Atom Computing
政府背书的中性原子量子计算龙头,私募估值为 $2.06B;但披露仍过少,不足以支撑按高确定性价格入场
Atom 是最可信的中性原子量子平台之一,但在公开经济指标稀薄的情况下,$2.06B 私募估值更适合继续研究,而不是作为高确信度入场价。
封面要素
公司概况
Atom Computing 是一家非上市中性原子量子计算公司,2018 年成立,总部位于 Berkeley, California。它的公开叙事已经把超过 1,200 个物理量子比特的规模、Microsoft 支持的逻辑量子比特商业化、DARPA Stage B 验证,以及 QuNorth 首个商业本地部署结合在一起, 使其成为更可信的后期量子硬件平台之一。已宣布资本超过 $300 million,Forge 报告的 D-1 估值为 $2.06 billion,这些都支撑公司的战略相关性; 但相对当前价格,运营模型仍明显披露不足。
- 成立时间
- 2018-01-01
- 创始人
- Ben Bloom, Jonathan King
- 创立地点
- Berkeley, California, USA
- 总部
- Berkeley, California, USA
- 产品
- Atom 销售或部署中性原子量子系统,包括一台接入 Microsoft 的逻辑量子比特机器,以及 QuNorth Magne 项目等本地部署。
- 客户
- 需要安全访问或高接触服务来使用先进量子系统的政府、主权、研究和企业买家。
- 商业模式
- 商业化本地部署量子系统及相关部署 / 支持工作,以逻辑量子比特能力和生态整合为主要价值驱动;任何经常性云收入或维护收入在公开资料中仍披露不足。
- 阶段
- Late-stage private (Series C / D-1)
- 融资情况
- 截至 2026 年 6 月,公开融资超过 $300 million,包括 $100 million Series C 和拟定的 $100 million Department of Commerce 支持; 随后 Forge 报告 D-1 定价对应 $2.06 billion 估值。
执行摘要
主要优势
- 中性原子规模、Microsoft 逻辑量子比特商业化和 DARPA Stage B,让这家私有量子公司拥有罕见可信的技术平台。
- QuNorth / Magne 销售是真实商业证明,说明 Atom 不只是发布研究里程碑,也能卖出本地部署系统。
- 累计融资超过 $300 million,再叠加 Department of Commerce 支持,相比许多量子同行明显缓解了近期融资压力。
- Forge 的 D-1 标记和保密申报信号表明,资本市场已把 Atom 视为可能的后期候选公司,而不是小型实验性初创。
主要风险
- 收入、利润率、订单积压和经常性收入占比披露仍太薄,无法有信心承销当前经济性。
- 商业证明仍集中在 QuNorth 这一旗舰部署上,而不是来自多元化的具名客户基础。
- $2.06B 估值已经内嵌高预期,给进度滑坡或后续需求疲软留下的容错很小。
- 公开量子可比公司波动仍很大,负面市场评论也仍把这个品类看成期权式资产,而不是软件式资产。
- 政府背书有战略价值,但本身不能证明客户需求可重复,也不能证明公司已具备 IPO 准备度。
未决问题
- 已签署的 D-1 条款清单、投资者名单和清算优先权堆叠。
- QuNorth 合同金额、收入确认节奏和 Microsoft 收入分成条款。
- 首个旗舰系统之外,第二个具名付费部署的证据。
- 经审计财务数据,以及 Forge 保密申报信号究竟仍有效还是历史信号的更清楚说明。
- 覆盖订单积压、毛利率、服务负担和经常性收入结构的运营指标。
目录
01公司概况
1.1 身份、阶段、版图与商业模式
Atom Computing 称自己使用光学捕获中性原子阵列,打造高度可扩展的基于量子门的量子计算机。已采纳证据在核心事实上一致,但并不覆盖每个运营细节的细微差别。 Forge 将公司总部列为 Berkeley, California,并把成立时间定为 2018 年;2023 年量子比特规模新闻稿显示其在 Boulder, Colorado 设有商业运营设施,2026 年官网则显示公司已经沿着产品、合作和招聘多线推进,而不是单纯的实验室项目。这种组合很重要,因为后续章节应把 Atom 视为一家拥有真实商业化触面的后期非上市深科技公司,而不是尚未成品化的研究剥离项目。最显眼的商业模式,是交付本地部署系统,并搭配具备逻辑量子比特能力的硬件、 战略伙伴集成,以及面向政府或企业的部署支持。与公开叙事围绕云市场访问展开的云优先竞争对手不同,Atom 现在公开强调 AC1000、逻辑量子比特, 以及主权或企业部署路径。因此,公司在身份和产品定位上看起来已经越过概念验证阶段,但在运营规模、客户广度和财务成熟度披露上仍有选择性。[CO001, CO002, CO003, CO004, CO005, CO006]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 / 备注 |
|---|---|---|---|---|
| 成立 | 2018 | 2018 | 高 | Atom 主页和 Forge 档案相互印证。 |
| 总部 | Berkeley,加利福尼亚州 | 2026-07-14 | 中 | Forge 给出了保留证据中最清晰的总部信息。 |
| 运营足迹 | Berkeley 总部,加上 Boulder 运营信号 | 2026-07-18 | 中 | 2023 年新闻稿提到 Boulder 运营;当前公开网站没有完整列出站点。 |
| 当前阶段 | 后期私营 / IPO 前 | 2026-07-18 | 中 | Forge 估值和保密递交信号提供支持,但没有公开 S-1。 |
| 核心架构 | 光学俘获中性原子 | 2026-07-18 | 高 | 主页、融资新闻稿和技术论文反复出现。 |
| 物理量子比特规模 | 1,200+ 全连接量子比特;1,225 位点阵列 / ~1,180 量子比特 | 2026-07-18 | 高 | 主页和 2023 年公告在 1,200+ 规模上吻合。 |
| 逻辑量子比特验证 | 24 个纠缠逻辑量子比特;在 28 个逻辑量子比特上计算 | 2024-11-19 | 高 | Microsoft 博客和 arXiv 论文相互印证。 |
| 最新宣布融资 | 累计超过 $300M;包括 $100M Series C 和计划中的 $100M DoC 支持 | 2026-06-16 | 高 | 公司和 PR Newswire 说法一致。 |
| 最新估值证据 | $2.06B Series D-1 估值 | 2026-07-14 | 中 | Forge 基于 COI 的市场数据证据;公司没有对应新闻稿。 |
| 具名商业客户 | QuNorth / Magne | 2025-07-17 | 高 | 客户侧和独立来源确认该交易。 |
| 当前收入 / ARR | null | 2026-07-18 | 高 | 没有保留来源披露当前收入或 ARR。 |
| 客户数 / 员工数 | null | 2026-07-18 | 高 | 没有保留来源披露可靠的全公司口径数量。 |
null 行表示保留的公开来源无法支持这些指标;缺口应带入后续尽调。
[CO001, CO003, CO005, CO009, CO014, CO017]Atom 目前的身份由中性原子硬件、逻辑量子位软件集成、主权客户和战略资本共同串起。
[CO005, CO010, CO012, CO020, CO024, CO027]公开证据在资本和里程碑指标上最强,在收入、客户、员工数等运营指标上最弱。
[CO001, CO009, CO014, CO020, CO024, CO032]1.2 创始人、领导层、投资人和治理可见度
公开创始人记录比 Atom 的融资记录更薄,但足以支撑后续章节。Forge 将 Benjamin Bloom 和 Jonathan King 识别为创始人; Atom 自身当前材料和外部伙伴材料则一致把 Ben Bloom 放在创始人 CEO 角色。资本可见度明显更好。公司和 PR Newswire 都披露,截至 2026 年 6 月, 累计融资超过 $300 million,其中包括 Third Point Ventures 领投、DCVC 和 Cisco Investments 参投的 $100 million Series C,以及另一笔 $100 million Department of Commerce 意向书。DCVC 自身评论更进一步,将政府支持描述为股权,这提高了 Atom 与政策对齐的战略重要性, 即便具体法律文件仍未公开。Forge 又补上一层,报告 2026 年 7 月 Series D-1 估值为 $2.06 billion、融资约 $89 million;但该估值来自基于 COI 的市场数据展示, 不是公司新闻稿披露的轮次信息,因此应视为强方向性证据,而不是完整披露的定价轮透明度。治理披露仍明显弱于融资披露:已采纳公开记录识别了领导者和投资人, 但没有给出完整董事会图谱、委员会结构或详细少数股东保护条款。[CO007, CO008, CO009, CO010, CO011, CO012]
| 人物 | 职务 | 证据 | 创始人-市场匹配 / 依赖 | 当前尽调备注 |
|---|---|---|---|---|
| Benjamin Bloom | 创始人兼 CEO | 公司和合作伙伴来源多次将 Bloom 标识为创始人兼 CEO | 高;Bloom 支撑技术可信度和对外叙事 | 关键人物依赖仍然显著。 |
| Jonathan King | 联合创始人 / 技术负责人 | Forge 将其列为创始人,技术论文作者名单也包含他 | 架构连续性上为高 | 公开高管头衔可见度弱于 Bloom。 |
| Microsoft 量子领导层 | 战略合作伙伴对口方 | Krysta Svore 和 Microsoft 博客把逻辑量子比特系统与 Atom 硬件联系起来 | 对通过逻辑量子比特技术栈商业化很重要 | 这是伙伴依赖,而不是内部治理。 |
| Third Point Ventures / DCVC 与 Cisco | 资本和生态利益相关方 | Series C 公告及支撑性报道点名 | 对融资、信号和网络访问重要 | 经济控制条款仍为私人信息。 |
公开证据在创始人兼 CEO 身份上很强,但完整高管名单和董事会构成仍薄弱。
[CO002, CO007, CO008, CO010, CO011, CO027]| 利益相关方 | 角色 | 控制 / 经济重要性 | 尽调要求 |
|---|---|---|---|
| Third Point Ventures | Series C 领投方 | 传递出对后期量子商业化的信心 | 核实董事会权利和跟投能力。 |
| DCVC | 长期投资方和政策放大器 | 把 Atom 接入深科技和联邦政策网络 | 确认所有权和治理影响力。 |
| Cisco Investments | 战略投资方 | 增加网络和企业生态可信度 | 厘清资本之外的商业合作范围。 |
| U.S. Department of Commerce 计划 | 计划提供 $100M 战略资本 | 可能降低融资风险,同时增加政策耦合 | 审查最终结构、里程碑和限制。 |
| Microsoft | 硬件-软件商业化伙伴 | 对逻辑量子比特产品化和 QuNorth 交付至关重要 | 厘清排他性、收入分成和路线图依赖。 |
| EIFO / Novo Nordisk Foundation 与 QuNorth | 客户背书生态利益相关方 | 提供首个商业本地部署验证和北欧市场滩头阵地 | 评估在主权支持旗舰销售之外的可复制性。 |
| 未披露 D-1 投资方 | 最新定价资本 | 帮助设定 Forge 报告的 $2.06B 标记 | 取得完整股权结构表、优先权和投资方身份。 |
多个利益相关方的经济条款、治理权利和持股比例没有完全公开。
[CO010, CO011, CO012, CO013, CO014, CO024]1.3 里程碑、合作伙伴与商业化证据
Atom 最强的公开证据来自里程碑。2023 年,公司宣布一个 1,225 位点阵列,装载大约 1,180 个量子比特,并称这是首个超过 1,000 个量子比特的通用门模型系统。2024 年和 2025 年,Microsoft 合作把讨论从原始物理规模推向逻辑计算,Microsoft 和 Atom 展示了一台商业机器, 可纠缠 24 个逻辑量子比特,并用 28 个逻辑量子比特进行计算。随后 DARPA 选择 Atom 进入 Quantum Benchmarking Initiative 的 Stage B, 让公司在政府主导、旨在验证实用规模路线图可信度的项目中获得可见角色。最重要的商业证据是 QuNorth。多个独立和客户侧来源称,Atom 与 Microsoft 将把 Magne——一套 Level 2 系统,包含 50 个逻辑量子比特和 1,225 个物理中性原子量子比特——交付给一个由 EIFO 和 Novo Nordisk Foundation 资助、丹麦方面拥有的倡议。 这很关键,因为它把 Atom 的故事从“有前景的物理”转成一笔具体的本地部署系统采购,背后有客户所有权、具名应用和区域战略意图。Cisco、NVIDIA 和 Nu Quantum 进一步拓宽了战略网络,但 QuNorth 和 Microsoft 仍是 Atom 将技术里程碑转化为可部署系统的最清晰证据。[CO017, CO018, CO019, CO020, CO021, CO022]
| 日期 | 事件 | 类型 | 金额 / 状态 | 参与方 | 影响 |
|---|---|---|---|---|---|
| 2018-01-01 | Atom 成立 | 成立 | 公司成立 | Ben Bloom;Jonathan King | 锚定后期私营公司时间线。 |
| 2023-10-24 | 宣布 1,225 位点 / ~1,180 量子比特系统 | 产品 | 跨过 1,000 量子比特门槛 | Atom Computing | 确立了中性原子规模里程碑。 |
| 2024-11-19 | 与 Microsoft 宣布商业逻辑量子比特系统 | 合作 | 24 个纠缠逻辑量子比特;28 个逻辑量子比特计算 | Microsoft;Atom | 叙事从物理规模转向可靠逻辑计算。 |
| 2025-05-06 | 宣布入选 DARPA Stage B | 监管 | 入选为期一年的 QBI Stage B 项目 | DARPA;Atom;Microsoft 支持 | 提升路线图可信度,也增加外部审视。 |
| 2025-07-17 | 宣布 QuNorth / Magne 交易 | 商业 | 计划 50 个逻辑量子比特;1,225 个物理量子比特 | QuNorth、Microsoft、Atom、EIFO 与 Novo Nordisk Foundation | 首个公开商业本地部署逻辑量子比特系统销售。 |
| 2025-07-17 | EIFO 和 Novo 承诺向 QuNorth 投入 €80M | 融资 | €80M | EIFO;Novo Nordisk Foundation | 打造主权支持的旗舰客户生态。 |
| 2026-05-21 | DCVC 公开描述 DoC 支持 | 监管 | 投资方描述为 $100M 股权持仓 | DCVC 与 U.S. Department of Commerce | 显示产业政策背书,也显示政策依赖。 |
| 2026-06-16 | 宣布 Series C 和累计融资 | 融资 | >$300M 累计融资;$100M Series C | Third Point Ventures、DCVC 与 Cisco Investments | 延长资产负债表跑道和商业化投入。 |
| 2026-06-17 | 主页宣布 Nu Quantum 合作 | 合作 | 战略性公用事业级 / 网络协作 | Atom;Nu Quantum | 显示计算核心之外更广的系统栈雄心。 |
