Skeleton Technologies
战略验证已有改善,但公开财务数据和股权结构细节仍不足以支撑明确的投资判断。
Skeleton 已在 AI 电力领域实现商业化,并获得战略层面的验证;但公开财务数据和股权结构细节仍过于稀薄,尚不足以支持按最新私募估值继续做投资判断。
覆盖范围与披露说明
公开证据截至 2026-10-18。报告对商业进展、制造足迹和战略验证的把握最强,但确切估值、收入质量、客户集中度和优先权结构细节仍不完整。
封面要素
公司概况
Skeleton Technologies 是一家总部位于塔林、处于后期阶段的私有储能公司,销售基于超级电容器的 GrapheneBBU 模块 / 机架和 SuperBattery 系统,面向 AI 数据中心、电网稳定、交通、工业和国防。公司已在美国和欧洲展现出商业化进展,但公开披露仍未完整说明估值、收入质量、客户集中度和股权结构经济性。
- 创始人
- Taavi Madiberk, Oliver Ahlberg
- 创立地点
- Tartu, Estonia
- 总部
- Tallinn, Estonia
- 产品
- GrapheneBBU 模块和机架、SuperBattery 电芯与系统,以及面向 AI 基础设施和其他关键任务应用的相关高功率储能产品。
- 客户
- AI 数据中心、电网运营商、工业 OEM 及其他关键基础设施买家
- 商业模式
- 销售硬件和系统,并由合作伙伴为基础设施客户主导部署和制造项目。
- 阶段
- late-stage private / pre-IPO
- 融资情况
- 2026 年 5 月首轮交割为 €33 million;公开报道的风险投资总额为 €392 million,随后出现 Siemens 主导、估值超过 $1 billion 的战略估值信号。
执行摘要
主要优势
- AI 数据中心和电网基础设施已有真实商业进展,包括在美国部署超过 100 MW。
- Siemens 领投轮以及其他工业领域投资人的参与,提供了战略层面的验证。
- Leipzig、Varkaus 以及美国扩建中的制造布局,为供应能力提供了可信度。
主要风险
- 公开财务状况仍不透明:未披露收入、毛利率、现金消耗、现金 runway 或积压订单。
- 商业业务集中在 Siemens 和北美,存在集中度风险。
- 多地制造以及质量 / 合规执行情况仍缺乏充分验证。
未决问题
- 2026 年 10 月战略轮的确切投后估值、优先股堆叠和各类别股份的经济条款。
- 按产品线拆分的绝对收入、毛利率、现金消耗、现金 runway 和积压订单。
- 已知超大规模云厂商 / 客户的集中度、合同期限和续约行为。
- 2025-2026 年制造扩张后的最新认证和质量证据。
目录
01公司概览
1.1 身份、足迹与阶段
指定来源池支持将 Skeleton 定义为一家围绕超级电容器和 SuperBattery 系统打造的高功率储能公司,而非普通电池供应商。公司和独立来源都持续将其业务与 AI 数据中心、电网应用、交通、工业和国防联系起来。其起源比当前法律主体情况更清晰:各来源一致认为公司于 2009 年在爱沙尼亚创立,但本来源池中关于当前总部的依据来自商业资料,而非公司官方披露。公司所处阶段的证据强于其法律结构:管理层公开讨论 IPO 准备、向超大规模云厂商交付以及美国制造扩张,这些共同支持“后期私有 / IPO 前”的定位。[CO001, CO002, CO003, CO006, CO015]
| 指标 | 数值 / 状态 | 日期 | 信心 | 缺口 / 备注 |
|---|---|---|---|---|
| 阶段 | 后期私有 / IPO 前 | 2026 | 中 | 基于 IPO 准备口径,而非公开申报状态。 |
| 创立年份 | 2009 | 2009 | 高 | 创立年份得到公司历史和独立报道交叉印证。 |
| 总部 | Tallinn, Estonia(仅商业资料) | 2026-10-16 | 中 | 指定官方来源集未重新说明当前总部。 |
| 已披露风险投资总额 | €392m | 2026-05-07 | 高 | 官方首轮交割公告中的数字。 |
| 最新估值信号 | >$1bn | 2026-10 | 中 | 据报道与 Siemens 主导的战略轮次有关;确切融资金额和稀释情况未披露。 |
| 北美收入占比 | 约一半收入 | 2026 | 中 | 公司声明,但没有绝对收入基数。 |
| 美国已部署系统 | >100 MW | 2026 | 中 | 来自美国扩张公告的公司声明。 |
| 员工数 | 201-500 employees | 2026-10-16 | 低 | 商业资料估算,并非官方披露。 |
| 客户数量 | 2026-10-18 | 缺口 | 指定公开证据提到部署和测试,但没有经过验证的客户数量。 | |
| 绝对收入 | 2026-10-18 | 缺口 | 本来源池仅披露区域收入占比,没有公司整体收入。 |
null 值标记指定证据池中没有公开支持的指标;信心水平不一,反映官方披露、独立报道和商业资料估算之间的差异。
[CO002, CO003, CO007, CO009, CO010, CO014, CO015]1.2 领导层、治理透明度与关键人物依赖
公开领导层证据范围有限,但前后一致。Taavi Madiberk 在公司的融资和扩张沟通中均以 CEO、联合创始人身份出现,外部报道则将他与联合创始人 Oliver Ahlberg 并列。本来源池没有提供董事会名单、详细控制权图谱,也没有创始人和一条招聘信息以外的完整高管名单。因此,创始人与市场的匹配度比治理韧性更容易得到支持。在董事会构成、继任规划和运营团队深度得到证实之前, Madiberk 仍是公开叙事的核心,公司在融资和商业定位上存在较明显的关键人物依赖。[CO004, CO005]
1.3 融资历史、投资者与当前阶段信号
在当前来源池中,资本形成是最清晰的阶段信号。Skeleton 宣布 2026 年 5 月首轮交割为 €33 million,称风险投资总额达到 €392 million,并将本轮定位为准备 2027 年美国 IPO。披露的新投资者包括 Axon Partners Group、SmartCap 和 Taiwania Capital;Tech Funding News 补充称,现有投资者 CBMM 也参与了首轮交割。截至 2026 年 10 月,Passive Components 和 toflow 均称 Siemens 主导的战略投资使估值超过 $1 billion,叙事也从单纯融资转向 AI 基础设施的战略分销和制造规模化。[CO007, CO008, CO009, CO011, CO015]
| 利益相关方 | 角色 | 控制权或经济重要性 | 公开证据 | 尽调要求 |
|---|---|---|---|---|
| Siemens | 战略投资者与商业伙伴 | 据报道,2026 年 10 月轮次给予 Skeleton 超过 $1bn 估值,并支持美国制造和硅谷研发。 | Passive Components 和 toflow 将 Siemens 与该战略轮次联系起来。 | 确认现金金额、治理权、独家范围和稀释。 |
| Axon Partners Group | 2026 年 5 月新投资者 | 被列为 IPO 前轮首轮交割参与者。 | 2026 年 5 月官方公告将 Axon 列为新投资者。 | 核查董事席位、持股比例和后续投资能力。 |
| SmartCap | 2026 年 5 月新投资者 | 被列为 IPO 前轮首轮交割参与者。 | 2026 年 5 月官方公告将 SmartCap 列为新投资者。 | 核查持股比例和任何治理保护。 |
| Taiwania Capital | 2026 年 5 月新投资者 | 管理层强调其对台湾芯片生态的战略相关性。 | 2026 年 5 月官方公告点名 Taiwania,并引用其 CEO 对供应链联系的表述。 | 将商业引荐潜力与被动资本区分开。 |
| Marubeni | 早期战略支持者 / 伙伴 | 历史资料提到 2021 年战略协议,后续融资摘要仍将 Marubeni 列为主要支持者。 | 公司历史和商业融资摘要都提到 Marubeni。 | 确认当前经济权益、权利和积极的市场拓展角色。 |
这是不完整的公开利益相关方视图,而非股权结构表。它只记录指定来源池明确点名的投资者或伙伴,不证明持股比例或董事会控制权。
[CO007, CO008, CO009, CO015]1.4 里程碑、规模信号与未解决的披露缺口
运营故事强于披露集合。Skeleton 的历史资料显示,公司自 2009 年创立以来持续建设,经历了多项重大制造里程碑,包括 2025 年莱比锡 Superfactory 开业以及开始向 AI 数据中心超大规模云厂商交付。美国扩张公告进一步称,公司在美国已部署超过 100 MW,北美客户贡献约一半收入。但同一来源池仍留下重要投资判断缺口:没有经过验证的绝对收入、官方客户数量、确认的当前董事会构成、Siemens 在 2026 年 10 月的确切现金投入,也没有披露核心性能声明的测试边界。因此,本章对公司的相关性和发展势头更有把握,对精确财务规模则不足。[CO010, CO011, CO012, CO013, CO014, CO016]
有公开证据支持的里程碑显示,公司先经历了长期的制造布局,随后在 2025—2026 年更明显地转向 AI 数据中心基础设施和 IPO 准备融资。
首个 ESA 客户 [CO016]
首款商业产品发布 [CO016]
首座大型德国工厂 [CO016]
Marubeni 战略协议 [CO016]
收购芬兰电池工厂 [CO016]
Leipzig Superfactory 开业 [CO013]
Siemens 牵头战略轮 [CO009]
1.5 图表
02市场分析
2.1 市场边界与现状替代方案
相关市场不是泛化的储能市场,也不是所有数据中心基础设施支出。Skeleton 保留的公开材料将核心需求定位为 AI 数据中心的关键任务电力系统,同时延伸至电网稳定、交通、工业和国防。按此口径,最相关的支出包括短时削峰、功率平滑,以及能快速应对 GPU 负载波动的备用电力硬件。传统锂离子备用系统、通用 UPS 架构和公用事业电网升级仍是主要现状替代方案,因此市场边界应由买家管理快速功率瞬变的需求界定,而不是由所有电池或服务器资本支出界定。[CM001, CM002, CM004, CM013]
| 细分 / 类别 | 包含支出 | 排除支出 | 买家 / 付款方 | 相关性 |
|---|---|---|---|---|
| AI 数据中心功率平滑 | 快速响应储能、控制和集成,用于削减 GPU 集群负载峰值 | 服务器芯片、冷却和长时电池建设 | 超大规模云厂商或运营商的基础设施及电气资本支出负责人 | 保留来源中增长最快的服务细分 |
| 数据中心备用电力 / BBU | 面向关键任务计算的短时电池备用单元、机架和模块 | 与同一采购动因无关的柴油发电和通用 UPS 更新 | 数据中心电气工程、设施和采购预算 | 与 GrapheneBBU 产品范围直接匹配 |
| 电网稳定与电能质量 | 平滑电网波动并推迟加固的快速响应系统 | 长时储能、大规模发电和输电建设 | 公用事业、电网伙伴和大型工业电力所有者 | 共享同一高功率价值主张的相邻市场 |
| 交通、工业和国防相邻市场 | 面向铁路、重型运输、工业脉冲负载和国防的高功率短时储能 | 乘用电动车电池和通用固定式储能 | OEM 项目负责人和关键任务运营团队 | 相关相邻市场,但不是 AI 数据中心核心规模估算口径 |
边界行将 Skeleton 的短时高功率用例与普通电池、服务器和发电资本支出区分开来;替代方案仍是现有电池备用系统和电网升级支出。
[CM001, CM002, CM004, CM013]2.2 买家路径与采用驱动因素
买家路径涉及多个利益相关方。在 AI 数据中心,使用者是现场电力和运营团队,但付款方通常是基础设施或电气资本支出预算负责人,他们希望增加算力,而不必等待规模巨大的电网升级。各来源对公开需求逻辑的描述一致:AI 数据中心负载峰值明显,数据中心用电需求上升,电网压力正成为部署瓶颈。Skeleton 自有材料进一步称,这一痛点使美国成为其增长最快的市场;公司在美国已部署超过 100 MW,约一半收入来自北美,从而把市场紧迫性直接连接到当前商业进展。[CM003, CM004, CM005, CM013]
| 细分市场 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 超大规模云厂商 / AI 训练园区 | 数据中心基础设施和电气工程团队 | 运行 GPU 集群的站点电力运营团队 | 基础设施资本开支和电力可用性预算 | 平滑短时、高幅电力尖峰,不必等待过度扩容的电网升级 | 数据中心基础设施副总裁或总监 | 并网瓶颈和不断上升的峰值负载成本 |
| 数据中心备用电源项目 | 关键任务设施和平台采购团队 | 负责正常运行时间和机架级韧性的运营团队 | 设施资本开支和韧性支出 | 用短时高功率机架替换或补充现有电池备用设备 | 设施工程或平台电力负责人 | 需要快速响应的备用电源和更高的机架功率密度 |
| 公用事业 / 电网稳定项目 | 公用事业电网现代化和项目开发团队 | 电网运营商和电能质量工程师 | 输配电项目预算 | 吸收快速波动,并推迟受压站点的部分电网加固 | 电网投资或电能质量项目负责人 | AI 园区或工业站点带来的本地负载波动 |
| OEM / 工业关键任务项目 | OEM 项目管理和采购团队 | 车辆、铁路、医疗或工业设备运营商 | 产品线或项目预算 | 提供突发功率、启动支持或电能质量稳定 | 产品工程或平台总经理 | 关键任务可靠性要求和循环寿命优势 |
买方、用户和付款方会随终端市场变化;现有来源支持多细分市场的采用路径,而不是单一且统一的预算负责人。
[CM003, CM004, CM011, CM013]2.3 受证据约束的规模估算与供给代理指标
公开证据支持多种规模观察角度,但没有一种是真正的 TAM、SAM 或 SOM。需求侧最有力的角度是数据中心用电量整体上升和 AI 专属负载增长;公司侧最有力的角度是已部署兆瓦数、北美收入占比和工厂产量。这些代理指标表明该品类具有商业意义且资本密集,但彼此不可相加,并混合了市场需求、已安装基础和制造能力。因此,投资者应将其视为采用和规模的边界标记,而非经过协调的市场模型替代品。[CM003, CM009, CM010, CM011, CM012, CM016]
| 发布方 / 视角 | 年份 | 地理范围 | 数值 | CAGR | 方法论 | 置信度 | 局限 |
|---|---|---|---|---|---|---|---|
| TFN / 引用 IEA 的需求视角 | 2025-2030 | 美国 | 数据中心用电量到 2030 年翻倍;AI 专属需求增至三倍 | 将第三方转述的宏观需求增长用作品类需求代理 | 中 | 不是 Skeleton 专属 TAM,且指定证据池未引用 IEA 原始表格 | |
| Skeleton 美国部署基数视角 | 2026 | 美国 | 已部署 100+ MW | 公司披露的装机基数代理,反映当前服务市场 | 中 | 装机基数不等于可触达的总需求 | |
| Skeleton 北美收入结构视角 | 2026 | 北美 | 约占收入一半 | 公司披露的收入占比代理,反映商业集中度 | 中 | 未披露绝对收入或分部毛利率 | |
| Skeleton Varkaus 供给视角 | 2025 | 芬兰 / 全球出货 | 1 GW 电池功率输出 | 用工厂产出代理近期向 AI、电网、出行和国防项目的供给 | 中 | 供给能力不等于买方预算或实际需求 | |
| Leipzig 产能视角 | 2025-2026 | 德国 | 年产最高 12M 个电芯 | 以 DCD 披露的铭牌电芯产能代理制造规模 | 中 | 未披露产能利用率,也未说明电芯如何转化为系统功率和收入 |
这些视角混合了宏观需求、部署基数、收入结构和工厂产出;它们只是市场规模的方向性约束,不是可相加的 TAM、SAM 或 SOM 数字。
[CM003, CM009, CM010, CM012, CM014, CM015, CM016]最有力的公开规模证据是需求、部署和供给代理指标的组合,而不是一套相互勾稽的 TAM/SAM/SOM 模型。
各层宽度仅表示层级,不代表数值比例。
各层按从宽泛需求背景到更窄的服务市场和供给代理指标排列;它们不可相加,且有意混用了不同单位。
[CM003, CM005, CM008, CM009, CM010, CM011, CM012, CM016]2.4 采用约束、矛盾与尽调缺口
主要约束不是市场需求不足,而是公开证明的精确度有限。指定来源池没有 AI 功率平滑或备用电力支出的独立预算模型,没有第三方基准验证 Skeleton 相对于现有电池系统的节能声明,也没有把电芯产量与系统级收入衔接起来的协调桥接。公开材料还分别使用 40% 和 45% 的节能表述,彼此未协调;莱比锡产能披露也没有把电芯换算为出货系统、利用率或收入。这些限制并不否定采用逻辑,但在管理层或客户提供协调一致的商业证据前,会降低对可用于估值的市场规模的信心。[CM006, CM007, CM008, CM009, CM015, CM016]
2.5 图表
03竞争对手
3.1 在指定证据中,解决方案类别比点名的同业更重要
指定来源池展现的竞争格局以类别为主,而非以供应商为主。官方和第三方描述反复将 Skeleton 定位于 AI 基础设施的高功率平滑、备用和电网稳定需求,并将其与传统锂离子电池或基础设施建设对比,而不是与一份完整列出的直接超级电容器竞争者名单对比。买家至少有四条可见路径解决同一问题:Skeleton 的机架到电网电力系统、Skeleton 的短时 SuperBattery 备用层、电池支持的替代方案,以及现有公用事业或设备升级。缺少清晰的点名同业集合本身就有信息量:公开定位更强调待解决任务的竞争和采购替代,而非供应商之间的正面比较。[CP001, CP007, CP008, CP009, CP013]
| 解决方案类别 | 类别 | 公开规模 / 支持 | 目标细分市场 | 已观察到的差异化 | 局限 |
|---|---|---|---|---|---|
| Skeleton AI 电力系统 | 直接方案 | 在美国已部署 100+ MW;约 50% 收入来自北美 | 超大规模云厂商、数据中心运营商、公用事业、OEM | 微秒级机架到电网响应,以及系统级集成 | 公开来源未披露标准定价或具名超大规模云厂商合同 |
| Skeleton SuperBattery / GrapheneBBU | 直接方案 | Varkaus 年产 1 GW;欧洲优先的 AI 电池工厂叙事 | 关键任务计算、AI 备用电源、电网稳定 | 以氧化铌 SuperBattery 材料为基础的短时备用电源 | 买方可见的持续时间、更换周期和合同条款未披露 |
| 电池备用替代方案 | 替代方案 | 指定证据池仅有品类级证据;未保留具名供应商 | AI 备用电源和正常运行时间工作负载 | 公开比较主要来自 Skeleton 与传统锂离子电池的对比 | 该证据池没有逐供应商的定价、持续时间或性能证据 |
| 公用事业 / 设备扩建 | 现状替代方案 | 走已装电网和电气设备路径,而非保留的产品来源 | 公用事业和数据中心基础设施所有者 | 新建电厂并升级设备,避免依赖现场缓冲 | 不是同类产品比较,且此处未披露资本开支周期 |
| 伙伴主导的基础设施渠道 | 现有市场渠道 | Siemens 投资、OEM 验证、供应商关系 | 通过已知渠道触达企业和基础设施买方 | 现有嵌入电力栈的龙头企业可以带来信任和渠道 | 独家权利细节仅由该证据池中的低信誉来源支持 |
| 具名直接超级电容同行 | 直接同行缺口 | 指定证据中未保留具名同行普查 | 最直接的供应商对供应商比较集合 | 这一缺口本身说明 Skeleton 的公开叙事重点是待解决任务的替代方案 | 当前证据池无法支持完整的具名同行矩阵 |
这是部分公开市场图景:各行反映指定预取来源中可见的明确解决方案类别和渠道路径,并非所有具名竞争者的完整普查。
