件
07风险
7.1 不透明与资本可信度是首要风险
现有证据足以说明 Roark 是一家真实运营的公司,但仅靠披露质量,还撑不起后期投资人的信心。Companies House 证明 Roark 仍处于 active 状态,给出了下一次申报节点,也显示了真实的股本申报序列和 2024 年账目。问题也在这些申报里:2024 年账目把英国和美国集团销售合并披露,申报链条没有给出清晰投资人名册,2026 年 4 月的 SH01 又增加了一层对普通股持有人重要的流动性优先权,却没有解释更完整的权利包。Resilience 和 Nordic9 从不同角度强化了同一个问题:$210 million Series B 和 $1.6 billion 估值的标题已经公开流传,但验证没有一起流传。Roark 留存的一手网页材料提供的是识别信息和地址,不是防务相邻后期融资通常需要的管理层级信任材料。因此,第一项风险的重点不在公司是否存在,而在外部投资人能否仅凭公开证据,可信地承销价格、治理杠杆和下行风险。[CR001, CR002, CR003, CR004, CR005, CR006]
Roark 最需要公开证明、却目前只给出部分或间接证据的地方,残余风险最高。
热力图是分析师对引用证据和缓解成熟度的综合,不是公司发布的风险登记册。
[CR027, CR039, CR040, CR042, CR045, CR046]7.2 权限、出口管制和资质门槛压缩了可用市场
官方政策材料显示,counter-UAS 买方环境只对满足权限、安全和资质规则的供应商变得更容易。FAA 公共安全指引仍把处置非法无人机界定为依赖权限的活动。JIATF-401 称,指挥官在 10 U.S.C. Code § 130i 下拥有更多灵活性,但这份灵活性同时要求训练合格的操作员、指定设施、跨机构数据共享,以及经过测试和验证的装备。DHS 和 CISA 又加了一道筛选:机构采购前应比较任务特定需求、场地条件和分层安全整合。OMB 和 BIS 再次抬高门槛,把无人系统采购同安全供应和本土能力逻辑绑定;FAA ARC 则建议先有性能标准、获批测试和供应商名单,再做大规模部署。这些规则都不是专门针对 Roark,但它们定义了 Roark 必须穿过的闸门。Roark 没有公开说明出口管制姿态、安全数据控制或资质状态,因此政策扩张并不会自动转化为公司的开放市场准入。[CR014, CR015, CR016, CR017, CR018, CR019]
| 风险 | 司法辖区 | 当前证据 | 可能性 | 严重性 | 缓解成熟度 | 剩余暴露 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| 权限受限的国土和公共安全部署 | 美国 | FAA 和 JIATF-401 来源显示,反无人机使用仍限定于特定权限和合格操作员 | 高 | 高 | 低到中 | Roark 只能服务具备法律权限和受训操作员的买方细分 | 要求提供买方权限映射、CONOPS,以及任何绑定受保护设施或资产的合同 |
| 已验证供应商和测试关卡风险 | 美国 / NATO | FAA ARC、DHS、CISA、Army 和 NATO 来源强调经过测试、验证或审查的系统 | 高 | 高 | 低 | 缺少第三方证明时,Roark 可能被最敏感的招标筛掉 | 要求提供测试报告、获批供应商状态,以及参与靶场或标准工作的证据 |
| 隐私和数据处理合规 | 美国 | DHS PIA 和 OMB 备忘录显示,隐私和信息安全控制是采购与部署的一部分 | 中 | 高 | 低 | 如果控制不成熟,WAMI、HSI 和 Plexus 数据处理可能拖慢或阻断采纳 | 要求提供隐私架构、留存政策、访问控制和事件处置流程 |
| 出口管制和安全供应暴露 | 英国 / 美国 | BIS 和 OMB 来源显示,安全采购、分类和筛查义务仍然有效 | 中 | 高 | 低 | 跨境销售或组件选择可能遭遇许可或安全来源约束 | 要求提供 EAR 和英国分类备忘录、筛查流程和出货历史 |
| 融资可信度和优先权悬而未决 | 英国 / 投资者侧 | 公开来源无法验证投资者身份,备案显示资本结构带有优先权 | 高 | 高 | 低 | 新投资者可能在权利不透明、外部验证薄弱的情况下承销 | 要求提供签署版股权结构表、交割文件、投资者名单和优先权补充协议包 |
各行按官方政策、备案和已披露资本证据评估严重性;这些判断不能替代法律顾问审查。
[CR008, CR009, CR010, CR014, CR015, CR016]政策和披露缺口会直接传导到客户准入、资本成本和下行保护。
这张图是基于引用政策和融资来源作出的因果综合,不是公司披露的运营模型。
[CR020, CR021, CR022, CR023, CR040, CR042]7.3 产品军备竞赛抬高了整合和执行脆弱性
Roark 公开讲述的产品故事很宽,但在证据很薄时,宽本身就是风险。Defense Tech News 和 DroneXL 描述的技术栈横跨 RF 与被动探测、边缘 AI、反无人机工具、WAMI 与高光谱覆盖,以及向 Plexus 供给数据的第三方飞手网络。战略上这很有野心,但也意味着许多依赖:传感器融合、数据治理纪律、外部操作员质量、现场支持、云或平台韧性,以及足够的工程深度来维持整条链路运转。留存记录中唯一硬性的员工数据,是 2024 年账目里的英国平均 49 名员工,不含承包商和国际员工。与此同时,披露更充分的同行正在发布更强的规模、捕获性能、部署数量、平台整合、融资深度和治理界面证据。Motorola 收购 D-Fend、Axon 整合 Dedrone 平台、Fortem 的部署主张、Shield AI 的自主能力证明、Epirus 的量产扩张,以及 DroneShield 的公开投资人界面,都说明这一品类正在走向更大、更经验证、更资本充足的技术栈。Roark 可能仍有切入口,但剩余风险在于,它的复杂度可能跑在公开证据所能证明的运营成熟度前面。[CR028, CR029, CR030, CR031, CR032, CR033]
| 失效模式 | 可能性 | 严重性 | 缓解成熟度 | 剩余暴露 | 未解决缺口 |
|---|---|---|---|---|---|
| Roark 探测性能、误报率或资质没有公开独立验证 | 高 | 高 | 低 | 采购和部署信心仍然脆弱 | 需要现场测试报告、操作员参考和获批供应商证据 |
| 复杂技术栈横跨 RF、边缘 AI、WAMI、HSI 和平台协同 | 中到高 | 高 | 低 | 集成故障可能在现场运营和支持负载中暴露 | 需要按子系统说明架构、正常运行时间和升级归属 |
| Plexus 和 WAMI/HSI 隐私或数据治理控制未公开 | 中 | 高 | 低 | 敏感买方可能要求目前尚不可见的证明 | 需要留存、访问、加密和事件控制文件 |
| 留存网页和信任文档单薄,削弱买方信心并拖慢伙伴尽调 | 高 | 中 | 低 | 即使产品有效,销售周期也可能拉长 | 需要在线文档、安全包和面向买方的合规材料 |
| 硬件加服务模式可能带来超出纯软件的履约、回款和支持负担 | 中 | 中到高 | 低到中 | 扩张可能同时压缩毛利和服务质量 | 需要披露 BOM、支持人员配置、应收账款账龄和质保 / 服务成本 |
剩余暴露是基于已披露证据的分析师判断,不是基于内部事件日志或质保数据。
[CR004, CR005, CR011, CR012, CR018, CR019]| 依赖 | 交易对手 / 生态 | 角色 | 集中度 | 失效场景 | 严重性 | 缓解 | 剩余暴露 |
|---|---|---|---|---|---|---|---|
| 第三方飞手和承包商网络 | 外部飞手和认证操作员 | 扩大 WAMI/HSI 和部分部署能力 | Unknown | 外部产能薄弱或参差不齐时,覆盖质量或部署速度下降 | 高 | 内部 SOP 和培训可以降低波动 | 在公开控制措施出现前保持高位 |
| 任务权限持有人 | 联邦、军方和公共安全买方 | 提供探测或反制无人机的法律权利 | 按细分为高 | Roark 可以宣传能力,但仍可能拿不到有资金和权限的项目 | 高 | 锁定 covered-facility 与 validated-marketplace 路径 | 资格状态未明前为高 |
| 安全或本土供应预期 | 政府采购与合规框架 | 决定哪些组件和系统可以买 | 中 | 组件或数据路径不合规会卡住采购 | 高 | 做物项分类并记录来源 | 无公开证据时为中高 |
| 多传感器与平台集成 | 内部子系统加数据源 | 交付承诺中的统一作战图景 | 高 | 子系统不兼容或延迟会打断运营叙事 | 高 | 发布参考架构和验证数据 | 证明前为高 |
| 资本与权利交易对手 | 不透明 SPV / 后期投资者 | 决定现金续航、治理杠杆和清算结果 | Unknown | 权利和治理动态不清,未来投资者只能盲目承销 | 高 | 披露已签署融资文件 | 投资人材料可见前为高 |
公开记录没有披露完整供应商、投资者或运营方名单,因此按依赖类别归集交易对手。
[CR010, CR015, CR016, CR020, CR021, CR026]| 角色 / 职能 | 依赖或缺口 | 发生可能性 | 严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| 创始人 / 控制人 | 单一董事和控股 PSC 让治理风险与关键人风险集中 | 高 | 高 | 增设独立董事会,明确授权边界 | 要求提供董事会纪要、组织架构图和继任计划 |
| 合规负责人 | 出口、隐私、安全和强权限市场需要专人负责,但公开信息看不到 | 中高 | 高 | 聘用或披露负责合规的责任人 | 要求列名合规负责人和汇报线 |
| 现场支持与集成工程 | 已披露 49 名英国员工,对照广泛网络和多系统主张显得偏薄 | 中 | 高 | 搭建区域支持和合作伙伴 QA 结构 | 要求按职能列示人员、承包商人数和事故升级模型 |
| 投资者与董事会监督 | 公开轮次叙事超过了公开治理证据 | 中 | 高 | 披露领投方、董事席位和保留事项 | 要求已签署治理文件包和董事会构成 |
各行聚焦会实质影响残余风险的职能,而不是做完整组织审查。
[CR006, CR007, CR013, CR031, CR042, CR043]Roark 公开运营叙事依赖几条外部或披露单薄的链路,且每条都可能独立失效。
节点反映公开材料披露或推断出的依赖类别;公司尚未发布这张图。
[CR028, CR029, CR030, CR031, CR032, CR041]7.4 缓释只有快速产出证据才有效
主要风险并非无法监控,但大多数风险需要管理层会议室里的证据,而不是更多叙事。因此,缓释路径很具体。Roark 可以拿出签署版 cap table、投资人文件,以及从已申报配股到所称融资轮次的清晰桥接,来降低披露风险。它可以展示获批测试结果、部署参考、安全数据控制,以及任何资质或供应商池进展,来降低采购和权限风险。它可以说明关键供应商、记录承包商或飞手控制,并证明 Plexus 与 WAMI/HSI 处理满足隐私和安全预期,来降低整合风险。它只有把这些运营证明同当前流动性、runway 和权利包披露结合起来,才能降低估值风险。关键投资人姿态是把时间视作风险放大器:如果下一次账目截止日、客户证明或政策资质里程碑过去后仍没有实证,公司就会从披露不足,变成在当前叙事价格下结构性不可承销。[CR040, CR042, CR045, CR046, CR047]
| 风险 | 可监测触发项 | 门槛 / 事件 | 行动含义 |
|---|---|---|---|
| 估值可信度 | 已签署融资证据 | 尽调已要求但仍无已签署 cap table、投资者名单或权利文件包 | 不承销价格;把估值叙事视为未验证营销 |
| 流动性与现金续航 | 下一期账目 / 管理层资料包 | 投资决策前没有当前现金、烧钱、订单积压或应收账龄支持 | 暂停,直到资产负债表可见 |
| 资质风险 | 测试和验证证明 | 无第三方测试报告、获批供应商进展或具名且验证过的部署 | 假设采购准入比 TAM 叙事更窄 |
| 安全供应与出口风险 | 合规资料包 | 对敏感买方没有物项分类、筛查流程或本土来源解释 | 只承销非敏感渠道,或退出 |
| 集成与运营规模 | 支持指标 | 覆盖范围主张扩张时,没有架构负责人地图、正常运行时间证据或承包商 QA 控制 | 下调服务可靠性和利润率假设 |
| 客户证明风险 | 具名生产环境参考 | 交割前没有具名客户、续约证明或量化部署结果 | 将商业耐久性视为未证实,投资论点不完整 |
每个触发项都设计成尽调中可由外部核查,而不是依赖笼统乐观判断。
[CR013, CR040, CR042, CR045, CR046, CR047]7.5 附录
08估值
8.1 建议与价格纪律
公开记录足以否定 Roark 是虚构公司,或 2025 年末融资标题毫无来源的说法;但它还不足以让外部投资人有信心承销标题价格。Companies House 申报显示了真实的资本形成,2025 年 12 月的实缴条目规模也足够大,使所称融资轮规模具备合理性。与此同时,同一套公开材料仍缺少签署版 cap table、完整投资人名册,以及能把合理性转化为价格确定性的权利包。因此,正确判断是继续研究,而不是买入。估值姿态是偏紧绷,但并非自动荒谬:申报链条把 2025 年末标记三角定位到与公开流传的 $1.6 billion 叙事同一数量级;但唯一已申报运营年度显示毛利为负,当前收入质量仍不透明。落到 IC 层面,Roark 只有在进入纪律明确、治理和稀释条款迅速浮出水面、管理层能证明 2026 年运营质量明显好于最后一份申报账目所暗示的水平时,才值得继续尽调。[CV016, CV017, CV021, CV023, CV024, CV025]
| 维度 | 当前观点 | 证据基础 | 决策含义 |
|---|---|---|---|
| 建议 | 继续研究 | 申报文件部分印证了融资规模,但投资者权利和当前运营质量披露仍不足 | 只有明确价格纪律并要求文件时,才继续尽调 |
| 信心 | 中 | 核心申报文件扎实,但估值叙事很大一部分仍依赖稀薄或设门槛的信息源 | 把判断用于方向参考,不用于精确定价 |
| 风险评级 | 高 | Cap table 不透明、优先权悬置和已申报毛利为负,都可能削弱下行保护 | 不要跳过治理或财务尽调 |
| 估值立场 | 偏高 | 一手证据支持大额融资确有发生,但公开证据还不足以支撑干净的后期成长溢价 | 要求折价,或拿到更强尽调证据 |
| 最佳下一步 | 已签署尽调包 | 承销新一轮前,需要 cap table、权利、当前利润率和客户证明 | 后续工作以管理层资料室材料为闸门 |
本表把业务质量与价格支撑拆开;建议有意跟随证据变化,而不是给出泛泛判断。
[CV021, CV023, CV025, CV028, CV055, CV056]| 视角 | 论点 | 支撑证据 | 什么会改变判断 |
|---|---|---|---|
| 多头论点 | Roark 可能已经完成一次合法的第三方估值跃升;其服务驱动的 C-UAS 平台扩张速度,也可能快于最近申报账目显示的水平 | 2025 年 12 月申报、Nordic9 标题、Tracxn 融资总额和大型网络叙事,都指向真实动能 | 需要当前收入结构、客户证明和已签署 cap table 披露来验证叙事 |
| 基准论点 | 融资确实发生了,但公开证据目前只支持低于流传标题的折价区间 | 一手申报证明现金和 cap table 有变动,但经济性和权利披露仍偏薄 | 如果能把 SH01 申报干净衔接到当前完全摊薄所有权,区间可能上修收窄 |
| 反论点 | 公开标题可能高估当前可投资价值,因为优先权条款、偏弱毛利率和不透明证明会压低普通股价值 | 2025 年 3 月反稀释条款、后续优先权层级和已申报毛利为负,都支持保持怀疑 | 更强利润率、具名客户和已核对权利栈会削弱反论点 |
| 证伪信号 | 如果管理层无法核对公开融资总额,或拿不出客户级验证,下一次定价可能低于叙事区间 | 公开数字漂移、数据库设门槛,目前留下太多未解问题 | 由第三方领投且文件透明的融资,可能正向重置信心 |
各论点刻意保持价格敏感:同一个运营故事,在一个价格可投,换一个价格就可能没有吸引力。
[CV021, CV024, CV025, CV026, CV028, CV051]建议来自几项因素的组合:融资信号真实、公开运营披露偏弱、权利不透明仍未解决。
这是分析师对证据链的综合,不是公司发布的决策树。
[CV021, CV023, CV025, CV028, CV049, CV050]Roark 在融资信号和市场相关性上得分最高,在证据质量和权利透明度上最弱。
分数是分析师概括信念维度的速记,不是统计输出。
[CV021, CV028, CV055, CV056, CV057]8.2 融资信号与 Cap Table 悬而未决
申报序列比新闻稿更重要,因为它是唯一不能被轻易当作宣传文案挥掉的一手证据。2025 年 3 月已经引入了保护投资人的优先权条款,包括反稀释。随后,2025 年 12 月的 SH01 显示,131 股 B 普通股实缴 £156.96 million;数日后又发生一次资本重组,实际上把该类别从 131 股重计为 131,000 股,同时保留名义价值。这个序列并不常见,但分析上有用:它隐含的每股价格指向约 £1.36 billion 的投后估值,发生在后续优先权和 2026 年 4 月发行之前,方向上与公开的 $1.6 billion 叙事一致。问题不在规模,而在缺失的桥接。外部人仍看不到哪些投资人买了哪些证券、精确的清算瀑布,或后续优先权层级如何影响普通股持有人。因此,本章把申报数学视作三角定位,而不是当前标记清晰无误的证明。[CV005, CV006, CV007, CV008, CV009, CV010]
| 事件 | 公开事实 | 对估值为何重要 | 残余不确定性 |
|---|---|---|---|
| 2025 年 3 月 SH01 | 81.0M 普通股加 19.0M 优先股,名义金额 £1.25;优先股类别包含反稀释和优先股息地位 | 说明所称 Series B 之前就存在投资者保护条款,并且会影响普通股有效价值 | 这些权利如何进入后续轮次文件,公开信息没有桥接 |
| 2025 年 12 月 15 日 SH01 | 131 股 B 普通股以 £156.96M 现金获配 | 大额融资事件发生的一手证据 | 未附投资者姓名或轮次备忘录 |
| 2025 年 12 月 23 日 SH01 | 股本基础重置为 1.131M 股,同时保留 £1,131 合计名义价值 | 可用于三角测算隐含 B 股价格和投后量级 | 单独来看不能证明完全摊薄 cap table 干净 |
| 2025 年 12 月 28 日 SH01 | 新增 221,339 股优先股,带股息和清算优先权 | 暗示普通股下行可能与标题中的企业叙事有实质差异 | 无公开清算瀑布或优先级地图 |
| 2026 年 4 月 4 日 SH01 | 新增 20,079 股 B 普通股;总股数升至 1.372M | 证实标题轮次窗口后 cap table 继续变动 | 该次发行的经济目的未公开解释 |
| 数据库 / 媒体 | Nordic9 称以 $1.6B 估值融资 $210M;Tracxn 称总融资 $287M;Resilience 强调验证缺口 | 说明轮次叙事传播很广,但尚未达到管理层级证据 | 融资总额和投资者身份在不同来源间仍有漂移 |
本表强调:稀释、证券类别和权利披露,与标题轮次金额同样重要。
[CV005, CV006, CV007, CV008, CV009, CV010]Roark 的估值最敏感于投资者如何权衡当前证明与 2025 年末融资信号。
数值是各情景中点,单位为十亿美元,不是可观察交易标记。
[CV051, CV052, CV053, CV054, CV058]8.3 可比组与情景区间
可比证据支持谨慎,而不是亢奋。Shield AI 和 Epirus 说明,防务自主和反无人机平台可以获得很大规模的私募融资,但两家公司披露的产品证明和战略支持方细节都比 Roark 目前更深。本组最清晰可用的定价可比,是 Motorola 收购 D-Fend:一家预计 2026 年收入 $185 million 的公司卖出 $1.5 billion,约等于 8.1x 前瞻收入。这个倍数已经计入了可观增长和战略稀缺性,但 D-Fend 同时有披露部署和具名买方。DroneShield 在 2026 年 6 月给出少见的公开纯标的参照,公开市值约 $1.77 billion;Leonardo DRS 和大型防务主承包商则提醒读者,公开市场对防务的胃口健康,但集中在规模更大、披露更充分的运营商身上。在这一背景下,接近流传 Roark 标题价的牛市情景只有在管理层能证明当前经济性和客户证据远超最后申报账目时才说得通。更可辩护的纯公开信息姿态,是把基准区间放在标题价下方,并设置明确计入不确定性的宽熊牛价差。[CV029, CV030, CV031, CV032, CV033, CV034]
| 可比公司 | 可见指标 | 估值或状态 | 相关性 | 相比 Roark 的局限 |
|---|---|---|---|---|
| Shield AI | 2025 年 3 月融资 | $240M 轮后估值 $5.3B | 有战略投资者参与的私营国防自主系统上界基准 | 自主能力证明和产品披露远强于 Roark |
| Epirus | 2025 年 3 月融资 | $250M Series D;总融资 >$550M | 显示反无人机资本强度和投资者胃口 | 引用新闻稿未给出硬性公开估值 |
| D-Fend Solutions | 2026 年 6 月 M&A | Motorola 同意以 $1.5B 收购;预期 2026 年收入 $185M | 直接品类可比中披露最充分的战略交易 | 退出价格反映具名买方、部署和已知收入基础 |
| Dedrone | 2024 年 5 月 M&A | 被 Axon 收购;价格未披露 | 验证战略买方对空域安全平台有胃口 | 无公开价格倍数 |
| Fortem Technologies | 2026 年 4 月战略融资 | Lockheed 投入 $25M,作为 Series B 首批款项 | 表明有主承包商背书的平台仍能吸引战略资本 | 未披露公司估值 |
| DroneShield | 2026 年 6 月公开市场 | 市值约 $1.77B | 少见的公开纯反 UAS 估值参考 | 公开市场流动性和披露与英国私营初创公司差异很大 |