| 2026-07-08 | Series D-1 融资出现在 Forge COI 数据中 | 融资 | $89.07M;$2.06B 估值 | 未披露投资方 | 最新估值信号仍由市场数据驱动,而非完整披露。 |
状态值反映保留公开证据中最强的材料,也保留融资数据来自市场数据展示、而非公司披露这一点。
[CO001, CO009, CO010, CO013, CO014, CO017]Atom 的公开记录显示了一条清晰路径:先刷新中性原子规模里程碑,再推进逻辑量子位商业化,并拿到主权支持客户验证。
[CO001, CO009, CO014, CO017, CO020, CO022]1.4 仍未知的事项及其重要性
本章的主要警示不是 Atom 是否真实存在的矛盾,而是经济和治理叙事仍有多少留在私下。已采纳来源对身份、融资、客户和里程碑叙事的支撑,足够用于后期非上市公司概览; 但它们没有披露收入、ARR、毛利率、员工数、客户数或 D-1 详细条款,也没有说明 Forge 的 IPO 信号意味着近期活跃上市路径,还是仅为历史保密流程的痕迹。 因此,关于量子时间线的独立怀疑评论仍然相关。Atom 在可见商业化上或许领先许多非上市同业,但若实用规模部署慢于预期,或首批客户项目过于集中,整个板块仍可能令人失望。 对后续章节而言,实际结论很清楚:公司概况可以把 Atom 的身份、融资基础和里程碑记录视为已确立事实,但必须继续把估值精度、可重复收入质量和多元化作为仍需尽调的问题, 而不是已确认的优势。[CO031, CO035, CO036, CO037, CO041, CO042]
1.5 图表
02市场分析
2.1 市场边界、纳入支出和替代品
Atom 应放在量子计算硬件与服务市场内分析,但即便在这个市场内,它的实际边界也比许多标题式 TAM 幻灯片更窄。相关支出不是“所有量子技术”,甚至也不是每一美元量子计算支出。 Atom 可触达的层,是通过云访问、本地部署系统、混合 HPC 集成和应用共创出售的中性原子量子能力,面向那些需要仿真或优化工作负载、且这些负载未来可能超过经典方法的买家。 这个边界包括 Aquila 和未来 Braket 访问、AIST 式本地部署路径,以及与算法开发和集成相关的服务工作。它不包括量子传感、量子通信或 QKD,也不包括多数后量子密码迁移预算, 因为这些类别可以快速增长而无需购买 Atom 硬件。因此,现状替代品不是“什么都没有”,而是经典 HPC、GPU 密集型 AI 仿真、经典优化软件和量子安全软件迁移。 BCG 提醒量子计算今天相对经典计算仍没有可触知商业优势,这一点很关键,因为它意味着 Atom 竞争的对象是快速改进的替代方案,而不只是其他量子供应商。中性原子仍然相关, 因为 AWS、Braket 和 Atom 都把这种模态与仿真和优化任务相连,而这些任务重视几何结构、可重构性和类全连接。[CM001, CM002, CM003, CM004, CM005]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 与 Atom 的相关性 |
|---|---|---|---|---|
| 中性原子量子硬件和系统 | 量子处理器、控制栈、本地安装、集成、支持 | 量子传感硬件、量子网络基础设施 | 国家实验室、主权项目、研究联盟、企业设计合作伙伴 | 这是 Atom 核心市场,因为 Aquila、Gemini 和 Libra 级系统就在这里销售或配置。 |
| 云端量子访问 | Braket 用量、高级云访问、指导、工作流开发、仿真时间 | 与量子无关的通用云计算支出 | 研究人员、企业 R&D 团队、大学、平台团队 | 这是重要入口,因为 Atom 当前商业化路径先从访问和共设计开始,再走向完整部署。 |
| 混合 HPC 与量子集成 | 共址部署、工作流编排、经典仿真、培训、基准设计 | 与量子项目无关的独立经典超算 | 国家实验室、超算中心、政府支持的研究枢纽 | 很关键,因为 AIST、NERSC、AWS、Pawsey 和 ICSC 都通过混合 HPC 用例来定位 Atom。 |
| 应用共设计与服务 | 用例选择、算法开发、集成、企业就绪工作 | 没有硬件或工作流绑定的通用战略咨询 | CIO/CTO 办公室、科学计算团队、创新预算 | 很重要,因为 BCG X 和 Deloitte 通过价值验证和路线图项目把 Atom 销售给企业。 |
| 邻近量子安全和通信支出 | 后量子准备工作坊、密码迁移规划、安全网络试点 | 直接中性原子计算收入 | 政府、银行、关键基础设施运营方 | 制造紧迫感和相邻预算,但其中大部分支出不会直接转化为 Atom 收入。 |
| 现状替代栈 | 经典 HPC、GPU AI、经典优化、量子启发式软件 | N/A | 现有 IT、R&D 或运营预算 | 这才是 Atom 必须击败的真正既有方案:它必须在 ROI、信任和工作流适配上胜出,而不能只靠技术新颖性。 |
纳入支出只限 Atom 相关的中性原子计算、混合部署和共设计渠道;排除类别是相邻或替代市场,而不是 Atom 的 TAM。
[CM001, CM002, CM003, CM004, CM005]2.2 证据约束下的规模测算:收入、价值池、主权预算和标杆采购视角
只有先收紧市场边界,公开市场规模测算才有用。短期第三方收入视角并不巨大:McKinsey 估计 2025 年量子计算公司收入刚过 $1 billion, 2028 年达到 $3.2 billion 至 $4.4 billion;QED-C 给出相似但并不完全一致的图景,2025 年为 $1.4 billion 至 $1.9 billion, 2028 年超过 $3 billion。这些差异已经足以说明,市场仍对定义高度敏感。BCG 对供应商在容错前能现实变现多少更加谨慎,将 2030 年供应商市场维持在仅 $1 billion 至 $2 billion,同时保留 2040 年 $90 billion 至 $170 billion 的更大供应商收入情景,以及 $450 billion 至 $850 billion 的经济价值区间。对 Atom 而言,最有决策价值的视角既不是宽泛的 2035 年价值池,也不是简单的 TAM 级联,而是主权预算和标杆采购视角:BCG 称公共订单已支撑超过半数市场, 英国战略承诺十年投入 £2.5 billion,美国再授权路径资助测试床和应用,DOE 希望到 2028 年拥有一个具备科学相关性的容错平台,而 Atom 自身已经在 AIST、NERSC、AWS 和 Italy 的 ICSC 项目中有具名证据。这足以说明市场真实存在,但若没有未公开定价、合同和利用率数据,还不足以发布清晰的中性原子特定 SAM 或 SOM。 因此,正确结论是 Atom 可触达的早期市场偏主权和仿真密集;长期上行取决于容错采用能否接近公司自己的时间表。[CM006, CM007, CM008, CM009, CM010, CM011]
| 发布方 / 视角 | 年份 | 地理范围 | 范围 | 数值 | CAGR / 斜率 | 方法论 | 置信度 | 关键限制 |
|---|---|---|---|---|---|---|---|---|
| McKinsey QT Monitor 2026 | 2025-2028 | 全球 | QC 公司收入 | 2025 年 $1.1B-$1.4B;到 2028 年 $3.2B-$4.4B | 47% CAGR (2024-2028) | 来自专家访谈、新闻搜索和 McKinsey 分析的收入模型综合 | 中-高 | 衡量厂商收入,而不是企业总价值或全部量子技术支出。 |
| McKinsey QT Monitor 2025/2026 | 2024-2035 | 全球 | QC 市场规模 / 用例价值 | 2024 年 $0.65B-$0.75B;到 2035 年市场 $43B-$72B;潜在价值 $1.3T-$2.7T | 只有商业化之后才出现阶跃变化 | 两个 Monitor 年份,结合收入和价值池视角 | 中 | 混合了直接市场规模和潜在经济价值;不是干净的 TAM 线。 |
| QED-C 市场预测 | 2025-2028 | 全球 | QC 细分收入 | 2025 年 $1.4B;到 2028 年 >$3B | ~30% 年增长 | 联盟市场分析加商业化调查 | 中 | 行业联盟视角可能比纯终端用户支出更接近厂商情绪。 |
| QED-C 行业现状 | 2025 | 全球 | QC 市场规模 | 2025 年 $1.9B | 30% 平均年增长 | 平衡计分卡式行业方法论 | 中 | 与 QED-C 市场预测页面并不完全一致,后者本身也显示定义敏感性。 |
| BCG 供应商市场视角 | 2030 | 全球 | 量子硬件和软件供应商收入 | $1B-$2B | 保守的 NISQ 时代爬坡 | 围绕 NISQ、广泛优势和容错阶段的情景分析 | 中-高 | 假设容错之前商业效用有限;可能低估乐观厂商情景。 |
| BCG 长期供应商视角 | 2040 | 全球 | 容错供应商收入 | $90B-$170B | 后置到 FT 时代 | 长期情景分析 | 低-中 | 取决于广泛容错采用;不能用作近端基准情景。 |
| 主权预算视角 | 2026+ | 美国 / 英国 / 盟友 | 公共资金、测试床和应用项目 | 英国 £2.5B / 10 年;美国 NQI 延期和 DOE 测试床;PQC 迁移强制要求 | 项目制,不基于 CAGR | 官方战略和项目文件 | 高 | 这些是赋能预算,不能全部直接转化为厂商收入。 |
| Atom 灯塔采购视角 | 2023-2026 | 日本 / 美国 / 意大利 / 全球云 | 具名访问和采购验证 | AIST 6.5B JPY 合同;NERSC QCAN;AWS Braket;ICSC 高级访问 | 管线建设,而非基于 CAGR | 具名客户和伙伴项目 | 中 | 有助于校准 SOM 现实性,但项目类型不同,不能加总成一个 TAM 数字。 |
这张表是受证据约束的观察框架,不是单一权威 TAM。数值有意混用收入、价值池、主权预算和具名采购视角, 因为公开资料里没有专门针对中性原子的 SAM 或 SOM 数据集。
[CM006, CM007, CM008, CM009, CM010, CM011]在证据约束下,视角堆叠展示:极大的长期价值池如何收窄为更小的近期收入层和 Atom 已验证需求层。
采用视角堆叠,而非严格的 TAM-SAM-SOM 级联,因为公开来源测量的是不同层级:价值池、供应商收入、细分市场收入和具名买方验证。
[CM007, CM010, CM011, CM013, CM015, CM016]公开区间显示,近期收入市场不大;但如果商业化和容错落地,长期上行空间很大。
中点是已发布低 / 高区间的算术中心,仅用于方便视觉比较;各行刻意覆盖不同时间窗口,因为没有单一公开来源覆盖 Atom 涉及的所有商业化窗口。
[CM006, CM007, CM008, CM009, CM010, CM012]2.3 按细分市场拆解买家、用户、付款方和采用路径
Atom 的买家地图比泛泛的“企业量子”故事更分层。当前最强买家类别是政府和国家实验室基础设施:付款方是主权或研究项目预算,用户是计算科学家或平台团队, 采用触发点是国家能力目标,而不是一年期企业 ROI 门槛。这就是 AIST、NERSC 和 ICSC 远比普通市场调研数据点重要的原因。第二层是企业共创:BCG X 和 Deloitte 都把 Atom 定位为企业、政府和 HPC 中心的合作伙伴,这些客户在大规模购买硬件时间前,需要先完成用例发现、基准设计和集成规划。在企业层里,制药和生命科学今天比金融更可信, 因为 Atom 有公开的生物学和药物发现项目;金融仍主要通过行业研究项目和 Atom 之外的混合试点可见。材料、化学和能源也在结构上重要,因为仿真是 QED-C、McKinsey、AWS 和 Pawsey 所共同指向的最有形短期用例集群。金融仍是长期目标,因为价值池很大、银行也在积极学习,但 Atom 自己的调查和公开银行证据都暗示其规模化更晚: 金融机构需要安全的混合环境、量子安全迁移计划,以及更强的实用性证据,之后才会成为重复买家。也就是说,Atom 的采用路径是先云端探索,再共创设计,再主权或实验室锚点部署, 最后才是更广泛的企业生产工作负载。[CM022, CM023, CM024, CM025, CM026, CM027]
| 细分市场 | 主要买方 | 用户 | 付款方 | 工作流 / 用例 | 预算所有者 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 政府 / 国家实验室 / HPC | 国家实验室、超算中心、主权量子项目 | 计算科学家、平台工程师、研究团队 | 政府拨款或项目预算 | 混合仿真、化学、材料、HEP、AI 关联研究工作流 | 项目办公室、实验室主任或部委支持的基础设施预算 | 能力建设任务,加上可信的混合 HPC-量子工作流访问。 |
| 医药与生命科学 | 药物发现团队、转化研究团队、设计合作伙伴 | 计算化学家、分子建模人员、生物信息学研究人员 | 研发创新预算 | 分子模拟、配体结合、蛋白质或生物学工作流 | 首席科学官、研发负责人或数字科学预算所有者 | 证明量子技术能改善一个高价值模拟任务,而这个任务目前靠经典方法只能做较差近似。 |
| 材料、化工与能源 | 材料研发团队、工业创新团队、应用科学实验室 | 材料科学家、化学家、工艺工程师 | 研发和先进工程预算 | 电池化学、材料发现、催化剂建模、工艺优化 | CTO、研发 VP 或业务单元创新负责人 | 拿出模拟或优化收益,并把收益挂到发现周期、吞吐量或可专利产出上。 |
| 金融服务 | 银行创新实验室、量化研究团队、安全负责人 | 量化人员、风险经理、安全架构师 | 技术、风险或转型预算 | 组合优化、Monte Carlo 加速、欺诈分析、PQC 准备 | CIO/CTO、首席风险官或网络安全预算所有者 | 安全的混合环境,加上量子在真实金融工作流上跑赢经典方法的证据。 |
| 国防与公共安全 | 国防机构、涉密研究项目、安全计算集成商 | 建模团队、安全工程师、任务分析师 | 国防或国家安全项目 | 安全通信规划、任务模拟、量子就绪项目 | 采购办公室或项目执行办公室 | 战略任务、PQC 紧迫性,以及供应商能处理主权和验证要求的可信度。 |
| 大学与研究联盟 | 国家级联盟、大学、共享研究基础设施 | 研究人员、博士后、研究生 | 公共资助或联盟资金 | 原型工作流、培训、基准测试、方法开发 | 项目 PI 或联盟主任 | 可负担的访问、培训支持,以及从试点工作走向可发表研究或国家能力建设的路径。 |
预算归属和触发因素基于留存的买方证据和市场研究;金融仍是 Atom 特定证据最弱的细分市场, 因为公开客户披露仍然缺失。
[CM022, CM023, CM024, CM025, CM026, CM027]Atom 目前最强的细分市场,把公共资金或研究使命、模拟密集型工作负载,以及可信的混合计算路径结合在一起。
矩阵排序的是相对就绪度,不是绝对市场规模;“Atom 当前验证”反映留存的公开证据,不含未披露管线数据。
[CM022, CM023, CM024, CM035, CM036, CM037]公开买方旅程先从云访问和共同设计开始,再进入主权或实验室锚定部署,最后才是更广泛的容错工作负载。
流程描述的是 Atom 目前展示得最清晰的公开路径;并非每个细分市场都会走完每一阶段,金融客户也可能比主权 HPC 买方更久卡在早期阶段。
[CM025, CM026, CM027, CM028, CM029, CM033]2.4 增长驱动、采用约束和估值相关性