[CP001, CP005, CP006, CP007, CP008, CP009, CP011, CP012, CP013]| 购买标准 | Skeleton AI 系统 | SuperBattery / GrapheneBBU | 电池备用替代方案 | 现状扩建 | 公开含义 |
|---|---|---|---|---|---|
| 应对 AI 尖峰负载 | 微秒级;独立报道低于 1 毫秒 | 面向关键任务计算的高功率备用电源 | 指定证据池没有显示该类别具备微秒级能力这一主张 | 电厂和设备升级靠扩建基础设施应对负载,而非即时缓冲 | 当痛点是快速负载波动而非纯粹续航时,Skeleton 更有优势 |
| 首要待解决任务 | 机架到电网平滑、电能质量、基础设施利用率 | 短时备用和正常运行时间保护 | 备用能源替代方案类别 | 增加或升级资产,提高容量和稳健性 | 比较集合针对具体工作负载,而非单一通用电池市场 |
| 化学体系 / 供应链叙事 | 官方美国扩张页面称不含锂、钴或锰 | 来自 Varkaus 的氧化铌 SuperBattery 电芯 | 指定证据提及传统锂离子电池和原有 LFP 电池资产 | 以设备为主的电网路径,而非电芯化学体系 | 供应链独立性是 Skeleton 采购叙事的一部分 |
| 可见部署证明 | 100+ kW 机架,且在美国已部署 100+ MW | 工厂年产 1 GW,且 AI 电池工厂已启动 | 该证据池未保留具名供应商证明 | 保留来源暗示已有装机基础,但没有量化 | Skeleton 的公开部署细节多于其保留的替代方案类别 |
| 渠道与信任支持 | 超大规模云厂商交付,以及 Siemens / GE / Amtrak / OEM 引用 | CBMM 和 InnoEnergy 围绕制造与材料提供的可信度 | 保留证据未知 | 熟悉的龙头企业路径,但没有保留的直接来源描述 | 买方对伙伴背书的看重程度可能不亚于原始技术指标 |
Unknown 表示指定公开证据池没有提供可支持的品类级证据;不表示该能力不存在于市场。
[CP001, CP003, CP005, CP006, CP007, CP008, CP009, CP010, CP011, CP013]这是对指定证据中明确竞争位置的序数图,比较快速响应匹配度与可见的渠道 / 制造杠杆,而非无依据的价格或市场份额估算。
- X: 快速平衡 AI 负载的匹配度(序数)
- low: 快速负载平衡匹配度低
- high: 快速负载平衡匹配度高
- description: Evidence-backed ordinal assessment from assigned public sources only
- scale: ordinal
- Y: 可见渠道 / 制造杠杆(序数)
- low: 可见杠杆有限
- high: 可见杠杆强
- description: Ordinal assessment based on retained public evidence rather than market-share data
- scale: ordinal
| 条目 | 快速平衡 AI 负载的匹配度(序数) (1–5) | 可见渠道 / 制造杠杆(序数) (1–5) | 说明 |
|---|---|---|---|
| 1. Skeleton AI 电力系统 | 5 | 4 | 微秒级机架到电网响应、100+ kW 机架设计点,以及在美国部署的 100+ MW。 |
| 2. Skeleton SuperBattery / GrapheneBBU | 4 | 3 | Varkaus 产能为 1 GW,并明确定位于任务关键型备用电力,而非长时储能。 |
| 3. 电池备用替代方案 | 2 | 3 | 保留证据主要按化学体系和产品类别,将 Skeleton 与传统锂离子备用电池作比较。 |
| 4. 公用事业 / 设备扩建 | 1 | 5 | 公司官方口径称,应对 AI 峰值负载也可以增建电厂、升级设备。 |
| 5. 合作伙伴主导的基础设施渠道 | 3 | 5 | 合作伙伴和 OEM 参考扩大了市场进入路径,但独家权利细节的可信度较低。 |
| 6. 主权供应定位 | 3 | 4 | 欧洲制造独立性和非锂定位是其采购叙事的一部分。 |
坐标是锚定来源定性证据的序数综合值,不是市场份额、价格或基准评分。
Skeleton AI 电力系统 [CP001, CP005, CP006]
Skeleton SuperBattery / GrapheneBBU [CP005, CP013]
公用事业 / 设备扩建 [CP009]
3.2 Skeleton 的公开切入口是响应速度加系统集成,而非已披露价格
在官方页面和独立报道中,Skeleton 都把自己呈现为系统供应商,而非零部件供应商。反复出现的差异化因素包括微秒级或亚毫秒级响应、功率电子与控制,以及从机架到电网连接稳定高度波动 AI 负载的能力。公开可见的规模也增强了可信度:莱比锡和 Varkaus 提供制造分量,官方及独立的美国扩张报道则称公司已部署超过 100 MW,北美客户贡献约一半收入。对投资判断同样重要的是来源池未披露的内容:没有买家可见的目录价格、标准合同结构,也没有公开 TCO 基准让买家清楚比较 Skeleton 与备用电池或现有基础设施方案。[CP002, CP003, CP005, CP006, CP010]
| 解决方案类别 | 公开价格 | 单位 / 合同模式 | 包含的能力 | 关键未知项 | 含义 |
|---|---|---|---|---|---|
| Skeleton AI 电力系统 | 未披露 | 覆盖从机架到电网的系统部署 | 在一个架构中整合电力电子、控制算法和储能技术 | 折扣、合同期限、软件 / 服务搭售、实际节省 | 公开证据更能支持能力判断,而不是清晰的 TCO 判断 |
| Skeleton SuperBattery / GrapheneBBU | 未披露 | 工厂支持的模块和备用电源系统 | 短时 AI 备用电源和高功率缓冲 | 单机架价格、维护周期、服务义务 | 买方尽调必须检验正常运行时间提升是否足以证明系统成本合理 |
| 电池备用替代方案 | 指定证据池未披露 | 备用电池解决方案类别 | 在备用角色上竞争,但未保留具名供应商的产品包细节 | 具名供应商、持续时间、安全权衡、集成成本 | 该证据池只能支持品类级比较 |
| 公用事业 / 设备扩建 | 指定证据池未披露资本开支 | 电厂、电网和设备升级 | 增加基础设施余量,而非本地缓冲 | 交付周期、边际电力成本、融资负担 | 这是现实的替代路径,尽管不是直接的产品 SKU |
该证据池没有 Skeleton 或具名替代供应商的买方可见目录价格,因此本展项比较的是产品组合和尽调未知项,而不是假装展示没有依据的价格对等。
[CP003, CP008, CP009, CP013]3.3 分销杠杆有潜力,但护城河持久性仍难证明
Skeleton 的公开信任故事不止是电芯化学。公司列举 Siemens、GE、Amtrak/Siemens Mobility US、BMW 和 IndyCar 等知名合作伙伴与参考客户关系;合作伙伴和公司来源强调欧洲制造独立性及跨大西洋供应链。这些信号有助于采购和监管接受度,尤其是基础设施买家不愿暴露于以中国为中心的关键材料供应。不过,公开材料对护城河持久性的说明仍不充分。一份可信度较低的销售情报资料称 Siemens 获得数据中心业务的全球独家销售权;若属实,这会扩大市场通路,但也把杠杆集中到一个渠道。德国、芬兰和计划中的美国扩产同时推进,执行风险仍在制造、客户获取和渠道协同,而不只是实验室性能。[CP004, CP011, CP012, CP014]
3.4 图表
04财务
4.1 收入模式与已披露进展
公开证据显示,Skeleton 通过销售高功率储能硬件和系统变现,而不是依靠经常性软件订阅。独立报道和公司材料描述了超级电容器电芯、模块、大型系统,以及 GrapheneGPU 或 GrapheneBBU 在 AI 数据中心和电网应用中的部署。最清晰的收入结构数据来自公司自述:Skeleton 称美国已部署超过 100 MW,北美客户贡献约一半公司收入。这是有意义的商业进展,但保留证据没有披露客户层面的合同金额、定价结构、订单积压或服务附加收入,因此当前收入质量仍需通过尽调判断,不能只依赖公开披露。[CI001, CI002, CI003, CI006, CI016]
| 收入来源 | 机制 | 单位 | 当前数值 / 状态 | 质量 | 尽调要求 |
|---|---|---|---|---|---|
| AI 数据中心电力系统 | 面向超大规模云厂商和 AI 工厂项目销售 GrapheneGPU / GrapheneBBU 类电力平滑和备用系统 | 已部署 MW / 系统出货 | 在美国已部署 '>100 MW';保留来源还描述了向美国主要超大规模云厂商交付 | 公司声称的市场牵引力有实际意义,但未公开合同金额或利润率 | 按客户拆分的收入、每 MW 平均售价、毛利率,以及续约或扩张模式 |
| 电网稳定系统 | 面向输电和电网应用销售基于超级电容器的系统,通常通过大型工业交易对手销售 | 项目 / 系统销售 | Leipzig 已向 Siemens、General Electric 和 Hitachi Energy 交付,用于欧洲电网应用 | 具名交易对手支持需求质量;未披露积压订单、合同期限和服务经济性 | 积压订单、维护搭售、付款条款和项目毛利率 |
| 超级电容器电芯和模块 | 在德国制造电芯、模块及配套系统,并销售至 OEM、工业和基础设施应用 | 电芯 / 模块 | Leipzig 的设计年产能最高为 12 million 个电芯,公司材料称已开始向 AI 数据中心超大规模云厂商交付 | 产能公开,但利用率和实际 €/cell 未披露 | 按产品系列拆分的良率、利用率、报废率和实际定价 |
| SuperBattery 产品 | 面向 AI、出行、电网、国防及伙伴集成解决方案的高功率电池产品 | GW 电池功率 / 电池包销售 | Varkaus 工厂已开业,披露投资 €50 million,电池功率输出为 one gigawatt | 资产已公开,但商业爬坡、保修负担和客户集中度未披露 | 订单、每电池包平均售价、保修准备金和营运资金状况 |
公开证据混合了公司声称和第三方佐证;缺失的商业字段是承保判断缺口,不代表没有业务。
[CI002, CI005, CI006, CI007, CI008]| 产品 | 价格 / 单位 / 合同 | 目录价与实际价格 | 折扣 / 未知项 | 来源 |
|---|---|---|---|---|
| AI 数据中心电力系统 | 保留来源未披露公开目录价或合同定价 | 实际平均售价、批量承诺、服务收入以及伙伴 / 渠道利润分成未披露 | 官方部署和融资来源描述了使用场景,但没有定价 | |
| 电网稳定系统 | 未披露公开目录价 | 项目范围、性能保证和维护定价未披露 | 官方工厂和交付来源披露了客户,但没有商业条款 | |
| SuperBattery 产品 | 未披露公开目录价 | 电池包级平均售价、保修条款、融资或租赁选项以及折扣安排未披露 | 官方芬兰发布和融资文章描述了产能及资金用途,但没有实际价格 |
价格单元格为空,表示保留的公开证据没有披露目录价或实际价格;不表示产品免费或无关紧要。
[CI016]公开证据显示,这是一条由硬件驱动的桥接链:从 AI 和电网电力瓶颈,经过工厂产出和渠道交付,最终到收入披露不完整。
| 节点 / 连接 | 节点 / 起点 | 终点 | 说明 |
|---|---|---|---|
| 节点 1 | AI 与电网电力瓶颈 [demand] | AI 工作负载造成剧烈功率波动,公用事业则面临快速响应的电网需求。 | |
| 节点 2 | 电芯、模块和电力系统 [products] | 保留的来源描述了超级电容器电芯、模块、更大型系统、GrapheneGPU 和 SuperBattery 产品。 | |
| 节点 3 | Leipzig 与 Varkaus 制造 [factories] | 已披露的欧洲工厂投资在 Leipzig 和 Varkaus 合计至少达到 €270 million。 | |
| 节点 4 | 直接 OEM 与合作伙伴渠道 [channels] | 已点名的交易对手包括 Siemens、General Electric、Hitachi Energy 和美国主要超大规模云厂商;Siemens 还拥有数据中心独家权。 | |
| 节点 5 | 装机基数与出货 [deployments] | Skeleton 称其在美国部署了超过 100 MW,并已向一家美国大型超大规模云厂商交付数百万颗电芯。 | |
| 节点 6 | 收入与毛利结果 [revenue] | 据称约一半收入来自北美,但实际定价、毛利率、现金消耗和订单积压仍未披露。 |
该流程是定性且有序的;由于保留的证据没有披露这些经济性,图中没有量化节点之间的转化率。
[CI001, CI002, CI003, CI006, CI009, CI015, CI016]4.2 制造足迹与资本密集度
Skeleton 的经济画像更接近资本密集型工业科技,而非轻资产软件。公司公开披露了 €220 million 的莱比锡 SuperFactory 和 €50 million 的 Varkaus SuperBattery 工厂;在美国扩产前,近期欧洲工厂投资至少达到 €270 million。莱比锡定位为高产量超级电容器工厂,年产量最高 12 million 个电芯;Varkaus 则为 AI、电网、交通和国防应用增加 1 GW 电池功率产能。这些资产可以改善供应控制和交付可信度,但也使利用率、良率、报废、保修计提和营运资本周转成为利润率判断的核心,而保留证据没有公开这些变量。[CI004, CI005, CI007, CI008, CI009, CI010, CI016]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调要求 |
|---|---|---|---|---|
| 北美收入结构 | 公司收入约 50% 来自北美客户 | 中 | 显示当前商业重心,以及潜在的客户或区域集中度 | 按地区拆分的收入和头部客户集中度 |
| 美国部署基数 | 在美国已部署并运行 '>100 MW' | 中 | 可作为商业牵引力和客户参考价值的装机基数代理 | 按客户群拆分的 MW、利用率和后续转化 |
| 最近披露的欧洲工厂资本开支 | Leipzig(€220 million)和 Varkaus(€50 million)合计至少 €270 million | 高 | 工厂资本开支会带来折旧、利用率风险和融资需求 | 资本开支计划、补贴、折旧政策和维护性资本开支 |
| Leipzig 产出能力 | 年产最高 12 million 个电芯 | 高 | 电芯吞吐量是固定成本吸收和毛利率潜力的关键驱动因素 | 利用率、良率、报废率和爬坡时间表 |
| Varkaus 产出能力 | 电池功率 one gigawatt | 高 | 电池爬坡会影响产品组合、库存需求和保修风险 | 订单、良率、保修计提和营运资金周转 |
| 毛利率、CAC、回本期、烧钱、现金跑道和积压订单 | 未公开披露 | 低 | 这些是评估收入质量、资本充足性和融资依赖的核心输入 | 月度管理账、客户群经济性、积压订单桥接表和融资契约包 |
该表混合了已披露的经营代理指标和明确缺失的承保判断输入;未发布指标应视为尽调要求,而不是可直接用于建模的事实。
[CI002, CI003, CI005, CI008, CI009, CI016]4.3 融资状况与依赖
最新保留的第三方融资报道称,Skeleton 完成了更大规模 IPO 前轮融资的首轮交割,金额为 €33 million,使计划于 2027 年赴美 IPO 前的风险投资总额达到 €392 million。同一批报道将新资本直接与美国生产能力及欧洲制造持续扩张联系起来,说明融资仍在支持工业规模化,而不只是为上市前的资产负债表润色。Batteries International 还报道 Siemens 进行了另一笔未披露金额的投资,并与数据中心独家渠道关系绑定。这可能加速销售,但也带来集中度和定价权风险,因为一条战略性市场通路集中在一个工业合作伙伴手中。[CI011, CI012, CI013, CI014, CI015]
4.4 财务结论与尽调阻碍
Skeleton 的公开案例最能证明商业相关性,最难证明投资判断的完整性。点名的交易对手、工厂开业、美国部署和当前融资都支持需求真实、公司已超越科研项目阶段。但缺失的恰恰是定价风险和下行保护最关键的数据:绝对收入、毛利率、现金消耗、资金续航、债务或项目融资义务、分批次销售效率、实际 ASP,以及按产品线划分的订单积压。在管理层提供这些数字前,更准确的定位是:这是一个有潜力但仍依赖融资的硬件平台,收入质量和利润率路径仍只能部分看清。[CI002, CI003, CI006, CI009, CI011, CI012, CI013, CI014, CI015, CI016]
4.5 图表
05产品与技术
5.1 已交付产品与用户任务
Skeleton 披露的产品并非泛泛的“储能”概念,而是分为 GrapheneBBU 模块、GrapheneBBU 机架和 D35/D60 SuperBattery 电芯。模块和机架定位为数据中心电池备用单元,电芯则面向数据中心、电网、汽车、重卡和航空航天领域的 OEM 或集成商。产品针对的是 AI 基础设施短促而剧烈的功率波动,而非长时储能:Siemens 将负载问题描述为同步 AI 芯片造成的毫秒至秒级波动,Skeleton 自有材料则将产品定位为桥接 / 调节层,可接入现有交流设施,并保留通往 800V 直流架构的路径。公开规格让顶层价值主张更具体:备用时间最高 90 秒,模块峰值功率 100 kW,机架峰值功率 1.2 MW;但这些规格仍比实际客户结果更容易验证。[CE001, CE002, CE003, CE004, CE005, CE008, CE009]
| 产品线 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| GrapheneBBU 模块 | 数据中心 OEM 或电力系统集成商 | 已公开定义 | 800 VDC 模块,集成 DC/DC 转换器和 BMS | 未公开装机基数或故障率 |
| GrapheneBBU 机架 | 超大规模云厂商或托管数据中心设施运营商 | 已公开定义 | 1.2 MW 峰值机架,带可配置 ±400V 接口,适用于改造部署 | 未公开参考架构或服务 SLA |
| D35 SuperBattery 电芯 | 数据中心或汽车集成商 | 已公开定义 | 小型化 6.3 kW/L 电芯,额定循环次数最高 50,000 次 | 未公开客户验证数据 |
| D60 SuperBattery 电芯 | 电网或重型设备集成商 | 已公开规格 | 7.1 kW/L 电芯,ESR 低于 D35 | 未公开现场寿命或电池包级 BOM |
状态反映保留的公开来源中已列明或已宣布的内容;这不是产量或现场可靠性的独立证明。
[CE001, CE002, CE003, CE004, CE005]| 用户任务 | 当前工作流痛点 | 公司方案 | 可量化收益 | 局限 |
|---|---|---|---|---|
| 吸收毫秒级 AI 负载尖峰 | 公用事业和 UPS 设备并非为同步的 AI 电力激增而设计 | 在设施电力与白区之间充当基于超级电容器的缓冲器 | 公司称平滑功率可将能耗削减最多 45% | 保留来源中没有独立的前后对比研究 |
| 为 IT 机架提供分钟级以内的备用电力桥接 | 传统电池 / 发电机组对短时备电而言体积庞大 | GrapheneBBU 模块和机架,备用时长最高 90 秒 | 模块峰值 100 kW、机架峰值 1.2 MW 均已公开规格 | 未公开衰减或维护曲线 |
| 将欧盟主权 AI 电力供应本地化 | 进口电池供应链可能削弱 AI 建设的韧性 | Leipzig 加 Varkaus 的 AI 电力产品制造布局 | Varkaus 被描述为 1GW 工厂,也是欧洲首座 AI 数据中心电池工厂 | 爬坡良率和质量指标未披露 |