| Leonardo DRS | 2026 年 6 月公开市场 | 市值约 $12.29B | 显示在位公司带入该品类的规模和披露强度 | 不是初创公司,也不是直接估值可比 |
| BAE / Northrop / Thales / SAAB / Rheinmetall(防务主承包商) | 2026 年 6 月公开市场 | 市值约 $28B 至 $74B,Lockheed 约 $117.8B,Motorola 约 $65.6B | 说明 2026 年防务股权市场胃口健康 | 这些是大型主承包商或多元防务可比,不是 Roark 的价格锚 |
Roark 没有披露足够的当前运营数据,无法拼出紧密的同类可比组,因此必须混用可比;战略 M&A 和公开市场可比主要用作边界标记。
[CV029, CV031, CV033, CV034, CV035, CV036]| 情景 | 估值区间(USD bn) | 核心假设 | 概率信号 | 关键下行 / 上行触发项 |
|---|---|---|---|---|
| 空头 | $0.3-$0.6 | 2024 年申报经济性在方向上仍有效,客户证明继续偏薄,优先权实质压低普通股价值 | 如果尽调无法核对融资总额,或无法证明利润率修复,该情景就有合理性 | 下一轮定价或二级交易明显低于 2025 年底叙事 |
| 基准 | $0.8-$1.2 | 2025 年底融资真实且具战略意义,但当前收入质量、留存和权利仍足以支持较大透明度折价 | 目前仅凭公开信息最能自洽的区间 | 管理层可用已签署文件和当前运营 KPI 收窄区间 |
| 多头 | $1.4-$1.8 | 2025 年 12 月估值标记合法,当前 2026 年经济性大幅改善,具名客户和重复部署验证网络主张 | 需要多个尽调事项兑现,才能从猜测转为可信 | 已签署 cap table、利润率证明和客户验证会显著上修判断 |
情景数值是分析师区间,不是可观察标记;锚点来自申报测算、已披露经济性、战略交易背景和公开反 UAS 估值参考。
[CV023, CV024, CV028, CV035, CV049, CV050]仅基于公开信息的承销区间很宽,因为精确下行保护和当前收入质量都仍未解决。
区间表示分析师在各情景下的投后估值带,单位为十亿美元。
[CV023, CV024, CV052, CV053, CV054]8.4 尽调要求与推翻论点的因素
能把判断从偏紧绷、继续研究推向合理的,不是又一篇摘要文章,而是管理层会议室里的证据。最快打开局面的材料,是签署版 cap table、清晰的权利摘要,以及当前收入质量证明,用来判断 Roark 是否已经长进了自己的叙事,还是只是设计出一个解释很薄的估值上调。客户引用、部署数量、续约行为和 backlog 质量同样重要,因为这家公司被按防务科技平台估值,而不是按一次性硬件转售商估值。如果融资总额漂移无法调和、优先权让新投资人的受偿顺位低于预期,或 2026 年经济性仍接近 2024 年申报中可见的负毛利图景,论点也会很快失效。对新投资人而言,决策规则很直接:只有公司补上这些证据缺口后才加价进入;否则,要么要求相对流传标题价有实质折扣,要么等下一次定价事件倒逼更高透明度。[CV018, CV025, CV027, CV052, CV053, CV054]
| 触发项 | 门槛或事件 | 如何传导到论点 | 行动含义 |
|---|---|---|---|
| Cap table 不透明持续 | 管理层无法在尽调中提供已签署 cap table 和权利摘要 | 价格支撑仍过度依赖叙事,无法描绘下行保护 | 不按流传标题推进 |
| 融资总额仍未核对 | 尽调后,公开口径和管理层口径的募资金额仍冲突 | 削弱对稀释测算和治理纪律的信心 | 重置跟踪,或在核对前回避 |
| 当前经济性类似 2024 年申报画像 | 2026 年数据中毛利率和现金转化仍然较差 | 削弱高倍数逻辑,并推向空头区间 | 要求实质低价,或停止 |
| 客户证明缺失 | 没有具名参考、续约数据或已验证部署数量 | 多头情景失去支撑高估值所需的运营证明 | 只假设基准或空头 |
| 权利栈优先级高于预期 | 优先权瀑布或附函实质损害普通股结果 | 即便企业价值正确,也可能高估新股权买到的东西 | 重新定价,或设计保护性条款 |
这些是投资流程否决标准,不是运营预测;每个触发项都设计成可在尽调中监测。
[CV025, CV026, CV027, CV028, CV052, CV053]| 主题 | 缺失证据 | 为何重要 | 负责人或尽调路径 |
|---|---|---|---|
| Cap table 与权利 | 已签署完全摊薄 cap table、SPA、附函和清算瀑布 | 需要把标题估值转换成真实下行保护和所有权经济性 | 公司律师和领投方文件 |
| 当前运营质量 | 2025-2026 年管理账、收入结构、毛利率桥接和现金续航 | 需要判断 2024 年毛利为负的画像是否已结构性修复 | CFO 资料包和董事会报告 |
| 客户证明 | 具名客户、部署数量、续约率和按细分列示的订单积压 | 需要验证网络和服务规模主张是否能支撑价格 | 销售负责人推荐加客户访谈 |
| 融资历史核对 | 按轮次融资历史逐项匹配每份 SH01 和数据库数字 | 需要核对 $210M、$260M、$287M 和 €57.6M 公开引用 | 财务团队核对备忘录 |
| 治理表面 | 董事会构成、保留事项,以及任何 SPV 治理特征 | 需要评估控制权、汇报纪律和后续融资流程质量 | 法律尽调加治理文件 |
每项要求都必须满足,才足以实质上调建议;没有一项只是装饰。
[CV018, CV025, CV027, CV028, CV057, CV058]免责声明
本报告仅供参考,完全基于截至 2026-06-20 可获得的公开信息,不构成投资建议。作出任何投资或商业决策前,均需独立完成财务、法律、技术和监管尽调。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Roark Aerospace trades legally as RO-ARK AEROSPACE LIMITED under UK company number 15319796. | 高 | SO002, SO003 |
| CO002 | Roark Aerospace's registered office is Monomark House, 27 Old Gloucester Street, London, WC1N 3AX, England. | 高 | SO002, SO003 |
| CO003 | Companies House lists Roark Aerospace as an active private limited company incorporated on 30 November 2023. | 高 | SO003, SO007 |
| CO004 | The company registered SIC codes 26120 (manufacture of loaded electronic boards) and 63990 (other information service activities n.e.c.). | 中 | SO003 |
| CO005 | Roark's archived legal page listed a Cambridge Science Park office and a New York office in addition to the London registered office. | 中 | SO002 |
| CO006 | Roark's archived homepage positioned the company as “Drone Detection as a Service | C-UAS as a Service.” | 中 | SO001 |
| CO007 | The archived homepage displayed NATO NCAGE U2D39 together with UEID Y5J7FXLHZAN6, DUNS 231246243, and VAT GB491393565. | 中 | SO001 |
| CO008 | Patton Atkinson-French is the only active director shown in Companies House records and was appointed on 30 November 2023. | 高 | SO004, SO005 |
| CO009 | Companies House lists Patton Atkinson-French as the sole person with significant control, holding more than 75% of shares and voting rights and the right to appoint or remove directors. | 中 | SO006 |
| CO010 | Companies House records describe Patton Atkinson-French as a British resident in England born in February 1985 with identity verification complete. | 高 | SO004, SO005 |
| CO011 | Ro-Ark's 2024 accounts say the company is principally engaged in the manufacture and service deployment of drone detection systems, counter-drone systems, and unmanned vehicles. | 中 | SO008 |
| CO012 | The 2024 accounts reported turnover of £36,906,898 for the period ended 30 November 2024. | 中 | SO008 |
| CO013 | The same accounts reported a pre-tax loss of £16,561,770 and gross loss of £12,418,074. | 中 | SO008 |
| CO014 | Ro-Ark reported total net assets of £79,346,325 at 30 November 2024. | 中 | SO008 |
| CO015 | The accounts disclosed an average of 49 UK-based employees, explicitly excluding contractors and international staff. | 中 | SO008 |
| CO016 | The accounts also disclosed a £16.42 million bond at 6.5% APR with term date March 2029. | 中 | SO008 |
| CO017 | A 13 March 2025 SH01 filing showed 81,000,000 ordinary shares and 19,000,000 preference shares allotted for cash at £1.25 nominal value per share, taking total nominal capital to £125,000,000. | 中 | SO009 |
| CO018 | A 15 December 2025 SH01 filing showed 131 B ordinary shares allotted for cash with an amount paid line of £156,963,450. | 中 | SO010 |
| CO019 | Follow-on SH01 filings on 23 December 2025, 28 December 2025, and 4 April 2026 indicate continued post-round cap-table adjustments across ordinary, B ordinary, and preference classes. | 高 | SO011, SO012, SO013 |
| CO020 | Roark's filing history shows 2024 accounts were filed on 20 June 2025 and multiple capital statements were filed across March 2025 to April 2026. | 中 | SO007 |
| CO021 | Nordic9 reported a $210 million Series B dated 18 December 2025, a $1.6 billion valuation, and total investments received of $260 million, but flagged the deal as “still pending verification.” | 中 | SO015 |
| CO022 | Resilience Media said Roark circulated a Boxing Day 2025 press release claiming a $210 million Series B at a $1.6 billion post-money valuation while naming only a U.S.-based investment bank via an SPV. | 中 | SO016 |
| CO023 | The visible portion of the Resilience Media story indicates the outlet could not readily verify Roark's extraordinary claims and sought further substantiation from management. | 中 | SO016 |
| CO024 | Tracxn describes Roark as a London-based Series B company founded in 2023 that has raised $287 million. | 中 | SO014 |
| CO025 | Tracxn also states that Roark had 244 employees as of December 2025 and had deployed more than 55,000 devices to over 3,000 end users. | 中 | SO014 |
| CO026 | Global Defense Tech News described Roark as one of the largest worldwide drone-detection and hyperspectral-imaging networks and said it had raised roughly €57.6 million since a 2022 founding. | 低 | SO017 |
| CO027 | DroneXL reported in January 2025 that Roark's Aero-Ark systems used ADS-B monitoring, Remote ID compliance, DJI packet collection and decryption, passive radar, and edge-AI classification. | 中 | SO018 |
| CO028 | DroneXL also reported that Roark sold OTC detection units from £4,500 and was expanding WAMI and HSI coverage to 30 U.S. cities and 12 U.K. cities through its Plexus platform. | 中 | SO018 |
| CO029 | The 2024 accounts note that turnover represents total group sales across both UK and US entities, implying an international group structure beyond the single UK legal entity. | 中 | SO008 |
| CO030 | Cushman & Wakefield's Q1 2026 UK regional offices report said Roark Aerospace leased 13,650 square feet in Cambridge. | 中 | SO019 |
| CO031 | NATO AIRCOM's June-July 2026 industry day put C-UAS at the center of alliance discussions around detection, tracking, mitigation, neutralisation, and interoperability. | 中 | SO020 |
| CO032 | Counter UAS Technology Europe 2026 highlighted demand for mobile and dismounted C-UAS, low-collateral defeat, and cost-effective anti-drone architectures shaped by Ukraine lessons. | 中 | SO021 |
| CO033 | Unmanned Airspace markets its 2026 directory as a procurement resource covering more than 550 companies and 1,000 programmes or systems, indicating a crowded C-UAS competitive landscape. | 中 | SO022 |
| CO034 | Unmanned Airspace said publicly announced government spending on counter-UAS systems exceeded $29 billion in the first three months of 2026. | 中 | SO023 |
| CO035 | DroneShield's investor-relations page shows a public counter-UAS pure-play with ongoing quarterly, annual, and investor-presentation disclosure, underscoring the disclosure gap between Roark and better-documented peers. | 中 | SO024 |
| CO036 | Shield AI's March 2025 $240 million round at a $5.3 billion valuation demonstrates that defense-autonomy investors were willing to back category leaders aggressively, but with far more detailed disclosure than Roark provided publicly. | 中 | SO025 |
| CO037 | Companies House shows Ro-Ark using the small-companies regime and notes that members did not require an audit under section 476 of the Companies Act 2006. | 中 | SO008 |
| CO038 | Companies House shows the next accounts for the period to 30 November 2025 were due by 31 August 2026 and that the latest confirmation statement was dated 21 May 2026. | 中 | SO003 |
| CO039 | Roark's public web footprint was unusually thin in mid-2026: the live site did not fetch cleanly from this environment, while the archived site exposed only minimal contact and slogan data. | 中 | SO001, SO002 |
| CO040 | Across the official site, filings, and third-party writeups, Roark consistently positioned itself around integrated detection, mapping, and command-platform services rather than a single sensor or effector product. | 高 | SO001, SO008, SO017, SO018 |