Atom 的市场有三个最重要驱动。第一,主权资金和公私测试床在主流企业 ROI 出现前,就持续创造非共识需求。第二,后量子密码和安全规划迫使政府及受监管机构现在就建立量子能力, 即便这些支出不会立即转成硬件采购。第三,市场正从炒作转向证据驱动采购;如果 Atom 的公开里程碑、伙伴栈和混合 HPC 定位强于同业,这反而利好 Atom。 但约束同样重要。BCG、Moody’s 和 IBM 都以不同方式指出,硬件就绪度仍是瓶颈,优势主张需要反复验证。Atom 自己的调查显示,只有少数机构已把应用有效规模化; QED-C 也提示人才和供应链瓶颈。经典计算竞争是另一个重大约束:AI 和 HPC 改进速度足以抬高每个仿真和优化用例所需的 ROI 门槛。BCG 的成本比较对估值尤其重要, 因为买家往往想要短回本周期,而量子计算时间仍比经典计算贵得多。换句话说,这个市场可以同时战略上很大、商业上很窄。估值上,投资人应更关注 Atom 的渠道证据、 采购质量和节奏纪律,而不是最终量子经济的抽象规模。[CM016, CM017, CM018, CM019, CM020, CM021]
| 驱动因素 / 约束 | 方向 | 时间 | 对 Atom 的含义 | 尽调追问 |
|---|---|---|---|---|
| 主权量子资金与测试床 | 顺风 | 当前活跃,持续到 2030+ | 即使广泛企业 ROI 尚未被证明,也能维持近期需求;最契合 Atom 的公开证据集。 | 按主权项目梳理 Atom 管线,而不是只看整体 TAM 叙事。 |
| 后量子密码和安全截止期限 | 顺风 | 2026-2035 迁移窗口 | 推动量子就绪预算和安全基础设施规划,政府和金融尤甚。 | 把相邻 PQC 预算与直接硬件转化假设拆开。 |
| 仿真优先的用例集中 | 顺风 | 近期 | 化学、材料和生物学与 Atom 的中性原子优势及公开合作伙伴组合匹配。 | 索要能把 Atom 硬件与买方 KPI 挂钩的基准证据,而不只是技术可行性。 |
| 混合 HPC 与云集成 | 顺风 | 近期至中期 | AWS、AIST、NERSC、Pawsey 和 ICSC 让买方更容易把 Atom 纳入既有计算栈。 | 检验混合工作流性能和延迟是否足够支撑客户反复使用,而不只是试点。 |
| 硬件就绪度与时间表不确定性 | 逆风 | 即刻且持续 | 如果 2028 年明显推迟,当前乐观情景的大部分会后移,而经典替代方案还在进步。 | 委托外部评估 Libra 架构时间表和里程碑依赖。 |
| 经典 AI 与 HPC 竞争 | 逆风 | 即刻且持续 | Atom 必须跑赢一个移动靶,尤其是 AI 和 GPU 持续改善的仿真与优化工作负载。 | 与具名经典基线对标,而不是泛泛主张“难以计算”。 |
| 人才短缺和供应链脆弱 | 逆风 | 即刻且持续 | 部署规模化需要稀缺的量子、光子和系统人才,专用组件也仍受供应约束。 | 检查招聘计划、合作伙伴依赖,以及超越灯塔项目时的组件瓶颈。 |
| ROI、信任和采购纪律 | 逆风 | 即刻且持续 | 市场已经转向证据驱动采购,买方签约前越来越要求经过验证的经济性、集成和治理。 | 索取云用户、设计合作伙伴和本地部署客户的定价、续约和转化数据。 |
方向从 Atom 视角判断。顺风提高买方意愿或资金可得性;逆风抬高证明门槛、拉长时间表, 或降低试点转生产合同的意愿。
[CM016, CM017, CM020, CM021, CM037, CM038]2.5 矛盾、未解问题和仍需尽调事项
本章最大的两个矛盾应被保留,而不是被平均掉。第一个是市场规模框架:短期收入视角集中在低个位数十亿美元,但长期价值池估计会跃升到数十亿或数千亿美元, 取决于来源衡量的是供应商收入、行业创造价值,还是更宽泛的量子技术伞。第二个是时间线:Atom 和 AWS 把 2028 年描述为可通过云访问的容错工作流起点, BCG 却仍把全面容错放在 2040 年之后。没有额外尽调,不能把两种观点同时当作基准情景。还有一些重要缺口。公开证据没有隔离出中性原子特定 SAM, 没有披露 Atom 的定价或部署经济性,也没有验证 Atom 在金融领域的具名客户,除了一般性行业实验。这些遗漏很重要,因为 Atom 的估值不太取决于“量子”在抽象意义上是否是大品类, 更取决于 Atom 能否在经典替代方案继续改进前,把一小组锚点渠道转成可重复商业合同。因此,实际尽调路径是要求披露垂直行业收入组合、定价和利用率假设、金融领域具名设计伙伴证据, 以及对 Libra 时间表的外部技术审查。在这些材料可得之前,市场逻辑应被视为战略方向强、在政府和 HPC 最强,但企业转化经济性仍不完整。[CM012, CM017, CM035, CM036, CM049, CM050]
2.6 图表
03竞争格局
3.1 格局与解决方案类别
Atom 并不在一个整齐的单一同业组里竞争。直接同业是中性原子队列,Atom Computing 和 Pasqal 都在出售同一个承诺:可扩展、基于量子门或混合的中性原子系统。 既有参照组更宽,也更有资本实力:IBM Quantum 和 Google Quantum AI 代表超导系统,IonQ 和 Quantinuum 代表离子阱,PsiQuantum 代表光子学, Rigetti 则是规模较小的超导全栈挑战者。D-Wave 不是同类的通用门模型对手,但对那些今天更在意生产工作流、而不是最终哪种架构赢得容错的优化密集型买家来说,它是真实替代品。 第二层替代品位于硬件供应商之上。AWS Braket 和 Azure Quantum 把多个后端聚合到一个采购界面之后,NVIDIA CUDA-Q 等混合 AI/HPC 平台让团队推迟对硬件的硬承诺。 这意味着 Atom 必须击败直接模态同业、更广泛的通用硬件替代方案,以及继续通过云经纪或经典 - 量子混合栈实验、而非标准化在单一硬件供应商上的默认选择。[CP001, CP002, CP003, CP006, CP009, CP012]
| 竞争者 / 替代方案 | 类别 | 规模 / 融资信号 | 目标买方 | 产品 / 访问范围 | 差异化与限制 |
|---|---|---|---|---|---|
| Atom Computing | 直接中性原子竞争者 | 2025 年融资 >$230M;云端与本地部署访问 | 研究实验室、HPC 中心、政府项目、探索仿真 / 优化 / ML 的企业 | AWS Braket 上的 Aquila、高级直接访问,以及本地部署中性原子系统 | 凭室温中性原子硬件、模拟仿真和逻辑量子比特路线图区隔;生态和渠道触达小于 IBM/IonQ/Quantinuum |
| Atom Computing | 直接中性原子竞争者 | 1,200+ 全连接量子比特和 $300M+ 融资在 2026 年宣布 | 追求大型门模型中性原子系统和逻辑量子比特项目的组织 | 直接中性原子系统和 Microsoft 关联的超级计算机路径 | 同一模态内规模信号强;公开商业化证据和定价仍薄 |
| Pasqal | 直接中性原子竞争者 | 公开上市前预计融资至少 €340M | 工业优化、金融、HPC 和云买方 | Pasqal Cloud、本地部署系统、Google Cloud 和 Azure 路线 | 工业和云布局强;确切经济性和容错成熟度仍在演进 |
| IBM Quantum | 既有通用硬件平台 | 30+ 台超过 100 量子比特的系统,2,300+ 可用量子比特,97% 正常运行时间 | 企业、HPC、研究机构和本地部署买方 | 通过 IBM Quantum Platform 提供 Open、按量付费、Flex、Premium 和本地部署计划 | 企业包装最透明、设备阵列最广;低温超导栈和纠错开销仍是真实约束 |
| Google Quantum AI | 既有研发对手 | Alphabet 支持的 Willow 项目,105 量子比特基准结果 | 顶尖研究合作者和旗舰科学伙伴 | 选择性 Willow Early Access 加 Cirq 软件生态 | 生态引力高、技术品牌强;近期采购入口弱,因为访问未公开 |
| IonQ | 商业离子阱竞争者 | 上市公司,2026 年 Q1 收入 $64.7M,并有激进的 2026-2030 路线图 | 企业、政府、云和数据中心运营商 | 直接云、预留、主流 SDK 支持,以及本地部署 Forte Enterprise 系统 | 商业重叠和渠道宽度很强;路线图野心走在当前 36 量子比特产品规模前面 |
| Quantinuum | 商业离子阱竞争者 | Honeywell 支持的 Helios 平台,提供 Azure 和直接订阅 | 企业、政府、医药、金融和先进研发买方 | 直接订阅、云服务、Azure 分发,以及本地部署 Helios 访问 | 高保真全栈方案,有公开经纪定价;偏企业的经济性可能放慢更广泛试验 |
| PsiQuantum | 光子长期竞争者 | 政府支持的 Chicago 和 Australia 公用事业级建设 | 主权、战略和长期企业买方 | 今天没有广泛公开算力;平台和软件故事瞄准公用事业级 FTQC | 深度制造逻辑和资本强度让它具备战略重要性;即时买方可及性有限 |
| Rigetti | 超导挑战者 | 全栈 Fab-1 制造商,有 108Q 云系统历史和 Novera 产品 | 研究、政府、HPC 关联实验室、硬件测试床 | 混合云服务,加上可立即交付的 9 量子比特本地部署 Novera QPU | 吸引需要硬件控制和低延迟混合工作流的买方;商业规模小于 IBM 或离子阱龙头 |
| D-Wave | 相邻 / 替代 | $20M FAU 系统协议和生产级 Leap 服务 | 优化密集型企业、公共部门、物流、制造 | Leap 量子云、混合求解器、本地部署 Advantage2 系统,以及门模型研发 | 今天用于优化任务的运营就绪替代方案最强;不是一对一的通用门模型竞争者 |
| AWS / Azure 经纪 + 混合 HPC 现状 | 云经纪与内部自建替代 | 超大规模云厂商分发,加上按量付费市场和经典 GPU 存量 | 想推迟硬件承诺的探索型买方、平台团队和预算所有者 | 经纪访问多个 QPU,加上 CUDA-Q 等混合开发栈 | 最容易多宿主并比较供应商;不能消除特定模态的性能差异 |
行覆盖直接中性原子同行、既有通用硬件平台、相邻替代和经纪化现状路线,这些都是 2026 年买方可以实际拿来与 Atom 比较的对象; 规模信号混合融资、设备阵列和商业化证据,而不是单一标准化指标。
[CP001, CP002, CP006, CP009, CP012, CP016]序数地图对比当前商业可访问性与架构差异化 / 长周期容错上行空间。
坐标轴是基于 2026 年留存公开证据得出的序数判断,依据包括访问模式、价格透明度和架构雄心,而不是来源报告的市场评分。
[CP002, CP006, CP009, CP012, CP016, CP021]3.2 定价、GTM 和信任比较
与许多量子初创公司相比,Atom 的商业触面异常清晰,因为 AWS 公布 Aquila 费率,Atom 也单独宣传高级直接访问和本地部署。即便如此,Atom 仍不是最容易购买的供应商。 IBM 仍是透明打包的基准,因为它在一个地方公开发布 Open plan、按量付费访问、更大的年度承诺和本地部署路径。IonQ 和 Quantinuum 是最接近的商业威胁, 因为两者都把直接云访问与企业部署路径结合起来,Quantinuum 还增加了通过 Azure 经纪的订阅,并公开月度定价。Google 的威胁不同:Willow 在技术上重要, 也有大型生态背书,但访问仍需提交提案,而不是像目录一样购买。Rigetti 和 D-Wave 各自增加了不同类型的信任信号——Rigetti 靠深度栈控制和可立即交付的本地部署 QPU, D-Wave 靠可用时间、混合求解器就绪度和真实签署的系统合同。实际采购中,买家比较的不只是架构和量子比特质量,还包括能否给产品定价、快速测试、接入 HPC, 并信任供应商能够支持多年期项目。[CP004, CP005, CP006, CP007, CP008, CP009]
| 购买标准 | Atom | IBM / Google | IonQ / Quantinuum | PsiQuantum / Rigetti | D-Wave / 云替代 |
|---|---|---|---|---|---|
| 公有云访问 | 是,通过 Amazon Braket 和高级直接选项 | IBM 是;Google 只有选择性研究访问 | 是,通过直接云和 Azure 或其他经纪路径 | Rigetti 是;PsiQuantum 今天没有广泛公开算力 | D-Wave 是,通过 Leap;AWS/Azure 经纪访问是核心替代优势 |
| 本地部署或专用访问 | 是,Atom 对外销售本地部署 HPC 安装和高级直接访问 | IBM 提供本地部署计划;Google 公开资料没有显示广泛本地部署方案 | 是:IonQ Forte Enterprise、Quantinuum 直接订阅和本地部署 Helios | Rigetti Novera 是;PsiQuantum 公用事业级站点尚不是普通客户部署 | 是:D-Wave Advantage2 本地部署;经纪云路线可以延后任何本地部署承诺 |
| 公开价格透明度 | Aquila 在 AWS 上价格透明度强;直接企业折扣未知 | IBM 在该组里价格透明度最强;Google 缺少公开定价 | 中:Azure 显示 IonQ 和 Quantinuum 定价,但直接条款仍要谈判 | 低:Rigetti 有一些 Azure 按时间计费信号和 Novera 销售;PsiQuantum 没有公开算力价格 | 低-中:D-Wave 页面省略简单费率卡;云经纪和经典栈暴露更透明的计量 |
| 开放开发者工具 | 中等:Braket 集成有帮助,但 Atom 不掌握主导 SDK 层 | 高:Qiskit 和 Cirq 形成很大开发者引力 | 高:IonQ 支持主流 SDK,Quantinuum 的 pytket 跨格式和后端 | 中-高:Rigetti 支持外部格式;PsiQuantum 仍更像平台逻辑,而不是广泛工具标准 | 中-高:D-Wave 支持 Python 工具;CUDA-Q 和经纪层拓宽替代栈 |
| 逻辑量子比特 / 容错叙事 | 高:路线图明确围绕逻辑能力和可部署容错系统 | 高:Google Willow 和 IBM 路线图都是旗舰容错叙事 | 高:IonQ 2026 逻辑目标和 Quantinuum Helios 路线图都很明确 | PsiQuantum 高;Rigetti 中 | D-Wave 通用 FTQC 低;经纪访问其他厂商则高 |
| 近期优化效用 | 中:模拟仿真和优化已上线,但类别采用仍早期 | 低-中:强项在研究和实验,而不是即时优化生产 | 中:企业试点和化学 / 金融项目真实存在,但仍处早期 | 低-中:Rigetti 偏实验;PsiQuantum 周期长 | 高:D-Wave 是最强近期替代,混合经典栈也能满足许多当前需求 |
| HPC / 混合集成 | 高:本地部署 / HPC 集成是 Atom 的核心信息 | IBM 本地部署和系统设计较强;Google 公开信息更偏研究 | 高:Forte Enterprise、Helios、Azure 和企业控制系统强调混合部署 | Rigetti 低延迟混合能力强;PsiQuantum 后续瞄准数据中心式基础设施 | 高:D-Wave 混合求解器和 NVIDIA 加速量子超级计算让替代逻辑具体可见 |
分组列是有意为之,因为领域过于碎片化,若不猜测无证据单元格,就无法干净地一厂商一列; 公共证据具有选择性时,单元格用定性表述,而不夸大成全面支持。
[CP002, CP007, CP010, CP016, CP019, CP020]| 提供商 / 路线 | 公开定价信号 | 合同模型 | 包含能力 | 未知项 / 注意事项 | 买方含义 |
|---|---|---|---|---|---|
| Atom / Amazon Braket | $0.30 / 任务、$0.01 / shot、$2,500 / 预留小时 | Braket 按量付费,加上高级直接访问和单独本地部署讨论 | Aquila 模拟仿真、优化、ML 实验,高级访问提供直接支持 | 直接企业折扣和已实现的高级访问经济性未公开 | 最容易定价的中性原子方案之一,降低评估摩擦 |