| 扩大北美部署 | 客户需要让工程能力、最终制造能力靠近数据中心需求 | Houston 工程设施和计划中的美国制造 | 公司称美国已有超过 100 MW 部署 | 计划于 1H26 实现的制造结果尚无独立验证 |
除非某行引用了合作伙伴或独立佐证,否则收益均为公司自述;等同于 null 的缺口反映公开证明缺失,而非功能为零。
[CE008, CE009, CE010, CE011, CE012, CE013, CE014]5.2 架构与运营足迹
保留来源允许我们把运营模式描述为分层堆栈,而不是营销式方框图。在产品层,模块和机架结合 DC/DC 转换、电池管理、液冷和高功率电芯;在材料层,Skeleton 将性能与专有 Curved Graphene 和 CBMM 氧化铌原料联系起来。在制造层,公司如今拥有更清晰的地理足迹:关于页面称莱比锡于 2025 年开始向 AI 数据中心超大规模云厂商交付,三个来源将 Varkaus 工厂描述为一项 €50 million / 1GW 的 SuperBattery 资产,休斯敦工程设施则被定位为美国制造的前站。这种组合说明公司确有 AI 电力基础设施运营平台,但也意味着执行取决于多地点爬坡质量、材料供应,以及能否把发布阶段公告转化为可重复的部署项目。[CE002, CE004, CE011, CE012, CE013, CE014, CE015]
| 层 / 流程 / 组件 | 作用 | 依赖 | 风险 |
|---|---|---|---|
| 电网 / 设施接口 | 吸收兆瓦级 AI 负载波动,支持改造部署 | 现有 AC 架构及未来迁移至 800V DC 的路径 | 没有公开的公用事业并网或参考设计验证 |
| GrapheneBBU 机架 | 提供站点级备用电力和功率缓冲 | 采用液冷电力电子器件,配备可配置的 ±400V 接口 | 未公开 MTBF 或维护间隔 |
| GrapheneBBU 模块 | 提供带控制电子器件的机架基础单元 | 集成 DC/DC 转换器和电池管理系统 | 公开来源未披露软件 / 控制栈细节 |
| SuperBattery 电芯 | 提供高功率电化学核心 | 基于 Graphene 的弧形 D35 和 D60 电芯设计 | 未保留电池包级安全测试或滥用测试材料 |
| CBMM 铌原料 | 支撑安全性、寿命和性能主张 | 获得 CBMM 的 NBXCELER 氧化铌 | 材料集中度和双供应源计划未披露 |
| 制造布局 | 为超大规模客户和北美客户扩大供给 | Leipzig 交付、Varkaus 爬坡以及 Houston / 美国建设 | 相比发布信息,良率、质量和进度方面的证明更薄弱 |
该表将产品组件与运营依赖放在一起,因为公开来源同时披露了两者;没有保留直接材料时,控制与合规问题仍明确列为缺口。
[CE002, CE004, CE006, CE007, CE011, CE014, CE015, CE016]Skeleton 披露的 AI 电力栈从设施集成一路延伸到电芯化学体系和制造布局。
- 电网 / 设施接口
- 应对毫秒至秒级的 AI 功率波动,无需彻底改造整个设施。
- 模块
- 现有 AC 架构改造
- 面向未来 800V DC 标准的路径
- 输出
- 降低公用事业侧扰动
- GrapheneBBU 机架
- 面向站点规模的白空间备用电力与平滑功率机柜。
- 模块
- 接口 — Configurable ±400V DC
- 电力 — 1.2 MW peak / 800 kW nominal
- 液冷
- 输出
- 机架级 AI 电力缓冲
- GrapheneBBU 模块
- 数据中心部署中短时备用电力的构件。
- 模块
- 集成 DC/DC 转换器
- 电池管理系统
- 备用时长 — Up to 90 seconds
- 输出
- 100 kW 峰值模块功率
- SuperBattery 电芯
- 以 D35 和 D60 规格销售的高功率电化学核心。
- 模块
- D35 — 6.3 kW/L; 148 Wh/L; up to 50,000 cycles
- D60 — 7.1 kW/L; 148 Wh/L; up to 50,000 cycles
- 输出
- 面向任务关键型环境的电芯级功率密度
- 材料体系
- 性能叙事与专有 Curved Graphene 和氧化铌材料投入绑定。
- 模块
- Curved Graphene
- CBMM NBXCELER 氧化铌
- 输出
- 安全性、寿命和充电速率定位
- 制造布局
- 商业化依赖欧洲工厂以及美国本地化。
- 模块
- 向 AI 数据中心超大规模云厂商交付的 Leipzig 超级工厂
- Varkaus 1GW SuperBattery 工厂
- 休斯顿工程团队和计划中的美国制造
- 输出
- 欧洲 / 美国扩产路径
5.3 差异化、成熟度与风险
Skeleton 的差异化在机械和场景证据最充分的地方最强:公开的高功率密度、为关键任务计算优化的短时备用、改造兼容性,以及借助 Siemens 合作渠道接入更广泛电气基础设施组合。随着公司增加工厂并声称美国已部署超过 100 MW,成熟度正在提高,但公开尽调记录仍落后于制造叙事。审阅的来源集没有提供独立前后性能研究、公开开发者 / 实践者信号,也没有状态页或认证页面等正式信任 / 安全 / 合规材料。这不否定技术主张,但意味着产品可靠性、控制和实施摩擦仍需直接尽调,而不能靠公开来源确认。[CE009, CE010, CE013, CE015, CE016]
5.4 图表
06客户
6.1 客户细分与付款方
公开证据显示客户基础多元,而非单一垂直买家。Skeleton 将电力系统定位于 AI 数据中心、电网、交通、工业和国防,官方披露还指向铁路、卡车、赛车、核聚变和公共部门用例。这种广度很重要,因为它说明公司并不只依赖投机性的 AI 需求。代价是不同细分市场的买家、使用者和付款方角色各异,跨细分可比性较弱,公司也没有发布清晰的客户数量或分部收入桥接。[CU001, CU008, CU009, CU010]
| 细分市场 | 买方 / 用户 / 付款方 | 用例 | 公开证明 | 收入 / 战略价值 | 缺口 |
|---|---|---|---|---|---|
| AI 数据中心 | 买方:超大规模客户 / 运营商;用户:设施和电力团队;付款方:基础设施资本开支所有者 | 机架到电网的功率平滑、UPS、削峰 | 官方 AI 页面、Siemens 独家渠道、美国超过 100 MW 的部署背景 | 增长最快的叙事;美国制造和 Silicon Valley 研发的锚点 | 没有具名超大规模客户、合同金额或续约披露 |
| 铁路和公共交通 | 买方:OEM 或运营商;用户:工程和运营团队;付款方:OEM 或交通所有者 | 列车电能质量和能源模块 | 通过 Siemens Mobility US 服务 Amtrak;IndyCar 和历史 OEM 参考 | 展示了 AI 之外在任务关键型交通领域的可信度 | 车队规模、扩张节奏和经济性未披露 |
| 商用车和工业 OEM | 买方:OEM / 工业客户;用户:工程和维护团队;付款方:OEM 或资产所有者 | 发动机启动和瞬态电力支持 | Škoda Electric 商业交易历史;未具名的美国大型卡车 OEM | 分散终端市场敞口 | 具名卡车 OEM 仍未披露 |
| 公共部门和前沿行业 | 买方:机构或企业;用户:技术团队;付款方:机构或项目赞助方 | 航天电容器和核聚变电力支持 | ESA 是首个客户;领先核聚变公司仅按类别指代 | 为高要求用例提供有用的技术验证 | 当前支出、续约和生产规模未披露 |
细分市场结合了公司发布的参考信息与外部报道;各行反映公开证明质量,而非可比的收入贡献。null 式缺口意味着底层分母或具名账户清单未公开。
[CU001, CU008, CU009, CU010, CU013]6.2 采用轨迹偏向北美与 AI 基础设施
指定来源对运营足迹的证明多于对客户 logo 数量的证明。Skeleton 称超过 100 MW 的系统已在美国部署并运行,约一半公司收入来自北美客户,这说明该地区已经具备战略重要性,而不只是目标。AI 基础设施推进也有资本支持:官方及外部 2026 年 10 月来源将超过 2 亿美元的扩张和硅谷研发与超大规模云厂商直接合作联系起来,Siemens 则为数据中心运营商带来庞大的既有企业通路。[CU002, CU003, CU006, CU007, CU008, CU011, CU013]
| 指标 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 美国部署基数 | 已部署并运行的系统超过 100 MW | 指定来源中的当前表述 | Skeleton 美国扩张公告 | 中 | 显示公司增长最快市场已有真实装机基础 | 未按客户、站点或产品系列拆分 |
| 北美收入占比 | 约占公司收入一半 | 指定来源中的当前表述 | Skeleton 美国扩张公告 | 中 | 区域集中度已经很高 | 没有客户数量、头部账户占比或留存数据 |
| AI 超大规模客户交付 | 2025 年开始交付 | 2025 | Skeleton 关于页面 | 中 | 暗示已进入生产商业化,而不只是试点 | 没有具名买方、出货量或续约状态 |
| 数据中心渠道覆盖 | Siemens 称其服务全球前十大数据中心供应商中的 9 家 | 2026-10-15 | Siemens 联合发布稿 | 中 | 可能加快企业级产品进入 AI 基础设施 | 这是 Siemens 的覆盖,不是 Skeleton 的赢单率或收入份额 |
轨迹指标混合了部署、区域收入占比、交付时点和合作伙伴覆盖;它们没有共同分母,不应解读为漏斗。公司发布的表述占主导,因为指定来源没有公布经审计的客户看板。
[CU002, CU003, CU006, CU008, CU011]Skeleton 的公开发展路径从早期技术验证走向渠道驱动的 AI 基础设施扩张,近期最强势的势头在北美。
这张旅程图根据指定来源呈现顺序和扩张逻辑;不代表转化率或规模相同的客户群。
[CU003, CU004, CU006, CU007, CU008, CU009, CU010, CU011, CU012, CU013]6.3 点名客户证明真实但不均衡
点名客户证据确实存在,但早期交通和公共部门案例强于最新的 AI 数据中心故事。公司历史页面称欧洲航天局是 Skeleton 的首个客户,并引用与 Škoda Electric 的大型商业交易;美国扩张公告又加入 Amtrak、IndyCar、一家大型卡车 OEM 和一名核聚变客户。相比之下,最具战略意义的 AI 客户仍只被描述为超大规模云厂商或领先运营商,而未点名,因此生产证据提升快于公开客户案例质量。[CU004, CU005, CU006, CU010, CU012]
| 客户 / 账户 | 细分市场 | 部署 / 用例 | 生产还是试点 | 结果 | 局限 |
|---|---|---|---|---|---|
| European Space Agency | 公共部门 / 航天 | 航天电容器开发 | 历史客户证明 | 公司历史资料称 ESA 是 Skeleton 的首个客户 | 历史参考;未披露当前支出、续约或产品状态 |
| Škoda Electric | 交通 OEM | 大规模商业交易 | 商业主张(2019 年) | 显示 OEM 采用已超出实验室测试 | 未披露合同规模、部署范围或当前状态 |
| Amtrak 列车(通过 Siemens Mobility US) | 铁路 / 交通 | 列车系统能源模块 | 声称已投入生产部署 | 具名美国参考客户将证明范围扩展到 AI 之外 | 未披露车队规模、合同期限或扩张数据 |
| AI 数据中心超大规模客户(未具名) | AI 基础设施 | 功率平滑 / 数据中心系统 | 声称自 2025 年起开始生产交付 | 当前战略上最重要的细分市场 | 终端客户仍未具名,因此证明只有一半 |
该清单有意保持不完整,因为指定来源没有提供具名客户或 AI 生产部署的完整列表。各行优先列出最清晰的公开证明点,并保留未具名战略账户的可见度,而不过度夸大确定性。
[CU004, CU005, CU006, CU010, CU012]6.4 持续性与集中度仍是最大的公开缺口
本章对留存质量的记录最少。指定来源均未披露 NRR、GRR、流失、续约率、合同期限或头部客户占比,因此持续性只能从持续制造扩张和反复出现的战略合作伙伴关系中推断,而无法直接衡量。集中度风险也确实存在:Siemens 同时是投资者、自动化伙伴和面向数据中心的全球独家销售商,而北美已经贡献约一半收入。这种组合可能加速增长,但也增加了对少数渠道和终端市场的依赖,而公开材料没有量化这种依赖。[CU003, CU007, CU012, CU013, CU014]
6.5 图表
07风险
7.1 风险排序与论点破坏信号
风险来自多线并行。Skeleton 正在工业化多条产品线,爬坡 €220 million 的莱比锡超级电容器工厂,扩大 €50 million 的 Varkaus SuperBattery 工厂,并把制造延伸至美国,同时为 2027 年美国 IPO 做准备。这种组合造成了明显错位:增长雄心可见,但公开运营披露仍有限。公开记录依赖公司控制来源的地方,剩余风险最高:AI 数据中心效率声明、客户集中度以及北美增长背后的确切经济性。实际止损标准很直接:如果管理层无法提供经审计的 2025-2026 年分部经济数据、独立验证的性能数据,以及超大规模云厂商和 OEM 项目的集中度指标,投资者就应认为 IPO 前叙事领先于证据。[CR003, CR004, CR005, CR007, CR008, CR012, CR014]
剩余风险集中在制造爬坡、披露不透明、商业集中度以及仍未验证的运营主张。
序数评级总结有证据支持的承保风险,并非模型化概率。
[CR001, CR003, CR004, CR005, CR008, CR009, CR010, CR012, CR014]7.2 监管、法律与披露风险主要在于仍未公开的内容
指定来源池没有显示正在进行的诉讼、召回或公开执法行动,因此当前核心法律风险是披露风险,而非已证实的法律违规。独立融资报道和保留的资料页确认 2026 年 5 月完成 IPO 前首轮交割,但公开记录仍缺少该融资的投前、投后估值、股份类别和优先权结构细节。这一点很重要,因为一家后期硬件公司可能在 headline 融资上显得风险已降,但清算权、客户承诺或资本支出义务仍可能隐藏明显下行风险。产品认证也有同样问题:保留来源描述了工厂开业和部署,却没有提供更新后的证书包、审计轨迹或产品审批记录,使外部投资者无法独立确认 2025 年扩张后的质量 / 合规成熟度。[CR005, CR006, CR007, CR012]
| 风险 | 管辖区 / 层面 | 公开状态 | 可能性 | 严重程度 | 缓释措施 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| IPO 前条款不透明 | 私募融资 / IPO 准备 | 2026 年 5 月首轮交割已公开,但投前、投后估值和优先权细节未公开。 | 高 | 高 | 新资本和 IPO 准备表明公司具备披露能力。 | 在审阅融资轮文件前维持高位。 | 获取 2026 年 5 月的条款清单、股权结构表、清算瀑布和附函摘要。 |
| 资质 / 合规证据缺口 | 德国、芬兰及未来的美国制造 | 保留来源描述了工厂和部署,但没有当前公开证书或审计材料包。 | 中 | 高 | 工业客户和工厂启动降低了从零起步的风险。 | 高,因为独立资质证明仍然缺失。 | 按站点和产品线索取 ISO、产品安全、客户资质及发布后的审计记录。 |
| IPO 准备披露缺口 | 美国上市 / 投资者报告 | 独立新闻提到计划于 2027 年在美国 IPO,但保留来源池中没有公开经审计的细分经济数据。 | 中 | 中 | 新投资者和工业合作伙伴可以推动报告规范。 | 在分享经审计的 2025-2026 年报告前维持中高位。 | 审阅经审计财务报表、治理路线图,以及任何招股书草案或准备备忘录。 |
各行按新晋后期投资者的视角以严重程度排序;没有公开案例不等于没有风险。
[CR005, CR006, CR007, CR012]7.3 运营风险来自同步扩张与仍未验证的性能声明
公开来源显示 Skeleton 已超越实验室阶段商业化,但也显示其运营故事仍有多少依赖公司自述。公司、合作伙伴和行业媒体材料都描述了快速工厂扩张、大型电网和 AI 基础设施客户,以及 GrapheneGPU 和 SuperBattery 系统带来的大幅效率提升。缺失的是把这些声明与规模化持续现场表现连接起来的独立运营桥梁。这一缺口很重要,因为公司不再销售早期试点,而是在要求市场为资本密集型制造和对正常运行时间敏感的基础设施项目定价。保留资料页显示员工数从 2024 年的 576 人降至 2026 年 3 月的 494 人,这本身不能证明运营承压,但确实引发了流程、质量和服务能力是否跟上实体足迹的问题。[CR001, CR003, CR004, CR009, CR010, CR012, CR013]
| 故障模式 | 证据信号 | 可能性 | 严重程度 | 缓解措施成熟度 | 剩余敞口 | 未解决缺口 |
|---|---|---|---|---|---|---|
| 多站点制造爬坡延误 | Leipzig、Varkaus 和计划中的美国制造接连扩张。 | 高 | 高 | 中 | 高 | 需要按站点提供良率、产能利用率和准时交付指标。 |
| 性能主张跑在第三方证明之前 | 40%-45% 的效率和 40% 的算力提升主张仍主要来自公司或合作伙伴放大。 | 中 | 高 | 低-中 | 高 | 需要独立基准数据和客户验证的结果研究。 |
| 任务关键型部署中的质量 / 资质延迟 | 产品面向超大规模客户、公用事业、交通和国防用例,而这些领域的审批周期很长。 | 中 | 高 | 中 | 中高 | 需要资质通过 / 失败记录、保修数据和现场事故率。 |
| 扩张期间的组织负荷 | 保留的业务概览显示,2026 年 3 月员工数低于 2024 年,但工厂布局扩大了。 | 中 | 中 | 低-中 | 中 | 需要按职能和站点提供招聘计划、人员流失和支持覆盖指标。 |
该表将执行风险与已证实事故分开;没有保留来源记录公开宕机或召回,因此未解决的数据请求很重要。
[CR001, CR003, CR004, CR009, CR010, CR012, CR013]7.4 依赖风险横跨超大规模云厂商、OEM 渠道、供应商与后期资本
Skeleton 的增长故事越来越集中在少数强势交易对手身上。官方和合作伙伴来源都把北美称为增长最快的市场,称约一半收入来自北美客户,并突出 Siemens、GE、Hitachi Energy、Marubeni 和 CBMM 等点名交易对手。这些关系是商业相关性的积极证明,但也增加了合作伙伴重新排序、认证延迟以及少数基础设施项目集中度带来的风险。融资模式进一步放大依赖:管理层显然在更大规模 IPO 前融资或最终上市前扩张,而公开来源仍未披露单位经济性、订单积压可见度或客户续约情况,外部投资者无法判断当前建设是自我强化还是仍依赖补贴。[CR002, CR005, CR008, CR009, CR010, CR011, CR014]
| 依赖 | 交易对手 | 作用 | 集中度信号 | 失败情景 | 严重程度 | 缓解措施 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| 北美需求集中 | 超大规模客户和电网客户 | 主要增长市场和大型项目需求来源 | 公司称约一半收入来自北美。 | AI 数据中心支出暂停或资质时间表延误。 | 高 | 欧洲制造布局和多垂直定位可缓和冲击。 | 在披露集中度和合同期限数据前维持高位。 |
| 渠道 / 商业化依赖 | Siemens、GE、Hitachi Energy、Marubeni、主要 OEM | 销售渠道、集成和参考账户 | 具名交易对手在公开证明点中占据重要位置。 | 关键合作伙伴重新排序优先级、推迟推广或重新协商经济条件。 | 高 | 多个具名交易对手降低对单一名称的依赖。 | 中高,因为项目集中度仍不透明。 |
