Cowboy Space Corporation
资金充足的轨道计算论点,B 轮融资估值 $2B,但尚无公开商业验证。
Cowboy Space 资金充足,但轨道计算仍停留在商业化前的投资逻辑:$2B B 轮融资确实落地,收入、客户和核心执行能力仍未得到证明。
覆盖范围与披露说明
截至 2026-10-17 对上一版 Cowboy Space 报告的更新。公开证据仍未补齐收入、客户、执行和股权结构证明。
封面要素
公司概况
Cowboy Space Corporation 是一家总部位于 San Carlos 的轨道基础设施初创公司,2026 年 5 月由 Aetherflux 更名而来,如今主打贯穿发射、轨道电力和在轨 AI 计算的垂直整合架构。公司宣布以 $2B 估值完成 $275M B 轮融资,但公开记录仍显示其处于商业化前阶段:核心 1 MW 轨道数据中心概念尚无披露收入、具名客户或运营证明。
- 总部
- San Carlos, California, USA
- 产品
- 一套垂直整合的轨道计算平台,计划用专门设计的运载工具把 AI 计算能力送入轨道,并用太空太阳能供电。
- 客户
- AI 基础设施买家、主权及国防计算用户,以及受益于地外处理能力的太空原生工作负载。
- 商业模式
- 当前未披露;公司所述长期模式是销售轨道计算容量,以及相关的发射和供电基础设施服务。
- 阶段
- Series B private company
- 融资情况
- 2026 年 5 月宣布以 $2B 估值完成 $275M B 轮融资;此前披露的股权融资约 $325M,追踪机构统计总额超过 $335M。
执行摘要
主要优势
- 机构投资人持续跟投,且估值达到 $2B,融资信号明确
- 产品架构具体,把发射、电力和计算整合在一起
- AI 基础设施瓶颈带来强劲的宏观需求驱动
主要风险
- 没有公开披露收入、ARR、客户或定价
- FCC 审查、热环境验证和发射执行仍未解决
- 估值跑在经营验证和股权结构透明度之前
未决问题
- 没有公开披露收入、ARR、订单积压、意向书或具名客户
- 热真空鉴定和抗辐射方案仍未披露
- 火箭发动机设计、推进测试、制造设施以及 FAA 发射许可证路径仍未披露
- 完整股权结构表、优先权安排以及任何债务或项目融资结构仍未披露
目录
01公司概览
1.1 身份与产品事实
Cowboy Space 如今将自己定位为垂直整合的轨道基础设施公司,而不只是太空太阳能概念。官网和 2026 年 5 月融资公告对后续章节可复用的稳定事实基本一致:公司 2024 年以 Aetherflux 名义成立,2026 年 5 月 11 日更名为 Cowboy Space,总部在 San Carlos,并在 Seattle 设有办公室;公司主张把发射、供电和在轨计算做成一套系统。产品层面最扎实的事实是架构,而非商业化:火箭上级预计会在轨道上变成一座 1 MW 数据中心,NVIDIA 是具名计算模块合作方。这足以定义商业模式,但仍只是设计目标,不是已证明的运营资产。[CO001, CO002, CO003, CO004, CO005]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 / 尽调备注 |
|---|---|---|---|---|
| 公司身份 | Cowboy Space Corporation(原名 Aetherflux) | 2026-05-11 | 高 | 更名日期已公开;确认是否仍有历史合同使用 Aetherflux 名称。 |
| 总部 / 办公室 | San Carlos 总部;Seattle 另设办公室 | 当前 | 高 | 公开来源确认办公地点,不确认各地点团队规模。 |
| 阶段 | B 轮私营公司 | 2026-05-11 | 高 | 当前阶段清晰,但董事会 / 控制权未公开说明。 |
| 最新估值 | $2B 投后,公司公告 | 2026-05-11 | 高 | 留存来源没有独立二级价格、要约收购或加价披露。 |
| 已披露融资额 | $325M 已披露股权 / $335M+ 追踪机构总额 | 2026-05 to 2026-10 | 中 | 各总额对追踪机构报告的赠款或非股权项目是否计入存在冲突。 |
| 收入 / ARR | 当前 | 低 | 索取经审计财务报表或 KPI 包;留存公开来源不支持任何数值。 | |
| 客户数 | 当前 | 低 | 索取已签客户清单、LOI 或积压订单表,不要从行业需求推断。 | |
| 员工人数 | 当前 | 低 | 索取组织架构或薪资快照;公开页面描述背景,不提供员工人数。 | |
| 首次自研发射目标 | 2028 年底前 / 2028 年目标 | 2026-05 | 中 | 仅为公开目标;确认推进、制造和许可关键路径。 |
null 表示截至 2026-10-17,留存公开来源集不支持精确数值;应按列出的尽调要求跟进,不要自行推断数字。
[CO001, CO002, CO006, CO007, CO008, CO009, CO013]1.2 领导层、治理与关键人物依赖
公开领导层仍高度集中在创始人兼 CEO Baiju Bhatt 身上。公司发布的融资材料、独立报道和追踪页面都把 Bhatt 放在产品论点、融资和更名叙事的中心,形成明显的关键人物依赖。留存来源中最有增量的团队证据偏法律和运营,而非广泛的高管搭建:SatNews 将 Joseph Yaffe 列为 COO 兼首席法务官,Silicon Valley Invest Club 列出了 2026 年 3 月和 4 月新增的顾问。这说明创始人周围的支持在扩大,但公开记录仍未披露完整董事会名单、表决结构或 B 轮后的控制权图谱。[CO010, CO014, CO015]
| 人员 | 职务 | 背景 | 创始人与市场匹配度或职能覆盖 | 关键人物依赖 |
|---|---|---|---|---|
| Baiju Bhatt | 创始人兼 CEO | Robinhood 联合创始人;更名、融资和产品定位的公开代言人 | 为轨道计算论点提供强资本市场信誉和第一性原理的产品主导 | 高:战略、融资叙事和外部信誉仍集中在 Bhatt 身上 |
| Joseph Yaffe | COO 兼首席法务官 | 2026 年 1 月从 Skadden 加入,此前负责该所 Palo Alto 办公室 | 为许可和执行补充法律、监管及运营深度 | 中:团队在扩充,但公开证据仍显示高管层面集中 |
| Matthew Seaford | 顾问委员会 | 追踪页面称其为安全与国防专家 | 体现公司对国防 / 政府定位的兴趣,而非日常运营能力 | 直接依赖度低;顾问角色无法补齐治理或执行缺口 |
| Brad Koenig | 顾问委员会 | 追踪页面列出其 Goldman Sachs 和 NIST-CHIPS 经历 | 增加政策、半导体和资本市场模式识别能力 | 直接依赖度低;顾问角色提供支持,不是已披露的控制职能 |
公开点名的领导层并不完整,且把运营职位与顾问增补混在一起;留存来源没有披露完整董事会、汇报结构或 B 轮后的治理文件包。
[CO010, CO014, CO015]1.3 融资历史与利益相关方结构
公开融资证据集中在 2026 年 5 月 B 轮,对总资本额的证据则弱得多。官方及独立报道都确认 Cowboy 宣布以 $2B 估值完成由 Index Ventures 领投的 $275M B 轮,但各来源对累计融资额存在差异:SatNews 引用约 $325M 的已披露股权融资,追踪类来源则显示三轮融资及一笔赠款合计超过 $335M。实际含义是,本轮融资真实且信号明确,但要做股权结构分析,仍需管理层直接材料来区分定价股权、赠款、创始人参与和任何侧袋载体。利益相关方的重要性也不止于投资人:NVIDIA 是技术可信度伙伴,FCC 则是近期监管把关者。[CO006, CO007, CO008, CO009, CO011]
| 利益相关方 | 角色 | 控制权或经济重要性 | 尽调要求 |
|---|---|---|---|
| Index Ventures | 领投方 | 同时领投 A 轮和 B 轮,是最清晰的持续外部资本信号 | 确认董事会席位、按比例配售权,以及为未来发射 / 基础设施资本开支提供过桥资金的意愿 |
| Baiju Bhatt | 创始人兼投资人 | 官方融资材料显示其既是创始人,也是参与投资的投资人,这可能集中控制权和经济利益 | 确认持股比例、表决结构,以及任何超额表决权或转让限制 |
| Breakthrough Energy Ventures / Construct / NEA / Interlagos | 老股机构支持者 | 重复参与说明论点背后有连续支持,但公开记录没有披露支票金额或条款 | 按投资人索取分配额、储备水平和任何内部人主导的轮次机制 |
| NVIDIA | 技术合作方 | 具名的 Space-1 合作方,即使没有披露收入,也为产品叙事提供重要可信度杠杆 | 确认关系属于战略营销、供应配额,还是更深层的商业承诺 |
| FCC | 监管把关者 | 星座许可和豁免处理决定轨道数据中心路线能否脱离纸面状态 | 确认申请状态、豁免获批概率,以及所需后续文件或修改 |
| 潜在客户 | 商业证明群体 | 留存来源没有点名已签客户、积压订单或有定价的需求,因此 go-to-market 验证仍弱 | 在评估收入时点前索取 LOI、设计合作方或已签约工作负载 |
本图谱有意混合资本提供方、战略技术伙伴、监管机构和需求侧交易对手,因为四者都会影响 Cowboy 当前阶段的控制权、执行速度或下行风险。
[CO005, CO006, CO009, CO012]1.4 里程碑与执行观察点
公开里程碑轨迹异常紧凑:Cowboy 从 2024 年成立,经历种子轮和 A 轮融资、2026 年初增补顾问委员会、2026 年 5 月更名及 B 轮融资,几天后又为 20,000 颗卫星的 “Stampede” 星座向 FCC 提交文件。这一顺序证明势头真实,但也把核心尽调问题从“有没有兴趣”提前推进到“项目能否执行”。留存的独立报道把首次自研发射定在 2028 年目标,并反复指出同一批未解决问题:真空中 1 MW 载荷的散热、大型星座的碎片暴露风险,以及仍未完成、还需监管跟进的卫星设计。[CO006, CO011, CO012, CO013, CO016]
公开记录显示品牌、融资和监管进展迅速加快,但执行验证仍未完成。
以 Aetherflux 名义成立公司 [CO001]
公开融资追踪记录出现种子轮 [CO011]
公开融资追踪记录出现资助 [CO011]
Index Ventures 领投 A 轮 [CO011]
Matthew Seaford 加入顾问委员会 [CO010]
Brad Koenig 加入顾问委员会 [CO010]
品牌重塑,并以 $2B 估值完成 $275M B 轮 [CO005, CO006]
为 Stampede 星座提交 FCC 申报 [CO012]
COO / CLO 职位公开出现 [CO014]
02市场分析
2.1 市场边界与现状替代方案
可用市场边界比“太空基础设施”更窄,又比单一卫星项目更宽。Cowboy 的公开叙事定义了一套垂直整合的轨道计算系统,其中发射工具、上级数据中心、轨道太阳能和光学网络都属于同一产品。这个定位很关键,因为现有替代方案不是另一家发射初创公司,而是争夺电网并网、冷却、土地和许可资源的地面 AI 容量扩张。因此,本章把轨道计算视为对 AI 基础设施瓶颈的回应,而非泛化的太空服务市场。纳入支出聚焦于计算容量交付及运营所需的物理系统,排除普通卫星通信、地球观测和来源未显示 Cowboy 直接服务的软件型 AI 类别。[CM001, CM002, CM003, CM007]
| 细分市场 / 类别 | 纳入支出 | 排除支出 | 买家 / 付款方 | 相关性 |
|---|---|---|---|---|
| 轨道 AI 计算系统 | 发射工具、上级计算节点、轨道太阳能、光学网络和运营 | 通用发射经纪、普通卫星通信带宽和纯软件 AI 工具 | 如果架构跑通,则为超大规模云厂商或主权计算基础设施预算 | Cowboy 公开架构定义的直接类别 |
| 地面 AI 容量替代方案 | 新建 AI 场地的电力、并网、土地、冷却、用水和许可成本 | 非基础设施企业软件预算 | 数据中心开发商、超大规模云厂商和基础设施采购团队 | Cowboy 声称正在绕开的现状基线 |
| 太空能量传输相邻市场 | 先于计算部署的共享太阳能和轨道能源架构 | 公用事业规模的地面发电市场 | 研发及任务开发预算 | 仅因同一平台支撑后续计算主张而相关 |