| IBM Quantum | 免费 Open 计划;PAYG $96/minute;Flex $72/minute;Premium $48/minute;本地部署仅报价 | 自助入门,加上合同计划和专用本地部署服务 | 量子算力访问、Qiskit Runtime、平台工具,以及可选加速器服务 | 实际企业折扣和本地部署定价不公开 | 同组最佳包装透明度,采购清晰度最快 |
| IonQ / Azure 和直接云 | Azure 公布 gate-shot 定价和最低执行费用;直接定价仍以报价为主 | 按需、预留、直接云和 Forte Enterprise 部署 | 多个 SDK、模拟器、直接支持和机架式企业硬件 | 非 Azure 直接条款和预留容量折扣未公开 | 与 Atom 高度重叠,适合想要实时访问加企业路径的买方 |
| Quantinuum / Azure 和直接订阅 | Azure Standard $125,000/month、Premium $175,000/month,用于 H2 访问 | 月度订阅、排队访问、直接云,以及本地部署 Helios 可用性 | H2 硬件、模拟器、软件栈、Azure 采购选项 | 非 Azure 直接定价、谈判折扣和利用率经济性不公开 | 企业就绪,但价格足以收窄买方范围 |
| Rigetti / Azure 或 Novera | Azure 按时间计费和直接硬件销售信号;这里没有保留 Rigetti 的广泛公开云价目卡 | 经纪运行时按量付费,Novera 直接购买 | 混合云访问、本地部署 9 量子比特 QPU、深度硬件控制 | 从 Novera 扩展到更广泛生产项目的真实成本仍不透明 | 吸引更看重控制和集成、而非交钥匙式托管访问的实验室 |
| Google Willow | 无公开价格信号保留 | 入围提案后,选定研究伙伴可早期访问 | 获批提案可访问顶尖硬件,加上 Cirq 生态 | 商业条款、容量和未来目录计划未知 | 竞争威胁来自战略和生态,而不是即时价格 |
| D-Wave Leap / Advantage2 | 保留的 D-Wave 页面没有简单公开费率卡 | 云访问、混合求解器和本地部署系统购买 | 退火系统、混合求解器,以及企业级正常运行时间 / 安全姿态 | 尽管系统销售被公开提及,确切商业标价仍不公开 | 替代路线,适合买方今天就需要运营化量子工作流 |
| 经纪云 / 混合 HPC 现状 | 通过超大规模云厂商或 GPU 消耗实现透明计量,但人员成本隐藏 | 云市场用量加内部工程时间 | 多供应商实验、仿真、混合工作流和推迟硬件承诺 | 工程负担和工作负载可移植性限制很少体现在标价里 | 为任何专用硬件供应商划出实际价格底线 |
这张表区分明确公开费率卡与合同驱动或未知经济性;由于多数量子硬件供应商仍私下谈判条款, 无证据单元格保留为未知或报价驱动,而不是强行标准化成虚假的横向可比。
[CP007, CP008, CP019, CP020, CP029, CP030]对比企业评估最看重的能力:访问、打包、工具和近期效用。
[CP002, CP007, CP010, CP016, CP019, CP023]3.3 切换成本、分发权力和伙伴访问
对“赢家通吃”护城河最强的反证,是访问层已经高度标准化。AWS Braket 把 Atom、IonQ 和 Rigetti 放在一个控制平面上;Azure Quantum 则发布 IonQ、Quantinuum、Rigetti 和 Pasqal 的伙伴专属产品。开放开发者工具强化了这个格局:Google 推 Cirq,Quantinuum 推 pytket, IonQ 支持主要 SDK,Rigetti 可从外部格式编译,NVIDIA 正推动一种混合量子 - 经典模型,把 QPU 视为更大加速系统的一部分。软件侧切换成本因此保持中等。 买家可以跑基准、多宿主,并推迟承诺。锁定效应仍存在,但主要是硬件特定和工作流特定,而不是应用层排他:中性原子模拟仿真、离子阱保真度、超导门速度、光子制造或退火式优化, 各自适配不同工作负载和控制假设。因此,分发权力向已经拥有云关系、企业支持模型或本地部署集成足迹的经纪商和既有厂商倾斜。Atom 的访问多样性帮助它留在候选名单上, 但尚未消除云市场鼓励的货比三家动态。[CP018, CP020, CP028, CP029, CP030, CP034]
3.4 护城河耐久性和反向证据
Atom 的护城河可信,但有条件。证据支撑最强的优势包括:已进入市场的真实中性原子产品、比多数非上市硬件初创公司更强的公开价格透明度、室温和 HPC 集成叙事, 以及明确绑定逻辑量子比特和可部署容错系统的路线图。这些特征很重要,因为许多竞争对手仍迫使买家在技术野心和具体访问之间二选一。但反向证据同样重要。 IEEE Spectrum 2026 年的框架是,行业首批纠错系统代表科学进展,而非广泛商业优势;因此,任何供应商——包括 Atom——都不应仅凭容错叙事就获得不受质疑的护城河。 中性原子也面临取舍:可扩展性和量子比特可移动性有吸引力,但操作速度慢于超导系统。与此同时,Atom 和 Pasqal 融资并扩展云或本地部署路径,直接同业正在变密; 更广泛的既有厂商也能在软件生态、渠道和客户教育上比 Atom 砸更多钱。因此,竞争结论是有利但不能自满:Atom 已足够差异化,值得关注,但其耐久性取决于能否把技术叙事转成对离子阱、 超导和云经纪替代方案的可重复胜利。[CP036, CP037, CP038, CP039, CP040, CP041]
| 护城河主张 | 主要威胁 | 严重性 | 当前证据 | 缓解 / 尽调追问 |
|---|---|---|---|---|
| 中性原子架构加室温 / HPC 适配 | Atom 和 Pasqal 等直接同行拥挤,再加上资本更足的离子阱和超导对手 | 高 | Atom 和 Pasqal 在扩展硬件、云访问和融资,IBM、IonQ 和 Quantinuum 的企业入口更宽 | 按工作负载索要当前相对 Atom、Pasqal、IonQ、Quantinuum 和 IBM 的胜率 |
| 公开 Braket 定价和多路线访问让 Atom 更容易被评估 | IBM 和 Azure 关联对手仍暴露更宽的计划阶梯或经纪比较界面 | 高 | IBM 发布完整阶梯,Azure 发布 IonQ/Quantinuum/Rigetti 计划,Braket 让多供应商比较保持简单 | 索要按渠道拆分的漏斗转化,以及公开透明度提高成交率的证据 |
| 本地部署和高级直接访问在部署后可提高切换成本 | IBM、IonQ、Quantinuum、Rigetti 和 D-Wave 等对手也销售专用或本地部署路线 | 中 | 专用部署正成为企业信任的基本门槛,不是 Atom 独有切入点 | 索取本地部署项目的参考客户、部署时间表和续约证据 |
| 逻辑量子比特路线图让 Atom 区别于纯实验叙事 | Google、IonQ、Quantinuum 以及整个品类都在提出相互竞争的 FTQC 主张,第三方质疑仍然很强 | 高 | Willow、IonQ 2026 年逻辑量子比特目标和 IEEE 的质疑,共同限制买方应计入多少路线图价值 | 要求里程碑标准独立,并绑定客户可用工作负载,而不只是内部路线图标签 |
| 可用工作负载能在完整 FTQC 前支撑溢价定价 | D-Wave 以及混合 AI/HPC 替代方案无需 Atom 专属锁定,也能满足近期优化或仿真需求 | 中 | Leap、混合求解器和 CUDA-Q 强化了延后专用硬件决策这一现状替代方案 | 对具体工作负载做基准测试,确认 Atom 是否在求解时间或结果质量上跑赢混合经典基线 |
风险等级衡量的是对 Atom 定价权和采购入围位置的冲击,不是量子计算这个品类失败的概率;本风险表聚焦最直接削弱 2026 年采购周期中差异化耐久度的威胁。
[CP031, CP032, CP033, CP037, CP039, CP041]一张紧凑评分卡,展示当前哪些竞争特质在帮助或削弱 Atom 的防御性。
取值基于留存公开证据综合得出的定性判断,证据涉及价格、渠道结构、技术路线权衡和商业化疑虑,而不是来自任何第三方竞争基准。
[CP030, CP034, CP035, CP036, CP041, CP042]3.5 图表
04财务情况
4.1 收入流、定价和公开牵引的现实
Atom 可见的变现触面真实存在,但公开披露仍很窄。公司及其伙伴生态明确围绕商业本地部署系统、具备逻辑量子比特能力的部署,以及对复杂政府、研究和企业买家的支持来定位 Atom。 QuNorth 是最干净的证据点,因为它是具名、客户背书、主权资助的系统项目,而不是泛泛的试点公告。但同一证据也凸显披露缺口:公开来源描述了买家、配置和 Magne 的战略重要性, 却没有披露 Atom 的确认收入、利润率结构或服务义务。其他地方也呈现同样模式。Microsoft 的商业机器公告证明 Atom 能卖进全栈打包方案,但没有说明该方案成本多少、 收入如何分成。诚实判断是,Atom 有收入触面和公开牵引,但缺少能让外部投资人区分一次性硬件收入与可重复平台经济性的运营披露。[CI001, CI002, CI003, CI004, CI005, CI012]
| 收入来源 | 机制 | 单位 | 当前价值 / 状态 | 质量 | 尽调问题 |
|---|---|---|---|---|---|
| 本地部署系统交付 | 旗舰硬件部署 | 系统销售 / 部署 | QuNorth 提供具名验证 | 战略价值高;公开可重复性证据弱 | 按项目索取管线和已签收入。 |
| 逻辑量子比特商业包 | 硬件叠加 Microsoft 软件栈 | 集成系统项目 | 商业包已公开发布 | 可能差异化,但经济模型不透明 | 澄清收入分成和支持模式。 |
| 政府 / 战略项目 | 政策和任务驱动型合作 | 项目或合同金额 | 在 DARPA / DoC 背景下可见 | 信号强,但本身不是直接收入 | 区分赠款、股权、合同和服务收入。 |
| 服务 / 集成支持 | 部署和技术支持 | 项目制工作 | 可能存在,但未披露 | 可能平滑收入,但未量化 | 索取服务组合和贡献毛利。 |
| 经常性软件 / 维护 | 可能存在,但未公开 | 订阅或支持 | null | 留存来源无法支撑 | 索取续约和维护指标。 |
值为 null 的行表示这些收入流在经济上可能存在,但留存公开证据无法支撑。
[CI001, CI002, CI003, CI004, CI034]| 价格 / 合同项目 | 标价与实际成交价 | 公开证据 | 影响 |
|---|---|---|---|
| AC1000 或本地部署系统价格 | Unknown | 未披露 | 无法清晰建模单位经济性。 |
| 集成 Microsoft 的逻辑量子比特套件 | Unknown | 商业报价已披露;价格未披露 | 经济模型可能取决于伙伴分成。 |
| QuNorth 系统合同给 Atom 带来的金额 | Unknown | 客户资金只披露到计划层面 | 客户验证不等于已确认收入。 |
| 支持 / 维护条款 | Unknown | 未披露 | 可能显著影响毛利和经常性。 |
留存来源披露了产品存在和客户项目,但未披露实际成交价格或标准合同条款。
[CI005, CI012, CI031]Atom 的变现看起来更多来自旗舰系统部署和战略项目,而不是广泛的公共云经常性使用。
[CI001, CI002, CI003, CI012, CI031]4.2 GTM 动作和单位经济代理指标
公开 GTM 动作看起来更偏咨询式和基础设施重,而不是软件式。Atom 最强证据来自战略项目——QuNorth、Microsoft、DARPA 和公共部门生态需求—— 而不是披露的客户队列或大规模云使用量。这意味着 CAC、回本周期和 NRR 等经典 SaaS 指标不只是缺失;它们也可能不是公司当前阶段的第一视角。 更好的代理视角是项目集中度和渠道质量。按这个标准,证据喜忧参半。QuNorth 是强旗舰,Department of Commerce 和 DARPA 背景暗示异常高的战略相关性。 但 Atom 尚未证明旗舰项目会转化为多元、可重复的商业合同基础。IonQ、Rigetti 和 D-Wave 的公开量子文件也强化了这一点的重要性:即便披露更多且有真实收入的公司, 仍可能承受重亏损、融资需求和不均衡收入结构。因此,Atom 的 GTM 故事足够可信,仍具可投资性;但仅凭公开数据,还不足以被评为高效或耐久。[CI016, CI017, CI018, CI019, CI020, CI021]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调问题 |
|---|---|---|---|---|
| 毛利率 | null | 高 | 判断硬件部署能否带来可持续贡献利润 | 按硬件、软件和服务索取毛利率拆桥。 |
| 烧钱速度 | null | 高 | 决定融资依赖和现金跑道 | 按路线图阶段索取月度烧钱和偏差。 |
| 现金跑道(月) | null | 高 | 是判断融资时点的核心 | 索取基准和冲刺里程碑下的现金跑道。 |
| 客户集中度 | 高可能性 | 中 | 旗舰部署可能夸大可复制性 | 按头部客户和管线阶段索取收入。 |
| 经常性收入占比 | Unknown | 高 | 区分平台经济性和一次性硬件销售 | 索取维护、软件和服务拆分。 |
| 销售周期 / CAC | Unknown | 高 | 咨询式深科技销售可能慢且昂贵 | 索取平均周期长度和技术销售人员数。 |
多数单位经济性指标仍未公开;这张表列出关键问题,而不是假装公开数据很精确。
[CI017, CI018, CI019, CI020, CI025, CI029]公开证据在已融资金额上最强,在运营指标上最弱,因此分析区间不得不很宽。
[CI004, CI010, CI017, CI025, CI029, CI030]Atom 财务模型最大缺口在于从旗舰部署到毛利、支持和经常性经济性的链条。
[CI003, CI005, CI012, CI017, CI025, CI031]4.3 资本充足性、成本结构和融资依赖
资本可见度是 Atom 财务画像中最强的一块。2026 年 6 月公告给出清晰标题——融资超过 $300 million——并点名 $100 million Series C 和拟定的 $100 million Department of Commerce 支持。DCVC 的评论把联邦部分描述为股权,进一步强化其重要性;Forge 2026 年 7 月 D-1 市场数据则暗示 Atom 能以数十亿美元估值获得额外资本。与此同时,披露的募资用途清楚表明 Atom 仍是重投入项目。资金将投向更高量子比特数量的系统、更好保真度、 控制系统、纠错、全球部署和团队增长。DARPA Stage B 又补充证据,说明路线图需要持续原型开发和风险压降支出。由于公开来源没有披露现金、烧钱速度或现金跑道, 正确立场是:Atom 当前融资状况好于许多非上市同业,但仍依赖融资。它的资本证据足以降低短期偿付担忧,但不足以排除在经济性变得广泛且可重复前继续融资的可能。[CI006, CI007, CI008, CI009, CI010, CI011]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调问题 |
|---|---|---|---|---|
| 已宣布累计资本 | >$300M | 高 | 表明资产负债表有实质支撑 | 与股权结构表和交割日期核对。 |
| Series C 轮金额 | $100M | 高 | 确认 2026 年有新增私募资本 | 获取完整条款和投资人权利。 |
| 美国商务部支持 | $100M 计划 / LOI | 高 | 降低战略风险并对齐政策 | 确认交割条件和拨款时点。 |
| 最新估值证据 | $2.06B Series D-1 轮 | 中 | 关键后期定价信号 | 获取条款清单、投资人名单和优先权。 |
| 准确现金余额 | null | 高 | 评估现金跑道所必需 | 索取最新现金和受限现金。 |
| 下一轮触发条件 | Unknown | 高 | 决定稀释风险 | 索取按里程碑制定的资本计划。 |
Forge 增加了有用的定价背景,但直接判断投资价值仍需私人融资文件。