| 特种材料依赖 | CBMM | SuperBattery 定位所需的氧化铌原料 | 公开材料将 SuperBattery 的安全 / 性能主张与 CBMM 合作联系起来。 | 原料、定价或排他性问题损害产品经济性或爬坡速度。 | 中 | Skeleton 强调非锂化学体系和多个制造站点。 | 中,因为供应商多元化未披露。 |
| 后期资本依赖 | IPO 前投资者基础 | 为扩张和潜在的美国上市路径提供资金 | €33 million 首轮交割已公开,但完整融资条款和未来融资需求仍不透明。 | 扩张速度超过现金创造能力,迫使公司以更弱条件进行下一轮融资。 | 高 | 大型工业投资者和 IPO 准备可以拓宽融资选择。 | 在披露现金 runway 和积压订单质量前维持高位。 |
各依赖项按交易对手或资本市场变化影响交付、增长或定价能力的速度排序。
[CR002, CR005, CR008, CR009, CR010, CR011, CR014]7.5 图表
08估值
8.1 建议与入场纪律
Skeleton 已跨过重要信号门槛:公司称 Siemens 主导的 2026 年 10 月投资使估值超过 $1 billion,独立报道则称同一交易还赋予 Siemens 销售 Skeleton 数据中心产品的独家权利。这一组合证明外部投资者的兴趣及其对 AI 电力基础设施的战略相关性,但也把商业杠杆集中在一个交易对手手中。同一批保留的公开材料仍未披露当前收入、毛利率、现金消耗、资金续航、已签订单积压或 2026 年 5 月的优先股堆叠。根据当前公开证据,正确结论是继续研究而非买入;信心中等、风险高,并应围绕最新私有估值保持价格纪律。[CV001, CV002, CV011, CV012, CV013]
| 指标 | 评估 | 决策含义 |
|---|---|---|
| 建议 | research-more | 战略验证有所改善,但价格支撑仍不充分。 |
| 置信度 | 中 | 有新的融资轮数据和进展信号,但公开运营披露仍然薄弱。 |
| 风险评级 | 高 | 制造执行、渠道集中度和融资结构不透明都可能改变结果。 |
| 估值立场 | 偏高 | 最新公开锚点高于 $1B,但缺乏足够的公开单位经济性支撑。 |
| 目标回报 / 持有 / 退出 | 仅凭公开证据无法支持 | 将 2027 年 IPO 表述作为情景标记,而不是作为已纳入投资判断的退出时钟。 |
| 决策含义 | 只有在加强尽调或价格更优时才推进 | 在评估投资进入前,索取股权表、运营指标和积压订单可见度。 |
面向读者的判断,综合引用的公开证据得出;并非公司自述评级。
[CV001, CV011, CV012, CV013, CV014]可见的商业和制造验证改善了故事,但最新公开估值指标仍然跑在已披露的投资判断数据之前。
| 节点 / 连接 | 节点 / 起点 | 终点 | 说明 |
|---|---|---|---|
| 节点 1 | 商业规模信号 [demand] | >100 MW deployed in the U.S.; roughly half of revenue from North American customers. | |
| 节点 2 | 产品验证信号 [product] | 2025 hyperscaler deliveries and company-claimed energy savings for AI clusters. | |
| 节点 3 | 制造准备度 [manufacturing] | Leipzig Superfactory plus the Varkaus 1 GW SuperBattery plant. | |
| 节点 4 | 最新估值锚点 [valuation] | October 2026 strategic round said to value Skeleton above $1B. | |
| 节点 5 | 门槛风险 [risks] | Exclusive Siemens data-center rights and missing revenue, burn, backlog, and preference data. | |
| 节点 6 | 推荐 [recommendation] | research-more / high risk / stretched until price or disclosure improves. | |
| 连接 1 | 商业规模信号 [demand] | 产品验证信号 [product] | 支持 |
| 连接 2 | 产品验证信号 [product] | 制造准备度 [manufacturing] | 通过……扩张 |
| 连接 3 | 制造准备度 [manufacturing] | 最新估值锚点 [valuation] | 强化 |
| 连接 4 | 最新估值锚点 [valuation] | 门槛风险 [risks] | 受到……制约 |
| 连接 5 | 门槛风险 [risks] | 推荐 [recommendation] | 导向 |
8.2 上行案例判断正确之处
看涨论点并非虚构。Skeleton 称美国已部署超过 100 MW 系统,北美客户贡献约一半收入;2025 年既有莱比锡 Superfactory 爬坡,也首次向 AI 数据中心超大规模云厂商交付。Varkaus SuperBattery 工厂又增加 1 GW 制造资产,管理层持续把功率平滑描述为 AI 集群部署的核心瓶颈。这些都是有意义的商业化和产能信号,支持 Skeleton 可能从有潜力的零部件供应商演变为战略基础设施供应商,尽管大部分性能和客户集中度证据仍来自公司,而非独立披露。[CV004, CV005, CV006, CV007, CV008, CV009]
| 论点 | 何种变化会改变判断 |
|---|---|
| 论点:战略验证 | 如果 Skeleton 披露 AI 基础设施部署带来的复购订单量、客户集中度和利润率走势,判断会改善。 |
| 论点:制造验证 | 如果管理层展示德国、芬兰及计划中的美国产能的产能利用率、良率和爬坡经济性,判断会改善。 |
| 反论点:估值不透明 | 如果投资者拿到经审计的 FY2025 和 2026 年初至今财务数据,以及当前 cap table waterfall 和清算优先权,判断会转弱。 |
| 反论点:渠道集中 | 如果 Skeleton 证明超大规模云厂商的直接拉动、多渠道分销,或数据中心业务的非独家经济性,判断会转弱。 |
这些行总结了主要的、有证据支持的论点,以及会改变投资判断的尽调事件。
[CV004, CV005, CV006, CV008, CV009, CV010, CV012]8.3 情景、退出准备度与门槛式尽调要求
情景区间很宽,因为公开证据把强烈的战略信号与薄弱的投资判断数据混在一起。上行情景需要超大规模云厂商试点转化为重复收入,Siemens 扩大而非限制分销,制造建设跟上 AI 基础设施需求。基准情景是,在收入质量和订单积压可见度改善前,Skeleton 值得密切跟踪最新公开估值,但不宜激进入场。悲观情景很直接:如果商业化延迟、IPO 时点推迟,或优先权条款比普通股股东预期的更保护投资者,当前估值可能迅速压缩。在把 2027 年 IPO 表述当作真实退出时钟前,投资者仍需股权结构、订单积压和单位经济性细节。[CV003, CV010, CV011, CV012, CV013, CV014]
| 情景 | 假设 | 估值 / 回报逻辑 | 关键风险 | 概率信号 |
|---|---|---|---|---|
| 多头 | 超大规模云厂商试点转为复购项目,美国产能爬坡,Siemens 合作扩大需求且不牺牲利润率。 | 由于战略稀缺性和产能验证都有改善,可以接受高于最新公开 >$1B 指标的价格。 | 执行延误、渠道锁定和 AI 电力支出放缓仍然重要。 | 需要独立的收入或订单积压披露,证明试点已经转为生产订单。 |
| 基准 | 需求仍然真实,但披露不完整,制造爬坡伴随正常摩擦。 | 值得在最新战略估值附近跟踪,但不应追高,因为验证进展优于价格透明度。 | 更大规模 IPO 准备轮中的单位经济性不透明和稀释风险。 | 当前公开证据最符合这一情景。 |
| 空头 | IPO 时点推迟、超大规模云厂商转化停滞,或隐藏的优先权条款让融资结构比普通股股东预期的更偏向保护投资者。 | 由于融资杠杆和披露风险会压过战略故事,入场价应低于最新公开估值。 | 下轮融资风险、利润率压缩和 cap table 悬置风险。 | 公开数据缺口使这一下行情景仍然成立。 |
由于保留的公开证据没有披露收入、利润率或现金流输入,情景分析只能定性,无法建立硬模型。
[CV001, CV003, CV011, CV012, CV013, CV014]8.4 图表
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;在作出任何投资决定前,应直接向管理层和一手文件核实。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Skeleton positions itself as a high-power energy-storage company spanning AI data centers, grids, mobility, industry, and defense through supercapacitor and SuperBattery systems. | 中 | SO001, SO004, SO007 |
| CO002 | Skeleton was founded in Tartu, Estonia in 2009. | 高 | SO004, SO007 |
| CO003 | A commercial profile lists Skeleton's headquarters as Tallinn, Estonia, but the assigned official sources do not independently restate the current headquarters. | 中 | SO005 |
| CO004 | Taavi Madiberk and Oliver Ahlberg co-founded Skeleton in 2009, and Taavi Madiberk is the public CEO/co-founder in the company's 2026 financing and expansion communications. | 高 | SO002, SO004, SO008 |
| CO005 | The assigned public pool does not disclose broader board composition or a full current executive roster beyond the founders and CEO visibility. | 中 | SO002, SO006, SO007 |
| CO006 | Skeleton's disclosed operating footprint spans Estonia-rooted origins, manufacturing in Germany and Finland, and active U.S. engineering and manufacturing expansion. | 高 | SO002, SO004, SO006, SO007 |
| CO007 | Skeleton announced a €33 million first close on 2026-05-07 and said disclosed total venture funding reached €392 million ahead of a planned 2027 U.S. IPO. | 高 | SO002, SO004 |
| CO008 | The May 2026 first close added Axon Partners Group, SmartCap, and Taiwania Capital, and Tech Funding News reported that existing investor CBMM also joined. | 高 | SO002, SO008 |
| CO009 | Independent October 2026 coverage said a Siemens-led strategic investment valued Skeleton at more than $1 billion and supported U.S. manufacturing plus Silicon Valley R&D. | 中 | SO003, SO005 |
| CO010 | Skeleton's U.S.-expansion release said more than 100 MW of systems were already deployed in the United States and that approximately half of revenue came from North American customers. | 中 | SO006 |
| CO011 | Passive Components says the Siemens-led financing did not disclose the exact amount raised or Siemens' individual contribution, limiting visibility into cash proceeds and dilution. | 中 | SO003 |
| CO012 | Passive Components says Skeleton's reported AI data-center energy-savings and computing-power improvements are manufacturer-stated system-level benefits without disclosed workload, baseline, or test-boundary detail in the announcement. | 中 | SO003 |
| CO013 | Skeleton's history page says the company opened the Leipzig Superfactory in 2025 and started deliveries to AI data center hyperscalers that year. | 中 | SO007 |
| CO014 | A commercial profile lists Skeleton in a 201-500 employee band, making headcount a low-confidence estimate rather than an official disclosure. | 低 | SO005 |
| CO015 | The 2026 financing narrative supports classifying Skeleton as a late-stage private company preparing for a potential U.S. IPO rather than as an early-stage venture. | 高 | SO002, SO003 |
| CO016 | Skeleton's history page traces key milestones from its first European Space Agency customer in 2011 through commercial launch in 2012, a major German factory in 2017, a Marubeni strategic agreement in 2021, and a Finland factory acquisition in 2023. | 中 | SO007 |
| CM001 | Skeleton defines its served market as mission-critical power systems across AI data centers, power grid, mobility, industry, and defense. | 中 | SM001, SM007 |
| CM002 | TFN frames traditional battery-backup vendors such as EnerSys and Vertiv as incumbents, implying Skeleton competes against established battery and UPS architectures rather than a greenfield category. | 中 | SM009 |
| CM003 | Skeleton says the U.S. is already a core market, with more than 100 MW deployed in country and approximately half of revenue generated from North American customers. | 高 | SM002, SM003 |
| CM004 | Retained sources indicate AI infrastructure is Skeleton's fastest-growing customer segment and that deliveries to AI data-center hyperscalers are already underway. | 高 | SM001, SM006 |
| CM005 | Skeleton's public materials consistently frame AI data-center load volatility and grid strain as the central adoption driver for its power-smoothing systems. | 高 | SM002, SM003, SM009 |
| CM006 | Skeleton claims its AI data-center solutions can reduce energy consumption by up to 45 percent. | 高 | SM002, SM003 |
| CM007 | DCD reported Madiberk's claim that Skeleton can deliver up to 40 percent energy savings and cut required grid capacity for a GPU cluster almost in half by smoothing AI load peaks. | 中 | SM001 |
| CM008 | The assigned pool does not reconcile Skeleton's public efficiency framing, which appears as 40 percent versus 45 percent energy reduction without an independent benchmark. | 中 | SM001, SM002, SM003, SM009 |
| CM009 | The retained public sources do not provide an independent TAM, SAM, or SOM for AI data-center power-smoothing and backup spend, forcing sizing to rely on deployment, revenue-share, and factory-output proxies. | 中 | SM001, SM002, SM003, SM004, SM006, SM007, SM008, SM009 |