| 太空到地面的数据传输 | 让轨道计算可用所需的光链路和地面连接 | 消费宽带和通用电信接入 | 平台运营商及购买服务质量的终端客户 | 必要的赋能环节,而非独立需求池 |
纳入 / 排除支出来自来源对架构和替代方案的描述;这是边界定义,不是量化收入模型。
[CM001, CM002, CM003, CM007]2.2 受证据约束的市场测算视角
公开证据提供了测算视角,却没有形成干净的市场模型。一篇独立行业文章称,超大规模云厂商计划在 2026 年投入近 $700B 建设数据中心,这可作为宏观需求信号,却不能直接视为 Cowboy 的可服务支出。公司及文件报道随后把披露机会收窄到拟建的 20,000 颗卫星星座:卫星位于晨昏太阳同步轨道,代表性节点重量为 20,000-25,000 千克,约有 1 MW 可用功率和略低于 800 个 GPU。这些数字描述的是雄心和节点设计,不是已签订单或实现的服务收入。结论是一个受证据约束的图景:问题规模显然很大,但 Cowboy 能赢得的份额尚无法量化,因为公开材料没有把宏观资本开支或物理架构连接到有定价的商业工作负载。[CM004, CM005, CM006, CM014]
| 发布方 / 视角 | 年份 | 地理范围 | 数值 | CAGR | 方法 | 置信度 | 局限 |
|---|---|---|---|---|---|---|---|
| The Robotics Media / 宏观需求视角 | 2026 | 全球 | 近 $700B 的超大规模云厂商数据中心容量支出 | 以贸易报告综合分析作为陆地 AI 资本支出压力的信号 | 中 | 宏观资本支出信号,不代表 Cowboy 的可服务市场 | |
| 新太空经济 / 星座上限 | 2026 | 全球 LEO | 在 700-1,000 km 晨昏 SSO 轨道上最多部署 20,000 颗卫星 | FCC 备案叙述和技术附件摘要 | 中 | 仍在审查中的方案,不是获批部署 | |
| SatNews / 节点容量视角 | 2026 | 单颗卫星 | 20,000-25,000 kg、1 MW,以及略低于 800 个 GPU | 对 Cowboy 已发布架构及 NVIDIA 模块配对的独立报道 | 中 | 工程目标,不代表实际运营容量 | |
| Satellite Today / 最低服务视角 | 2026 | 首次部署 | 最少一颗卫星即可开展商业运营 | 贸易报道引用的公司监管论点 | 中 | 最小可行部署,不证明客户需求 |
公开证据混合了资本支出、星座规模和单节点设计目标;现有留存来源没有披露定价、客户转化,或同口径的 TAM/SAM/SOM 模型。
[CM004, CM005, CM006, CM012, CM014]公开数字指标从宏观 AI 算力需求一路延伸到 Cowboy 称可启动服务的最小部署规模,但它们并不能组成可比的 TAM/SAM/SOM 框架。
各层宽度仅表示层级,不代表数值比例。
这是一个从宏观需求到最小部署、按证据约束排序的观察框架。各层单位不同,不能相加为 TAM、SAM 或 SOM。
[CM004, CM005, CM006, CM012, CM014]2.3 买家、用户与采用路径
来源暗示了可能的客户类别,却没有点名客户。公开评论把超大规模云厂商和 AI 云服务商视为最直接受电力和并网约束影响的预算方;最合理的早期工作负载则是对延迟不敏感或太空原生的应用,例如国防、地球观测、主权计算、灾害响应和科学处理。这与已披露的推进路径一致:2026 年下半年进行太空到地面的能量传输演示,2027 年上半年完成端到端计算任务,并在 2028 年底前争取首次集成自研火箭和数据中心发射。因此,采用故事是分阶段、靠可信度推进的:公司必须先证明架构能工作,更多买家才会把轨道计算视为基础设施采购选项。[CM008, CM009, CM010, CM012]
| 细分市场 | 买方 | 使用者 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 超大规模容量缓解 | 超大规模云厂商基础设施组织 | ML 平台和计算容量团队 | 中央数据中心资本支出委员会 | 当陆地互联排队拖慢新建项目时增加容量 | 基础设施和容量规划负责人 | 电力、制冷或许可审批出现多年瓶颈 |
| 政府及国防韧性计算 | 机构或主承包商项目办公室 | 任务分析团队 | 项目预算发起人 | 在轨道上运行对延迟容忍或由韧性需求驱动的工作负载 | 任务或项目经理 | 对主权计算或难以被干扰的计算能力有战略需求 |
| 太空原生数据处理 | 太空运营商或分析服务商 | 在轨和地面处理团队 | 任务运营预算 | 把轨道传感器或自主运行工作负载放在更靠近源数据的位置处理 | 产品或任务运营负责人 | 仅依靠地面处理带来的高下行链路成本或运营延迟 |
| 科研或灾害响应推理 | 公共部门或科研联盟买方 | 科学家或应急响应分析师 | 资助、任务或公共用途资金来源 | 围绕轨道感知或远程运营开展突发推理 | 项目发起人 | 延迟容忍度比最低计算成本更重要的用例 |
买方、付款方和触发因素均根据留存来源所述用例及约束叙述推断得出,因为现有留存来源没有点名已签约客户或商业定价。
[CM008, CM009, CM010]2.4 限制近期市场兑现的因素
对市场叙事最有力的反作用是,核心工程和监管约束在每个独立来源中都清晰可见。真空散热、辐射耐受、光学网络可靠性、发射经济性、碎片暴露和未完成的卫星设计,都还是闸门问题,而不是已解决的实现细节。即使更乐观的报道也把市场描述为取决于发射频率和技术闭环,而非一个已有披露需求的现成采购类别。这使 Cowboy 的市场论点在战略上有吸引力,但仍处于论证阶段:公司解释了客户为什么可能在意,却没有公开证据证明系统脱离概念阶段后谁会付费、合同条件是什么或服务水平能达到何种程度。[CM011, CM012, CM013, CM014]
2.5 展示材料
03竞争格局
3.1 竞争集合:替代方案比直接同行更清晰
现有资料更有力地支持一个宽泛的替代方案集合,而不是一份紧凑的具名同行清单。Cowboy 自己的材料把公司定位为发射、供电和轨道计算一体化架构;独立报道则反复把真正的现有替代方案指向受电网、用水和许可瓶颈约束的地面 AI 基础设施,而不是成熟的轨道竞争者。资料中最清晰的外部具名参照来自 Caplight 的可比公司列表,其中有 Starcloud、Relativity Space 和 Orbital Inc.,但披露的运营细节不足,不能视为已验证的正面竞争者。实际来看,这里的公开记录说明买家现在可以买到地面计算,也能观察到仍在申请阶段的轨道新进入者,却还不能把 Cowboy 与一组披露充分的轨道供应商直接比较。[CP001, CP006, CP010, CP014]
| 替代方案 | 类别 | 公开规模 / 融资信号 | 目标客户或工作负载 | 已披露差异化 | 当前证据的关键局限 |
|---|---|---|---|---|---|
| Cowboy Space | 直接进入者 / 目标公司 | $275M B 轮,估值 $2B;拟议服务尚未成为获批运营网络 | 受电网或数据中心约束阻碍的陆地 AI 计算买方 | 发射载具、轨道电力和在轨计算的一体化技术栈 | 指定来源池中没有获批运营网络或公开定价 |
| 陆地超大规模数据中心运营商 | 现有企业 / 现状替代方案 | 独立报道称,超大规模云厂商计划在 2026 年支出近 $700B 扩充容量 | 通用 AI 训练和推理工作负载 | 现成的运营容量,采购流程也已明确 | 电力、用水和许可约束已经在收紧陆地扩张 |
| SpaceX 轨道数据中心备案 | 直接新兴同行 | 已受理备案;引用文章称最多可部署一百万颗卫星 | 未来轨道计算和网络工作负载 | 已披露的系统设想规模远大于 Cowboy | 在指定来源池中仍处于备案阶段,不是商业产品 |
| Google Project Suncatcher | 可能进入者 / 邻近研究项目 | 指定来源池中仅有研究或论文 | 未来的太空 AI 基础设施方案 | 大型科技公司的兴趣验证了这一类别 | 指定来源池中没有披露商业化方案或买方证明 |
| Starcloud | 直接或邻近可比公司 | Caplight 上的嵌入相似度为 87% | 指定来源池中未披露 | Caplight 标示其商业模式与 Cowboy 相似 | 来源池缺少定价、客户和运营细节 |
| Relativity Space / Orbital Inc. | 邻近太空基础设施可比公司 | Caplight 上的嵌入相似度分别为 83% 和 80% | 太空基础设施工作负载;本来源池未披露轨道计算方案 | 显示资本密集度和太空技术栈复杂度存在重叠 | 相似度信号不证明产品范围等价 |
仅限指定来源池。相似度行来自 Caplight;等同于 null 的表述表示指定来源池没有披露该字段,不代表公司不存在该信息。
[CP001, CP006, CP008, CP009, CP010, CP013, CP014]3.2 差异化来自整合,但切换摩擦仍让现状占优
Cowboy 的差异化不在软件广度或渠道,而在架构。公司称发射工具、供电系统和计算载荷共同设计,使上级成为轨道上的 1 MW 数据中心。如果轨道供电经济性、散热管理和发射频率都能达到商业规模,这可能形成护城河。但同一套整合也给当前买家带来很高的切换负担:与地面超大规模容量不同,资料中的产品还不是获批并运营的服务,没有公开定价、客户案例或合同结构。因此,Cowboy 的竞争姿态是:概念和雄心比普通航空航天初创公司更有差异化,但今天仍不如它希望取代的地面现状方案可替代。[CP002, CP007, CP010, CP015]
| 采购标准 | Cowboy Space | 陆地超大规模替代方案 | 指定来源池中的轨道进入者 | 含义 |
|---|---|---|---|---|
| 发射与计算的一体化所有权 | 是;公司把发射载具和数据中心载荷设计成一体化飞行器 | 否;陆地容量与发射相互独立 | 不一;SpaceX 备案和 Google 研究暗示轨道雄心,但已披露方案细节稀少 | Cowboy 的核心差异化在架构,而非商业层面 |
| 独立于电网的论点 | 是;轨道太阳能是其核心卖点 | 否;仍受陆地电力和许可审批约束 | 对轨道进入者而言,概念上是 | Cowboy 针对的是陆地的真实瓶颈,但目前仍只是投资论点 |
| 公开监管进展 | FCC 方案审查中 | 陆地许可路径已确立,但本来源池没有详细覆盖 | SpaceX 备案已受理;Google 证据仍处于研究阶段 | 监管成熟度让现状计算方案优于轨道进入者 |
| 公开运营证明 | 来源池中没有披露已在运营的轨道数据中心网络 | 作为一类方案确实存在,但本工作者没有单一具名运营商证据 | 指定来源池没有证明正在运营的轨道计算服务 | 买方切换风险仍指向现有替代方案 |
| 公开定价 / 产品包装披露 | 指定来源池未知 | 指定来源池未知 | 指定来源池未知 | 当前比较是战略性的,不涉及合同或单位经济性 |
| 公开客户证明 | 指定来源池中没有具名客户胜利 | 运营需求显然存在,但客户细节不在本工作者的来源范围内 | 指定来源池中没有具名买方证明 | 目前还没有通过输赢证据验证持久性 |
不受支持的单元格标为 unknown 或受来源限制,而不是自行猜测。该矩阵比较指定来源池中的披露和能力信号,不是经过审计的产品对等性。
[CP002, CP007, CP008, CP009, CP010, CP014, CP015, CP016]在所分配的样本中,相比地面现状,Cowboy 的一体化堆栈野心较高,但披露的商业成熟度较低。
- X: 一体化堆栈所有权(序数)
- low: 模块化/外包堆栈
- high: 端到端自有堆栈
- description: Ordinal score from assigned-pool evidence on how much of launch, power, and compute the alternative appears to control.
- scale: ordinal
- Y: 披露的商业成熟度(序数)
- low: 概念或申报阶段
- high: 可实际采购的运营选项
- description: Ordinal score from assigned-pool evidence on whether the alternative is a disclosed operating purchase option versus a proposal or study.