[CI006, CI007, CI008, CI009, CI010, CI025]在公开经济性尚未广泛成熟前,Atom 新资金要覆盖多条硬件密集型支出线。
[CI006, CI008, CI013, CI014, CI015, CI026]4.4 财务结论和尽调阻塞项
平衡结论是,Atom 看起来在商业上认真、在战略上有融资能力,但根本上仍披露不足。正向案例有分量:公司已经融资,有客户,旗舰部署真实存在,公共部门支持降低了眼前的生存风险。 警示同样核心。公开来源仍没有给投资人一个干净视图,说明收入组合、经常性与一次性经济性、毛利率、现金跑道或客户集中度。因此,常规自下而上的基本面模型还无法用公开证据搭建。 尽调路径应聚焦直接向管理层索取:收入桥、在手订单、递延收入、利润率拆解、烧钱速度、现金跑道、支持经济性,以及 Department of Commerce 和 D-1 结构的法律细节。 在取得这些材料前,Atom 的财务章节支撑的结论是战略动能强,但基本面透明度明显不完整。[CI021, CI025, CI026, CI027, CI028, CI029]
4.5 图表
05产品与技术
5.1 以客户工作流定义产品
Atom 的公开产品触面最好理解为系统与栈的组合,而不是单一 SKU。官网聚焦 AC1000 和“逻辑量子比特时代”;融资新闻稿称公司正在为企业和政府客户部署商业本地系统。 这已经暗示至少三层产品:核心中性原子硬件、通过 Microsoft 软件栈启用逻辑量子比特,以及为需要安全本地访问的客户提供高接触部署或支持。QuNorth 的 Magne 项目把这一点具体化。客户和独立来源描述的是一个全栈系统:Atom 提供中性原子硬件,Microsoft 提供软件、中间件和云连接。因此,工作流证据更少指向低摩擦云消费, 更多指向面向化学、材料、邻近药物发现和优化工作负载的定制部署。这对承保很重要,因为商业触面有价值,但必然服务较重:Atom 卖的是强大量子系统, 以及让复杂买家真正用起来的集成工作。[CE001, CE002, CE003, CE017, CE018, CE019]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| AC1000 / 旗舰系统 | 政府、企业和研究买方 | 公开定位活跃 | 将中性原子规模与逻辑量子比特叙事结合 | 需要更完整的技术数据表和价格披露。 |
| 中性原子硬件核心 | 量子工程师和终端客户 | 已运行 | 大型阵列和由原子移动带来的连接性 | 需要制造和现场服务细节。 |
| Microsoft 逻辑量子比特栈 | 高阶技术用户 | 商业包已发布 | 加入量子比特虚拟化、AI 和 Azure 集成 | 需要商业条款和支持边界。 |
| QuNorth / Magne 部署 | 北欧研究和产业用户 | 建设中 / 旗舰部署 | 50 个逻辑量子比特,并有主权背景支持访问 | 需要准确交付里程碑和验收标准。 |
| Nu Quantum 集成工作 | 未来系统集成商 | 早期战略合作 | 将叙事延伸到网络 / 实用规模栈 | 需要具体产品路线图和客户用例。 |
状态仅反映留存公开证据;公开材料清楚展示旗舰模块,但没有完整 SKU 目录。
[CE001, CE002, CE017, CE018, CE021, CE033]| 用户任务 | 现有工作流 | Atom 方案 | 已衡量 / 声称收益 | 限制 |
|---|---|---|---|---|
| 量子化学研究 | 经典仿真叠加量子探索 | 通过 QuNorth / Microsoft 使用逻辑量子比特系统 | 更高端科学工作流和未来数据集生成 | 目前没有公开 ROI 指标。 |
| 材料发现 | HPC 和建模团队 | 搭载 Microsoft 栈的本地部署量子硬件 | 更强的混合科学计算路径 | 仍高度依赖旗舰项目。 |
| 药物发现相邻研究 | 学术和企业 R&D | 支持逻辑量子比特的系统叠加 AI/HPC 集成 | 支持深度技术研究和算法开发 | Atom 单独没有公开生产案例研究。 |
| 优化和供应链实验 | 产业创新团队 | 通过客户或伙伴项目获得全栈系统 | 探索复杂优化问题 | 商业可复制性尚未验证。 |
工作流证据更能支撑目标用例,对可衡量业务结果或定价支撑较弱。
[CE019, CE020, CE025, CE034]Atom 目前的产品路径从旗舰硬件出发,接上合作伙伴软件,再以高接触部署进入高级工作负载。
[CE002, CE017, CE018, CE019, CE020]5.2 架构和运营模型
已采纳技术证据显示出一致的架构逻辑。Atom 的量子比特是光学陷阱中的单个中性原子,并可按需移动以产生相互作用。2023 年公告通过 1,225 位点阵列、约 1,180 个量子比特,确立了物理层规模,并同时提出长相干和中途测量主张。Microsoft 2024 年公告及后续 arXiv 论文随后把分析从物理规模转向可靠逻辑计算。 两者合起来描述了 24 个纠缠逻辑量子比特、在 28 个逻辑量子比特上计算、一个 256 量子比特中性原子处理器、通过原子移动实现全连接,以及把部分错误来源转化为可检测原子损失的擦除转换技术。 实际上,这意味着 Atom 的运营模型不只是硬件制造,而是硬件加上解码器感知软件、纠错编排和 AI/HPC 集成。这个组合差异化明显,因为它给 Atom 提供了一条从大型物理阵列走向有用逻辑计算的公开路径;但也意味着最弱的一层——硬件保真度、解码器质量或编排——都可能成为扩展瓶颈。[CE004, CE005, CE006, CE007, CE008, CE009]
| 层 / 组件 | 角色 | 依赖 | 风险 |
|---|---|---|---|
| 中性原子量子比特 | 物理计算基底 | 激光俘获和原子操控 | 保真度和可复现性。 |
| 大型原子阵列 | 扩大物理量子比特数量 | 控制和校准系统 | 规模化后的运营漂移。 |
| 原子移动 / 连接性 | 支持相互作用和逻辑操作 | 精准运动控制 | 错误累积和编排复杂度。 |
| 错误检测 / 擦除转换 | 提升逻辑性能 | 成像、解码器逻辑、软件 | 损耗处理和解码器质量。 |
| Microsoft 量子比特虚拟化 | 用物理量子比特创建逻辑量子比特 | 合作伙伴软件栈 | 伙伴依赖。 |
| AI / HPC 集成 | 运行混合工作流和科学发现闭环 | Azure Elements 和经典计算 | 经济和运营复杂度。 |
架构的差异化来自硬件规模与软件主导的逻辑量子比特控制之间的互动,而不是任何单一层本身。
[CE004, CE005, CE007, CE010, CE011, CE013]Atom 的产品价值取决于多个层协同:从中性原子硬件,到逻辑量子位软件,再到混合计算集成。
[CE004, CE013, CE014, CE025]Atom 要放大规模,需要硬件保真度、解码器质量、伙伴软件和客户部署就绪度同步推进。
[CE007, CE011, CE013, CE016, CE030]5.3 部署、路线图和差异化
Atom 今天最强的部署证据来自本地部署和伙伴主导项目。QuNorth 提供最清晰的旗舰项目,DARPA Stage B 则是判断路线图是否超越营销的最清晰外部基准。 NERSC 的中性原子征集也显示,高性能计算买家现在就在围绕这类系统建设,而不是只押注遥远未来。相对同业,Atom 的公开差异化不只是“更多量子比特”。 IBM 和 Google 强调路线图和可通过云访问的超导系统,Quantinuum 强调离子阱硬件和自有全栈平台,QuEra 强调模拟和中性原子访问。已采纳来源显示,Atom 的优势在于物理规模叠加 Microsoft 支持的逻辑量子比特商业化。这是很强的差异点,但不是产品成熟的保证。许多最重要的当前证据仍落在高接触项目中, 而不是广泛自助使用;因此,正确解读是 Atom 拥有通向实用性的差异化路径,而不是实用性已经被广泛产品化。[CE015, CE016, CE017, CE018, CE021, CE022]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 影响 | 来源 |
|---|---|---|---|---|
| 2023-10 | 1,225 站点 / ~1,180 量子比特系统 | 已宣布 | 确立物理规模领先主张 | PR Newswire / Forbes |
| 2024-11 | 24 个纠缠逻辑量子比特 / 28 逻辑量子比特计算 | 已宣布 | 将叙事推进到可靠逻辑计算 | Microsoft / arXiv |
| 2025-05 | DARPA Stage B 入选 | 进行中项目 | 外部路线图审查 | PR Newswire / DARPA |
| 2025-07 | QuNorth / Magne 系统 | 客户项目 | 旗舰本地部署验证 | Novo / QCR / TQI |
| 2026-06 | Nu Quantum 合作 | 战略延伸 | 显示更广的系统栈野心 | Atom 主页 |
路线图最强的证据在里程碑和伙伴关系;更大的缺口仍是商业运营指标。
[CE005, CE008, CE015, CE017, CE021]Atom 的成熟度在物理规模和旗舰部署证据上最高,在公开企业运营披露上最低。
[CE017, CE020, CE027, CE028, CE030]5.4 信任、质量、合规和未解缺口
技术证据令人印象深刻,但企业保障证据相对薄。已采纳公开来源大量讨论量子比特数量、逻辑量子比特里程碑和战略部署,却很少涉及正式可用时间、事件响应、定价透明度, 或 SOC 2、ISO 27001 等认证。这并不意味着存在隐藏缺陷;它只说明买家和投资人不应仅凭物理进展推断运营成熟度。反向解读很直接:如果支持流程、 安全控制和可重复部署经济性尚未公开,一家公司可以技术领先,同时运营仍处早期。关于量子时间线的独立评论也强化了这一谨慎态度。因此,Atom 的产品和技术姿态作为技术平台具备可投资性, 作为商业化故事也有前景;但在把该产品视为企业级标准基础设施前,直接尽调仍应索取可靠性指标、支持条款和部署手册。[CE027, CE028, CE029, CE030, CE036, CE037]
5.5 图表
06客户情况
6.1 按买家、用户、付款方和用例细分客户基础
Atom 可见客户基础数量不广,但类型相当清楚。最强具名证据来自一个主权背书倡议:QuNorth,由 EIFO 和 Novo Nordisk Foundation 资助, 预期用户是北欧研究人员和产业界,核心用例是化学、材料、生物学和优化。政府和公共研究渠道与企业渠道同样重要。DARPA 的 Stage B 角色让美国政府成为重要路线图客户或赞助方; NERSC 和更广泛的国家量子倡议也显示,公共实验室和政策背书机构正在围绕中性原子系统建设工作流。企业需求以叙事形式存在——材料科学、制药、能源、物流和工业优化—— 但 QuNorth 生态之外的公开具名生产用户仍很稀疏。因此,正确的细分结论是,Atom 当前最强的是主权、研究和高接触技术买家;企业垂直仍通过伙伴主导和旗舰项目渠道浮现。 围绕其他量子供应商的公开市场和私募市场报道也表明,买家仍在评估少数高知名度平台,而不是几十家可互换供应商,这提高了每个 Atom 旗舰参考案例的战略重要性。[CU001, CU002, CU003, CU004, CU005, CU006]
| 客群 | 买方 / 用户 / 付款方 | 用例 | 规模 / 战略价值 | 缺口 |
|---|---|---|---|---|
| 主权 / 北欧计划 | 买方:QuNorth;用户:北欧研究人员和产业界;付款方:EIFO + Novo | 面向化学、材料、生物、优化的逻辑量子比特计算 | 最高等级的公开客户证据 | Atom 未披露收入。 |
| 美国政府项目 | 买方 / 赞助方:DARPA | 路线图验证和实用规模基准测试 | 战略重要性高 | 本身不是直接产品收入。 |
| 国家实验室 / HPC 买家 | 用户:研究团队和平台运营方 | 中性原子工作流开发 | 重要的近期需求信号 | Atom 具体合同数量未公开。 |
| 企业科研买家 | 用户:制药、材料、能源研发团队 | 仿真和发现工作流 | 长期上行空间大 | 具名生产买家稀少。 |
| 优化 / 工业用户 | 用户:先进工业团队 | 路由、规划和优化 | 潜在扩张路径 | 公开证据仍薄。 |
分群依据留存的具名证据和政策需求信号,而不是已披露客户名单。
[CU001, CU004, CU005, CU007, CU011, CU012]Atom 最清晰的买方路径,是从战略资格筛选走到旗舰部署,再进入更广的生态访问。
[CU007, CU021, CU023]具名实证最强的是旗舰部署质量,最弱的是广度和留存披露。
[CU001, CU007, CU013, CU015, CU030]6.2 采用轨迹和具名客户证据
公开采用轨迹更多由具体里程碑衡量,而不是客户数量。QuNorth 是关键事件,因为它把 Atom 的产品故事转成一个具名商业本地部署,带有客户资金、逻辑量子比特配置和明确时间安排。 Magne 系统预计在 2027 年初前后投入运行,这意味着当前证据在承诺和配置上很强,但在实际使用上仍有一部分是未来时。QuNorth 之外,最强的邻近证据来自政策和项目对齐: DARPA QBI、NERSC 的中性原子征集,以及支持量子能力建设的政策框架。这些证据有价值,因为它们展示需求路径,但它们不等同于一份已披露的付费生产账户名单。 已采纳公开来源也没有给出总装机基础、客户队列或多个企业账户的部署节奏。因此,采用路径应被解读为真实但集中:Atom 已跨过从理论到旗舰客户证据的门槛, 但尚未在公开层面跨过广泛组合可见度的门槛。这也意味着下一次更新应专门测试 Magne 是否已从承诺推进到可衡量用户活动和重复工作负载需求。[CU001, CU002, CU007, CU008, CU014, CU015]
| 指标 | 值 / 状态 | 日期 | 来源 | 置信度 | 含义 |
|---|---|---|---|---|---|
| 首个商业本地部署订单 | QuNorth / Magne | 2025-07-17 | Atom + Novo + QCR | 高 | 商业化已跨过门槛,证据明确。 |
| Magne 投运时间 | 2027 年初 / 2026-27 新年 | 2025-07-17 | TQI / DCD / Novo | 高 | 部分使用证据仍是未来时。 |
| 客户系统中的逻辑量子比特 | 50 个逻辑量子比特 | 2025-07-17 | QCR / Novo | 高 | 客户系统定位在先进能力层级。 |
| 客户系统中的物理量子比特 | 1,225 个物理中性原子量子比特 | 2025-07-17 | QCR | 高 | 展示旗舰部署的配置规模。 |
| 公开客户总数 | null | 2026-07-18 | 留存公开来源 | 高 | 客户广度仍未披露。 |
具名客户证据强;广泛采用指标仍属私有信息。
[CU001, CU002, CU003, CU014, CU028, CU030]| 客户 / 利益相关方 | 分群 | 部署 / 用例 | 生产 / 试点 | 结果 / 局限 |
|---|---|---|---|---|
| QuNorth | 主权支持的计划 | 面向北欧学界和产业界的 Magne 逻辑量子比特系统 | 面向生产的旗舰部署 | 最强公开证据;使用仍在爬坡。 |
| DARPA | 政府项目 | 实用规模路线图基准 | 项目制 / 验证 | 重要可信度信号,不是标准客户账户。 |
| NERSC 生态 | 国家实验室 / HPC | 中性原子工作流和提案活动 | 评估 / 工作流阶段 | 显示买家兴趣,不是已披露的 Atom 合同。 |
| Microsoft 生态 | 渠道 / 商业化合作伙伴 | 全栈逻辑量子比特系统交付 | 经合作伙伴进入生产的路径 | 关键渠道,但本身不是终端客户。 |
本表把真正终端客户证据,与验证渠道、商业化合作伙伴分开。
[CU001, CU007, CU008, CU019, CU034]公开证据从广泛市场需求收窄到少数具名旗舰部署。