| CM001 | Roarks archived homepage positioned the company as "Drone Detection as a Service | C-UAS as a Service." | 中 | SM010 |
| CM002 | DroneXL said Roarks Aero-Ark offer combined ADS-B, Remote ID, DJI packet collection and decryption, passive radar, and edge-AI-enabled sentry or portable units. | 中 | SM011 |
| CM003 | Defense Tech News described Roark as combining sensors, counter-drone tools, and WAMI/HSI surveillance through the Plexus AI platform across land, air, and maritime environments. | 低 | SM012 |
| CM004 | NATO AIRCOM framed the 2026 C-UAS challenge as covering the full chain from detection and tracking to mitigation and neutralisation. | 中 | SM006 |
| CM005 | The U.S. Army said the allied marketplace initiative begins with counter-drone systems before potentially expanding to drones, radars, sensors, tanks, helicopters, and other capabilities. | 高 | SM001, SM002 |
| CM006 | DefenseScoop reported that the FY26 Pentagon request calls out $13.4 billion for autonomy and autonomous systems. | 中 | SM003 |
| CM007 | The same FY26 briefing cited $3.1 billion for counter-UAS across the services. | 中 | SM003 |
| CM008 | Broad autonomy budgets overstate Roarks immediate market if used as direct TAM because counter-UAS is only one subset of those programs. | 中 | SM003, SM010, SM006 |
| CM009 | Defense News reported that NATO plans to curate 18 counter-UAS systems across nine use cases with both purchase and lease options. | 中 | SM004 |
| CM010 | At Eurosatory 2026, the U.S. Army and representatives from nine allied and partner countries signed a letter of intent to streamline counter-UAS acquisition and training. | 高 | SM001, SM002 |
| CM011 | Army leaders said the marketplace model is meant to expose vendors to operator feedback and help governments scale the systems that prove effective. | 高 | SM001, SM002 |
| CM012 | Defense News said NATO wants standardized testing, innovation ranges, and faster pre-completed contracting so nations can procure qualified systems immediately. | 中 | SM004 |
| CM013 | AIRCOM Industry Day 2026 makes interoperability and operational relevance central buyer criteria for future C-UAS capability development. | 中 | SM006 |
| CM014 | Counter UAS Europe 2026 centered its agenda on mobile and deployable systems, non-kinetic options, collateral-damage minimization, and battlefield lessons from Ukraine. | 中 | SM009 |
| CM015 | Defense News reported that low-cost interceptor economics are resetting C-UAS procurement logic, including a Lithuanian purchase of 48 Merops systems at about $15,000 per shot. | 中 | SM005 |
| CM016 | The same Defense News report said the Army bought 13,000 Merops interceptors in eight days after Iranian drone attacks on U.S. positions. | 中 | SM005 |
| CM017 | Fortem and Lockheed Martin claimed that Fortems software-centric counter-UAS platform can reduce cost per engagement by more than 80 percent while integrating with allied networks. | 中 | SM020 |
| CM018 | Unmanned Airspace said publicly announced government C-UAS contracts exceeded $29 billion in the first months of 2026. | 中 | SM007 |
| CM019 | The same Unmanned Airspace article said there are now more than 70 interceptor-drone manufacturers competing for procurement attention. | 中 | SM007 |
| CM020 | The Unmanned Airspace 2026 directory says it tracks more than 550 company listings and 1,000 programme or system descriptions. | 中 | SM008 |
| CM021 | Shield AI raised $240 million at a $5.3 billion valuation to deploy Hivemind Enterprise across OEMs, governments, and companies building autonomy products. | 中 | SM013 |
| CM022 | Epirus raised an oversubscribed $250 million Series D that lifted total venture funding above $550 million to scale Leonidas production and supply chain resilience. | 中 | SM014 |
| CM023 | Motorola Solutions agreed to acquire D-Fend for $1.5 billion and said D-Fend was on pace for $185 million of 2026 revenue with deployments in more than 30 countries. | 中 | SM015 |
| CM024 | D-Fend says its installed base spans nearly 30 countries and customers across defense, homeland security, law enforcement, airports, prisons, critical infrastructure, and major events. | 中 | SM016 |
| CM025 | Axon says Dedrone is used for airports, critical infrastructure, public events, and defense operations and is now being bundled into Axons broader public-safety stack. | 中 | SM018 |
| CM026 | Fortem says its systems have named deployments with U.S. Customs and Border Protection, the U.S. Presidential Inauguration, and operations in Ukraine. | 中 | SM019 |
| CM027 | Leonardo DRS markets layered Group 1 to 3 counter-UAS systems and says its tactical radars can track nano drones up to 18 kilometers away. | 中 | SM021 |
| CM028 | DZYNE markets portable, mobile, and fixed layered counter-UAS products and says its systems are fielded in more than 50 countries. | 中 | SM022 |
| CM029 | SRC says it has fielded counter-UAS technology with the U.S. Army and Air Force and has also been used for VIP protection and large public events. | 中 | SM023 |
| CM030 | DroneShield presents itself as the only publicly listed company exclusively dedicated to counter-UAS and says the counter-drone market is now valued above $10 billion. | 中 | SM024 |
| CM031 | Across the cited sources, recurring buyer segments include armed forces, airports and airfields, critical infrastructure operators, public-safety agencies, and high-profile event or VIP security users. | 中 | SM015, SM016, SM018, SM019, SM021, SM022, SM023 |
| CM032 | Budget ownership differs by segment, with defense ministries funding battlefield systems, airport or aviation operators funding airfield continuity, and site operators or public-safety agencies funding domestic protection use cases. | 中 | SM001, SM004, SM015, SM016, SM018, SM023 |
| CM033 | Interoperability and test validation are becoming first-order buying filters rather than optional extras in coalition counter-UAS procurement. | 高 | SM001, SM002, SM004, SM006, SM020, SM021, SM023 |
| CM034 | Purchase-plus-lease or service-oriented procurement paths are becoming more plausible in this market, which fits Roarks marketed DDaaS / C-UAS-as-a-service positioning better than one-off hardware-only sales. | 中 | SM004, SM010 |
| CM035 | Roarks public offer appears more naturally aligned to deployable surveillance, recurring monitoring, and fixed-site layered defense than to prime-led kinetic or directed-energy programs. | 中 | SM010, SM011, SM012, SM021, SM022, SM023 |
| CM036 | Roark has not publicly matched the named-customer proof published by peers such as Fortem, D-Fend, Dedrone or SRC. | 中 | SM015, SM016, SM018, SM019, SM023, SM025 |
| CM037 | A hot market does not guarantee Roark can win share because buyers increasingly expect qualification evidence, interoperability, and operator references before scaling vendors. | 中 | SM001, SM004, SM006, SM015, SM019, SM025 |
| CM038 | The $3.1 billion DoD line, the $13.4 billion autonomy line, the >$29 billion public contract tally, and the >$10 billion or >$12 billion vendor market narratives are not directly comparable TAM estimates because they use different geographies, time windows, and category definitions. | 中 | SM003, SM007, SM020, SM024 |
| CM039 | No open-web source in this chapter isolates a clean Roark SAM or SOM inside counter-UAS, so any company-specific share assumption still requires management evidence. | 低 | SM003, SM004, SM007, SM008, SM010, SM025 |
| CM040 | Resilience Media argued that Roarks scale and funding narrative could not be readily verified, which raises the bar for converting sector demand into company-specific underwriting confidence. | 中 | SM025 |
| CP001 | Roark's archived homepage described the company as “Drone Detection as a Service | C-UAS as a Service.” | 中 | SP001 |
| CP002 | Roark's public trade coverage disclosed Aero-Ark hardware starting prices of £2,150 to £4,450, DDaaS from £95 per user per month, and rapid deployment from £185 per day. | 中 | SP002, SP003 |
| CP003 | Roark's clearest public wedge is low-friction pricing and service packaging rather than a uniquely documented integration or procurement moat. | 高 | SP001, SP002, SP003, SP004 |
| CP004 | Shield AI competes as adjacent autonomy infrastructure more than as a direct low-cost drone-detection service vendor. | 中 | SP005, SP006 |
| CP005 | Shield AI announced a $240 million strategic round in March 2025 at a $5.3 billion valuation. | 中 | SP005 |
| CP006 | Shield AI's homepage frames its product around mission autonomy in contested and degraded environments, reinforcing a platform rather than point-solution posture. | 中 | SP006 |
| CP007 | Epirus announced a $250 million Series D in March 2025, bringing total venture funding above $550 million. | 中 | SP007 |
| CP008 | Epirus positions itself around high-power microwave and layered short-range air defense, making it a different control-point competitor from Roark's detection-led offer. | 中 | SP007, SP008 |
| CP009 | D-Fend announced a $31 million financing round in December 2024 and said revenue had been growing more than 60% year over year with deployments across nearly 30 countries. | 中 | SP009 |
| CP010 | D-Fend's EnforceAir page emphasizes detection, tracking, identification, RF-cyber takeover, safe landing, and open-API integration across tactical, vehicular, stationary, and man-portable deployments. | 中 | SP010 |
| CP011 | Motorola Solutions agreed to acquire D-Fend for $1.5 billion in June 2026 and cited expected 2026 revenue of $185 million with thousands of deployments across more than 30 countries. | 中 | SP011 |
| CP012 | Motorola's planned acquisition gives D-Fend access to a much larger public-safety and enterprise distribution platform than a standalone startup usually has. | 中 | SP010, SP011 |