| CM010 | Skeleton positions the Varkaus SuperBattery factory as producing one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. | 中 | SM007 |
| CM011 | Skeleton's GrapheneBBU rack is specified at up to 90 seconds of backup duration, 1.2 MW peak power, and 800 kW nominal power. | 中 | SM008 |
| CM012 | Fresh 2026 funding was explicitly linked to scaling AI data-center energy storage and expanding manufacturing capacity in Europe and the U.S. | 中 | SM004, SM005, SM009 |
| CM013 | The retained sources support a buyer path that spans hyperscaler and data-center operators, utilities and grid partners, and OEM or industrial programs rather than a single budget owner. | 高 | SM001, SM003, SM006, SM008 |
| CM014 | Skeleton's company history says the Leipzig Superfactory opened and deliveries to AI data-center hyperscalers began in 2025. | 中 | SM006 |
| CM015 | DCD reported that the Leipzig facility has annual capacity of up to 12 million cells. | 中 | SM001 |
| CM016 | Retained public sources do not bridge Leipzig's cell-capacity disclosure to utilization, shipped system power, or revenue, weakening capacity-based market-sizing precision. | 中 | SM001, SM003, SM007, SM008 |
| CP001 | Skeleton's clearest public wedge in AI infrastructure is microsecond or sub-millisecond response for power-quality stabilization from rack level to the grid. | 高 | SP001, SP005, SP006 |
| CP002 | Skeleton publicly claims its AI power systems can improve energy efficiency or reduce energy consumption by roughly 40 to 45 percent in target deployments. | 中 | SP001, SP003, SP005, SP006 |
| CP003 | Public product descriptions show Skeleton selling a system stack that spans power electronics, control algorithms, cells or modules, and full power systems rather than a component-only offer. | 高 | SP001, SP006 |
| CP004 | Public evidence shows Skeleton commercializing across AI data centers, the power grid, mobility, industry, defense, and transportation-linked applications rather than a single end market. | 高 | SP002, SP004, SP007, SP010 |
| CP005 | Skeleton is scaling supply through Leipzig supercapacitor production, a one-gigawatt Varkaus SuperBattery site, and planned U.S. manufacturing capacity. | 高 | SP002, SP003, SP004, SP007 |
| CP006 | Official and independent U.S. expansion coverage both describe North America as a major market, citing more than 100 megawatts deployed in the country and roughly half of revenue from North American customers. | 高 | SP003, SP005 |
| CP007 | Skeleton's differentiation narrative explicitly contrasts its offer with conventional lithium-ion batteries on response speed, cycle life, fire risk, and critical-material dependence. | 中 | SP003, SP006, SP008 |
| CP008 | The assigned evidence frames the most visible substitute as battery-backed backup power rather than a named direct supercapacitor peer set. | 中 | SP004, SP006, SP008 |
| CP009 | Skeleton's own U.S. expansion framing says buyers can also respond to AI power spikes by building more power plants and upgrading equipment, making infrastructure buildout a status-quo competitive path. | 中 | SP003 |
| CP010 | Procurement trust is part of the competitive story because official and partner sources emphasize European manufacturing independence and transatlantic supply chains. | 中 | SP003, SP007 |
| CP011 | Skeleton also leans on proof-by-association with established infrastructure and OEM names including Siemens, GE, Amtrak or Siemens Mobility US, BMW, and IndyCar. | 中 | SP002, SP003, SP010 |
| CP012 | A low-reputation sales-intel profile says Siemens' October 2026 investment included exclusive global data-center sales rights, which would strengthen distribution while concentrating go-to-market leverage if accurate. | 低 | SP009 |
| CP013 | Varkaus is positioned as short-duration backup for mission-critical computing, which narrows direct competition toward uptime and power-quality solutions rather than long-duration energy arbitrage. | 中 | SP004, SP008 |
| CP014 | Because Skeleton is ramping German, Finnish, and planned U.S. capacity while expanding AI deployments, competitive execution risk still depends on manufacturing and channel coordination as much as on cell performance. | 中 | SP002, SP003, SP004, SP007 |
| CI001 | Public evidence describes Skeleton as monetizing supercapacitor cells, modules, battery products, and larger power systems rather than a pure recurring software subscription. | 中 | SI004, SI010 |
| CI002 | Skeleton says more than 100 MW of systems are deployed and operating in the United States. | 中 | SI005 |
| CI003 | Skeleton says approximately half of company revenue is generated from North American customers. | 中 | SI005 |
| CI004 | Retained sources tie Skeleton's Leipzig SuperFactory to €220 million of investment. | 高 | SI002, SI003 |
| CI005 | Leipzig is designed for annual output of up to 12 million cells. | 高 | SI002, SI003 |
| CI006 | Retained sources report that Leipzig is already supplying Siemens, General Electric, Hitachi Energy, and major U.S. hyperscalers. | 高 | SI002, SI003 |
| CI007 | Skeleton says it invested €50 million in its Varkaus SuperBattery factory. | 高 | SI001, SI007 |
| CI008 | Skeleton says the Varkaus factory produces one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. | 高 | SI001, SI007 |
| CI009 | The two disclosed European factory investments sum to at least €270 million before any U.S. manufacturing buildout. | 高 | SI002, SI007 |
| CI010 | Skeleton said it would establish U.S. manufacturing capacity for AI data-center solutions in the first half of 2026. | 高 | SI005, SI009 |
| CI011 | Independent funding coverage reports that Skeleton closed a €33 million first close of a larger pre-IPO round. | 中 | SI009, SI010 |
| CI012 | Independent funding coverage reports total venture funding of €392 million. | 中 | SI009, SI010 |
| CI013 | The latest retained funding reports say the new capital is intended to expand manufacturing capacity in Europe and the United States. | 中 | SI009, SI010 |
| CI014 | Batteries International reports that Siemens made a further undisclosed investment in Skeleton. | 中 | SI001 |
| CI015 | Batteries International reports that Siemens has exclusivity to sell and scale Skeleton's products in the data-center market. | 中 | SI001 |
| CI016 | The retained public source set does not disclose absolute revenue, realized pricing, gross margin, CAC, payback, burn, runway, debt obligations, or backlog. | 中 | SI001, SI002, SI005, SI009, SI010 |
| CE001 | Skeleton publicly markets SuperBattery cells, modules, and racks for AI data centers, grid, automotive, heavy trucks, and aerospace customers. | 中 | SE003 |
| CE002 | GrapheneBBU modules are disclosed as liquid-cooled battery backup units with an integrated DC/DC converter, battery management system, and 800 VDC interface. | 中 | SE003 |
| CE003 | GrapheneBBU modules are specified for up to 90 seconds backup duration, 100 kW peak power, 66 kW nominal power, and 97% peak efficiency. | 中 | SE003 |
| CE004 | GrapheneBBU racks are disclosed with a configurable ±400V DC interface, reverse polarity protection, liquid cooling, and 1.2 MW peak / 800 kW nominal power. | 中 | SE003 |
| CE005 | The D35 SuperBattery cell is disclosed at 1.8V, 6.3 kW/L power density, 148 Wh/L energy density, and up to 50,000 cycles. | 中 | SE003 |
| CE006 | The D60 SuperBattery cell is disclosed at 1.8V, 7.1 kW/L power density, 148 Wh/L energy density, 0.7 mOhm ESR, and up to 50,000 cycles. | 中 | SE003 |
| CE007 | Siemens describes Skeleton's supercapacitor solution as an electrical shock absorber for megawatt-scale AI load swings that can fit existing AC data-center architectures while preserving a path to 800V DC. | 中 | SE001 |
| CE008 | Skeleton says its systems can smooth AI data-center demand and reduce energy consumption by up to 45 percent. | 中 | SE004, SE006 |
| CE009 | Skeleton says more than 100 MW of systems are deployed in the United States and that North American customers generate about half of company revenue. | 中 | SE004, SE006 |
| CE010 | The Houston engineering facility is presented as the lead-in to U.S. manufacturing capacity for AI data-center solutions in the first half of 2026. | 中 | SE004, SE006 |
| CE011 | Three retained sources describe Varkaus as a €50 million SuperBattery factory with 1GW of output for AI data centers, grid stability, mobility, and defense. | 高 | SE002, SE005, SE008 |
| CE012 | Multiple 2025 sources describe Varkaus as Europe’s first AI data-center battery factory or first site producing high-power batteries for AI data centers. | 高 | SE002, SE005, SE008 |
| CE013 | Skeleton's about-page timeline says the Leipzig superfactory started deliveries to AI data-center hyperscalers in 2025. | 中 | SE007 |