[CU011, CU013, CU021, CU030]6.3 耐久性、扩张和渠道依赖
客户章节在耐久性上明显变弱。公开来源没有披露 NRR、GRR、流失、续约、合同期限或客户集中度,分析师只能依赖代理指标。最佳代理指标是旗舰本身的质量: QuNorth 不是模糊试点,Microsoft-Atom 系统也明确面向学术界和产业界。即便如此,一个旗舰并不是多元装机基础。渠道也适用同样逻辑。Microsoft 是重要商业化伙伴, 政府相关项目是重要验证渠道,但这些优势也会带来依赖。如果客户管线过度依赖主权或伙伴塑形需求,而重复企业转化仍然缓慢,它可能看起来比实际更强。合理乐观情景是从锚点部署向相邻化学、 材料和优化客户落地后扩张。合理悲观情景是集中度:少数战略项目占据注意力,而企业采用比预期慢得多。今天的公开证据支持前者作为一种可能, 但还不是已经确立的模式。另一个警示是竞争性买家教育:Google 和 Quantinuum 已经可以把客户指向可见项目和技术栈文档,因此 Atom 不仅要靠物理赢, 还要让买家相信自己确实能搭起可重复工作流。具名参考、部署后案例研究和续约证据会在下一次更新中特别重要。[CU015, CU016, CU017, CU018, CU019, CU020]
| 指标 | 值 / 状态 | 分群 | 置信度 | 尽调要求 |
|---|---|---|---|---|
| NRR | null | 全部 | 高 | 索取账户级留存数据。 |
| GRR | null | 全部 | 高 | 索取续约和降购数据。 |
| 合同期限 | null | 旗舰部署 | 高 | 索取期限安排和服务义务。 |
| 重复使用 | 仅代理指标 | 旗舰项目和计划型买家 | 中 | 跟踪 Magne、DARPA 或 HPC 项目是否扩大范围。 |
| 客户背书质量 | QuNorth 较高,其他较低 | 具名账户 | 中 | 索取更多直接客户背书。 |
留存仍是 Atom 客户故事中最大的公开信息缺口之一。
[CU015, CU017, CU021, CU032]| 扩张驱动因素 | 集中度风险 | 影响 | 尽调路径 |
|---|---|---|---|
| 借 QuNorth 落地后扩张 | 旗舰项目主导外界认知 | 高 | 索取 Magne 之外的主权客户管线。 |
| Microsoft 渠道撬动 | 合作伙伴依赖 | 高 | 澄清管线归属和排他性。 |
| 政府支持需求 | 采购摩擦和政策依赖 | 中高 | 审查从项目参与转化为付费部署的情况。 |
| 化学 / 材料工作负载 | 用例可信度仍在形成 | 中 | 索取具名设计伙伴和转化证据。 |
| 优化 / 工业工作负载 | 商业紧迫性不确定 | 中 | 索取 ROI 和基准研究。 |
扩张上行空间真实存在,但当前公开证据仍把 Atom 的客户证明集中在少数战略渠道。
[CU018, CU019, CU020, CU025, CU026, CU035]Atom 仍需证明的关键转变,是从旗舰部署走向可重复的多账户需求。
[CU017, CU021, CU025, CU031, CU032]6.4 图表
07风险
7.1 按严重性排序的风险概览
Atom 的最高等级风险不是已知丑闻,也不是某一场监管冲突;真正的风险在于,这条雄心很大的多代容错路线图,可能比当前公开叙事暗示的来得更晚、成本更高,或更难商业化复制。公司手里有可信资产:真实云访问、点名的 AIST 合同、DARPA 阶段门、以及很深的 NVIDIA/AWS 生态。但这些资产也把下行风险集中起来。Libra 一旦延误,gigaquop 里程碑如果后移,或者公共项目客户验证没能扩展成可复制企业需求,收入集中和融资压力会很快恶化。阅读本章应把它看成一个耦合系统。出口管制会拖慢招聘和合作;政府客户安全要求会拖慢采购;AWS 和 NVIDIA 依赖会压缩利润率或排期弹性;客户集中会让每一次项目延误都比多元化软件公司更痛。因此,热力图把路线图压缩、客户集中、以及合规与依赖之间的相互作用,放在普通创业噪音之上。[CR001, CR003, CR006, CR011, CR016, CR020]
| 排名 | 风险 | 可能性 | 影响 | 缓释成熟度 | 剩余敞口 | 投资含义 |
|---|---|---|---|---|---|---|
| 1 | 从 Libra 到 gigaquop 计划,容错路线图延期或表现不达标 | 高 | 严重 | 中 | 高 | 在容错交付日期和逻辑系统性能反复兑现之前,把 Atom 视为里程碑驱动型投资。 |
| 2 | 客户集中在政府、主权和 HPC 灯塔账户 | 高 | 高 | 中低 | 高 | 在当前公开客户集合之外出现具名非项目型续约或第二套付费系统之前,不要把可重复企业需求写入投资模型。 |
| 3 | 量子技术出口管制和承包商合规负担 | 高 | 高 | 中 | 高 | 除非管理层能拿出干净的合规资料包,否则模型中应纳入招聘放慢、跨境交易放慢和采购摩擦上升。 |
| 4 | 分发、经济性和解码对 AWS 与 NVIDIA 平台的依赖 | 中高 | 高 | 中 | 中高 | 如果某个战略平台伙伴调整优先级、延迟或更改商业条款,应假设进度或利润率下行。 |
| 5 | 融资不透明和持续资本密集 | 高 | 高 | 中 | 高 | 没有现金跑道桥接安排、烧钱披露和下一轮下行情景测算,不要假设当前资本基础足够。 |
| 6 | 对 DARPA、NQCC、NERSC 和 AIST 的项目阶段依赖 | 中高 | 高 | 中 | 中高 | 把公共项目胜出当作可信度信号,而不是广泛商业多元化的证据。 |
| 7 | 专业 QEC 和部署人才短缺 | 高 | 中高 | 中低 | 中高 | 要求管理层展示 QEC、平台工程和现场支持的留才与招聘深度。 |
| 8 | 法务 / IP / 安全认证可见度缺口 | 中 | 中高 | 低 | 中高 | 把没有公开争议或认证这件事视为开放尽调问题,而不是风险已解决的证据。 |
严重性排序综合截至 2026-07-05 的留存公开记录,并把路线图压缩、集中度和合规负担权重放得高于叙事动能。
[CR001, CR006, CR011, CR016, CR029, CR035]剩余风险视角显示,Atom 最重的风险敞口集中在路线图压缩、集中度和合规牵连,而不是已知诉讼或召回。
[CR009, CR011, CR016, CR030, CR035, CR036]7.2 监管、法律和主权风险
Atom 面临的法律风险,首先是政策边界问题,而不是已经摆到台面的诉讼问题。美国出口管制如今覆盖量子计算机、组件、软件和技术,已发布规则明确考虑部分外籍人员访问的视同出口和再出口可见性。这一点很关键,因为 Atom 正是依赖国际研究人才、跨境合作者和主权客户项目的公司。联邦和关键基础设施采购的重要性,也高于普通创业公司法律模板所暗示的水平。OMB、CISA、NIST、白宫以及后续法律评论都指向同一个方向:后量子迁移和承包商网络安全义务,正在成为联邦生态的运营要求,而不是抽象的长期主题。英国则通过 NQCC 和 National Security and Investment Act 框架加入主权视角。与此同时,公开法律案卷和专利检索工具虽然存在,但保留的公开证据仍无法给出经律师确认的诉讼或自由实施结论。因此,投资人应把法律风险按“合规执行 + 尽调不完整”来承销,而不是把它当成健康证明。[CR016, CR017, CR018, CR019, CR020, CR021]
| 规则 / 事项 | 管辖区 | 当前状态 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| 量子出口管制和视同出口报告 | 美国 / 全球 | 量子物项自 2024-09-06 起生效;视同出口报告 / 通用许可证框架已发布 | 高 | 高 | 行业规则公开,法律评论也已成熟,但 Atom 的具体分类历史未公开 | 高 | 获取产品 ECCNs、技术管制计划、年度视同出口报告,以及任何未决或被拒的许可证申请。 |
| 联邦承包商网络安全和 PQC 迁移义务 | 美国联邦 / 关键基础设施 | NSM-10、OMB M-23-02、NIST、CISA、白宫和法律评论都显示,承包商要求在收紧 | 中高 | 高 | 公开标准存在,但 Atom 专属证明不可见 | 中高 | 索取 NIST 800-171 对齐情况、漏洞披露政策,以及任何政府客户安全问卷或 SSP。 |
| 敏感量子交易中的主权与外国投资审查 | 英国和跨境交易 | NSI 指引及 BIS/CFIUS 背景显示,敏感技术交易可能被审查 | 中 | 中高 | 框架可见;具体交易的审查历史不可见 | 中 | 审阅关于客户、投资者和收购情景的英美律师备忘录。 |
| 诉讼、专利清查与自由实施可见度 | 美国及其他适用司法辖区 | 公开案卷和专利工具存在,但本章未留存已厘清的 Atom 特定争议结果 | 中低 | 中高 | 公开检索工具降低搜索摩擦,但不能替代律师审查 | 中 | 用每个实体名称变体检索 PACER、州法院、ITC、PTAB 和专利转让记录;审查 FTO 意见。 |
| 国家安全系统的量子安全营销限制 | 美国国家安全买家 | NSA 明确警告,除非克服限制,否则不要依赖 QKD/QC 保护 NSS | 中低 | 中 | 指引公开;是否适用于 Atom 提案取决于具体表述 | 中低 | 审查提案措辞、方案说明和任何联邦安全主张是否夸大。 |
各行按公开法律和监管风险在剩余投资判断中的重要性排序,而不是按新颖度排序。本表是公开记录登记,不是律师确认的合规备忘录。
[CR016, CR017, CR018, CR019, CR020, CR021]7.3 运营、伙伴和客户集中叠加
从运营看,Atom 已经比许多量子创业公司更像一家真实公司,因为它不只是承诺未来机器;它已经在 Braket 上支持 Aquila,为 Libra 云交付做准备,推进公共项目部署,并在容错硬件到来之前共设计应用。这份广度本身也是核心风险。云分发把访问和经济性绑定到 AWS。解码器和混合工作流性能要经过 NVIDIA 及周边经典基础设施。公共验证经过 DARPA,客户验证则高度依赖 AIST、NERSC 和 NQCC。每个节点都有价值,但没有一个冗余。AIST 部署证明客户确实愿意购买本地系统;NERSC 和 NQCC 证明机构相关性;AWS 证明分发可以触达用户。但同一组证据也揭示集中:大多数具名验证来自政府、主权或 HPC 组织,这些买方能接受长周期和政策目标。Atom 一旦错过里程碑,损害会快速传导,因为客户验证、伙伴信心和路线图可信度都压在少数交易对手和用例上。传导和依赖图关注的正是这种相关性,而不是孤立的技术轶事。[CR006, CR008, CR009, CR010, CR011, CR012]
| 失效模式 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 未解决缺口 |
|---|---|---|---|---|---|
| Libra 与 gigaquop 路线图压缩 | 高 | 严重 | 中 | 高 | 公开目标要求多个世代的里程碑连续成功。 |
| 解码器和经典集成瓶颈 | 中高 | 高 | 中 | 中高 | QEC 吞吐依赖 NVIDIA 侧加速,以及 QPU 之外的软件协同设计。 |
| 云端加本地部署的现场支持负担 | 中高 | 高 | 中 | 中高 | Atom 必须同时支持 Braket 接入、高级访问,以及主权或 HPC 安装。 |
| 公开安全和正常运行时间保障仍薄 | 中 | 中高 | 低 | 中高 | 留存来源宣传安全访问,但未提供第三方保障材料或公开 SLA 证据。 |
| 政府项目采购准备资料包 | 中 | 中高 | 中低 | 中高 | 承包商合规要求收紧的速度,快于 Atom 公开保障材料的扩展。 |
运营行把路线图、集成、服务和安全保障失效模式拆开,因此 Atom 必须逐项证明缓释,而不能只靠泛泛技术故事。
[CR006, CR008, CR010, CR027, CR028, CR029]| 依赖项 | 交易对手 | 角色 | 集中度 | 失效情景 | 严重性 | 缓释 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| 云访问与分发 | Amazon Web Services / Braket | 主要公共访问渠道,也可能承载未来 Libra | 高 | Braket 的定价、优先级或时间表只要变化,就会影响访问成本和公开 GTM 可信度 | 高 | 既有分销关系深,路线图承诺也可见 | 高 |
| 混合 QEC 与解码器加速 | NVIDIA | 解码器、校准和加速超算合作伙伴 | 高 | 经典计算延迟或集成拖期,会拖慢容错进展和混合用例落地 | 高 | 合作仍在深度推进,且与装机基础高度相关 | 中高 |
| 技术验证与分阶段支持 | DARPA QBI | 政府验证方和资金 / 支持节点 | 中高 | 无法推进或里程碑变化,会同时削弱可信度和项目支持 | 高 | 分阶段外部验证已经到位 | 中高 |
| 标杆部署与主权参考客户 | AIST / ABCI-Q | 最大的公开具名本地部署合同,也是日本参考站点 | 高 | 延期、表现不达标或无法复用,会同时伤害收入证明和国际可信度 | 高 | 已落地的大型项目,且贴近 GPU 超算场景 | 高 |
| 公共项目渠道与英国存在感 | NQCC | 测试床合同和英国主权能力合作伙伴 | 中 | 项目延期或主权摩擦会拖慢英国侧验证和部署叙事 | 中高 | 已承诺实物交付 | 中 |
| 科研 HPC 访问与用户培育 | NERSC | 面向 Aquila 和 Gemini 的美国国家实验室式访问渠道 | 中 | 科研使用未能转化为持久的付费生产需求 | 中 | 项目扩大用户访问面,并强化反馈闭环 | 中 |
| 资本与战略背书 | Google、SoftBank Vision Fund 2、NVentures、Valor、QVT、Safar 等 | 资本、信号价值和生态入口 | 中 | 收入还没打开前,后续融资条款或战略优先级可能变化 | 高 | 投资人组合质量高,且近期融资已关闭 | 中高 |
依赖风险集中在少数平台、公共项目和战略资助方身上;这些节点会同时影响技术可信度和商业证明。
[CR008, CR009, CR010, CR011, CR013, CR014]主要下行链条是路线图或合规摩擦拖慢客户实证,进而带来融资压力和估值压缩。
[CR017, CR019, CR029, CR036, CR037, CR040]Atom 公开可见的执行路径,靠少数渠道、计算合作伙伴、公共项目和主权标杆账户支撑。
[CR008, CR010, CR011, CR013, CR036, CR041]7.4 财务、人才和投资逻辑破裂标准
财务风险和执行风险仍紧密绑在一起。2025 年融资显著改善了 Atom 的位置,但公开证据仍没有显示烧钱速度、现金、现金跑道,或 AIST 及其他具名项目背后的当前收入集中度。这意味着投资人无法干净地把技术进展和融资风险切开。商业环境也没有量子热潮顶点叙事暗示的那么宽容。Atom 自己的 2026 年市场报告描述了验证驱动的采购、许多买方预算持平、主权过滤器,以及由量子纠错牵头的专业人才短缺。Atom 想从公开技术领先走向可扩展的企业采用时,卡住它的正是这些摩擦。人才风险因此和硬件风险一样重要:Atom 必须招聘并留住稀缺的 QEC、平台和现场支持人才,同时管理更多政府、主权和伙伴接口。缓释逻辑仍然可信——DARPA 验证、强伙伴、重要公共项目,以及超过 $230 million 的近期资本——但这些缓释都只是部分有效。投资逻辑破裂点不只是一次实验失手;路线图延误、复购客户失败、惩罚性融资、合规材料缺失任意叠加,都可能把一家技术可信的公司变成永远往后推的商业化故事。[CR001, CR002, CR005, CR015, CR030, CR031]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释因素 | 尽调路径 |
|---|---|---|---|---|---|