- scale: ordinal
| 条目 | 一体化堆栈所有权(序数) (1–10) | 披露的商业成熟度(序数) (1–10) | 说明 |
|---|---|---|---|
| 1. Cowboy Space | 9 | 4 | |
| 2. 地面超大规模替代方案 | 3 | 10 | |
| 3. SpaceX 轨道数据中心申报 | 8 | 5 | |
| 4. Google Project Suncatcher | 6 | 2 | |
| 5. Caplight 可比公司群 | 5 | 4 | |
| 6. 轨道数据中心样本组 | 6 | 3 |
序数评分有证据支撑,但不是来源原生指标。Cowboy 的一体化得分较高,因为公司称火箭和数据中心是一体化飞行器。地面替代方案的成熟度得分最高,因为所分配的报道把地面算力视为现有基准。引用的证据显示,SpaceX 和 Google 在品类验证上高于 Cowboy,但仍处于提案或研究阶段。Caplight 可比公司群和 SatNews 样本组只是稀疏的同业集合,并非经过审计的商业同类公司。
[CP002, CP005, CP006, CP008, CP009, CP010, CP014, CP015]3.3 护城河耐久性仍处于论证阶段,商业证明很薄
在现有证据中,最持久的护城河要素是概念层面的垂直整合,而不是经验证的商业锁定。独立报道强调稀缺技术人才的竞争;负面报道则突出尚未解决的散热和碎片问题,以及未完成的卫星设计。这些事实并不否定战略,但意味着在 Cowboy 把架构转化为获批运营、客户赢单和披露充分的商业方案之前,护城河耐久性仍停留在论证阶段。因此,本章只支持有限判断:Cowboy 在概念和雄心上有差异化,但当前公开记录太薄,无法与具名同行比较实际定价、买家偏好或多归属行为。[CP005, CP011, CP012, CP016]
| 方案类别 | 公开价格 / 合同模式 | 包含的能力 | 披露状态 | 竞争含义 |
|---|---|---|---|---|
| Cowboy 轨道计算服务 | 指定来源池未公开披露 | 轨道电力、计算载荷和发射架构 | 仅有商业化前披露 | 目前很难证明价格 / 性能优势 |
| Cowboy 一体化发射载具 | 未披露为独立方案 | 上面级成为在轨数据中心 | 已披露概念和路线图;缺少产品包装条款 | 一体化可能成为护城河,但也会加深买方对单一供应商的依赖 |
| SpaceX 轨道数据中心方案 | 指定来源池未公开披露 | 规模非常大的拟议 NGSO 轨道数据中心系统 | 备案阶段披露 | 在价格或客户证明出现前,规模雄心已经可见 |
| Google Project Suncatcher | 指定来源池未公开披露 | 未来太空 AI 基础设施研究概念 | 研究阶段披露 | 代表期权价值和信号,而不是当前商业竞争 |
| 陆地超大规模替代方案 | 指定来源池未报价 | 正在运营的地面 AI 计算容量 | 商业上真实存在,但本工作者的来源没有记录价格 | 即使没有来源支持的价目表细节,现状方案也比 Cowboy 更容易采购 |
| Caplight 可比公司群 | 指定来源池未公开披露 | 基于相似度的集合:Starcloud、Relativity Space、Orbital Inc。 | 仅用于识别可比公司 | 同行集合已经存在,但经济基准比较仍需针对各同行开展新的尽调 |
本表记录指定来源池对产品包装和定价的披露。反复出现的“未披露”是有证据支持的信息缺口,不是为假定定价预留的占位符。
[CP006, CP008, CP009, CP012, CP014, CP015, CP016]3.4 展示材料
04财务
4.1 收入模式清晰可读,但尚未兑现收入
公开材料对产品概念的描述一致:Cowboy Space 想用垂直整合架构销售轨道 AI 计算容量,让发射工具的上级进入轨道后成为 1 MW 数据中心。这给投资人和读者提供了理解潜在收入来源的具体机制,但仍是概念阶段的机制,而不是已披露的商业引擎。现有资料没有客户合同、价目表、GPU 小时价格、最低承诺、利用率假设,也没有把发射和卫星成本连接到实际服务利润率的毛利桥。公开里程碑仍是工程里程碑:能量传输演示、光通信集群任务和首次自研火箭发射,而非披露的收入里程碑。[CI003, CI004, CI011, CI012]
| 收入来源 | 机制 | 单位 | 当前价值 / 状态 | 质量 | 尽调问题 |
|---|---|---|---|---|---|
| 轨道 AI 计算服务 | 出售在轨 Stampede 卫星 / 数据中心节点的计算容量 | 定价单位未披露 | 公开描述的产品概念;没有公开收入或定价 | 公司描述的模式,尚未证明经济效益 | 披露已签约工作负载、计费单位、最低承诺和利用率假设 |
| 发射支持的部署 | 专有发射载具将计算节点送入轨道,使其成为火箭上面级 | 收入项目未披露 | 首次专有火箭发射目标为 2028 年 | 运营里程碑,不是已证明的收入 | 澄清发射只是内部赋能,还是也会作为可计费服务提供 |
| 太空到地面电力传输演示 | 旨在验证基础子系统的演示任务 | 演示里程碑 | 首次卫星任务计划在 2026 年晚些时候开展 | 技术验证步骤,不代表商业牵引力 | 提供商业化路径、交易对手和预期变现时间 |
| 光学数据传输 / 处理 | 未来的集群任务支持高速光通信和在太空处理 | 服务定价未披露 | 已讨论 2027 年集群任务;没有公开客户承诺 | 仅是产品路线图信号 | 提供 SLA 目标、数据传输经济性和早期设计合作伙伴证据 |
公开来源描述了机制和里程碑,但没有披露实际定价、客户合同或毛利转化。
[CI003, CI004, CI012]公开证据描述了兑现收入的物理路径,但定价、签约需求和利润转化仍未披露。
| 节点 / 连接 | 节点 / 起点 | 终点 | 说明 |
|---|---|---|---|
| 节点 1 | AI 电力和数据中心瓶颈 [demand] | 所保留报道把地面电力和电网约束作为需求前提。 | |
| 节点 2 | 一体化发射器加算力载荷 [integrated-vehicle] | Cowboy 称火箭上面级和数据中心载荷按一体化飞行器设计。 | |
| 节点 3 | 1 MW 轨道数据中心节点 [orbital-node] | 报道称每个上面级进入轨道后都会成为一座 1 MW 数据中心。 | |
| 节点 4 | 轨道计算和数据处理服务 [service] | 公开来源描述了预期的 AI 和数据处理工作负载,而非已入账收入。 | |
| 节点 5 | 客户定价和利用率 [monetization] | 没有公开价目表、合同期限、订单积压或利用率披露。 | |
| 节点 6 | 扣除发射和运营成本后的毛利 [gross-profit] | 公开来源没有披露利润率、现金消耗或现金续航转化。 | |
| 连接 1 | AI 电力和数据中心瓶颈 [demand] | 一体化发射器加算力载荷 [integrated-vehicle] | 驱动 |
| 连接 2 | 一体化发射器加算力载荷 [integrated-vehicle] | 1 MW 轨道数据中心节点 [orbital-node] | 部署 |
| 连接 3 | 1 MW 轨道数据中心节点 [orbital-node] | 轨道计算和数据处理服务 [service] | 使能 |
| 连接 4 | 轨道计算和数据处理服务 [service] | 客户定价和利用率 [monetization] | 通过……转化 |
| 连接 5 | 客户定价和利用率 [monetization] | 扣除发射和运营成本后的毛利 [gross-profit] | 减去未披露成本结构 |
这是公开机制桥接图,不是经测量的收入瀑布图;未披露的定价和成本节点被明确列出,而不是由我们估算。
[CI003, CI004, CI011, CI012, CI013]4.2 定价、客户和牵引力仍是尽调要求
财务评估受限的原因不是收入数据互相矛盾,而是收入数据缺失。Caplight 公开页面披露融资元数据,却明确没有公开收入、ARR 或盈利数据;其他资料同样止步于融资轮次和产品叙事。没有留存来源点名客户、提供积压订单或 LOI 数量、量化利用率,或区分标价与实际定价。因此,现有证据可以描述 Cowboy 说它将如何变现,却无法说明买家是否已承诺付费、合同条件是什么,或早期规模能达到多少。[CI007, CI011, CI012]
| 方案 / 指标 | 价格 / 单位 / 合同 | 标价与实际实现价格 | 折扣 / 未知项 | 来源 |
|---|---|---|---|---|
| 轨道计算容量 | 未公开披露 | 未知 | 没有公开价目表、期限或承诺量安排 | Caplight 公开页面及指定新闻报道 |
| 客户承诺 / 积压订单 / LOI | 未公开披露 | 未知 | 留存来源池中没有具名客户或已签需求指标 | 指定新闻报道和 Caplight 公开页面 |
| 电力传输演示 | 仅有演示任务;没有公开商业定价 | 不适用 | 没有披露从演示转化为经常性收入的情况 | Space.com 里程碑报道 |
| 发射服务经济性 | 未公开披露 | 未知 | 没有公开区分内部发射成本和对外可计费服务 | 来源:Business Wire、Space.com、The Robotics Media |
本表有意记录不可获得的定价信息,而不是臆测来源未披露的 SaaS 式或发射式变现条款。
[CI007, CI011, CI012]4.3 资本充足度取决于穿越未解决的工程与监管闸门
最新融资显著延长了 Cowboy 推进轨道计算论点的能力,但本身并没有回答下一个投资判断问题:公司最终需要多少资本,才能建造发射工具、大型轨道数据中心卫星,以及配套的地面、制造和监管体系。独立报道把 FCC 文件视为审核中的提案,而非获批运营网络;SatNews 则把发射—计算架构描述为一场同时解决散热、发射、光链路和单位经济性瓶颈的押注。公开来源对累计融资额也有分歧,且没有披露现金余额、烧钱速度、债务或项目融资义务。证据足以称公司有资金开展试验,却不足以精确评估现金跑道或稀释风险。[CI001, CI002, CI005, CI006, CI008, CI009, CI010, CI013, CI014, CI015]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调问题 |
|---|---|---|---|---|
| 最新一轮 / 估值 | $275M B 轮,估值 $2B | 高 | 设定新资本的进入价格,但不代表运营表现 | 确认投后口径、清算优先权和任何结构化条款 |
| 累计融资额 | 公开来源在约 $325M 和 $335M 之间存在冲突 | 低 | 影响稀释历史和截至目前现金的口径 | 逐轮核对募集所得、补助、创始人资本及任何非股权融资 |
| 收入运行率 / ARR | 中 | 核心投资判断指标不可得 | 提供月度收入、年化运行率,以及经常性与非经常性收入构成 | |
| 客户数量 / 积压订单 / LOI | 中 | 没有已承诺交易对手就没有需求证明 | 提供已签客户、设计合作伙伴、积压订单及 LOI 转化假设 | |
| 毛利率 / 贡献毛利率 | 中 | 没有服务成本和发射分摊就无法评估单位经济性 | 拆分发射、卫星、电力、通信和运营成本结构 | |
| 现金消耗 / 现金跑道 / 债务义务 | 中 | 不能仅凭融资额判断融资充足度 | 提供当前现金、月度消耗、债务 / 项目融资及下一轮融资触发条件 |
null 表示指定公开来源没有披露该指标;置信度列反映证据质量,不代表业务质量。
[CI001, CI006, CI007, CI009, CI010, CI014, CI015]4.4 财务结论
本章财务结论是混合的。Cowboy Space 确实有融资势头、具名机构支持者,以及能够追溯到合理未来收入机制的产品故事。但支持传统投资判断的所有核心运营指标,在现有资料中都不存在:没有收入、没有 ARR、没有披露的客户承诺、没有实际定价、没有烧钱数据,也没有毛利证据。本次更新的正确解读是:资本可得性是论点成立的必要条件,却不是论点已经商业验证的证明。[CI001, CI004, CI012, CI013, CI014]
4.5 展示材料
05产品与技术
5.1 集成式产品论点
Cowboy Space 披露的产品不是独立卫星平台,也不是传统的搭载载荷。公司和独立来源始终描述一套把运载工具开发、在轨供电和 AI 计算整合到同一架构中的垂直系统,火箭上级本身计划留在轨道上,充当数据中心节点。这种具体性很重要:它让公司拥有明确的产品定义,而非宽泛的市场论点。已宣布的计算合作方是 NVIDIA,节点目标为 1 MW,同一体化飞行器还要同时负责发射频率和交付计算能力。实际含义是,当前产品差异化主要来自航天器、发射和供电之间的系统级整合,而不是专有软件。[CE001, CE002, CE005, CE010]
| 模块 / 资产 | 主要使用者 | 公开状态 / 成熟度 | 已披露差异化 | 尽调缺口 |
|---|---|---|---|---|
| 轨道计算节点 | 潜在 AI 计算买方 | 概念 + 备案阶段 | 上面级成为在轨 1 MW 数据中心 | 需要热真空、正常运行时间和基准测试证据 |
| 专用发射载具 | Cowboy 部署运营 | 载具开发中;没有公开飞行件 | 产品本身就包含发射频次设计,而不是外包 | 需要发动机、推力、设施和集成测试数据 |
| 光学传输 + Ka 波段 TT&C | 网络运营和监管机构 | 架构已披露;服务尚未公开运营 | 光学优先方案旨在减少对拥塞 RF 路径的依赖 | 需要链路预算、地面网络和故障切换细节 |
| 电力传输演示卫星 | 内部工程团队 + 未来客户 | 首颗卫星计划在 2026 年晚些时候发射 | Cowboy 能部署太空电力硬件的最早公开证明 | 需要演示载荷配置和结果 |
综合官方融资 / 更名公告及独立备案报道;成熟度标签区分已备案架构和已飞行能力。
[CE001, CE002, CE004, CE008, CE009, CE010]公开来源描述了一套垂直整合堆栈:结合轨道选择、光学传输、NVIDIA 计算载荷和专用发射器,上面级会成为计算节点。
- 轨道 + 电力边界
- Stampede 被描述为晨昏太阳同步星座,规模设计旨在持续获得日照。
- 条目
- 最多 20,000 颗卫星
- 700-1,000 km 壳层
- 输出
- 摆脱地面电网排队、由太阳能供电的计算
- 计算节点
- 每个上面级计划作为 1 MW 轨道数据中心运行。
- 模块
- NVIDIA Space-1 Vera Rubin 模块
- 机载 GPU 略少于 800 个
- 输出
- 每个节点 1 MW
- 传输 + 控制
- 光链路是主要通信路径,Ka 波段 TT&C 留作特定任务阶段和紧急情况使用。
- 模块
- 光学星间链路
- 光学星地链路
- 18.8-19.3 GHz 下行 TT&C
- 28.6-29.1 GHz 上行 TT&C
- 演示路径
- 近期首个在轨里程碑是电力波束演示卫星,而不是完整商业计算部署。
- 条目
- 首颗卫星于 2026 年稍晚发射
- 星地电力波束演示
- 输出
- 最早的公开验证步骤
- 发射一体化
- 发射器和数据中心按一个系统设计,而不是让载荷搭乘通用火箭。
- 条目
- 专用发射器
- 上面级就是载荷
- 首次自有发射目标:2028 年底
- 输出
- 若开发成功,可能形成内部发射节奏
- 合规边界
- 这套架构仍取决于 FCC 的公开审查,以及之后完成最终航天器设计文件包。
- 条目
- 申请待审
- 预计之后修改许可证
- 输出
- 监管路径仍是产品成败关键
5.2 运营模式与部署路径
对于如此早期的公司,公开运营模式也异常具体。FCC 文件报道将 Stampede 描述为最多 20,000 颗卫星组成的星座,运行在 700 到 1,000 千米的晨昏太阳同步轨道,使用光链路作为主要数据通道,并在特定任务阶段和紧急情况下使用 Ka 波段遥测、跟踪和指挥;监管材料还暗示服务可以从单颗卫星起步,再逐节点扩张,无需等待星座达到完整规模。纸面上,这降低了最低可行部署门槛,但也意味着早期产品故事高度依赖第一颗演示卫星、首个轨道计算节点和首次自研发射能否大体按披露时间表落地。[CE003, CE004, CE006, CE007, CE008, CE009]
| 用户任务 | 当前工作流 | Cowboy 方案 | 公开宣称的收益 | 局限 |
|---|---|---|---|---|
| 在地面电网受限时增加 AI 算力 | 建设或租用地面数据中心容量 | 在 LEO 运行 1 MW 轨道算力节点 | 节点本身无需建设地面电网 | 尚无公开成本、SLA 或客户验证 |
| 处理临近空间的 AI 工作负载 | 先下传数据,再在地面处理 | 使用基于 NVIDIA 的在轨算力载荷 | 让算力靠近太阳能和源自太空的数据链路 | 尚无公开吞吐量或工作负载基准数据 |
| 在发射资源稀缺时扩张算力 | 等待第三方发射清单的可用性 | 自有火箭,并把上面级做成载荷 | 如果运载器成熟,或可把发射节奏收归内部 | 发射硬件尚未得到公开验证 |
| 在完整星座建成前进入市场 | 大型网络通常会等到舰队规模更大 | 先从一颗卫星开始,随后随发射扩张 | 首次商业运营的纸面门槛更低 | FCC 审查及后续许可证修改仍然关键 |
这些好处仅限于公开来源明确声称的内容;客户经济性和运营指标仍未披露,属于待补的尽调问题。
[CE002, CE004, CE005, CE006, CE007, CE010]5.3 成熟度、依赖项与控制
该产品概念比多数论证阶段的太空初创公司更具体,但成熟度仍低,因为关键工程和监管依赖项都悬而未决。公开来源称卫星设计尚未完成;FCC 要求提供包括太阳能板在内的物理尺寸后才会推进申请;发射工具的发动机细节、推力、测试件和设施也尚无公开披露。独立评论还在聚焦同一批硬技术瓶颈:在真空中排出兆瓦级废热、在拥挤的轨道带安全运营大型星座,以及证明一体化发射加节点模式能按初创公司的时间表制造和获批。从尽调角度看,本章呈现的是一套具体架构,公开验证有限,而不是已经为投资判断完成去风险的产品。[CE008, CE011, CE012, CE013, CE014]
| 层级 / 流程 | 公开角色 | 关键依赖 | 公开风险 |
|---|---|---|---|
| 轨道 + 电力包络 | 使用晨昏太阳同步轨道,持续获取太阳能 | FCC 批准 + 最终航天器尺寸 | 监管机构仍在审查尺寸和规模 |
| 发射上面级 | 搭载节点并作为产品留在轨道上 | 新运载器开发及后续复用路径 | 尚无公开的发动机、推力或测试场地披露 |
| 算力载荷 | 搭载基于 NVIDIA Space-1 的 AI 基础设施 | 辐射、热控和封装鉴定 | 尚无公开基准或鉴定结果 |
| 网络控制 | 日常传输使用光链路,TT&C 阶段使用 Ka 波段 | 指向精度、地面接入和应急运营 | 尚无公开可靠性或故障切换数据 |
| 安全 + 合规 | 满足碎片、天文学和许可要求 | FCC 回复、后续许可证修改及处置设计 | 申请仍在审理,卫星设计尚未完成 |
本表将已披露架构与尚未解决的工程、合规依赖分开;这不代表整套系统已具备生产条件。
[CE003, CE004, CE005, CE011, CE012, CE013, CE014]5.4 展示材料
06客户
6.1 客户论点指向狭窄但高价值的早期细分市场
最有力的客户故事聚焦细分市场,而非具体账户。 Cowboy 把轨道计算定位为回应地面电力、土地、用水和许可瓶颈的方案,因此最相关的买家是需要异常高能耗 AI 容量的机构,或已经运营太空资产、其数据能从在轨处理受益的机构。 独立市场分析明确指出,早期轨道计算需求更可能来自地球观测运营商、国防与安全机构、空间站或月球基础设施用户,以及测试特殊工作负载的 AI 开发者,而非广泛的企业云市场。 这把第一批客户收窄到预算充足且确有任务需求的账户,也意味着近期 go-to-market 路径会比较集中,而不是广泛、分散地获取客户。[CU001, CU002, CU008, CU010]
| 分群 | 买方 / 用户 / 付费方 | 用例 | 规模 / 战略价值 | 缺口 |
|---|---|---|---|---|
| 超大规模 AI 和前沿模型构建者 | 基础设施团队 / ML 平台团队 / 大额算力预算 | 轨道容量可用后,承接高功率训练或推理工作负载 | 理论需求很大,但未披露具体客户或工作负载 | 指定证据池中没有合同、部署或采购标准的证据 |
| 国防和政府机构 | 项目办公室 / 任务用户 / 政府预算 | 弹性离网算力、主权算力或任务专用在轨处理 | Robotics Media 将政府合同胜出列为待观察里程碑,而非已签合同 | 此处未保留具体机构、授标或采购记录 |
| 地球观测及其他空间资产运营商 | 卫星运营商 / 载荷团队 / 任务运营商 | 在更靠近数据源的位置处理数据,缓解下传或时延约束 | 独立市场分析将其列为较有可能的早期轨道算力客户群之一 | 未披露 Cowboy 专属的部署、试点或参考客户 |
| 空间站和月球基础设施运营商 | 基础设施运营商 / 任务用户 / 项目预算 | 为地外运营提供算力和存储,满足特殊能源或位置约束 | 可能价值高,但属于由任务需求驱动的窄众市场 | 尚无公开管线、发射计划或客户验证将 Cowboy 与该分群连接起来 |
各行根据保留的公开来源归纳可能的买方群体;它们反映分群匹配度,不代表已验证的成交客户数或收入结构。