| CP013 | Axon's 2024 deal announcement described Dedrone as having hundreds of commercial, government, and military customers worldwide. | 中 | SP012 |
| CP014 | Dedrone's homepage says its AI-driven platform is hardware agnostic, uses more than 18 million images, and supports more than 30 proven integrations across security requirements. | 中 | SP014 |
| CP015 | Axon says Dedrone is now part of its public-safety platform and ties the technology to drone-as-first-responder, public spaces, airports, critical infrastructure, and defense uses. | 中 | SP013 |
| CP016 | Fortem positions SkyDome as an integrated end-to-end counter-UAS system that combines radars, interceptors, and management software. | 中 | SP019 |
| CP017 | Fortem argues that many counter-UAS deployments fail because multi-vendor components lack compatibility and integration, implying that system integration itself is a competitive moat. | 中 | SP019 |
| CP018 | Fortem's homepage publishes performance markers including more than 85% fixed-wing capture rate, more than 95% rotary target capture rate, and roughly 5,000 UAVs captured successfully. | 中 | SP019 |
| CP019 | Fortem's February 2025 update cites CBP use, inauguration security, Ukraine deployments, and the claim that DroneHunter is the only approved low-collateral drone-on-drone interceptor in U.S. airspace. | 中 | SP015 |
| CP020 | Lockheed Martin invested $25 million in Fortem in April 2026 as the initial tranche of Fortem's Series B and tied the company into the Sanctum counter-UAS ecosystem. | 中 | SP016 |
| CP021 | Fortem said in October 2025 that it received purchase orders for 12 C-UAS systems from U.S. allies in Europe and the Middle East. | 中 | SP017 |
| CP022 | Fortem's March 2026 critical-infrastructure announcement tied the company to Lockheed integration, DHS venue protection, and a U.S. Army support contract, reinforcing stronger procurement proof than Roark publicly shows. | 中 | SP018 |
| CP023 | DroneShield describes itself as the only publicly listed company exclusively dedicated to counter-UAS. | 中 | SP020 |
| CP024 | DroneShield's homepage says the company is trusted by 34-plus government agencies worldwide and offers on-the-move, dismounted, fixed-site, and software solutions. | 中 | SP021 |
| CP025 | Leonardo DRS's homepage publishes 2025 company figures of $3.6 billion revenue, $8.7 billion backlog, and 7,300 employees, placing it far above startup scale. | 中 | SP025 |
| CP026 | Leonardo DRS's counter-UAS page markets detect-identify-track-defeat coverage for Group 1–3 threats and says its tactical radars can track nano drones up to 18 kilometers away. | 中 | SP022 |
| CP027 | Leonardo DRS said in April 2026 that it won first place in a DoD counter-drone competition with its Ring electronic-warfare system and highlighted MLIDS and M-SHORAD roles. | 中 | SP023 |
| CP028 | Leonardo DRS's maritime M-MEP announcement shows the company can package detection through defeat into a modular maritime layer for ships, ports, littoral infrastructure, and expeditionary forces. | 中 | SP024 |
| CP029 | DZYNE says its counter-UAS systems are fielded in more than 50 countries and span portable, mobile, and fixed-site layered defense. | 中 | SP026, SP027 |
| CP030 | SRC says it has been developing and testing counter-UAS technology since 2005 and has fielded systems for the U.S. Army, Air Force, VIP events, airfields, and critical infrastructure. | 中 | SP028 |
| CP031 | Unmanned Airspace's 2026 directory tracks more than 550 companies and 1,000 programs or systems, indicating a very crowded supplier landscape. | 中 | SP029 |
| CP032 | The U.S. Army's 2026 marketplace language shows that buyers are trying to expand access to systems already proven on current battlefields and fit for interoperability and training. | 中 | SP030 |
| CP033 | Roark's competitive set spans direct service-led challengers, specialist C-UAS vendors, and incumbent or prime-style alternatives rather than a single startup cohort. | 高 | SP001, SP005, SP008, SP010, SP014, SP019, SP021, SP022, SP027, SP028 |
| CP034 | Most larger peers cited in this chapter do not publish official list pricing on the reviewed surfaces, so public comparison is mostly between packaging models rather than like-for-like dollar prices. | 高 | SP006, SP008, SP010, SP014, SP019, SP021, SP022, SP025, SP027, SP028 |
| CP035 | Roark's public pricing is unusual among peers and implies a detection-led wedge rather than a full integrated base-defense program sale. | 高 | SP002, SP003, SP010, SP014, SP019, SP021 |
| CP036 | Open APIs and hardware-agnostic integration claims from D-Fend and Dedrone reduce absolute lock-in, but once customers validate an integrated stack the operational switching cost still becomes meaningful. | 高 | SP010, SP014, SP019 |
| CP037 | Fortem, D-Fend, Dedrone, DroneShield, and Leonardo DRS all publish stronger deployment, transaction, or ecosystem trust signals than Roark does in open sources. | 中 | SP011, SP013, SP018, SP020, SP023, SP025, SP004 |
| CP038 | Roark's public disclosure on named customers, repeat deployments, and procurement wins is materially thinner than the better-documented competitors in this chapter. | 高 | SP004, SP011, SP013, SP018, SP020, SP023 |
| CP039 | Resilience Media explicitly questioned whether Roark's unicorn-scale claims could be independently verified. | 中 | SP004 |
| CP040 | Roark faces commoditization risk from below because accessible detection hardware and managed monitoring are easier to copy than validated integrated effectors or prime-led channels. | 中 | SP003, SP019, SP021, SP022, SP028 |
| CP041 | Roark also faces displacement risk from above because primes and newly consolidated specialists can combine stronger channels with broader integrated stacks. | 中 | SP011, SP013, SP016, SP025 |
| CP042 | On public evidence alone, Roark looks capable of carving out a detection-led service niche, but not yet of proving a durable moat against better-capitalized integrated competitors. | 中 | SP003, SP004, SP010, SP014, SP019, SP025 |
| CI001 | Roark filed FY2024 turnover of £36,906,898. | 中 | SI001 |
| CI002 | Roark filed FY2024 cost of sales of £49,324,972 and gross loss of £12,418,074. | 中 | SI001 |
| CI003 | Roark filed FY2024 administrative expense of £3,094,908 and interest payable of £1,072,500. | 中 | SI001 |
| CI004 | Roark filed a FY2024 pre-tax loss of £16,561,770. | 中 | SI001 |
| CI005 | Roark reported cash at bank and in hand of £6,109,227 at 30 November 2024. | 中 | SI001 |
| CI006 | Roark reported stocks of £13,581,223 at 30 November 2024. | 中 | SI001 |
| CI007 | Roark reported debtors of £87,109,226, including £36,109,654 due after more than one year, at 30 November 2024. | 中 | SI001 |
| CI008 | Roark reported creditors due within one year of £17,409,012 at 30 November 2024. | 中 | SI001 |
| CI009 | Roark disclosed a £16,420,000 bond at 6.5% APR with term date March 2029. | 中 | SI001 |
| CI010 | Roark reported total net assets of £79,346,325 and fixed assets of £7,375,661 at 30 November 2024. | 中 | SI001 |
| CI011 | Roark's turnover policy says turnover equates to total group sales received across both UK and US entities. | 中 | SI001 |
| CI012 | Roark disclosed an average of 49 UK-based employees in FY2024 and said this excluded contractors and international staff. | 中 | SI001 |
| CI013 | Companies House shows Roark's next accounts for 30 November 2025 are due by 31 August 2026. | 高 | SI002, SI003 |
| CI014 | The 13 March 2025 SH01 filed 81,000,000 ordinary and 19,000,000 preference shares at £1.25 each for cash, for £125,000,000 aggregate nominal capital. | 高 | SI002, SI004 |
| CI015 | The March 2025 preference shares carried block voting rights, full-ratchet anti-dilution, and preferential dividend status. | 中 | SI004 |
| CI016 | The 15 December 2025 SH01 shows 131 B ordinary shares allotted for cash with amount paid of £156,963,450. | 高 | SI002, SI005 |
| CI017 | The B ordinary shares filed in December 2025 carried a liquidity preference over the ordinary class. | 中 | SI005 |
| CI018 | The 23 December 2025 SH01 recapitalized Roark to 1,000,000 ordinary shares and 131,000 B ordinary shares with no cash paid on that allotment. | 中 | SI006 |
| CI019 | The 28 December 2025 SH01 added 221,339 preference shares for £338.84 cash and brought total shares to 1,352,339. | 中 | SI007 |
| CI020 | The 4 April 2026 SH01 added 20,079 B ordinary shares for USD 1,390. | 中 | SI008 |
| CI021 | After the 4 April 2026 filing, Roark's recorded share capital included ordinary, B ordinary, and preference classes totaling 1,372,418 shares. | 高 | SI002, SI008 |
| CI022 | Roark's homepage marketed the company as "Drone Detection as a Service | C-UAS as a Service." | 中 | SI018 |
| CI023 | Roark's archived legal page listed London, Cambridge, and New York addresses. | 中 | SI019 |
| CI024 | sUAS News published public list pricing of £2,150 for Aero-Ark CRI, £3,200 for Aero-Ark WiRED, £4,450 for a combined unit, £95 per user per month for Detection as a Service, and £185 per day for Rapid Deploy. | 中 | SI014 |
| CI025 | DroneXL reported Roark's OTC systems starting at £4,500 excluding taxes and bundled with Plexus access. | 中 | SI013 |
| CI026 | DroneLife reported that Roark offered Aero-Ark both for sale and as a service and said the service had units installed in 70 cities worldwide. | 低 | SI012 |
| CI027 | Taken together, public trade coverage frames Roark as pursuing both hardware revenue and recurring service revenue rather than a single pure-software model. | 高 | SI012, SI013, SI014, SI018 |
| CI028 | PitchBook and Crunchbase both track Roark as a profile in mainstream venture databases, but the open web does not expose usable current financial detail from either source. | 中 | SI016, SI017 |