| CE014 | Skeleton links SuperBattery performance and durability to proprietary Curved Graphene plus CBMM niobium-oxide inputs. | 高 | SE002, SE008 |
| CE015 | Skeleton says its technology avoids lithium, cobalt, manganese, and other critical raw materials as part of the manufacturing and supply narrative. | 中 | SE006 |
| CE016 | The retained public product and news surfaces do not disclose a status page, security/privacy program, or third-party compliance certification for the AI-power products. | 低 | SE003, SE006, SE007, SE008 |
| CU001 | Skeleton publicly positions its customer base across AI data centers, grids, mobility, industry, and defense rather than around a single end market. | 中 | SU003, SU007 |
| CU002 | Skeleton says more than 100 MW of its systems are deployed and operating in the United States. | 中 | SU005 |
| CU003 | Skeleton says approximately half of its revenue is generated from North American customers. | 中 | SU005 |
| CU004 | Skeleton's company history page identifies the European Space Agency as its first customer in 2011. | 中 | SU006 |
| CU005 | Skeleton says it signed large-scale commercial deals with global OEMs including Škoda Electric in 2019. | 中 | SU006 |
| CU006 | Skeleton says it started deliveries to AI data center hyperscalers in 2025. | 中 | SU006 |
| CU007 | Siemens gained exclusive global rights to sell and scale Skeleton's high-power storage products for data-center customers. | 高 | SU001, SU008 |
| CU008 | Siemens says it is relied upon by nine of the top ten global data-center providers. | 中 | SU002 |
| CU009 | Skeleton's public AI data-center offer is centered on power smoothing, UPS, and short-duration resilience rather than long-duration energy shifting. | 高 | SU003, SU004 |
| CU010 | Skeleton's U.S. expansion release names IndyCar, a large global truck OEM, Amtrak via Siemens Mobility US, and a leading nuclear fusion company as customer or deployment references. | 中 | SU005 |
| CU011 | Official and external October 2026 sources tie more than $200 million of operational scale-up and Silicon Valley R&D to direct work with hyperscalers in the United States. | 高 | SU001, SU008 |
| CU012 | External coverage notes that Skeleton has not publicly named any hyperscaler customers. | 中 | SU001 |
| CU013 | Channel concentration risk is elevated because Siemens is simultaneously a strategic investor and the exclusive global seller for Skeleton's data-center products. | 高 | SU001, SU002, SU008 |
| CU014 | The assigned sources do not disclose NRR, GRR, churn, renewal rates, or contract lengths for any Skeleton customer segment. | 中 | SU001, SU005, SU006, SU008 |
| CR001 | Skeleton markets its AI-infrastructure power systems as responding in microseconds and improving energy efficiency by up to 45 percent. | 中 | SR003 |
| CR002 | Skeleton says the U.S. is its fastest-growing market, with more than 100 MW deployed there and roughly half of revenue generated from North American customers. | 中 | SR006 |
| CR003 | Independent and partner sources say the Leipzig SuperFactory represents about €220 million of investment, targets up to 12 million cells annually, and is already supplying major industrial and hyperscaler customers. | 中 | SR002, SR005 |
| CR004 | Company and independent sources say the Varkaus SuperBattery factory carries roughly €50 million of investment and 1 GW of annual output for AI-data-center and grid applications. | 高 | SR002, SR008, SR009 |
| CR005 | Independent news sources say Skeleton announced a €33 million first close that brought total venture funding to about €392 million ahead of a planned 2027 U.S. IPO. | 中 | SR004, SR010 |
| CR006 | The retained public funding materials do not disclose pre-money or post-money valuation terms for the May 2026 pre-IPO round. | 中 | SR001 |
| CR007 | The retained public sources disclose factory openings, capital raised, and strategic partners, but not revenue, gross margin, cash burn, runway, or backlog by product line. | 中 | SR001, SR004, SR006 |
| CR008 | Skeleton's public AI-data-center growth story is concentrated around hyperscaler, grid, and North American infrastructure demand rather than a broadly disclosed customer mix. | 高 | SR006, SR007, SR008 |
| CR009 | Skeleton and partner sources present a European manufacturing and no-lithium-or-cobalt raw-material stack as the main mitigation against supply-chain and energy-security risk. | 高 | SR005, SR006 |
| CR010 | Public materials link SuperBattery safety and performance positioning to a partnership with CBMM and its niobium-oxide inputs. | 高 | SR008, SR009 |
| CR011 | Named counterparties such as Siemens, GE, Hitachi Energy, Marubeni, and CBMM appear prominently in the retained sources, implying meaningful partner-channel and supplier dependence in commercialization. | 高 | SR005, SR006, SR008 |
| CR012 | The retained pool does not provide independent benchmarking or audited customer outcome data for Skeleton's 40 to 45 percent AI-data-center efficiency claims. | 中 | SR003, SR005, SR006 |
| CR013 | The retained tear sheet shows headcount of 576 in 2024 versus 494 in March 2026 while the company simultaneously pursued multiple factory launches. | 低 | SR001 |
| CR014 | The retained public sources do not reveal customer concentration, contract duration, or renewal behavior for the hyperscaler and industrial programs cited in Skeleton's growth narrative. | 中 | SR006, SR007 |
| CV001 | On 2026-10-14, Skeleton announced a Siemens-led strategic investment that it said brought the company's valuation above $1 billion and would fund U.S. manufacturing expansion and Silicon Valley R&D. | 高 | SV002, SV003 |
| CV002 | AI Weekly reported that Siemens received exclusive rights to sell Skeleton's data-center products as part of the October 2026 deal. | 中 | SV002 |
| CV003 | Multiple May 2026 funding reports said Skeleton's €33 million pre-IPO first close brought total venture capital raised to €392 million and preceded a larger round ahead of a planned 2027 U.S. IPO. | 中 | SV001, SV004, SV005 |
| CV004 | Skeleton said more than 100 MW of systems are deployed in the United States and that roughly half of company revenue comes from North American customers. | 中 | SV006 |
| CV005 | Skeleton's about page says the company opened its Leipzig Superfactory and started deliveries to AI data-center hyperscalers in 2025. | 中 | SV007 |
| CV006 | Skeleton opened its Varkaus, Finland SuperBattery factory in November 2025 with €50 million of total investment and one gigawatt of output for AI data centers and other markets. | 中 | SV008 |
| CV007 | The retained official materials show operating manufacturing assets in Germany and Finland plus an active U.S. manufacturing expansion program, reducing pure scale-up risk versus an unbuilt-platform story. | 高 | SV006, SV007, SV008 |
| CV008 | Skeleton claims its AI data-center power systems can deliver up to 40% energy savings and double-digit GPU-performance improvement by smoothing rapid load fluctuations. | 中 | SV003 |
| CV009 | Tech Funding News reported that Skeleton delivered millions of supercapacitor cells to a major American hyperscaler and had more clients testing the technology in 2026. | 中 | SV004 |
| CV010 | The visible public comparison set in retained coverage points to EnerSys, Vertiv, and Capacitech Energy as adjacent alternatives rather than to a clean public-equity peer set. | 中 | SV004 |
| CV011 | The May 2026 financing coverage disclosed new capital and IPO intent but did not publish a post-money valuation, share-price-by-class, or liquidation waterfall for the pre-IPO tranche. | 中 | SV001, SV004, SV005 |
| CV012 | The retained public evidence describes demand, funding, and manufacturing plans, but it does not disclose current revenue, gross margin, cash burn, runway, or signed backlog by product line. | 中 | SV001, SV004, SV005, SV006 |
| CV013 | Because the latest public valuation marker is above $1 billion while the operating disclosure set remains thin, a disciplined investor would need either better diligence access or a more favorable entry price before underwriting a buy call. | 中 | SV002, SV003 |
| CV014 | Skeleton's repeated 2027 U.S. IPO target shows exit ambition, but public evidence still reads as pre-underwriting signaling rather than IPO-ready disclosure. | 中 | SV001, SV004, SV005 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Skeleton Technologies | Power systems for AI infrastructure | Skeleton | |
| SO002 | Skeleton Technologies | Skeleton Technologies Announces First Close of Pre-IPO Funding Round with Initial €33 Million Investment | Skeleton Technologies, a global AI infrastructure and grid power systems provider, today announced the first close of a larger funding round at €33 million, raising its total venture capital funding to €392 million in preparation for its planned initial public offering (IPO) in the United States in 2027. |