[CU001, CU002, CU008, CU010]Cowboy 从市场拉动走向持续客户扩张,仍要跨过多道商业化前验证关卡。
[CU002, CU004, CU005, CU006, CU010, CU012]6.2 公开采用证明仍由伙伴和里程碑驱动
现有证据池没有出现具名付费客户、已签合同金额、积压订单、LOI、客户数披露或量产部署。 但它展示了一串商业化前的证明点:与 NVIDIA 合作、FCC 文件勾勒的单星起步方式、计划于 2026 年下半年开展的能量传输演示,以及独立报道把与超大规模云客户和政府客户签约视为值得观察的里程碑,而不是已经锁定的赢单。 这套组合很有意义,因为它证明了生态兴趣和商业化路径,但与已验证的使用、产生收入的部署或可供参考的客户结果仍有本质区别。[CU003, CU004, CU006, CU007, CU012]
| 指标 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 已具名付费客户 | 指定公开证据中未披露 | 2026-10-17 | 对指定证据池的跨来源审阅 | 中 | 客户验证仍停留在商业化前 | 客户数、合同金额、积压订单和意向书 |
| 已具名生产部署 | 未披露 | 2026-10-17 | 对指定证据池的跨来源审阅 | 中 | 尚无可验证的线上工作负载证据 | 部署日期、工作负载、利用率和结果 |
| 合作伙伴验证 | 已披露与 NVIDIA 的合作 | 2026-05-11 至 2026-05-13 | Business Wire;GeekWire | 中 | 验证生态兴趣及算力技术栈的关联 | 商业条款、工作负载承诺,以及该伙伴是否同时是买方 |
| 最早的公开运营里程碑 | 2026 年稍晚发射首颗卫星,用于无线输电演示 | 2026-05-11 至 2026-05-13 | Business Wire;GeekWire | 中 | 显示技术里程碑排在客户增长之前 | 演示能否转化为客户使用或已签需求 |
| 最早的商业运营说法 | 服务可从一颗在轨卫星开始 | 2026-05-15 | Satellite Today | 中 | 暗示在达到完整星座规模前,可采取低发射量路径 | 目标客户、预期工作负载及初始规模下的单位经济性 |
| 潜在合同里程碑 | “‘与超大规模和政府客户签约’被列为近期里程碑” | 2026 | The Robotics Media | 中 | 表明目标买方类别,但不代表已锁定业务 | 客户名称、采购阶段、合同规模和时间 |
本表将已披露里程碑与缺失的需求指标分开,避免把商业化前验证误当成客户牵引力。
[CU003, CU004, CU005, CU006, CU007, CU012]| 客户 / 对手方 | 分群 | 部署 / 用例 | 生产还是试点 | 结果 | 局限 |
|---|---|---|---|---|---|
| NVIDIA | 技术合作伙伴 / 赋能供应商 | 计划用于低地球轨道算力的 Space-1 Vera Rubin 模块 | 合作伙伴集成计划 | 表明算力技术栈已有具名生态合作方 | 仅是合作伙伴验证;不能证明 Cowboy 有付费终端客户或已部署工作负载 |
| 未具名的超大规模客户 | 潜在企业买方 | Robotics Media 将未来赢得超大规模客户合同列为待观察里程碑 | 潜在 | 确认超大规模客户是目标买方类别 | 未披露具名客户、合同、时间表或生产使用证据 |
| 未具名的政府客户 | 潜在公共部门买方 | Robotics Media 将未来赢得政府客户合同列为待观察里程碑 | 潜在 | 暗示 Cowboy 认为政府是可行的早期分群 | 未披露具名机构、采购载体、授标金额或运营项目 |
覆盖范围有意保持不完整,因为指定公开证据只有一个具名合作伙伴,没有具名付费客户或生产客户。
[CU003, CU004, CU006, CU012]6.3 耐久性未经证明,首批合同集中度可能很高
公开证据目前没有耐久性指标:现有资料没有披露 NRR、GRR、流失率、合同期限、群组表现或续约证据。 这很重要,因为最合理的早期买家是一小批技术能力强的账户,它们的采购周期取决于发射可靠性、散热表现和监管进展,而不是普通 SaaS 的扩张机制。 SatNews 的怀疑性解读在这里很有参考价值:公司仍需证明发射频率、散热管理和碎片经济性,客户论点才会变得耐久。 如果这些障碍被跨越,Cowboy 可能拿下高价值灯塔客户;在此之前,客户集中度风险无法量化,但可能偏高,因为早期市场专业且规模小。[CU005, CU009, CU010, CU011]
6.4 展示材料
07风险
7.1 监管批准是第一道硬闸门
监管问题不是 Cowboy 是否有轨道计算愿景,而是当前申请能否在没有重大延误或重新设计的情况下通过非常规审查。本资料中的独立报道和申请评论在三点上相互一致:Stampede 仍是申请,不是获批运营系统;公司要求对标准 NGSO 部署机制作出多项豁免;外部评论者已要求 FCC 面临特别严格的审查,因为该方案把超大型星座、光链路和新颖的轨道数据中心用例结合在一起。因此,首要法律风险从普通文书工作转向开创先例的审查、豁免结果,以及 Cowboy 能否在商业进度变得关键前,把仍在演变的技术记录变成可获许可的系统。[CR002, CR003, CR004, CR011, CR012, CR015]
| 风险 | 司法辖区 / 监管面 | 当前状态 | 可能性 | 严重性 | 缓解措施成熟度 | 残余风险 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| FCC 批准以及豁免被拒或延期 | 美国 FCC Part 25 / Stampede 申请 | 申请待审;豁免事项仍属于当前在审文件的一部分 | 高 | 高 | 低 | 高 | 调取当前案卷历史、逐项豁免的法律备忘录,以及 2026 年 5 月以来的机构问题。 |
| 系统级碎片 / PMD / 失事审查 | 美国 FCC 及更广泛的轨道安全审查 | SWF 表示,公开记录仍缺少完整规模的量化分析 | 中 | 高 | 低 | 高 | 要求提供最新碎片模型、PMD 假设、失事分析和分阶段授权计划。 |
| 天文学 / 环境 / 治理外部性 | FCC、商务部、跨机构政策 | 公开资料已指出外部性,但该证据池看不到最终框架 | 中 | 中 | 低 | 中 | 要求管理层提供天文学缓解方案、任务授权观点和 NEPA 立场。 |
| 设计未完成导致的许可证修改风险 | FCC 许可流程 | 独立报道称设计尚未完成,服务前还需后续修改 | 高 | 中 | 低 | 高 | 审查工程冻结日期、修订节奏,以及哪些设计变量仍会影响许可证。 |
本表仅基于公开资料,汇总 SWF 2026 年 7 月的 FCC 意见及 2026 年 5 月对 Cowboy 申请的报道;不包括非公开监管沟通和后续案卷事件。
[CR002, CR003, CR004, CR011, CR012]在公开证据仍缺少已授予权限、完整设计闭环或经演示的发射与运营路径之处,剩余风险最高。
[CR011, CR013, CR014, CR015]7.2 发射与计算一体化设计叠加执行风险
Cowboy 的核心产品架构也是主要的运营风险放大器。公开报道描述了一套系统,其中发射工具上级变成 1 MW 数据中心载荷,因此发射、载荷、散热和轨道运营无法彼此独立去风险。如果奏效,它可能形成更强护城河;但也意味着一次进度延误就可能同时拖慢论点的多个部分。独立来源仍指出真空散热尚未解决、卫星设计未完成,首次自研发射也只以 2028 年为目标。高度依赖光链路的架构减少了对日常频谱的依赖,却引入了可用性和地面运营纪律要求,目前大多仍停留在概念描述,缺少运营数据证明。[CR005, CR006, CR007, CR008, CR009, CR010, CR011, CR013, CR015]
| 失效模式 | 证据 | 可能性 | 严重性 | 缓解措施成熟度 | 残余风险 | 未解决缺口 |
|---|---|---|---|---|---|---|
| 真空中 1 MW 的热排散失败 | 独立评论仍将热负荷问题视为未解决 | 高 | 高 | 低 | 高 | 需要热真空数据、散热器假设和算力降额包络。 |
| 一体化运载器与载荷进度延误 | 火箭上面级和轨道数据中心按一体化运载器设计 | 高 | 高 | 低 | 高 | 需要将推进、结构、制造和测试里程碑与 2028 年发射目标对齐。 |
| 光链路和地面运营可靠性不足 | 该架构降低了对 RF 的依赖,但仍需要指向、天气规划和地面多样性 | 中 | 中 | 低 | 中 | 需要备用通信方案、地面站地图和服务可用性假设。 |
| 首飞和星座运营经验缺口 | SWF 表示 Cowboy 尚未运营过星座,也尚未完成首次演示任务飞行 | 中 | 高 | 低 | 中 | 需要演示后的遥测数据、运营复盘,以及首次任务解决重大设计未知项的证据。 |
评级仅反映保留公开来源明确提出的问题;本表不涉及私有测试数据、安全审计或制造就绪度资料包。
[CR005, CR006, CR007, CR008, CR009, CR010, CR011]7.3 资本、交易对手和证明点定义终止标准
B 轮融资显著提高了存活能力,但不能证明 Cowboy 已经解决当前公开讨论的完整架构所面临的融资或需求风险。现有记录支持近期开发推进,却不能得出公司已经为自研发射器、轨道数据中心载荷,以及向 20,000 颗卫星概念扩张所需的基础设施完成融资的结论。对 Nvidia 模块、发射准备度、FCC 时间表,以及客户是否会接受一个公开仍处于演示前阶段的服务,依赖度都很高。就投资而言,最清晰的终止标准是可衡量的:没有可信演示读数、没有明确的设计冻结和许可路径、没有发射频率计划,或接近 2028 年时仍没有客户和融资证明。[CR001, CR009, CR010, CR013, CR014, CR015]
| 依赖 | 对手方 / 系统 | 角色 | 集中度 | 失效场景 | 严重性 | 缓解措施 | 残余风险 |
|---|---|---|---|---|---|---|---|
| 算力模块平台集中度 | NVIDIA Space-1 Vera Rubin 模块 | 核心 AI 载荷模块 | 单一平台集中度高 | 模块、路线图或鉴定延误会推迟载荷就绪 | 高 | 鉴定替代方案,并将载荷目标限定在可采购范围内 | 高 |
| 发射就绪度和节奏 | Cowboy 运载器及任何临时发射伙伴 | 将数据中心载荷送入轨道 | 高 | 即使存在需求,运载器延误也会导致没有商业服务 | 高 | 锁定分阶段发射计划及临时伙伴备选方案 | 高 |
| 监管时间表 | FCC 及可能的跨机构审查面 | 运营和扩张的许可证门槛 | 高 | 批准时间会打乱融资和商业里程碑 | 高 | 推进分阶段授权,并加强案卷沟通 | 高 |
| 商业验证和融资深度 | 客户及资本市场 | 需求验证和资产负债表支持 | 高 | 公开时间表推进速度快于客户验证或融资明晰度 | 高 | 披露意向书、积压订单、融资计划和分阶段部署经济性 | 高 |
本表将外部对手方与门控系统放在一起,因为公开记录通过集中依赖描述 Cowboy 的风险,而非披露供应商名单或客户管线。
[CR001, CR006, CR008, CR009, CR010, CR013, CR014]7.4 展示材料
08估值
8.1 当前融资背景清晰,但真正的估值支撑仍很薄
公开融资数据很直接:Cowboy 在 2026 年 5 月宣布以 $2B 估值完成 $275M B 轮,第三方数据库也把公司列入独角兽阵营。更不清楚的是,这一估值有多少来自已观察到的商业牵引力,又有多少来自投资人愿意为定义品类的基础设施论点提供资金。Dealroom 和 Silicon Valley InvestClub 大体一致认为公司已融资数亿美元、估值约 $2B,但公开记录仍让累计融资、此前各轮的确切条款和二级市场定价保持不透明。因此,这一估值信号作为市场出清后的融资事件是真实的,但作为持久内在价值的证据仍然薄弱。公开来源也没有披露完整优先股堆叠或任何债务 / 项目融资负担,限制了当前价格下的入场纪律。[CV001, CV002, CV003, CV011, CV012, CV013]
8.2 关键缺口是商业化和执行证明,而非叙事质量
Cowboy 的故事很容易理解:AI 数据中心需求正在挤压地面电力和许可资源,因此如果垂直整合的轨道计算系统能跑通,就可能产生价值。问题在于,公开记录仍显示故事与投资判断证据之间存在证明缺口。Silicon Valley InvestClub 明确报告没有公开收入、ARR 或盈利披露。Cowboy 曾描述 2026 年下半年的首颗卫星演示,而独立报道把首次自研 1 MW 发射目标推到 2028 年。FCC 路径也处于早期:New Space Economy 认为 Stampede 应被视为审核中的提案,而不是已批准网络。SatNews 更直接地描述下行风险,强调在论点能凭运营基本面而非雄心获得投资价值前,散热、碎片暴露和单位经济性风险是核心证明负担。[CV004, CV005, CV006, CV007, CV008, CV009]
| 视角 | 正方论点 | 反方论点 / 改变观点的因素 |
|---|---|---|
| 市场驱动因素 | AI 电力和并网约束构成将算力移出地球的合理宏观理由。 | 只有当 Cowboy 将这一宏观叙事转化为有客户支撑的需求,而不是停留在行业叙事上,观点才会改善。 |
| 资本支持 | Index 领投的 $275M、$2B 估值 Series B 轮表明顶级投资者愿意大规模为这一论点提供资金。 | 如果核心里程碑完成前的后续融资只能以平估值或下调估值完成,观点就会恶化。 |
| 产品架构 | 一体化火箭加 1 MW 轨道数据中心的设计如果能跑通,可能形成差异化护城河。 | 只有公开测试结果和发射进展证明这套架构经得住工程现实,判断才会改善。 |
| 商业验证 | 员工人数增长和持续招聘表明 Cowboy 正在为快速执行补充人手。 | 如果当前里程碑窗口结束后仍看不到公开收入、试点或具名客户,判断会恶化。 |
| 监管 / 执行负担 | Cowboy 已与 FCC 接洽并公开系统细节,说明它不只是在推销抽象概念。 | 如果申报停滞、豁免受到质疑,或碎片/热学问题拖慢项目,判断会恶化。 |
这些行对照说明公司为何可能获得估值溢价,以及在溢价具备投资价值前必须拿出的具体证据。
[CV001, CV005, CV006, CV007, CV008, CV009, CV010, CV015, CV016]建议维持观察,因为宏观叙事和资本支持仍不足以抵消商业化验证缺失及尚未解决的执行风险。
| 节点 / 连接 | 节点 / 起点 | 终点 | 说明 |
|---|---|---|---|
| 节点 1 | AI 电力瓶颈 [market-need] | 宏观驱动因素确实存在。 | |
| 节点 2 | 2026 年融资验证 [financing-proof] | $275M B 轮,估值 $2B。 | |
| 节点 3 | 技术范围激进 [technical-scope] | 20,000 颗卫星的 Stampede 计划和 1 MW 轨道节点。 | |
| 节点 4 | 商业验证缺失 [commercial-gap] | 没有公开收入、ARR 或具名客户证据。 | |
| 节点 5 | 剩余执行风险 [execution-risk] | FCC 审查、热学否决、碎片和发射时间仍未解决。 | |
| 节点 6 | 观察/等待证据 [recommendation] | 只有客户、里程碑和股权结构清晰度改善后,估值才应重估。 | |
| 连接 1 | AI 电力瓶颈 [market-need] | 2026 年融资验证 [financing-proof] | 解释投资者兴趣 |
| 连接 2 | 2026 年融资验证 [financing-proof] | 技术范围激进 [technical-scope] | 为架构提供资金 |
| 连接 3 | 技术范围激进 [technical-scope] | 商业验证缺失 [commercial-gap] | 仍缺少需求验证 |
| 连接 4 | 技术范围激进 [technical-scope] | 剩余执行风险 [execution-risk] | 带来工程和许可负担 |
| 连接 5 | 商业验证缺失 [commercial-gap] | 观察/等待证据 [recommendation] | 阻止买入建议 |
| 连接 6 | 剩余执行风险 [execution-risk] | 观察/等待证据 [recommendation] | 维持严格的价格纪律 |
8.3 公开证据支持持续跟踪,上行空间受里程碑制约
乐观情景是 Cowboy 将资本、人才和 NVIDIA 的配合转化为一个不寻常但可守的轨道基础设施平台,在现有企业或相邻发射玩家关闭窗口前建立优势。基准情景没那么戏剧性:公司保持资金充足且具有战略吸引力,但下一轮证据来自演示、客户承诺和监管进展,而不是公开财务证明。悲观情景是 $2B 私募估值已经计入了相当一部分未来成功,而公开太空公司往往要在运营证明深得多之后才达到类似水平;与此同时,最难的工程和许可问题仍未解决。综合来看,公开证据支持继续跟踪,而不是买入:公司重要到值得密切观察,但以现有公开证明看,当前入场价格已经假设了过多未来成功。[CV010, CV014, CV015, CV016]
| 情景 | 假设 | 估值/回报逻辑 | 概率信号 |
|---|---|---|---|
| 乐观 | Cowboy 按时完成首次演示,保住 2028 年发射路径,并披露可信的轨道计算客户需求。 | 当前估值可以守住甚至继续上升,因为公司将把一个投机性品类转化为由里程碑支撑的基础设施期权。 | 下一轮周期必须拿出客户、技术里程碑和监管进展的公开证据。 |
| 基准 | Cowboy 资金充足、技术野心不减,但商业化和监管验证慢于融资叙事。 | 当前估值偏满,但未必失真;投资者会等证据,而不是追价。 | 这最符合今天融资支持强、经营披露少的组合。 |
| 悲观 | 热学、碎片、许可或发射执行明显延期,商业化仍未得到验证。 | 当前估值容易受伤,因为它已触及公开市场航天公司的估值区间,却没有相匹配的经营验证。 | 平轮或下轮融资、演示失败或监管受挫,都会让这条路径从警示变成现实。 |
这些情景是定性的,因为公开证据没有披露精确目标价所需的收入、单位经济性或股权结构条款。
[CV004, CV005, CV006, CV008, CV014, CV015, CV016]8.4 下一轮尽调应聚焦真正能推动价格的证明
最有用的尽调不是再次复述 Cowboy 的市场叙事,而是寻找能压缩当前估值背后不确定性的证据。第一优先级是商业证明:具名设计合作方、试点、积压订单或早期收入,证明需求不只是全行业 AI 基础设施热情。第二是技术证明:热真空结果、运载工具进展,以及一体化上级数据中心架构通过真实测试闸门的证据。第三是资本结构透明度:条款清单、清算优先权、按比例配售权,以及任何优先于新股权的债务或项目融资债权。论点破坏点可能是下轮降价融资、重大 FCC 挫折,或无法把 2026-2028 年里程碑路径转化为客户支持的执行。[CV006, CV009, CV013, CV016]
8.5 展示材料
免责声明
本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开,任何投资决策前都应直接向管理层及一手文件核实。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Cowboy Space was founded in 2024 as Aetherflux and publicly rebranded to Cowboy Space Corporation on May 11, 2026. | 高 | SO002, SO003, SO006 |
| CO002 | Retained public sources place Cowboy Space's headquarters in San Carlos, California, and identify Seattle as an additional office. | 高 | SO001, SO006 |
| CO003 | Cowboy describes itself as a vertically integrated orbital infrastructure company spanning launch vehicles, space-based power, and in-orbit compute for AI workloads. | 高 | SO001, SO003 |
| CO004 | Management's core architectural claim is that the rocket's upper stage and the data-center payload are designed as a single vehicle, with each upper stage intended to become a one-megawatt data center in orbit. | 中 | SO001, SO002, SO004 |
| CO005 | Cowboy says it is collaborating with NVIDIA to deploy Space-1 Vera Rubin modules in low Earth orbit. | 中 | SO002, SO003, SO006 |