| CI029 | Tracxn describes Roark as a Series B company and reports $287 million of total funding. | 中 | SI015 |
| CI030 | Nordic9 reported a $210 million Series B at a $1.6 billion valuation and said total investments received were $260 million, while labeling the deal pending verification. | 中 | SI011 |
| CI031 | Resilience said Roark circulated a release claiming a $210 million Series B at a $1.6 billion post-money valuation and named only a US-based investment bank via an SPV. | 中 | SI010 |
| CI032 | The public record reviewed here does not identify a full investor roster or complete financing terms for the late-2025 round. | 高 | SI010, SI011, SI015, SI016, SI017 |
| CI033 | Roark had real 2024 revenue but the only filed year shows negative gross profit, so the public record does not support a mature margin story. | 中 | SI001 |
| CI034 | Roark's FY2024 receivables were more than twice annual turnover and year-end cash was only £6.1 million. | 中 | SI001 |
| CI035 | The combination of £13.6 million of stock, £87.1 million of debtors, and only £6.1 million of cash implies capital intensity beyond a pure software model. | 中 | SI001 |
| CI036 | Roark's FY2024 current assets were driven more by receivables than by cash. | 中 | SI001 |
| CI037 | Roark used non-equity capital as well as share issuance, as shown by its bond and finance-lease liabilities. | 中 | SI001 |
| CI038 | Roark's public record does not disclose ARR, backlog, gross margin by product line, CAC, payback, churn, or customer concentration. | 高 | SI001, SI003, SI010, SI011 |
| CI039 | Because filed turnover aggregates UK and US group sales, the public accounts do not let investors underwrite revenue by legal entity or product line. | 高 | SI001, SI019 |
| CI040 | The direct roark-aerospace.com root returned a 404 and the direct /legal page fetch also failed, weakening current first-party disclosure reliability. | 高 | SI018, SI020, SI031 |
| CI041 | Shield AI publicly announced a $240 million financing at a $5.3 billion valuation in March 2025. | 中 | SI021 |
| CI042 | Fortem publicly disclosed a $25 million Lockheed-led tranche and named specific orders and contracts in 2025 and 2026. | 高 | SI022, SI023, SI024, SI025 |
| CI043 | Motorola said it would acquire D-Fend for $1.5 billion and expected D-Fend to generate $185 million of 2026 revenue. | 中 | SI026 |
| CI044 | Axon and Dedrone publicly describe deployments across public safety, airports, critical infrastructure, defense operations, and hundreds of customers worldwide. | 中 | SI027, SI028, SI030 |
| CI045 | Peer disclosure makes Roark's missing customer, backlog, and margin data more material rather than less. | 高 | SI021, SI022, SI023, SI024, SI026, SI027, SI028, SI029 |
| CI046 | Roark's next major official financial checkpoint is the filing of 30 November 2025 accounts due by 31 August 2026. | 高 | SI002, SI003 |
| CI047 | The March 2025, December 2025, January 2026, and April 2026 equity filings prove material capital formation occurred even if the unicorn headline remains partly unverified. | 高 | SI002, SI004, SI005, SI006, SI007, SI008 |
| CI048 | The public record is insufficient to calculate current monthly burn or runway after the late-2025 and early-2026 financings. | 高 | SI001, SI002, SI003, SI010, SI011 |
| CI049 | The £156.96 million cash amount filed on 15 December 2025 is directionally consistent with a very large round but does not fully reconcile to Nordic9's claimed $210 million headline. | 高 | SI005, SI011 |
| CI050 | The 23 December and 28 December 2025 filings show cap-table restructuring around the financing, including major changes in nominal values and class counts with minimal new cash consideration. | 高 | SI006, SI007 |
| CI051 | The 4 April 2026 B ordinary allotment for USD 1,390 appears economically trivial relative to the December 2025 financing signal. | 中 | SI008 |
| CI052 | Fortem said Q3 2025 orders from U.S. allies in Europe and the Middle East roughly doubled versus both the prior quarter and the year-ago quarter. | 中 | SI023 |
| CI053 | Fortem said DHS awarded it a multimillion-dollar FIFA World Cup 2026 contract and the U.S. Army awarded a three-year $18 million counter-drone contract. | 中 | SI024 |
| CI054 | Dedrone says its platform supports over 30 proven integrations and draws on more than 18 million images for AI-based drone detection. | 中 | SI028 |
| CE001 | Roark's archived homepage positioned the company around "Drone Detection as a Service | C-UAS as a Service." | 中 | SE001 |
| CE002 | The archived homepage listed NATO NCAGE U2D39, UEID Y5J7FXLHZAN6, DUNS 231246243, and VAT GB491393565. | 中 | SE001 |
| CE003 | Roark's archived legal page listed London, Cambridge, and New York addresses plus UK registered number 15319796. | 中 | SE002 |
| CE004 | DroneLife described Aero-Ark as a self-contained AI-powered mobile unit that can operate standalone or inside another detection platform. | 中 | SE003 |
| CE005 | DroneLife said Aero-Ark combines forward-scatter detection and packet sniffing. | 中 | SE003 |
| CE006 | DroneLife said the forward-scatter method uses radio-wave changes against LEO and MEO satellites across VHF, UHF, L-band, and GNSS protocols. | 低 | SE003 |
| CE007 | DroneLife said packet sniffing scans 2.4 GHz and 5 GHz channels to identify and track drones. | 中 | SE003 |
| CE008 | DroneLife said Aero-Ark exposes LTE, SATCOM, PoE, and WiFi connectivity and can convert packets into ASTERIX formats. | 中 | SE003 |
| CE009 | DroneLife said authorized clients can deploy counter-UAS measures on a manual or automated basis. | 中 | SE003 |
| CE010 | DroneLife said Roark offered rapid deployment units and an as-a-service model with claimed coverage in 70 cities worldwide. | 低 | SE003 |
| CE011 | sUAS News said Aero-Ark WiRED uses forward-scatter analysis from onboard receivers ingesting LEO and MEO satellite signal data and an onboard FPGA for AI processing. | 中 | SE004 |
| CE012 | sUAS News priced Aero-Ark WiRED at £3,200 ex VAT and said it claims 360-degree coverage up to 120 nautical miles depending on topology and placement. | 低 | SE004 |
| CE013 | sUAS News said Aero-Ark CRI scans 2.4 GHz and 5.8 GHz and includes packet detection on SATCOM, LTE, and LoRa with a claimed range up to 8 km. | 低 | SE004 |
| CE014 | sUAS News priced the combined WiRED and CRI device at £4,450 ex VAT, Detection as a Service at £95 per user per month ex VAT, and Rapid Deploy at £185 per day ex VAT. | 中 | SE004 |
| CE015 | sUAS News said Roark offered optional C-band communications modules and network-disruptor elements for users permitted to deploy them. | 中 | SE004 |
| CE016 | sUAS News said Roark described Detection as a Service as limited in scope but targeted 100 cities globally by the end of 2024. | 低 | SE004 |
| CE017 | DroneXL said the standard Aero-Ark stack includes ADS-B monitoring, Remote ID compliance, DJI packet collection and decryption, and passive radar. | 中 | SE005 |
| CE018 | DroneXL said Roark built an urban Aero-Ark variant using programmable metasurfaces and conical horn transceiver units. | 低 | SE005 |
| CE019 | DroneXL said Roark marketed individual or co-unit systems in portable and sentry configurations with on-device AI processing and classification. | 中 | SE005 |
| CE020 | DroneXL said all Roark detection systems include access to the Plexus data platform, which integrates airspace and ground-level monitoring in real time. | 中 | SE005 |
| CE021 | DroneXL said Roark's WAMI and HSI services covered 30 U.S. cities and 12 U.K. cities. | 低 | SE005 |
| CE022 | DroneXL said Roark's pilot program offered Part 107 sponsorship, free WAMI and HSI cameras, and Plexus access to third-party drone pilots. | 低 | SE005 |
| CE023 | DefenseTech News said Roark integrates sensors, counter-drone tools, and decision support into platforms deployable across land, air, and maritime operations. | 低 | SE006 |
| CE024 | DefenseTech News said Roark's product lineup includes DDaaS, counter-drone weapons, robotics deployment, and continuous imaging through WAMI and hyperspectral cameras. | 低 | SE006 |
| CE025 | DefenseTech News said Plexus AI handles automated threat identification and response coordination. | 低 | SE006 |
| CE026 | DefenseTech News said Roark's command platform operates at the network edge and scans radio frequencies, decrypts communications, and analyzes Doppler shifts. | 低 | SE006 |
| CE027 | Companies House shows Ro-Ark Aerospace Limited was incorporated on 30 November 2023. | 中 | SE007 |
| CE028 | Roark's 2024 filed accounts reported 49 average UK employees, which is the only staffing datapoint in the retrieved evidence set. | 中 | SE009 |
| CE029 | No retrieved Roark source in this run published field reliability, false-positive, uptime, or MTBF metrics. | 中 | SE001, SE002, SE003, SE004, SE005, SE006 |
| CE030 | No retrieved Roark source in this run published a public security, privacy, or product certification page for Plexus, WAMI or HSI, or Aero-Ark. | 中 | SE001, SE002, SE003, SE004, SE005, SE006 |
| CE031 | No retrieved Roark source in this run published named customer case studies or deployment references for the DDaaS or WAMI network. | 中 | SE001, SE002, SE003, SE004, SE005, SE006 |
| CE032 | SRC says mature C-UAS stacks combine radar, electronic warfare, and EO/IR sensors through open interfaces and modular hardware. | 中 | SE010 |
| CE033 | SRC markets those modular systems to military force protection, critical infrastructure, airfields, and VIP or event security. | 中 | SE010 |
| CE034 | Leonardo DRS markets detect-identify-track-defeat systems for Group 1 to 3 UAS and says its tactical radars can track nano drones up to 18 km while handling hundreds of targets. | 中 | SE011 |