| SO003 | Passive Components | Siemens Backs Skeleton’s US AI Power Manufacturing Expansion | The announcement does not provide the workload profiles, comparison baseline, measurement boundary or detailed test conditions behind those figures. |
| SO004 | Vestbee | Skeleton Technologies lands €33M as hyperscaler demand accelerates for AI power solutions | Founded in 2009 by Taavi Madiberk and Oliver Ahlberg, Skeleton Technologies develops high-power energy storage systems based on supercapacitors and its proprietary Curved Graphene technology. |
| SO005 | toflow.ai | What Funding Has Skeleton Technologies Raised? Full Round History | toflow.ai | Valuation at RoundAbove $1B |
| SO006 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SO007 | Skeleton Technologies | About Skeleton Technologies | Power Systems for AI Infrastructure | Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SO008 | Tech Funding News | Eyeing a 2027 US IPO, Skeleton Technologies secures €33M to power AI data centres with supercapacitors — TFN | New investors are Axon Partners Group, SmartCap, and Taiwania Capital, while existing investor CBMM also joined. |
| SM001 | Data Center Dynamics | Supercapacitor developer Skeleton opens Leipzig factory in Germany, targets AI data center market | While the company’s technology is used across both the automotive sector and electrical grids, its fastest-growing customer segment is AI infrastructure. |
| SM002 | Passive Components | Skeleton Technologies Expands in U.S. to Power AI Data Centers | Data center electricity demand in the United States is expected to more than double by the end of this decade, with AI data centers representing one of the fastest-growing segments. |
| SM003 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SM004 | Vestbee | Skeleton Technologies lands €33M as hyperscaler demand accelerates for AI power solutions | The funding round is intended to help Skeleton Technologies scale its high-power energy storage solutions for AI data centers, expand manufacturing capacity in Europe and the United States. |
| SM005 | Enlit World | AI firm Skeleton secures investment to tackle data centre power constraints | AI firm Skeleton secures investment to tackle data centre power constraints | Enlit World |
| SM006 | Skeleton Technologies | About Skeleton Technologies | Power Systems for AI Infrastructure | Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SM007 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power | Varkaus SuperBattery factory produces one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. |
| SM008 | Skeleton Technologies | SuperBattery | Skeleton | Battery Backup Unit (BBU) for data center backup power, based on Skeleton’s proprietary SuperBattery energy storage technology. ... Backup duration Up to 90 seconds ... Peak power 1.2 MW ... Nominal Power 800 kW |
| SM009 | Tech Funding News | Eyeing a 2027 US IPO, Skeleton Technologies secures €33M to power AI data centres with supercapacitors — TFN | In 2025, data centre electricity use rose by 17%, with AI centres growing even faster. The International Energy Agency expects data centre electricity use to double by 2030, and AI-specific demand to triple. |
| SP001 | Skeleton Technologies | Power systems technology for AI infrastructure | Skeleton power systems respond in microseconds to stabilize high-density AI compute loads, improving power quality, efficiency, and infrastructure utilization from the IT rack to the electrical grid. |
| SP002 | Skeleton Technologies | About Skeleton Technologies | Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SP003 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SP004 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power | The new facility is Europe’s first site producing high-power batteries for AI data centers, and the first globally to industrialize backup power systems for mission-critical computing and emerging sectors such as nuclear fusion. |
| SP005 | Data Center Dynamics | Supercapacitor developer Skeleton opens first US engineering facility in Houston, Texas | The system is a high-power energy-storage device designed to handle sharp load fluctuations in grids and AI data centers in under a millisecond. |
| SP006 | Vestbee | Skeleton Technologies lands €33M as hyperscaler demand accelerates for AI power solutions | Unlike conventional lithium-ion batteries, Skeleton’s supercapacitors store energy electrostatically rather than through chemical reactions, enabling them to charge and discharge within microseconds, withstand significantly more cycles, and operate with lower fire risk. |
| SP007 | InnoEnergy | Skeleton Technologies opens SuperBattery factory in Finland | By localising advanced battery production in Finland, Skeleton helps reinforce European manufacturing independence at a time when demand for high-power storage is accelerating across AI infrastructure, mobility, defence, and grid stability. |
| SP008 | Batteries International | Skeleton opens Finland’s 1GW SuperBattery site to power AI | Batteries International reported last year that CBMM and British battery developer, Echion, had opened a niobium-based anode material production plant in Araxá, Brazil which they said was poised to revolutionize the lithium ion battery industry by enabling ultra-fast charging and enhanced performance. |
| SP009 | toflow.ai | What Funding Has Skeleton Technologies Raised? Full Round History | Siemens invested more than $100 million in Skeleton Technologies in October 2026, becoming a minority shareholder and securing exclusive global rights to sell and scale Skeleton's supercapacitor products within its electrical infrastructure offering for data centers. |
| SP010 | Wikipedia | Skeleton Technologies | Skeleton offers supercapacitor-based energy storage systems for wind power applications, which help control wind power plant output and provide ancillary services to the power system. |
| SI001 | Batteries International | New Siemens investment in Skeleton’s supercap tech to boost AI data centre services | Siemens has made a further undisclosed investment in supercapacitor firm Skeleton Technologies as part of a new ‘exclusive partnership’ for the German industrial group to expand in the AI factories sector... Under the partnership, announced on October 14, Siemens will have exclusivity in selling and scaling Skeleton’s products in the data centre market. |
| SI002 | Skeleton Technologies | Skeleton opens €220 million Leipzig factory to stabilise Europe’s electrical grid and AI infrastructure | Leipzig, Germany - 28th November 2025: Skeleton Technologies has officially opened its €220 million SuperFactory in Markranstädt, near Leipzig... Opening the Leipzig supercapacitor facility follows the recent opening of the company's SuperBattery facility in Varkaus, Finland. The Leipzig SuperFactory will create 420 jobs in Saxony and is designed for an annual output of up to 12 million cells. |
| SI003 | InnoEnergy | Skeleton Technologies opens new SuperFactory in Germany | Designed for an annual output of up to 12 million cells, the facility is already supplying Siemens, General Electric, and Hitachi Energy for European grid applications, as well as major US hyperscalers for AI data centres, and represents an investment of €220 million. |
| SI004 | Battery Tech Network | Skeleton Technologies - Battery-Tech Network | Skeleton Technologies has raised over €300M from investors including Marubeni Corporation, Siemens Financial Services, the founders of Adyen, the European Investment Bank, Germany’s Federal Ministry for Economic Affairs and Energy, FirstFloor Capital, Taavet Hinrikus, and Bengt Wahlqvist. |
| SI005 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SI006 | Skeleton Technologies | About Skeleton Technologies | Power Systems for AI Infrastructure | Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SI007 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power | Skeleton Technologies, Europe’s leading manufacturer of high-power energy storage solutions, announced the official opening of its SuperBattery factory in Varkaus, Finland, and total investment of 50 million euros in the factory... Skeleton currently employs more than 65 people in Finland and in Varkaus alone, will grow to over 200 employees by the end of 2029... Varkaus SuperBattery factory produces one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. |
| SI008 | Enlit World | Skeleton opens Europe’s largest supercapacitor factory in Germany | Skeleton opens Europe's largest supercapacitor factory in Germany | Enlit World |