| CO006 | On May 11, 2026, Cowboy announced a $275 million Series B at a $2 billion valuation led by Index Ventures, with IVP, Blossom Capital, and SAIC named as new investors alongside returning backers. | 高 | SO002, SO003, SO004 |
| CO007 | SatNews reports that disclosed equity raised to date stands at roughly $325 million when it adds the May 2026 Series B to prior disclosed equity rounds. | 中 | SO004 |
| CO008 | Tracker-style sources show $335 million-plus across three rounds and list a February 2025 grant in the historical funding record. | 中 | SO006, SO007 |
| CO009 | Public totals for capital raised are not fully aligned, so cover metrics should distinguish the roughly $325 million disclosed-equity reading from the $335 million-plus tracker total rather than present one unqualified number. | 中 | SO004, SO006, SO007 |
| CO010 | Silicon Valley Invest Club lists Matthew Seaford as a March 2026 advisory-board addition and Brad Koenig as an April 2026 advisory-board addition. | 中 | SO006 |
| CO011 | Public funding trackers place Cowboy's pre-Series B financing history at Seed in October 2024, a Grant in February 2025, and a Series A led by Index Ventures in April 2025. | 中 | SO006, SO007 |
| CO012 | Independent reporting says Cowboy asked the FCC in mid-May 2026 for authority to operate the Stampede constellation, a proposal for up to 20,000 satellites that was still under review rather than approved at publication. | 高 | SO004, SO005, SO008 |
| CO013 | Independent coverage frames Cowboy's first proprietary launch as a 2028 target or before the end of 2028, not as a completed milestone. | 中 | SO004, SO005 |
| CO014 | SatNews identifies Joseph Yaffe as Cowboy's COO and Chief Legal Officer and says he joined the company in January 2026 after leading Skadden's Palo Alto office. | 中 | SO009 |
| CO015 | The retained public record still makes Baiju Bhatt the clear strategic and narrative center of the company, creating meaningful founder key-person dependence. | 中 | SO003, SO004, SO006 |
| CO016 | Adverse independent coverage repeatedly flags the same unresolved execution risks: thermal rejection for a one-megawatt payload in vacuum, debris exposure from a very large constellation, and an unfinished satellite design that still needs public validation. | 中 | SO004, SO009 |
| CM001 | Cowboy's disclosed product boundary is an orbital-compute stack that combines launch, upper-stage compute, orbital power, and optical networking rather than a generic satellite service. | 中 | SM002, SM004 |
| CM002 | The status-quo substitute in the retained sources is terrestrial AI-capacity expansion constrained by power, cooling, water, land, and permitting rather than another orbital-compute vendor. | 中 | SM001, SM003, SM007 |
| CM003 | Public evidence supports a market thesis built around infrastructure bottlenecks and deployment speed instead of a differentiated application-layer AI product. | 中 | SM001, SM003, SM005 |
| CM004 | An independent trade article says hyperscalers are planning nearly $700 billion of data-center capacity spending in 2026. | 中 | SM003 |
| CM005 | Independent reporting describes Stampede as a proposed system of up to 20,000 satellites in dawn-dusk sun-synchronous orbit between 700 and 1,000 kilometers. | 中 | SM005, SM006, SM007 |
| CM006 | Public reporting describes each representative orbital node as a 20,000-25,000 kilogram platform with roughly 1 megawatt of usable power and just under 800 onboard GPUs. | 中 | SM001, SM006 |
| CM007 | Cowboy says the rocket upper stage becomes the data center in orbit, making launch capacity part of the compute product rather than an external dependency. | 中 | SM002, SM004 |
| CM008 | The retained sources do not name commercial customers, but they imply early demand from hyperscalers, government or defense programs, and space-native operators that value resilient or in-orbit compute. | 低 | SM001, SM003, SM005 |
| CM009 | The disclosed adoption path is a space-to-Earth power-beaming demo later in 2026, an end-to-end compute mission in the first half of 2027, and a first integrated in-house launch targeted before the end of 2028. | 中 | SM001, SM002, SM004 |
| CM010 | The most plausible early workload classes in retained independent reporting are latency-tolerant or space-native applications such as defense systems, Earth observation, disaster response, scientific computing, and sovereign AI infrastructure. | 低 | SM001, SM005 |
| CM011 | Independent sources consistently identify thermal rejection, radiation tolerance, optical-network reliability, launch economics, and long-term system reliability as adoption constraints for orbital compute. | 中 | SM001, SM005, SM008 |
| CM012 | The retained sources imply a phased market in which one satellite could start commercial operations before a full network exists, even though the constellation itself remains a proposal under FCC review. | 中 | SM005, SM007 |
| CM013 | An adverse August 2026 SatNews article says shedding a megawatt of waste heat in vacuum is unsolved and that Cowboy's satellite design is not finished. | 中 | SM008 |
| CM014 | Public evidence is insufficient to isolate a comparable TAM, SAM, or SOM because the disclosed numbers mix macro capex, constellation count, node capacity, and valuation without pricing, customer conversion, or service-level assumptions. | 中 | SM001, SM003, SM005, SM007 |
| CP001 | Cowboy Space says it is building vertically integrated orbital infrastructure spanning launch vehicles, space-based power, and in-orbit compute. | 高 | SP001, SP004 |
| CP002 | Cowboy Space says each upper stage in its proposed architecture becomes a 1-megawatt data center in orbit. | 高 | SP001, SP004, SP008 |
| CP003 | Cowboy Space said in May 2026 that it planned its first satellite launch later that year to demonstrate space-to-Earth power beaming. | 中 | SP001 |
| CP005 | The assigned independent coverage frames Cowboy's competition for engineering talent against SpaceX, Blue Origin, and other in-space startups rather than against an established orbital-compute vendor set. | 中 | SP003, SP008 |
| CP006 | Caplight lists Starcloud, Relativity Space, and Orbital Inc. as Cowboy Space comparables with embedding-similarity scores of 87%, 83%, and 80% respectively. | 中 | SP006 |
| CP007 | Cowboy's FCC filing should be treated as a proposal under review rather than an approved or operating orbital data-center network. | 中 | SP007 |
| CP008 | NewSpace Economy reports that SpaceX had an orbital data-center application accepted for filing in February 2026 involving a proposed NGSO system of up to one million satellites. | 中 | SP007 |
| CP009 | NewSpace Economy reports that Google's Project Suncatcher paper added another large technology company to the debate over future space-based AI infrastructure. | 中 | SP007 |
| CP010 | The assigned reporting says hyperscalers are planning to spend nearly $700 billion on data-center capacity in 2026 while terrestrial sites face power, water, and permitting constraints. | 中 | SP003, SP007 |
| CP011 | Independent reporting in the pool says Cowboy's thesis still faces unresolved thermal, debris, and unfinished-satellite-design risk. | 中 | SP007, SP008 |
| CP012 | Silicon Valley Invest Club says no publicly reported revenue, ARR, or profitability figures were found in the provided sources. | 中 | SP004 |
| CP013 | Public sources in the assigned pool show Cowboy disclosed a $275 million Series B at a $2 billion valuation in May 2026. | 高 | SP001, SP003, SP004 |
| CP014 | Because the assigned pool shows no public price book, customer contracts, or operating network for Cowboy, the best-evidenced current substitute remains terrestrial AI data-center capacity rather than a proven orbital-compute vendor. | 中 | SP003, SP004, SP007 |
| CP015 | Cowboy's main moat claim is integrated launch, power, and compute ownership, but that same integration concentrates launch, licensing, and thermal execution into one critical path. | 中 | SP001, SP007, SP008 |
| CP016 | The assigned pool does not provide enough disclosed pricing or customer proof to build a hard win-loss benchmark against named peers. | 低 | |
| CI001 | Cowboy Space announced a $275 million Series B at a reported $2 billion valuation in May 2026. | 高 | SI001, SI002, SI003, SI008 |
| CI002 | Index Ventures led the Series B and the disclosed syndicate included IVP, Blossom Capital, SAIC, Breakthrough Energy Ventures, Construct Capital, Andreessen Horowitz, NEA, Interlagos, and founder Baiju Bhatt. | 高 | SI001, SI003, SI008 |
| CI003 | Public sources describe Cowboy's revenue concept as a vertically integrated system in which the launch vehicle's upper stage becomes a one-megawatt data center in orbit. | 高 | SI001, SI002, SI003, SI004, SI008 |
| CI004 | Space.com reported that Cowboy plans a power-beaming demonstration mission later in 2026, an optical-communications cluster mission in 2027, and a first proprietary rocket launch in 2028. | 中 | SI002 |
| CI005 | New Space Economy described the Stampede FCC filing as a proposal under review rather than an approved or operating orbital data-center system. | 中 | SI005 |
| CI006 | SatNews reported that Cowboy filed for up to 20,000 satellites and that disclosed equity to date stands at roughly $325 million. | 中 | SI004 |