| CE035 | Leonardo DRS said its maritime C-UAS package combines radar, EO-IR sensors, networking, the SAGEcore software platform, and a platform-integration kit for ships and autonomous surface vessels. | 中 | SE012 |
| CE036 | DZYNE describes a layered architecture that moves from RF sensing into AI and ML identification and GNSS-spoofing mitigation across portable, mobile, and fixed products. | 中 | SE013 |
| CE037 | D-Fend EnforceAir says its RF-cyber system detects, identifies, locates, tracks, and can take control of rogue drones without depending on Remote ID. | 中 | SE014 |
| CE038 | D-Fend says EnforceAir offers open API integration, multi-use deployment kits, and autonomous or manual modes. | 中 | SE014 |
| CE039 | Dedrone says its AI-driven autonomous C2 uses behavior-model filters, neural networks, an 18-million-image base, and more than 30 integrations. | 中 | SE015 |
| CE040 | Fortem says mature end-to-end C-UAS systems link radars, interceptors, and manager software and are already deployed at venues, military bases, airports, and infrastructure. | 中 | SE016 |
| CE041 | Fortem's Lockheed contract shows competitor offerings are sold as layered systems tied into a larger mission-management software stack. | 中 | SE017 |
| CE042 | Epirus says credible short-range air-defense scale needs open architecture, manufacturing expansion, supply-chain resilience, and one-to-many kill-chain economics. | 中 | SE018 |
| CE043 | DroneShield claims 34 or more government agencies and stresses operator-simplicity, layered defense, and software-enabled detect-track-defeat workflows. | 中 | SE019 |
| CE044 | The Counter UAS Europe 2026 agenda shows the practitioner community is focused on mobile deployable C-UAS, collateral damage, low-cost economics, and autonomous sensor-to-effector chains rather than open-source software ecosystems. | 中 | SE020 |
| CE045 | AIRCOM Industry Day 2026 says NATO is emphasizing interoperability, operational relevance, and the full chain from detection and tracking to mitigation and neutralization. | 中 | SE021 |
| CE046 | Resilience Media said Roark's public narrative and funding or product claims were hard for the outlet to verify. | 中 | SE022 |
| CE047 | D-Fend's December 2024 financing release documented a near-30-country installed base, strong growth, and partner integration efforts for a rival product family. | 中 | SE023 |
| CE048 | Defense News said NATO's vendor pool will pre-qualify static, deployable, and fully mobile systems across point, area, and border defense and then subject them to standardized testing. | 中 | SE024 |
| CE049 | The PR Newswire market article shows promotional investor interest in integrated AI-powered sensing and counter-UAS stacks, but it is not strong execution proof for Roark itself. | 低 | SE025 |
| CE050 | Shield AI frames resilient autonomy as mission critical in contested or degraded environments where communications are lost and operators are overloaded. | 中 | SE026 |
| CE051 | Dedrone says a full counter-drone system should combine RF sensors, radar, PTZ cameras, mobile detection units, mitigation technology, and a software platform that ties them together. | 中 | SE027 |
| CE052 | DZYNE says its counter-UAS products are layered detect-track-identify-defeat systems designed for portable, mobile, and fixed-site deployments. | 中 | SE028 |
| CE053 | The JIATF-401 marketplace says validated counter-UAS systems will be compared using real-world performance data and are expected to integrate sensors, effectors, and mission command into interoperable networks. | 中 | SE029 |
| CE054 | The U.S.-U.K. declaration says common data standards will be incorporated into JIATF-401 marketplace requirements, increasing interoperability pressure on vendors. | 中 | SE030 |
| CE055 | IDGA's 2026 Counter-UAS Summit markets itself as a forum where military, federal, allied, airport, and infrastructure stakeholders align on procurement priorities and technology gaps. | 中 | SE031 |
| CE056 | The 2024 ALSSA counter-unmanned systems study says the largest current gaps are policy, authorities, integration, and procurement rather than a lack of tactical doctrine. | 中 | SE032 |
| CE057 | SRC says Silent Archer uses an open-architecture, sensor-agnostic design, supports multiple configurations, and complements kinetic weapon systems in a layered defense. | 中 | SE033 |
| CE058 | SRC says the Army's LIDS family combines radar, electronic warfare, direction finding, and EO/IR camera systems to detect, track, identify, and defeat Group 1 to 3 UAS. | 中 | SE034 |
| CU001 | Roark’s archived homepage positioned the company around Drone Detection as a Service and C-UAS as a Service. | 中 | SU001 |
| CU002 | Roark’s archived legal page listed London, Cambridge, and New York addresses. | 中 | SU002 |
| CU003 | DRONELIFE reported that Aero-Ark was offered for direct sale, Rapid Deployment service, and an as-a-service model. | 中 | SU004 |
| CU004 | DRONELIFE reported that Roark’s as-a-service model had units installed in 70 cities worldwide. | 中 | SU004 |
| CU005 | sUAS News reported that Roark’s detection-as-a-service coverage was limited in scope but targeted 100 cities globally by the end of 2024. | 低 | SU005 |
| CU006 | DroneXL reported that Roark’s WAMI and HSI services covered 30 U.S. cities and 12 U.K. cities. | 中 | SU006 |
| CU007 | DroneXL reported that Roark offered pilot incentives including RPAS qualifications, free cameras, and Plexus access. | 中 | SU006 |
| CU008 | Tracxn claimed that Roark had more than 55,000 deployed devices and more than 3,000 end users across government, military, and private-sector accounts. | 低 | SU007 |
| CU009 | The retrieved evidence set does not independently corroborate Tracxn’s 55,000-device or 3,000-end-user claims with named references or official operating metrics. | 中 | SU001, SU004, SU005, SU006, SU007 |
| CU010 | The retrieved public Roark and independent sources do not identify a named Roark production customer. | 中 | SU001, SU004, SU005, SU006, SU015 |
| CU011 | The retrieved public sources do not separate Roark pilot accounts from production accounts. | 中 | SU004, SU005, SU006, SU007 |
| CU012 | The supportable Roark buyer archetypes are government or military, public safety or security, critical infrastructure, event-driven users, and broader private monitoring customers. | 中 | SU001, SU004, SU005, SU006, SU007 |
| CU013 | Roark’s public DDaaS price point of £95 per user per month lowers upfront adoption cost relative to owned hardware. | 中 | SU005 |
| CU014 | Roark’s public Rapid Deploy price point of £185 per day fits temporary or event-driven use cases better than long procurement cycles. | 中 | SU005 |
| CU015 | The U.S. Army says its marketplace model is intended to let vendors offer products to soldiers and scale the systems that work. | 中 | SU008, SU009 |
| CU016 | DVIDS says the counter-UAS marketplace is being expanded for allies and partners. | 中 | SU009 |
| CU017 | JIATF-401 says its marketplace uses validated systems and an existing IDIQ vehicle so customers can order faster. | 中 | SU023 |
| CU018 | The U.S.-U.K. counter-drone standards agreement says vendors that adopt the standards will have a more direct path into the JIATF-401 marketplace. | 中 | SU024 |
| CU019 | Defense News reported that NATO planned to select 18 C-UAS solutions across nine use cases. | 中 | SU011 |
| CU020 | Defense News reported that NATO intended both purchase and lease options for its C-UAS marketplace contracts. | 中 | SU011 |
| CU021 | NATO AIRCOM’s 2026 industry day centered on interoperability, operational relevance, and full-spectrum C-UAS capabilities. | 中 | SU010 |
| CU022 | DefenseScoop reported that the FY26 DoD budget request included $3.1 billion for counter-UAS. | 中 | SU013 |
| CU023 | Unmanned Airspace reported that publicly announced government C-UAS contracts exceeded $29 billion in the first months of 2026. | 中 | SU014 |
| CU024 | DHS says its C-UAS Purchasing Tool is designed for first responders and security agencies with expanded authorities. | 中 | SU019 |
| CU025 | DHS says the FY2026 grant cycle awarded an initial $250 million to the National Capital Region and the 11 states hosting FIFA World Cup 2026 matches. | 中 | SU019 |
| CU026 | FAA says public safety agencies are best positioned to deter, detect, and investigate unauthorized or unsafe UAS operations. | 中 | SU021 |
| CU027 | FAA says DOJ, FAA, DHS, and FCC issued guidance to help non-federal public and private entities understand the laws governing detection and mitigation technologies. | 中 | SU020 |
| CU028 | The homeland C-UAS policy fact sheet says commanders can extend defensive actions beyond the physical fence line of an installation. | 中 | SU025 |
| CU029 | The homeland C-UAS policy fact sheet says federal agencies can share UAS-related data more broadly under the updated framework. | 中 | SU025 |
| CU030 | War guidance says trained and certified contractor personnel can be used as C-sUAS operators. | 中 | SU023 |
| CU031 | DoD’s counter-unmanned strategy emphasizes detection, active and passive defenses, and integrated open modular solutions delivered with speed and scale. | 中 | SU026 |
| CU032 | CRS says FY2025 NDAA counter-UAS provisions addressed training, threat libraries, and additional interceptors. | 中 | SU027 |
| CU033 | ALSSA says the largest current gaps in counter-unmanned systems are policy, authorities, integration, and procurement rather than doctrine. | 中 | SU027 |
| CU034 | Dedrone says it protects more than 1,130 sites, 53 airports, and 187 public safety agencies. | 中 | SU016 |