| 商业化与合作伙伴前线领导力 | 公开叙事仍靠少数高管管理云、主权和政府关系 | 中 | 高 | 最近融资和项目胜利说明领导层在运转,但管理梯队深度并未完全公开 | 索取继任计划、授权分工,以及每个主要合作伙伴或政府账户的负责人。 |
| QEC 与解码器人才 | 容错推进需要稀缺的量子纠错和系统人才 | 高 | 高 | 资本和生态合作能缓解压力,但公开市场仍认为专业人才短缺 | 索取 QEC 和系统岗位的留存、关键岗位空缺、招聘周期数据。 |
| 现场部署与支持运营 | 本地部署和高级访问需要服务、QA 和升级处理流程,不能只靠研发厚度支撑 | 中高 | 高 | AIST 与云端运营证明公司已有部分能力,但公开服务指标有限 | 审阅现场工程、客户成功、事故响应和质保归属的组织架构。 |
| 合规、法务与安全运营 | 出口、采购和主权客户义务需要专职运营负责人 | 中 | 高 | 标准已经公开,但 Atom 内部具名控制负责人未公开 | 索取合规负责人名单、外部律师沟通节奏和政策例外日志。 |
| 董事会与治理可见度 | 公开材料尚未展示完整委员会地图或少数股东保护框架 | 中 | 中高 | 近期融资有帮助,但治理透明度仍然不完整 | 索取董事会构成、委员会章程、投资人权利和票据转换治理条款。 |
这里的人才风险不太像某位创始人离职,更像 Atom 能否足够快地给容错、客户交付和合规配齐运营系统。
[CR005, CR033, CR037, CR043, CR046]| 风险 | 可监测触发项 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 路线图执行 | 公开里程碑节奏 | Libra 明显滑到 2028 年之后,或 gigaquop 路径失去可信的 2028-2029 年窗口 | 暂停估值扩张,重设整套时间模型基准。 |
| 客户可复用性 | 具名付费标杆客户 | 到下一次重大融资事件时,除当前公开组合外仍没有第二套具名付费系统,或等量级主权 / HPC 合同 | 将当前客户证明视为集中式期权价值,而不是可复用商业引擎。 |
| 出口与承包商合规 | 管理层合规材料包 | 尽调中管理层无法展示 ECCN 映射、视同出口控制和政府客户网络安全姿态 | 升级合规审查,并停止假设全球招聘或采购没有摩擦。 |
| 平台依赖 | AWS 或 NVIDIA 承诺 | 关键路径上出现重大合作伙伴重新排序、云托管延期或解码器集成拖期 | 提高进度和利润率折扣,并把基准情景收窄到现有代际产品。 |
| 资本充足性 | 现金跑道与下一轮条款 | 无法明确覆盖下一组里程碑的现金跑道,或出现附带惩罚性优先级的平轮 / 下轮融资 | 重新测算稀释、下行控制和完成路线图的能力。 |
| 人才与团队 | 关键岗位连续性 | 关键 FTQC、QEC 或合作伙伴前线负责人流失,且没有明确继任者 | 立即上调执行风险,并要求证明管理梯队深度。 |
| 安全保障 | 政府或关键基础设施尽调结果 | 对敏感部署拿不出可信的正常运行时间、CUI 处理或安全认证材料包 | 在保障证据出现前,将公共部门和受监管行业扩张视为延期。 |
否决条件强调可观察事件——进度滑坡、集中度、合规失败、融资条款和人员连续性——而不是叙事信心的变化。
[CR015, CR029, CR030, CR031, CR032, CR033]08估值
8.1 建议和价格纪律
Atom 值得认真放进观察名单,但公开证据不足以证明当前价格明显便宜。Forge 2026 年 7 月基于 COI 的数据给出 Atom 具体的 $2.06 billion Series D-1 估值,并显示投资人在 6 月 Series C 后仍愿意继续融资。这很重要,因为这个价格不是只靠聊天室外推堆出来的传闻。即便如此,投资逻辑仍主要由里程碑驱动,而不是运营基本面。Atom 有首单验证、逻辑量子比特商业化验证、Microsoft 分发可信度和不寻常的联邦支持,但仍未公开披露收入、毛利率或经常性收入占比。正确决策因此是继续研究:把 Atom 视为高质量深科技候选标的,但其估值已经为技术稀缺性、政策相关性,以及 QuNorth 会成为多个真实系统中的第一个、而不是一次性旗舰项目这一信念,嵌入了溢价。[CV001, CV003, CV004, CV008, CV009, CV010]
| 维度 | 评估 | 证据基础 | 决策含义 |
|---|---|---|---|
| 建议 | 继续研究 | 技术和政策信号强,但 $2.06B 估值下经济性披露不完整 | 保留主动尽调权,而不是给出已充分承销的高确信判断 |
| 置信度 | 中 | Forge 价格较具体,但收入质量和 D-1 条款仍未公开 | 不要把精确度外推到已抓取记录之外 |
| 风险评级 | 高 | 资本密度、旗舰客户集中和上市可比公司波动仍然重要 | 先承销下行,再承销上行 |
| 估值立场 | 偏高 | 当前估值在量子可比公司范围内说得通,但相对已披露基本面偏贵 | 要求证明商业化正在铺开 |
| 回报门槛 | 稀释前,2x 毛回报需要可信路径到 >$4.1B,3x 毛回报需要到 >$6.2B | 当前价格给执行失误留下的余地少于更早期入场 | 只有出现可复用部署,才支持激进价格 |
决策姿态对价格敏感,因为公开证据在估值标记和战略背书上更强,在经常性收入或股权结构表细节上更弱。
[CV001, CV017, CV043, CV044, CV045, CV046]| 论点 | 投资逻辑证据 | 反向逻辑或风险 | 什么会改变判断 |
|---|---|---|---|
| 技术稀缺性 | 逻辑量子比特商业化、DARPA Stage B 和中性原子规模让 Atom 具备稀缺性 | 稀缺性不等于持久变现 | 展示另一个具名部署或经常性收入信号 |
| 政府支持 | DoC 和 DARPA 参与降低偿付与信号风险 | 政策背书仍可能与终端市场宽度不足并存 | 清楚区分拨款、股权和真实客户需求 |
| 商业证明 | QuNorth 证明有人愿意购买本地部署系统 | 单个主权旗舰客户可能夸大可复用性 | 增加第二个具名买家,或证明云端到系统购买的转化 |
| 相对估值 | 按抓取估值看,Atom 仍低于 IonQ、Quantinuum 和 PsiQuantum | 比更大的可比公司便宜,不代表按自身基本面就便宜 | 展示收入规模或更清晰的 IPO 准备度 |
| IPO 可选性 | Forge 显示保密申报措辞和 D-1 融资动能 | 这里看不到公开 S-1、经审计数据或路演证据 | 披露经审计财务或清晰上市时间表 |
| 入场纪律 | 如果执行能守住可比公司平价,当前估值可能成立 | 估值快速上调叠加披露稀薄,给拖期留下的空间很小 | 承销价格前,先看 D-1 条款和下行保护 |
每一行都是承销视角,不是预测;这张表明确列出要把结论上调,哪些条件必须成真。
[CV003, CV009, CV011, CV020, CV033, CV035]投资建议维持「继续研究」:高溢价信号真实存在,但当前价格仍跑在公开经济披露前面。
该流程总结因果逻辑,不是数学模型。
[CV001, CV009, CV011, CV012, CV033, CV043]8.2 融资背景和可比公司栈
理解 Atom 的价格,最好看融资动能和相对估值,而不是近端盈利能力。Forge 显示,Atom 在 2026 年 6 月的 C-1 估值为 $714.4 million,到 2026 年 7 月 D-1 标记升至 $2.06 billion,同时还显示一笔 $89.07 million 融资。一家公开收入画像仍基本未披露的公司遇到这样的重估,幅度很陡。相对可比公司解释了市场为何仍可能容忍。IonQ 已经有可观收入,公开市值接近 $13 billion;Quantinuum 2026 年 6 月 IPO 目标约 $14.3 billion,并募资 $1.68 billion;PsiQuantum 仍凭公用事业级雄心支撑 $7 billion 私营估值;D-Wave 和 Rigetti 等上市公司即便收入基数有限,也继续以数十亿美元市值交易。因此,Atom 在量子公司群中并不算被错分,但它确实需要持续执行,才能守住一个比公开运营披露更快到来的溢价重估。[CV001, CV002, CV014, CV016, CV017, CV018]
| 情景 | 假设 | 估值 / 回报逻辑 | 关键风险 | 概率信号 |
|---|---|---|---|---|
| 乐观 | QuNorth 按时落地,第二个买家出现,IPO 窗口保持打开 | 估值可能向享有上市溢价的同业靠拢,并越过 >2x 毛回报门槛 | 执行、制造和公开市场情绪仍然关键 | 系统销售可复用,收入进展可见 |
| 基准 | Atom 交付旗舰里程碑并保持战略价值,但披露仍有限 | 下一轮融资只是小幅越过当前估值,并保留期权价值 | 商业宽度仍窄,股权结构条款更重要 | 交付成功,但缺少广泛收入证明 |
| 悲观 | 旗舰项目拖期,或收入仍过于集中,同时量子板块倍数降温 | 私募估值向平轮或下轮区间压缩 | 客户集中、融资条款和市场情绪一起恶化 | 交付失手、需求转弱或融资结构带惩罚性 |
情景逻辑按里程碑推演,而不是基于 DCF;公开收入和利润率披露仍不足以支撑常规经营模型。
[CV017, CV036, CV038, CV039, CV040, CV048]| 可比公司 | 指标 | 倍数 / 估值 / 状态 | 相关性 | 局限 |
|---|---|---|---|---|
| Atom Computing | 未上市 D-1 估值 | ~$2.06B (Forge, Jul 2026) | 本章当前承销锚点 | COI 得出的估值比收入披露更具体 |
| IonQ | 上市市值加 2026 年收入指引 | 市值约 $12.98B;FY2026 指引 $260M-$270M | 展示收入可见时,公开市场愿意给量子硬件 / 软件叙事的定价 | 上市溢价内含波动性,也有更成熟的披露体系 |
| Quantinuum | 2026 年 IPO 估值 | 估值最高约 $14.3B;IPO 募资约 $1.68B | 更成熟全栈量子平台的高端基准 | Atom 若没有更强收入证明,未必能拿到 IPO 热度 |
| PsiQuantum | 未上市估值 | $1B 轮融资后估值约 $7B | 对公用事业级雄心有参考价值的远景型基准 | 光子架构和规模大得多的融资与 Atom 不同 |
| QuEra | 未上市中性原子融资 | 融资 >$230M;保留来源未披露估值 | 在融资深度和云合作叙事上最接近的同模态同行 | 没有保留下来的公开估值标记可直接比较 |
| Oratomic | 2026 年中性原子 Series A | $300M Series A 轮;审阅的免费来源中估值不透明 | 显示投资人对中性原子挑战者兴趣强 | 阶段、收入成熟度和免费来源估值透明度都不同 |
| D-Wave / Rigetti | 上市市值 | 市值约 $6.20B 和约 $4.69B | 较低的上市量子基准,提醒投资人市场仍愿意为稀缺资产付费 | 架构和产品组合不同于 Atom 的中性原子路径 |
可比集合是代表性而非穷尽;直接私募估值标记若未出现在保留的免费来源中,本表会明确标为不透明或仅有融资信息。
[CV001, CV018, CV019, CV021, CV022, CV024]示意性估值敏感性展示:相对当前 D-1 标记,更多商业化证据可能如何改变合理估值判断。
敏感性分析仅作示意,把估值与商业化里程碑绑定,而不是完整预测模型。
[CV017, CV037, CV038, CV039, CV040, CV047]8.3 商业化、政府背书和 IPO 路径
商业化和政府支持合起来看,Atom 的溢价更容易辩护。QuNorth 销售意味着 Atom 不再只是研究叙事;它有具名本地部署客户项目、具体部署目标,以及与 Microsoft 集成的产品故事。DARPA Stage B 和美国商务部更广泛的量子投资项目,也进一步降低了 Atom 只是一个没有战略赞助的孤立科学项目的概率。但这些积极因素也容易被过度解读。政府资金降低偿付压力,也可能加快部署,但它不能证明可复制客户基础;单一主权旗舰项目也不等于广泛商业采用。因此,Forge 的保密递交标签应解读为可选性,而不是 Atom 已经马上具备 IPO 条件的证据。更清晰的公开市场路径是:把 QuNorth 转化为可见的已交付收入,增加另一个具名系统买方,并在要求公开市场投资人用 Quantinuum 式热情给故事定价之前,证明逻辑量子比特商业化可以扩展到一个旗舰项目之外。[CV003, CV006, CV007, CV008, CV009, CV010]
| 准备度项目 | 当前公开证据 | 正向解读 | 仍缺少什么 |
|---|---|---|---|
| 具名付费客户 | QuNorth / Magne 交易已公开 | 证实本地部署需求存在 | 合同金额、收入确认时点和服务经济性 |
| 产品包装 | Microsoft 与 Atom 提供带逻辑量子比特的商用机器 | 说明产品不只是实验室演示 | 标准定价和部署节奏 |
| 第三方技术验证 | DARPA Stage B 参与已公开 | 给路线图增加外部可信度 | 证明 DARPA 可信度能转化为更广泛购买 |
| 战略资本支持 | DoC / CHIPS 支持和 D-1 融资已公开 | 降低近期融资压力 | 最终法律结构、投资人权利和限制 |
| IPO 信号 | Forge 显示保密申报措辞 | 说明管理层已探索公开市场准备 | 公开申报、经审计报表和时间表清晰度 |
准备度只按可公开引用的证据评估;多项信号方向上正面,但对公开市场承销仍不完整。
[CV003, CV008, CV009, CV010, CV011, CV035]Atom 在技术可信度和战略背书上得分最高,在公开经济披露和广泛商业化实证上得分最低。
评分把留存证据压缩成投委会速记,并非第三方评级。
[CV009, CV011, CV012, CV035, CV036, CV045]8.4 情景承销和尽调门槛
情景分析应保持简单,并对证据敏感。乐观情景下,Atom 按计划交付 QuNorth,把 Microsoft 和政府可信度转化为另一个具名客户,并让 IPO 窗口保持足够久,使市场按规模化战略平台而不是单项目硬件押注来估值。基准情景下,公司仍能融资,也保持战略相关性,但公开披露只逐步改善,下一轮只以有限幅度高于当前标记完成。悲观情景下,旗舰项目延误,客户验证继续集中,或者更广泛的量子板块估值倍数压缩到足以让一家技术可信的公司也撑不住 2026 年 7 月溢价。这个下行并非理论风险:反向市场评论已经把纯量子上市股票框定为押注突破的期权,且回撤风险很重。务实做法是要求能直接改变概率的尽调材料——D-1 条款、优先股堆叠、QuNorth 经济性、收入确认,以及另一个付费部署证据。在这些材料出现之前,正确姿态是有纪律的好奇,而不是估值自满。[CV017, CV029, CV030, CV033, CV034, CV038]
| 主题 | 缺失证据 | 重要性 | 负责人 / 尽调路径 |
|---|---|---|---|
| D-1 经济性 | 条款清单、投资人名单、优先股堆叠、清算条款 | 判断当前估值是干净的普通股价值,还是结构很重的价格 | 领投方 / 律师审阅已签署融资文件 |
| QuNorth 收入确认 | 合同金额、里程碑和支持义务 | 区分头条客户证明和真实经济性证明 | 管理层财务访谈和客户合同尽调 |
| 第二个付费部署 | 另一笔具名系统销售,或持久的云端到本地部署转化 | 会实质降低单一旗舰客户集中度 | 管线审阅加客户推荐访谈 |
| 政府支持条件 | 关闭时间、限制,以及支持是股权、拨款还是与里程碑绑定 | 改变稀释测算和治理含义 | 审阅 DoC 文件和董事会材料 |
| IPO 准备度 | 保密申报措辞背后的经审计报表、投行和时间假设 | 检验公开市场可选性是真实存在,还是只是探索 | 管理层问答和公开申报跟踪 |
| 经营模型 | 硬件毛利、维护收入和服务负担 | 决定 Atom 能否越过定制化旗舰项目,规模化复制 | 收入桥、毛利拆解和部署单元经济性材料 |
问题刻意问得很具体,因为 Atom 的估值比它的经济性和资本结构细节更容易观察。