| SI009 | Tech Funding News | Eyeing a 2027 US IPO, Skeleton Technologies secures €33M to power AI data centres with supercapacitors — TFN | Skeleton Technologies, an energy storage company based in Estonia and Germany that we met at TBB, has closed the first part of its pre-IPO funding round at €33 million. This brings its total venture capital raised to €392 million... The fresh funding will help Skeleton expand into the US market, including setting up production capacity there. It will also support further manufacturing growth in Europe as the company prepares for its 2027 IPO. |
| SI010 | Vestbee | Skeleton Technologies lands €33M as hyperscaler demand accelerates for AI power solutions | Skeleton Technologies, a global AI infrastructure and grid power systems provider, has announced a €33 million first close of a larger funding round, bringing its total venture capital funding to €392 million ahead of a planned U.S. IPO in 2027... The funding round is intended to help Skeleton Technologies scale its high-power energy storage solutions for AI data centers, expand manufacturing capacity in Europe and the United States. |
| SE001 | Siemens AG | Siemens invests in Skeleton Technologies expanding its data center portfolio | The solution can be deployed in existing AC architectures and retrofit projects, while providing a pathway toward future 800V DC standards. |
| SE002 | Battery Tech Expo Europe | Skeleton Technologies Opens SuperBattery Factory in Finland, Providing One Gigawatt of AI Power | Varkaus SuperBattery factory produces one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. |
| SE003 | Skeleton Technologies | SuperBattery | Integrated DC/DC converter and battery management system (BMS) |
| SE004 | Data Center Dynamics | Supercapacitor developer Skeleton opens first US engineering facility in Houston, Texas | The US is already a core market for the company, with more than 100MW of systems deployed nationwide and around half of its revenue generated from North American customers. |
| SE005 | Batteries International | Skeleton opens Finland’s 1GW SuperBattery site to power AI | The firm said on November 11 the plant in Varkaus, with a 1GW annual capacity, is Europe’s first producing high-power batteries for AI datacenters. |
| SE006 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SE007 | Skeleton Technologies | Power Systems for AI Infrastructure - About Skeleton Technologies | Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SE008 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power | Skeleton’s SuperBattery technology is powered by CBMM’s NBXCELER™ niobium oxides, developed in partnership with CBMM, the global leader in the production and commercialization of niobium products. |
| SU001 | ESS News | Siemens leads '9-figure' investment into Skeleton Technologies for AIDC power solutions - Energy Storage | Siemens is taking a minority stake and gains exclusive global rights to sell and scale Skeleton’s high-power storage products for data center customers. ... It has not named any hyperscaler customers. |
| SU002 | Electronicsmedia | Siemens Skeleton Technologies Partnership to Boost AI Data Centers | Siemens supports customers across the entire data center ecosystem from the utility connection to the white space and is a trusted partner relied upon by nine of the top ten global data center providers. |
| SU003 | Skeleton Technologies | Power systems technology for AI infrastructure | Skeleton | Skeleton power systems are engineered to stabilize high-density AI compute environments in real time, supporting everything from rack-level power fluctuations to grid-scale infrastructure challenges. |
| SU004 | Passive Components | Skeleton Releases Graphene‑Based UPS for AI Data Centers | The system is specified to enable a 40% increase in computing power while requiring up to 44% less grid connection capacity, which is highly relevant where grid access is constrained or subject to strict connection limits. |
| SU005 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SU006 | Skeleton Technologies | Power Systems for AI Infrastructure - About Skeleton Technologies | 2011 Development of space capacitors for our first customer, the European Space Agency ... 2019 Large-scale commercial deals signed with global OEMs, including Škoda Electric ... 2025 Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SU007 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power | Skeleton currently employs more than 65 people in Finland and in Varkaus alone, will grow to over 200 employees by the end of 2029, reinforcing Finland’s leadership in advanced manufacturing and clean technologies. Varkaus SuperBattery factory produces one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. |
| SU008 | Skeleton Technologies | Skeleton Technologies secures strategic investment led by Siemens to expand US manufacturing | The partnership will address hyperscaler demand from the grid to the rack, while bringing Skeleton’s technology and manufacturing closer to its largest AI infrastructure customers in the United States. |
| SR001 | Silicon Valley Invest Club | Skeleton Technologies tear sheet | |
| SR002 | Energy-Storage.news | Skeleton Tech begins manufacturing high-power solutions to AI boom in Europe | Skeleton has invested around €50 million (US$58.16 million) in its 1GW annual production capacity SuperBattery factory in Finland. The Leipzig SuperFactory required around €220 million investment and will produce around 12 million cells annually. |
| SR003 | Skeleton Technologies | Power systems technology for AI infrastructure | Skeleton power systems respond in microseconds to stabilize high-density AI compute loads, improving power quality, efficiency, and infrastructure utilization from the IT rack to the electrical grid. |
| SR004 | Data Center Dynamics | Supercapacitor co. Skeleton Technologies raises €33m as it targets 2027 IPO | The company, which develops graphene-based supercapacitors for the utility and AI data center markets, says the larger-than-expected round takes its total private funding to €392 million ($462m). |
| SR005 | EIT InnoEnergy | Skeleton Technologies opens new SuperFactory in Germany | Designed for an annual output of up to 12 million cells, the facility is already supplying Siemens, General Electric, and Hitachi Energy for European grid applications, as well as major US hyperscalers for AI data centres, and represents an investment of €220 million. |
| SR006 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | The U.S. is already a core market for Skeleton, with more than 100 MW of systems deployed and operating across the country. Approximately half of the company’s revenue is generated from North American customers. |
| SR007 | Skeleton Technologies | About Skeleton Technologies | Skeleton opens Leipzig Superfactory - the most modern supercapacitor factory in the world, and starts deliveries to AI data center hyperscalers. |
| SR008 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power | Varkaus SuperBattery factory produces one gigawatt of battery power for AI data centers, grid stability, mobility, and defense. |
| SR009 | Batteries International | Skeleton opens Finland’s 1GW SuperBattery site to power AI | The firm said on November 11 the plant in Varkaus, with a 1GW annual capacity, is Europe’s first producing high-power batteries for AI datacenters. |
| SR010 | Vestbee | Skeleton Technologies lands €33M as hyperscaler demand grows | The funding round is intended to help Skeleton Technologies scale its high-power energy storage solutions for AI data centers, expand manufacturing capacity in Europe and the United States. |
| SV001 | Business20Channel TV | Skeleton Technologies €33M Pre-IPO 2026: Supercapacitors Target AI Data | |
| SV002 | AI Weekly | Siemens Takes Exclusive Data-Center Rights to Skeleton Technologies, Lifts Supercapacitor Maker Past $1B | |
| SV003 | Skeleton Technologies | Skeleton Technologies secures strategic investment led by Siemens to expand US manufacturing | |
| SV004 | Tech Funding News | Eyeing a 2027 US IPO, Skeleton Technologies secures €33M to power AI data centres with supercapacitors | |
| SV005 | Data Center Dynamics | Supercapacitor co. Skeleton Technologies raises €33m as it targets 2027 IPO | |
| SV006 | Skeleton Technologies | Skeleton Technologies Accelerates U.S. Expansion to Address AI-Driven Grid Strain | |
| SV007 | Skeleton Technologies | About Skeleton Technologies | Power Systems for AI Infrastructure | |
| SV008 | Skeleton Technologies | Skeleton Technologies Opens SuperBattery Factory in Finland Providing One Gigawatt of AI Power |