| CI007 | Caplight's public company page shows fundraising metadata but no publicly reported revenue, ARR, or profitability figures. | 中 | SI006 |
| CI008 | Caplight states that the full shareholder register is not public and that filing records require access, leaving preference-stack and dilution terms undisclosed on the public page. | 中 | SI006 |
| CI009 | Silicon Valley Invest Club lists Cowboy Space at a $2 billion May 11, 2026 valuation with $275 million for the Series B and $335 million total funding raised. | 低 | SI007 |
| CI010 | Silicon Valley Invest Club estimates dilution at about 20% for the Series A and about 14% for the Series B using disclosed round sizes and post-money valuations. | 低 | SI007 |
| CI011 | No retained public source discloses customer count, backlog, LOIs, utilization, or realized pricing for Cowboy's proposed orbital compute services. | 中 | SI002, SI006, SI007 |
| CI012 | The public evidence supports a pre-commercial revenue narrative, but it does not disclose contracted workloads, price cards, or gross-margin data. | 中 | SI001, SI002, SI003, SI006, SI008 |
| CI013 | New Space Economy argues that Cowboy's architecture still depends on unproven assumptions about solar power, thermal rejection, optical communications, and launch economics at scale. | 中 | SI005 |
| CI014 | The assigned public evidence does not support cash-runway underwriting because neither burn nor cash on hand is disclosed alongside the Series B proceeds. | 中 | SI003, SI006, SI007 |
| CI015 | Third-party public datasets do not agree on total capital raised, with SatNews citing roughly $325 million disclosed equity to date and Silicon Valley Invest Club listing $335 million total funding raised. | 低 | SI004, SI007 |
| CE001 | Cowboy publicly defines its product as vertically integrated orbital infrastructure spanning launch vehicles, space-based power, and in-orbit compute. | 高 | SE001, SE007, SE008 |
| CE002 | Cowboy says the rocket upper stage and data-center payload are one vehicle, and each upper stage is intended to become a 1-megawatt data center in orbit. | 高 | SE001, SE002, SE005 |
| CE003 | Public filing coverage describes Stampede as a proposed constellation of up to 20,000 satellites in dawn-dusk sun-synchronous orbits between 700 and 1,000 kilometers. | 中 | SE003, SE004, SE006 |
| CE004 | Public sources say Stampede would rely primarily on optical links while using Ka-band telemetry, tracking, and command for selected mission phases and emergencies. | 中 | SE003, SE004 |
| CE005 | Cowboy and independent summaries identify NVIDIA Space-1 Vera Rubin modules as the announced compute partner for the in-orbit AI stack. | 高 | SE001, SE008 |
| CE006 | Cowboy says commercial operations can begin with as few as one satellite in orbit and then scale with later launches. | 中 | SE004 |
| CE007 | Independent reporting says Cowboy expects each satellite to mass 20,000 to 25,000 kilograms and generate 1 megawatt for just under 800 onboard GPUs. | 高 | SE002, SE005, SE009 |
| CE008 | Public reporting says Cowboy targets the first proprietary launch carrying a 1-megawatt orbital compute payload before the end of 2028. | 高 | SE002, SE007, SE009 |
| CE009 | Cowboy's public materials say the first satellite later in 2026 is meant to demonstrate space-to-Earth power beaming rather than full commercial service. | 高 | SE001, SE008 |
| CE010 | Cowboy's disclosed differentiation is vertical integration across launch cadence, orbital power supply, and compute payload design rather than dependence on third-party launch availability. | 中 | SE001, SE002, SE007 |
| CE011 | In September 2026, the FCC's Space Bureau asked Cowboy for physical dimensions including solar panels before the application could proceed in review. | 中 | SE006 |
| CE012 | Public sources say Cowboy's satellite design was still unfinished at filing time and would require a later license modification before service begins. | 中 | SE003, SE005, SE009 |
| CE013 | Independent commentary explicitly flags megawatt-scale heat rejection in vacuum and constellation-scale debris exposure as unresolved product risks. | 中 | SE005, SE006, SE009 |
| CE014 | SatNews reported in September 2026 that Cowboy had not publicly named the rocket engine, thrust figure, launch site, or a flown test article. | 中 | SE006 |
| CU001 | Cowboy's public customer thesis is built around terrestrial AI infrastructure constraints rather than around a disclosed installed customer base. | 中 | SU003, SU006, SU007 |
| CU002 | Independent market analysis says the most plausible early orbital-compute customers are Earth-observation operators, defense and security agencies, space-station users, lunar-infrastructure firms, and AI developers with unusual energy or location constraints. | 中 | SU008 |
| CU003 | The assigned customer evidence pool does not disclose named paying customers, customer count, backlog, LOIs, or production deployments for Cowboy. | 中 | SU001, SU002, SU003, SU004, SU006, SU007 |
| CU004 | Cowboy's disclosed collaboration with NVIDIA is partner proof for the compute stack but not customer proof for demand. | 中 | SU003, SU007 |
| CU005 | Cowboy says commercial operations could begin with as few as one satellite in orbit and then scale with subsequent launches. | 中 | SU006 |
| CU006 | Independent coverage lists contract wins with hyperscale and government customers as milestones to watch, implying those wins are targets rather than contracts already disclosed. | 中 | SU002 |
| CU007 | Cowboy's nearer-term public milestones are a later-2026 power-beaming satellite and engineering buildout rather than disclosed customer go-lives. | 中 | SU003, SU007 |
| CU008 | The near-term orbital-data-center market remains limited and specialized rather than a broad replacement for terrestrial cloud demand. | 中 | SU008 |
| CU009 | SatNews argues Cowboy still must prove launch cadence, thermal management, and debris economics before the customer thesis looks durable. | 中 | SU005 |
| CU010 | Because the most plausible early buyer set is narrow and mission-driven, Cowboy's first contracts would likely carry elevated concentration risk. | 中 | SU005, SU008 |
| CU011 | Public evidence in the assigned pool does not disclose retention metrics such as NRR, GRR, churn, contract length, or cohort behavior. | 中 | SU002, SU003, SU007 |
| CU012 | Cowboy's public adoption proof is still thesis-led because the retained evidence shows a filing, financing, partnerships, and hiring, but not live usage evidence. | 中 | SU001, SU003, SU004, SU007 |
| CR001 | Cowboy publicly disclosed a $275 million Series B at a $2 billion valuation in May 2026. | 高 | SR002, SR006 |
| CR002 | Stampede is still an FCC application under review rather than an approved operating constellation. | 高 | SR001, SR004 |
| CR003 | Cowboy's filing asks for multiple waivers because the system expects to rely mainly on optical links and limited Ka-band TT&C rather than routine RF service links. | 高 | SR001, SR008 |
| CR004 | Secure World Foundation urged the FCC to deny or defer Cowboy's waiver requests until the applicant provides a sufficiently detailed technical record to assess risks at full scale. | 中 | SR001 |
| CR005 | SWF said Cowboy had not previously operated a satellite constellation and its first demonstration mission was not slated until October 2026. | 中 | SR001 |
| CR006 | Public coverage describes Cowboy's architecture as an integrated system in which the launch-vehicle upper stage becomes the orbital data-center payload. | 中 | SR005, SR006, SR007 |
| CR007 | SatNews framed one-megawatt heat rejection in vacuum as an unsolved engineering problem for Cowboy's thesis. | 中 | SR003 |
| CR008 | Independent coverage says Cowboy's optical-link-heavy architecture still requires precise pointing, weather planning, ground-station diversity, and strong operational controls. | 中 | SR004, SR008 |
| CR009 | Cowboy's first proprietary launch is publicly targeted for 2028. | 中 | SR003, SR005, SR008 |
| CR010 | EverExcel said commercial-scale GPU data-center operations are targeted from 2028 as Cowboy's proprietary launch vehicle matures. | 中 | SR002 |
| CR011 | Independent coverage says Cowboy's satellite design is unfinished and will require a license modification before service begins. | 中 | SR003, SR007 |
| CR012 | SWF says Cowboy still needs system-level analysis and disclosure for cumulative collision risk, post-mission disposal performance, and aggregate interference effects. | 高 | SR001, SR004 |
| CR013 | Public sources show disclosed capital and a launch target, but they do not show that full-scale launch cadence, payload supply, and customer delivery risk have already been closed. | 中 | SR001, SR003, SR006, SR008 |
| CR014 | The current public record is still missing enough evidence to size burn, full-system capital intensity, and commercial proof with confidence. | 中 | SR001, SR002, SR004, SR006 |
| CR015 | Residual exposure remains highest where Cowboy still lacks public demonstration, final design closure, or granted authority. | 中 | SR001, SR003, SR004, SR007, SR008 |
| CV001 | Cowboy Space announced a $275M Series B at a $2B valuation in May 2026, led by Index Ventures with participation from IVP, Blossom Capital, SAIC, and prior investors. | 高 | SV001, SV004 |