| CU035 | Dedrone publishes named case studies including NASCAR, the Tulsa State Fair, and correctional-agency deployments. | 中 | SU016 |
| CU036 | Dedrone says it does not publicly name most customers but offers confidential customer references by industry. | 中 | SU016 |
| CU037 | Fortem publicly names deployments with U.S. Customs and Border Protection, the Presidential Inauguration, Ukraine, and New Jersey public-safety stakeholders. | 中 | SU018 |
| CU038 | DroneShield publicly targets military, critical infrastructure, public safety, government agencies, and venues or stadiums. | 中 | SU022 |
| CU039 | The retrieved public sources do not disclose Roark’s net revenue retention, gross revenue retention, churn, or renewal rates. | 中 | SU001, SU004, SU005, SU006, SU007, SU015 |
| CU040 | The retrieved public sources do not disclose Roark’s contract length, SLA terms, or customer-satisfaction metrics. | 中 | SU001, SU004, SU005, SU006, SU007 |
| CU041 | The retrieved public sources do not quantify Roark’s top-customer revenue share or channel dependence. | 中 | SU001, SU003, SU004, SU005, SU006, SU007 |
| CU042 | Roark’s customer validation is below the public disclosure standard shown by named-deployment competitor pages and case studies. | 中 | SU015, SU016, SU018, SU022 |
| CU043 | Roark’s low-friction pricing and service packaging fit trial or niche deployments better than fully evidenced program-of-record sales given the current proof set. | 中 | SU005, SU011, SU019, SU023, SU024 |
| CU044 | The most valuable next diligence items are named customer references, contract documents, cohort metrics, and concentration data. | 中 | SU015, SU016, SU019 |
| CR001 | Companies House shows Ro-Ark Aerospace Limited was incorporated on 30 November 2023 and has its next accounts due by 31 August 2026. | 中 | SR003 |
| CR002 | Companies House explicitly says it does not check the accuracy of information filed on the company overview page. | 中 | SR003 |
| CR003 | Roark's filing history shows material capital events in March 2025, December 2025, January 2026, and April 2026 rather than a single fully transparent round document. | 中 | SR004 |
| CR004 | Roark's 2024 filed accounts reported £36.9 million of turnover and a £16.56 million pre-tax loss for the period ended 30 November 2024. | 中 | SR007 |
| CR005 | The 2024 accounts say turnover equated to total group sales across both UK and US entities, limiting how precisely outsiders can map reported revenue to one legal entity or product mix. | 中 | SR007 |
| CR006 | Companies House lists Patton Atkinson-French as Roark's sole current director. | 中 | SR005 |
| CR007 | Companies House says Patton Atkinson-French holds 75% or more of shares and voting rights and can appoint or remove directors, concentrating formal control. | 高 | SR005, SR006 |
| CR008 | Resilience Media reported that Roark publicized a $210 million Series B at a $1.6 billion post-money valuation while naming only a U.S.-based investment bank via an SPV. | 中 | SR001 |
| CR009 | Nordic9 also listed a $210 million Series B and $260 million total funding but labeled the deal still pending verification. | 中 | SR002 |
| CR010 | Roark's April 2026 SH01 filing says B ordinary shares carry a liquidity preference over the ordinary share class. | 中 | SR008 |
| CR011 | Roark's archived legal page disclosed only a contact email, two UK addresses, one U.S. address, and the UK registration number rather than a substantive compliance or trust package. | 中 | SR009 |
| CR012 | Roark's archived homepage positioned the company around Drone Detection as a Service and C-UAS as a Service and exposed NCAGE, UEID, DUNS, and VAT identifiers. | 中 | SR010 |
| CR013 | The open web shows real financing activity but still does not publicly verify an investor roster or an audited bridge from filed allotments to the unicorn valuation headline. | 中 | SR001, SR002, SR004, SR008 |
| CR014 | FAA public-safety guidance says law-enforcement and public-safety agencies must understand safe drone operations and their legal authority when handling unauthorized UAS activity. | 中 | SR013 |
| CR015 | JIATF-401 said updated homeland guidance now operates under 10 U.S. Code § 130i, expands defensive perimeters, and treats unauthorized surveillance of designated facilities as a threat. | 中 | SR014 |
| CR016 | JIATF-401 said the updated policy also authorizes trained and certified contractor personnel as counter-UAS operators and links commanders to a marketplace of tested and validated materiel. | 中 | SR014 |
| CR017 | The DHS C-UAS Purchasing Tool says first responders need mission-specific rubrics to compare vendor offerings before selecting detection, tracking, identification, and mitigation systems. | 中 | SR015 |
| CR018 | CISA guidance says buyers should set detection requirements from threat, facility, airspace, and resource context and then integrate detection into layered security plans. | 中 | SR016 |
| CR019 | The DHS C-UAS privacy assessment says testing, evaluation, and operational deployment must mitigate privacy risks and protect personally identifiable information. | 中 | SR031 |
| CR020 | OMB's secure-UAS memo says insecure foreign-manufactured UAS can create data-breach, hidden-feed, and supply-chain risks for federal operations. | 高 | SR017, SR030 |
| CR021 | OMB's memo says agencies may consider particular domestic sources and even noncompetitive acquisition procedures when domestic UAS capability is necessary. | 中 | SR017 |
| CR022 | The FAA ARC recommended minimum performance standards, approved third-party testing, and a continually evolving approved list of technologies or vendors for UAS detection and mitigation systems. | 中 | SR018 |
| CR023 | The FAA ARC also said airports lack mitigation authority today and need clear approval paths, operator training, and privacy protections before broader deployment. | 高 | SR018, SR013 |
| CR024 | CRS described counter-UAS authorities, training, threat-library work, and enterprise assessments as active NDAA subjects, showing that usable authority remains legislated rather than automatic. | 中 | SR019 |
| CR025 | BIS highlights Export Administration Regulations, item classification, country guidance, and screening lists as live compliance tools for export-sensitive technology. | 中 | SR030 |
| CR026 | Roark's public materials do not disclose an export-control program, domestic-source plan, or named security architecture despite its UK and U.S. defense-facing positioning. | 中 | SR009, SR010, SR017, SR030 |
| CR027 | The combination of sparse company disclosures and policy-heavy buyer requirements creates a disclosure-openness gap at exactly the point where Roark needs higher trust. | 中 | SR009, SR010, SR015, SR017 |
| CR028 | Defense Tech News described Roark as combining sensors, counter-drone tools, and decision support into unified land, air, and maritime platforms. | 中 | SR012 |
| CR029 | DroneXL said Roark's OTC stack included ADS-B monitoring, Remote ID, DJI packet decryption, passive radar, and on-device AI. | 中 | SR011 |
| CR030 | DroneXL said Roark expanded WAMI and hyperspectral coverage to 30 U.S. cities and 12 U.K. cities while recruiting third-party pilots with certifications, cameras, and Plexus access. | 中 | SR011 |
| CR031 | Roark's 2024 accounts disclosed an average of 49 UK-based employees and explicitly excluded contractors and international staff from that count. | 中 | SR007 |
| CR032 | The public record does not identify Roark's core suppliers, contract manufacturers, cloud providers, or redundancy arrangements. | 中 | SR007, SR009, SR010 |
| CR033 | Motorola said D-Fend had thousands of deployments in more than 30 countries and expected 2026 revenue of about $185 million in a $1.5 billion acquisition. | 中 | SR024 |
| CR034 | Axon said Dedrone served airports, critical infrastructure, public events, and defense operations and is now part of Axon's wider public-safety platform. | 中 | SR025 |
| CR035 | Fortem said its SkyDome business had momentum across government and allied use cases, illustrating that credible peers publish more operating proof than Roark does. | 中 | SR026 |
| CR036 | Shield AI said Hivemind Enterprise is built from autonomy software already proven in GPS- and communications-denied environments. | 中 | SR027 |
| CR037 | Epirus said it raised $250 million to scale Leonidas production capacity and supply-chain resilience for layered defense. | 中 | SR028 |
| CR038 | DroneShield maintains an investor-relations surface as a public counter-UAS pure play, underscoring the disclosure standard available from better-documented peers. | 中 | SR029 |
| CR039 | Compared with those peers, Roark has not publicly provided equivalent named deployments, validated performance metrics, or governance-grade disclosure. | 中 | SR001, SR011, SR012, SR024, SR025, SR026, SR027, SR028, SR029 |
| CR040 | The official procurement path is expanding, but it increasingly rewards validated, secure, authority-aware vendors rather than companies with only broad marketing claims. | 中 | SR014, SR015, SR016, SR018, SR020, SR021, SR022 |
| CR041 | Roark's public story depends on stitching together RF detection, edge processing, WAMI and HSI feeds, and a partner-assisted operating network rather than on a single simple product. | 中 | SR011, SR012 |
| CR042 | The combination of an unverified headline valuation, undisclosed investor roster, and preference-bearing share classes creates a financing-credibility risk for new common-equity investors. | 中 | SR001, SR002, SR008 |
| CR043 | Single-controller governance, thin public trust documentation, and an undisclosed supplier stack make key-person and operational-resilience risk hard to underwrite. | 中 | SR005, SR006, SR009, SR010 |