| CV002 | Dealroom publicly lists Cowboy Space with a 2026 valuation range of $1–2.5B and disclosed equity funding of $100–500M. | 中 | SV002 |
| CV003 | Silicon Valley InvestClub lists Cowboy Space's latest valuation as $2B and total funding raised as $335M+. | 中 | SV005 |
| CV004 | Silicon Valley InvestClub says no public revenue, ARR, or profitability figures were found for Cowboy Space in the provided sources. | 中 | SV005 |
| CV005 | Cowboy said it plans to launch its first satellite later in 2026 to demonstrate space-to-Earth power beaming. | 高 | SV001, SV004 |
| CV006 | Independent coverage says Cowboy is targeting its first proprietary rocket launch carrying a 1 MW orbital compute payload by the end of 2028. | 中 | SV003 |
| CV007 | New Space Economy says the Stampede FCC filing should be treated as a proposal under review, not as an approved constellation or operating orbital data center network. | 中 | SV006 |
| CV008 | New Space Economy and SatNews describe Stampede as a plan for up to 20,000 satellites in 700 to 1000 km sun-synchronous orbits that rely mainly on optical links with limited Ka-band TT&C. | 中 | SV006, SV007 |
| CV009 | SatNews argues that Cowboy still has to prove thermal rejection, debris management, and unit economics before orbital data centers are investable at scale. | 中 | SV007, SV008 |
| CV010 | Dealroom shows 61 employees, 7 open positions, and 90.6% one-year team growth. | 中 | SV002 |
| CV011 | Caplight's public page exposes Cowboy's round chronology and investor names, but valuation details and filing documents are not available on the open page. | 中 | SV009 |
| CV012 | Silicon Valley InvestClub estimates dilution of about 20% in the Series A and about 14% in the Series B from disclosed round sizes and post-money valuations. | 中 | SV005 |
| CV013 | Public sources still do not disclose Cowboy's full shareholder register, liquidation preferences, or any debt or project-finance structure senior to new equity. | 中 | SV005, SV009 |
| CV014 | Silicon Valley InvestClub's public IPO yardsticks place AST SpaceMobile near ~$1.4B, Rocket Lab near ~$4.1B, and Planet Labs near ~$2.8B at IPO, so Cowboy's private $2B mark already sits inside the range of scaled public space businesses. | 中 | SV005 |
| CV015 | Given a disclosed $2B valuation, no public revenue disclosure, and unresolved regulatory and engineering proof points, public evidence supports a track rather than buy stance at the current price. | 中 | SV001, SV005, SV006, SV007 |
| CV016 | A more constructive entry would require public evidence of customer or revenue proof, successful demo and launch milestones, and clearer regulatory progress on Stampede. | 中 | SV003, SV005, SV006, SV007 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Cowboy Space Corporation | Cowboy Space Corporation | Orbital Data Centers for AI | By integrating a megawatt-class data center directly into the upper stage of our launch vehicle, we eliminate redundant mass and dedicate every possible kilogram to high-performance AI compute. It's a 1-megawatt data center with active thermal management and integrated compute, designed as a single vehicle from the outset. |
| SO002 | Cowboy Space Corporation | Introducing Cowboy Space: A $275 Million Series B | Cowboy Space | On May 11, 2026, we announced $275 million in Series B financing at a $2 billion valuation and introduced our new name: Cowboy Space Corporation. |
| SO003 | Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | Cowboy Space Corporation, the orbital infrastructure company founded by Robinhood co-founder Baiju Bhatt, today announced $275M in Series B financing at a $2B valuation led by Index Ventures, with participation from new investors IVP, Blossom Capital, and SAIC, alongside existing investors Breakthrough Energy Ventures, Construct Capital, Andreessen Horowitz, NEA, Interlagos and Baiju Bhatt. |
| SO004 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | Dissipating one megawatt of thermal load in vacuum is not a trivial engineering exercise, and a 20,000-satellite shell at 700–1,000 km has a debris liability that the optical-only architecture mitigates but does not eliminate. Cowboy’s own FCC filing concedes the satellite design isn’t done. |
| SO005 | Via Satellite | Cowboy Space Files for 20,000-Satellite Constellation for Data Center Services | Cowboy Space followed up its $275 million Series B round earlier this week by filing for a 20,000 satellite constellation to provide data center services from space. |
| SO006 | Silicon Valley Invest Club | Cowboy Space Corporation — $2B · Baiju Bhatt | In May 2026, Cowboy Space raised a $275M Series B at a $2B valuation , rebranded from Aetherflux, and filed with the FCC for a 20,000-satellite orbital data center constellation . |
| SO007 | Caplight | Cowboy Space Corporation Valuation & Funding History | Caplight | Round Date Amount Raised Valuation Investors Sources Series B May 11, 2026 Series A Apr 2, 2025 Grant Feb 1, 2025 Seed Oct 1, 2024 |
| SO008 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | As of May 15, 2026, the filing should be treated as a proposal under review, not as an approved constellation, completed system, or operating orbital data center network. |
| SO009 | SatNews | Silicon Valley Space Week Speaker Focus: Joseph Yaffe, Cowboy Space Corporation – SatNews | Joseph Yaffe is Chief Operating Officer and Chief Legal Officer of Cowboy Space Corporation.... Yaffe joined in January, when the company was still Aetherflux.... Shedding a megawatt of waste heat in vacuum is unsolved, a constellation of that size at that altitude carries real debris exposure, and Cowboy’s own FCC filing concedes the satellite design is not finished. |
| SM001 | Data Center Frontier | Cowboy Space and the Case for Orbital AI Data Centers | Hyperscale AI campuses can wait years for transmission upgrades and utility interconnections. |
| SM002 | Yahoo Finance | Cowboy Space Corporation eyes orbital data centers | rather than having the upper stage of the rocket, which traditionally carries a payload to space. The architecture that we're building actually has the upper stage of the rocket transformed into the data center in space. |
| SM003 | The Robotics Media | Cowboy Space Raises 5M Series B to Launch AI Data Centers Into Orbit | Hyperscalers are planning to spend nearly $700 billion on data center capacity in 2026 alone, and terrestrial sites are running into power, water and permitting walls. |
| SM004 | Cowboy Space Corporation | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | The company is developing a constellation of Low Earth Orbit satellites that harness solar energy, a purpose-built launch vehicle to deliver them and compute payloads designed specifically for the space environment. The rocket's upper stage and data center payload are designed as a single vehicle ... In orbit, each upper stage becomes a 1-megawatt data center. |
| SM005 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | the documents present Stampede as a system of up to 20,000 satellites in dawn-dusk sun-synchronous orbits between 700 and 1000 km altitude. |
| SM006 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | Each Stampede satellite is expected to weigh 20,000 to 25,000 kilograms, generate one megawatt of usable power from dawn-dusk sun-synchronous orbits between 700 and 1,000 kilometers, and host just under 800 GPUs. |
| SM007 | Satellite Today | Cowboy Space Files for 20,000-Satellite Constellation for Data Center Services | it can begin providing commercial operations with as few as one satellite in orbit |
| SM008 | SatNews | Silicon Valley Space Week Speaker Focus: Joseph Yaffe, Cowboy Space Corporation | The case against starts with thermal. Shedding a megawatt of waste heat in vacuum is unsolved, a constellation of that size at that altitude carries real debris exposure, and Cowboy's own FCC filing concedes the satellite design is not finished. |
| SP001 | Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | The rocket's upper stage and data center payload are designed as a single vehicle — not as a workhorse and its freight — eliminating redundant mass and optimizing power and compute delivered to orbit. In orbit, each upper stage becomes a 1-megawatt data center. |
| SP002 | SpaceNexus | Cowboy Space Corporation: Orbital Data Center | |
| SP003 | The Robotics Media | Cowboy Space Raises 5M Series B to Launch AI Data Centers Into Orbit | Cowboy is competing for engineers against SpaceX, Blue Origin and a flotilla of in-space servicing startups, including efforts like Redwire's MANUS lunar robotic arm work. |
| SP004 | Silicon Valley Invest Club | Cowboy Space Corporation — $2B · Baiju Bhatt | |
| SP005 | Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | Cowboy is also collaborating with NVIDIA to deploy NVIDIA Space-1 Vera Rubin Modules, bringing state-of-the-art AI infrastructure to the Low Earth Orbit environment. |
| SP006 | Caplight | Cowboy Space Corporation Valuation & Funding History | |
| SP007 | NewSpace Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | As of May 15, 2026, the filing should be treated as a proposal under review, not as an approved constellation, completed system, or operating orbital data center network. |