| CR044 | Roark's public materials still do not identify a named production customer or validated deployment reference strong enough to offset integration and execution risk. | 中 | SR010, SR011, SR012, SR024, SR025 |
| CR045 | If Roark cannot produce validated test results, named qualified deployments, and secure-data controls, the procurement pathways described by official sources are likely to stay only partially accessible. | 中 | SR015, SR016, SR017, SR018, SR022 |
| CR046 | No retained public source discloses current cash, runway, or a post-2025 balance-sheet bridge, so the next official checkpoint remains the accounts due on 31 August 2026. | 中 | SR003, SR004, SR007 |
| CR047 | The thesis breaks if upcoming disclosures fail to substantiate liquidity, secure-supply compliance, named validated deployments, and the economic logic behind the claimed valuation. | 中 | SR003, SR017, SR018, SR024 |
| CV001 | Roark Aerospace is an active UK company incorporated on 30 November 2023. | 中 | SV001 |
| CV002 | Roark's filed 2024 accounts report £36,906,898 of turnover. | 中 | SV003 |
| CV003 | The same 2024 accounts report a pre-tax loss of £16,561,770. | 中 | SV003 |
| CV004 | The 2024 accounts imply negative gross profit because cost of sales exceeded turnover. | 中 | SV003 |
| CV005 | A March 2025 SH01 allotted 81,000,000 ordinary shares and 19,000,000 preference shares at £1.25 nominal value per share. | 中 | SV004 |
| CV006 | That March 2025 preference class carried block voting rights, full-ratchet anti-dilution, and preferential dividend status. | 中 | SV004 |
| CV007 | The 15 December 2025 SH01 shows 131 B ordinary shares allotted for cash with £156,963,450 paid. | 中 | SV005 |
| CV008 | The 23 December 2025 SH01 restated the capital base to 1,131,000 total shares while preserving £1,131 aggregate nominal value. | 中 | SV006 |
| CV009 | The 28 December 2025 SH01 added 221,339 preference shares with dividend and liquidity preference rights. | 中 | SV007 |
| CV010 | The 4 April 2026 SH01 added 20,079 B ordinary shares and lifted the total share count to 1,372,418. | 中 | SV008 |
| CV011 | The PSC register still shows Patton Atkinson-French controlling more than 75% of shares and voting rights. | 中 | SV009 |
| CV012 | Companies House lists Patton Atkinson-French as the sole active officer. | 中 | SV010 |
| CV013 | Roark's archived homepage described the company as Drone Detection as a Service and C-UAS as a Service. | 中 | SV015 |
| CV014 | DroneXL reported Roark's OTC systems starting at about £4,500 excluding taxes. | 中 | SV016 |
| CV015 | DefenseTech.News described Roark as operating a large drone-detection network and cited roughly €57.6 million raised since founding. | 中 | SV017 |
| CV016 | Nordic 9 reported a $210 million Series B at a $1.6 billion valuation and explicitly labeled the deal pending verification. | 中 | SV011 |
| CV017 | Resilience Media said Roark circulated a release claiming a $210 million Series B at a $1.6 billion post-money valuation. | 中 | SV012 |
| CV018 | Resilience reported that the only investor named in the circulated round narrative was a US-based investment bank via an SPV. | 中 | SV012 |
| CV019 | Tracxn tracks Roark as a Series B company and says it has raised $287 million in funding. | 中 | SV013 |
| CV020 | PitchBook publicly exposes Roark as a valuation and funding profile but keeps the substantive financial fields behind a gate. | 中 | SV014 |
| CV021 | Taken together, the December 2025 filings prove that a very large financing event occurred even though they do not identify the full investor roster. | 中 | SV005, SV006, SV007, SV012 |
| CV022 | If the £156.96 million paid line is paired with the later 131,000 B-ordinary count after the 1,000-for-1 redenomination, the implied price is roughly £1,198 per B share. | 中 | SV005, SV006 |
| CV023 | Using the 1,131,000 post-23-December total share count, that implied B-share price indicates a post-money value near £1.36 billion before later preference and April issuance. | 中 | SV005, SV006 |
| CV024 | A post-money value near £1.36 billion is directionally consistent with a publicly circulated $1.6 billion late-2025 valuation headline. | 中 | SV005, SV006, SV011 |
| CV025 | The public record still does not provide a signed cap table, full investor list, or complete rights package for the claimed Series B. | 中 | SV012, SV014, SV002 |
| CV026 | The March 2025 anti-dilution and preference terms indicate that dilution and liquidation preferences matter to any new investor even before the claimed Series B. | 中 | SV004 |
| CV027 | The December 2025 and April 2026 filings show that Roark's capital stack changed repeatedly over a four-month span, complicating precise fully diluted valuation work. | 中 | SV005, SV006, SV007, SV008 |
| CV028 | Because Roark's only filed revenue year shows negative gross profit, a premium software-style revenue multiple would require unproven future margin repair. | 中 | SV003 |
| CV029 | Shield AI announced a $240 million funding round in March 2025 at a $5.3 billion valuation. | 中 | SV018 |
| CV030 | Shield AI's disclosed use of proceeds and deployed autonomy proofs set a higher evidence bar than Roark's current public package. | 中 | SV018, SV012 |
| CV031 | Epirus announced a $250 million Series D in March 2025 that brought total venture funding above $550 million. | 中 | SV019 |
| CV032 | Epirus did not disclose a public valuation in the cited release, limiting its usefulness as a hard pricing comp. | 中 | SV019 |
| CV033 | D-Fend announced a $31 million financing round in December 2024. | 中 | SV020 |
| CV034 | Motorola agreed in June 2026 to acquire D-Fend for $1.5 billion and said D-Fend expected full-year 2026 revenue of $185 million. | 中 | SV021 |
| CV035 | Motorola's announced D-Fend price equates to roughly 8.1x expected 2026 revenue. | 中 | SV021 |
| CV036 | Axon's acquisition of Dedrone confirms strategic buyer appetite for airspace-security platforms even when price is undisclosed. | 中 | SV022 |
| CV037 | Lockheed Martin's $25 million investment in Fortem in April 2026 shows strategic capital continuing to flow into counter-UAS platforms tied to prime ecosystems. | 中 | SV023 |
| CV038 | DroneShield describes itself as the only publicly listed company exclusively dedicated to counter-UAS. | 中 | SV024 |
| CV039 | DroneShield's public market capitalization was about $1.77 billion in June 2026. | 中 | SV025 |
| CV040 | Leonardo DRS's public market capitalization was about $12.29 billion in June 2026. | 中 | SV026 |
| CV041 | Motorola Solutions' public market capitalization was about $65.59 billion in June 2026. | 中 | SV027 |
| CV042 | L3Harris Technologies' public market capitalization was about $54.92 billion in June 2026. | 中 | SV028 |
| CV043 | Lockheed Martin's public market capitalization was about $117.80 billion in June 2026. | 中 | SV029 |
| CV044 | Thales' public market capitalization was about $54.81 billion in June 2026. | 中 | SV030 |
| CV045 | Northrop Grumman's public market capitalization was about $74.07 billion in June 2026. | 中 | SV031 |
| CV046 | BAE Systems' public market capitalization was about $71.68 billion in June 2026. | 中 | SV032 |
| CV047 | SAAB AB's public market capitalization was about $28.35 billion in June 2026. | 中 | SV033 |
| CV048 | Rheinmetall's public market capitalization was about $64.41 billion in June 2026. | 中 | SV034 |
| CV049 | Compared with DroneShield's roughly $1.77 billion public market value, Roark's $1.6 billion headline already assumes near-peer credibility despite much thinner disclosure. | 中 | SV011, SV012, SV025 |
| CV050 | Compared with Motorola's 8.1x expected-revenue D-Fend acquisition, Roark's headline price is hard to justify from only £36.9 million of filed turnover and negative gross profit. | 中 | SV003, SV021 |
| CV051 | A base-case Roark valuation is better framed as a broad range than as a point estimate because the financing signal is stronger than the operating disclosure. | 中 | SV003, SV005, SV006, SV012 |
| CV052 | A bear case of roughly $0.3 billion to $0.6 billion reflects downside if 2024 revenue quality proves weak, preference rights bite harder than expected, or current operating proof does not catch up with the round narrative. | 中 | SV003, SV004, SV007, SV012 |
| CV053 | A base case of roughly $0.8 billion to $1.2 billion reflects some credit for the late-2025 financing signal but still discounts the opaque rights stack and margin uncertainty. | 中 | SV003, SV005, SV006, SV012, SV025 |
| CV054 | A bull case of roughly $1.4 billion to $1.8 billion requires the December 2025 financing to represent a legitimate third-party mark and for management to substantiate stronger 2026 operating quality than the filed 2024 accounts show. | 中 | SV003, SV005, SV006, SV011, SV013 |
| CV055 | The most defensible recommendation on current public evidence is research-more rather than buy, because price support exists in the filings but governance, customer proof, and current economics are still under-disclosed. | 中 | SV003, SV005, SV006, SV012, SV013, SV021 |
| CV056 | The valuation stance is stretched rather than clearly expensive because primary filings partially triangulate the magnitude of the round even while the narrative remains thinly sourced. | 中 | SV005, SV006, SV011, SV012 |
| CV057 | The most plausible way for Roark to move from stretched to fair is to deliver a signed cap table, current revenue-quality evidence, named customers, and a transparent preference bridge. | 中 | SV003, SV012, SV014 |
| CV058 | Without that diligence package, any 2026 entry at or above the circulated late-2025 headline should be treated as a high-burden underwriting decision rather than a routine growth-round follow-on. | 中 | SV011, SV012, SV014 |