| SP008 | SatNews | Silicon Valley Space Week Speaker Focus: Joseph Yaffe, Cowboy Space Corporation | The case against starts with thermal. Shedding a megawatt of waste heat in vacuum is unsolved, a constellation of that size at that altitude carries real debris exposure, and Cowboy's own FCC filing concedes the satellite design is not finished. |
| SI001 | Unite.AI | Cowboy Space Raises $275M Series B at $2B Valuation to Build Orbital AI Data Centers | |
| SI002 | Space.com | Cowboy Space raises $275 million to launch AI data centers on brand-new rocket | |
| SI003 | Cowboy Space Corporation via Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | |
| SI004 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | |
| SI005 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | |
| SI006 | Caplight | Cowboy Space Corporation Valuation & Funding History | |
| SI007 | Silicon Valley Invest Club | Cowboy Space Corporation — $2B · Baiju Bhatt | |
| SI008 | The Robotics Media | Cowboy Space Raises 5M Series B to Launch AI Data Centers Into Orbit | |
| SE001 | Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically Integrated Orbital Data Centers and Rockets | The rocket's upper stage and data center payload are designed as a single vehicle — not as a workhorse and its freight — eliminating redundant mass and optimizing power and compute delivered to orbit. In orbit, each upper stage becomes a 1-megawatt data center. |
| SE002 | TechCrunch | There aren't enough rockets for space data centers — Cowboy Space raised $275 million to build them | Cowboy Space plans to build its data centers directly into the second stage of its rocket. ... The company expects each satellite to have a mass of 20,000 to 25,000 kilograms and to generate 1 MW of power for just under 800 onboard GPUs. |
| SE003 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | Together, the documents present Stampede as a system of up to 20,000 satellites in dawn-dusk sun-synchronous orbits between 700 and 1000 km altitude. ... The Technical Annex says Stampede will principally rely on optical laser links among satellites, between satellites and the ground... |
| SE004 | Satellite Today | Cowboy Space files for 20,000-satellite constellation for data center services | Instead, the company said it can begin providing commercial operations with as few as one satellite in orbit, and it plans to expand compute capabilities with subsequent launches. |
| SE005 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | The company's upper stage is the data center payload. ... The satellite design, Cowboy told the commission with admirable candor, is unfinished and will require a license modification before service begins. |
| SE006 | SatNews | The FCC Asked Cowboy Space How Big Its Satellites Are | Provide the physical dimensions of the system's space stations, in particular, the maximum cross-sectional area, including the dimensions of the solar panels. |
| SE007 | The Robotics Media | Cowboy Space Raises $275M Series B to Launch AI Data Centers ... | The full architecture is a constellation of low-Earth-orbit satellites that harness solar energy, a purpose-built launch vehicle to lift them, and compute payloads designed specifically for the radiation, thermal and power constraints of space. |
| SE008 | Silicon Valley Invest Club | Cowboy Space Corporation — $2B · Baiju Bhatt | Cowboy Space is developing Stampede — a constellation of Low Earth Orbit satellites that harness solar energy to run on-orbit GPU data centers — alongside a purpose-built launch vehicle whose upper stage becomes a 1-megawatt data center once in orbit. |
| SE009 | SatNews | Silicon Valley Space Week Speaker Focus: Joseph Yaffe, Cowboy Space Corporation | Shedding a megawatt of waste heat in vacuum is unsolved, a constellation of that size at that altitude carries real debris exposure, and Cowboy's own FCC filing concedes the satellite design is not finished. |
| SU001 | Data Center Dynamics | Cowboy Space submits FCC application for a 20,000 satellite orbital data center constellation | The design of the satellites remains unfinished, as do the company’s rockets, which themselves must be approved by the Federal Aviation Administration’s Office of Commercial Space Transportation. |
| SU002 | The Robotics Media | Cowboy Space Raises 5M Series B to Launch AI Data Centers Into Orbit | Near-term milestones include a first Galactic Brain test node, contract wins with hyperscale and government customers, and clarity on whether Cowboy will partner with existing launch providers for early flights or wait for its own vehicle. |
| SU003 | Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | Cowboy is also collaborating with NVIDIA to deploy NVIDIA Space-1 Vera Rubin Modules, bringing state-of-the-art AI infrastructure to the Low Earth Orbit environment. |
| SU004 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | |
| SU005 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | The harder questions are the ones Forrester raised in March and the ones our cooling piece sketched. Dissipating one megawatt of thermal load in vacuum is not a trivial engineering exercise, and a 20,000-satellite shell at 700–1,000 km has a debris liability that the optical-only architecture mitigates but does not eliminate. |
| SU006 | Satellite Today | Cowboy Space Files for 20,000-Satellite Constellation for Data Center Services | Instead, the company said it can begin providing commercial operations with as few as one satellite in orbit, and it plans to expand compute capabilities with subsequent launches. |
| SU007 | GeekWire | Cowboy Space raises $275M as it seeks 40-60 employees for new satellite and rocket hub in Seattle | |
| SU008 | New Space Economy | Orbital Data Center Companies Building Space-Based Compute Infrastructure | Early orbital customers are more likely to include Earth observation operators, defense and security agencies, space station users, lunar infrastructure firms, and AI developers testing models under unusual energy and location constraints. |
| SR001 | Secure World Foundation | SWF Filing: Comments on Cowboy Space's Orbital Data Center Application | Cowboy Space has not previously operated a satellite constellation at any scale, and its first demonstration mission is not slated until October 2026. |
| SR002 | EverExcel | EverExcel Participates in Cowboy Space Corporation's $275 Million Series B | The company plans to launch its first satellite later this year to demonstrate wireless power beaming from orbit to Earth, with commercial-scale GPU data centre operations targeted from 2028 as its proprietary launch vehicle programme matures. |
| SR003 | SatNews | Silicon Valley Space Week Speaker Focus: Joseph Yaffe, Cowboy Space Corporation | The case against starts with thermal. Shedding a megawatt of waste heat in vacuum is unsolved, a constellation of that size at that altitude carries real debris exposure, and Cowboy’s own FCC filing concedes the satellite design is not finished. |
| SR004 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | As of May 15, 2026, the filing should be treated as a proposal under review, not as an approved constellation, completed system, or operating orbital data center network. |
| SR005 | The Robotics Media | Cowboy Space Raises $275M Series B to Launch AI Data Centers Into Orbit | Building a brand-new launch vehicle to a 2028 debut, while also fielding compute payloads, is one of the more ambitious schedules in the sector. |
| SR006 | Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | The company plans to launch its first satellite to orbit later this year to demonstrate space-to-Earth power beaming. |
| SR007 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | The FCC filing requests multiple waivers, most notably because Stampede would rely primarily on optical inter-satellite and downlink communications rather than the congested radio bands every other megaconstellation is fighting over. The satellite design, Cowboy told the commission with admirable candor, is unfinished and will require a license modification before service begins. |
| SR008 | Satellite Today | Cowboy Space Files for 20,000-Satellite Constellation for Data Center Services | Cowboy Space is asking for a waiver from typical FCC schedule deployment rules that require an operator to deploy a certain number of satellites by a certain date with a bond and milestone framework in place. |
| SV001 | Cowboy Space Corporation via Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | |
| SV002 | Dealroom | Cowboy Space Corporation — Unicorn company profile | Dealroom | |
| SV003 | The Robotics Media | Cowboy Space Raises 5M Series B to Launch AI Data Centers Into Orbit | |
| SV004 | Cowboy Space Corporation via Business Wire | Cowboy Space Corporation Raises $275M Series B For Vertically-Integrated Orbital Data Centers and Rockets | |
| SV005 | Silicon Valley InvestClub | Cowboy Space Corporation — $2B · Baiju Bhatt | |
| SV006 | New Space Economy | Cowboy Space Seeks FCC Approval for Stampede Orbital Data Centers | |
| SV007 | SatNews | Cowboy Raises $275M and Files for 20,000 Orbital Data Centers, Forcing the ODC Thesis Into View | |
| SV008 | SatNews | Silicon Valley Space Week Speaker Focus: Joseph Yaffe, Cowboy Space Corporation – SatNews | |
| SV009 | Caplight | Cowboy Space Corporation Valuation & Funding History | Caplight |