[CV017, CV035, CV036, CV037, CV047, CV048]悲观、基准、乐观区间框定当前证据能支撑的范围,并不假装知道未来收入的精确数字。
区间是基于当前可比标的和里程碑结果的情景估计,不是经审计的市场标记。
[CV017, CV038, CV039, CV040, CV048]8.5 图表与附录
免责声明
这份 report-meta 文件仅反映截至 2026-07-18 在各章 YAML 中审阅过的公开证据。Atom 仍是私营公司,因此投资建议和估值判断对未披露的融资条款、客户经济性和收入质量数据高度敏感;这些信息未出现在留存的公开来源中。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Atom Computing was founded in 2018. | 高 | SO001, SO005 |
| CO002 | Forge identifies Benjamin Bloom and Jonathan King as Atom Computing founders. | 中 | SO005 |
| CO003 | Forge lists Atom Computing headquarters in Berkeley, California. | 中 | SO005 |
| CO004 | Atom maintains an operating footprint in Colorado in addition to its Berkeley headquarters, including a Boulder commercial operations facility cited in the 2023 qubit announcement. | 中 | SO006 |
| CO005 | Atom positions itself as a builder of highly scalable gate-based quantum computers using optically trapped neutral atoms. | 高 | SO001, SO002 |
| CO006 | The company markets the AC1000 system and the logical-qubit era as the centerpiece of its current commercial message. | 中 | SO001 |
| CO007 | Ben Bloom is publicly identified as founder and CEO of Atom Computing. | 高 | SO002, SO004, SO007 |
| CO008 | Jonathan King appears as a founder in Forge’s company profile and as an author on Atom’s logical-qubit technical paper. | 中 | SO005, SO019 |
| CO009 | Atom said on 2026-06-16 that it had raised more than $300 million in total funding. | 高 | SO002, SO003 |
| CO010 | Atom’s announced Series C round was $100 million and was led by Third Point Ventures. | 高 | SO002, SO003 |
| CO011 | DCVC and Cisco Investments were disclosed participants in the Series C financing. | 高 | SO002, SO003 |
| CO012 | The 2026 funding announcement also referenced a signed Letter of Intent with the U.S. Department of Commerce for $100 million. | 高 | SO002, SO003 |
| CO013 | DCVC characterized the Department of Commerce support as a $100 million equity stake in Atom Computing. | 中 | SO004 |
| CO014 | Forge reports Atom at a $2.06 billion Series D-1 valuation in July 2026 based on Certificate of Incorporation data. | 中 | SO005 |
| CO015 | Forge lists a July 8, 2026 Series D-1 financing of about $89.07 million with undisclosed investors and 22,701,858 shares outstanding. | 中 | SO005 |
| CO016 | Forge’s funding table shows an approximately $99.67 million Series C-1 financing on 2026-06-16 and a reported $714.4 million post-money figure for that round. | 中 | SO005 |
| CO017 | Atom announced in 2023 that it created a 1,225-site atomic array populated with an average of 1,180 qubits. | 中 | SO006, SO020 |
| CO018 | Atom said the 2023 system was the first universal gate-based quantum computer to cross the 1,000-qubit threshold. | 中 | SO006, SO020 |
| CO019 | Atom highlighted 40-second coherence times as part of the 2023 next-generation-system milestone package. | 中 | SO006 |
| CO020 | Microsoft and Atom announced a commercial machine capable of entangling 24 logical qubits and computing with 28 logical qubits. | 高 | SO008, SO019 |
| CO021 | The Microsoft-Atom system was offered for order in late 2024 with delivery expected in 2025. | 中 | SO008 |
| CO022 | DARPA selected Atom for Stage B of the Quantum Benchmarking Initiative. | 高 | SO007, SO017, SO018 |
| CO023 | DARPA’s QBI asks participants to map the R&D plan, risks, and prototypes needed to reach utility-scale quantum systems. | 中 | SO007, SO018 |
| CO024 | Atom said it sold its first commercial on-premises quantum computer to QuNorth in 2025. | 高 | SO002, SO003, SO012 |
| CO025 | QuNorth plans to operate the Magne system as a 50-logical-qubit, 1,225-physical-qubit machine based in Copenhagen. | 高 | SO009, SO011, SO012 |
| CO026 | Novo Nordisk Foundation and EIFO committed €80 million to establish QuNorth and acquire Magne. | 高 | SO009, SO011, SO012, SO013 |
| CO027 | Microsoft contributes the quantum software, middleware, and cloud connectivity for Magne while Atom delivers the neutral-atom hardware. | 高 | SO009, SO011, SO014 |
| CO028 | Atom’s homepage and 2026 funding release both describe strategic collaborations with Cisco and NVIDIA. | 中 | SO001, SO002, SO015, SO016 |
| CO029 | Atom and Nu Quantum announced a June 2026 partnership focused on utility-scale quantum computing and networking. | 低 | SO001, SO001 |
| CO030 | Public materials present Atom as a late-stage private company with IPO preparation signals but without a public S-1 filing available in retained sources. | 中 | SO001, SO005 |
| CO031 | Forge labels Atom as a company with an IPO mentioned and a confidential filing reference, but public-source detail on the filing itself remains thin. | 中 | SO005 |
| CO032 | The homepage directly advertises 1,200+ fully connected qubits and entry into the logical-qubit era. | 高 | SO001, SO002 |
| CO033 | The 2026 funding announcement says Atom is building commercial on-premises systems for enterprise and government customers worldwide. | 高 | SO002, SO003 |
| CO034 | The QuNorth transaction is the strongest public proof that Atom has moved from research-only posture to commercial system delivery. | 高 | SO002, SO009, SO011, SO012 |
| CO035 | Independent sector commentary still warns that commercially relevant quantum computing may be at least a decade away, which tempers milestone enthusiasm for Atom. | 中 | SO021, SO022 |
| CO036 | Public sources do not disclose Atom’s current revenue, ARR, customer count, or company-wide headcount. | 高 | SO001, SO002, SO005 |
| CO037 | Public sources do not disclose the exact cash terms, preferences, or investor identity for Atom’s July 2026 Series D-1 financing. | 中 | SO005 |
| CO038 | Federal and congressional quantum initiatives in 2026 indicate that public-sector demand and strategic policy support remain meaningful tailwinds for Atom. | 中 | SO004, SO023, SO024, SO025 |
| CO039 | NERSC’s 2026 neutral-atom proposal call shows national-lab buyers are actively creating workflows for neutral-atom quantum systems. | 中 | SO023 |
| CO040 | The 2026 company homepage shows Atom’s news cadence centered on funding, the Nu Quantum alliance, and a June 2026 quantum-error-correction technical perspective. | 中 | SO001 |
| CO041 | Atom’s valuation evidence is strongest in COI-derived Forge data rather than in a contemporaneous company press release for the D-1 round. | 高 | SO002, SO005 |
| CO042 | The Department of Commerce support reduces capital-risk perception but also ties Atom more closely to U.S. industrial-policy execution. | 中 | SO003, SO004, SO024 |
| CO043 | Atom’s most visible commercialization path in public sources is still concentrated in strategic government, research, and sovereign ecosystem deployments rather than broad enterprise volumes. | 高 | SO002, SO009, SO011, SO023 |
| CO044 | Whether Atom can turn technical leadership into diversified repeat revenue remains an open diligence question rather than a publicly settled fact. | 低 | |
| CO045 | Whether the confidential-filing signal on Forge represents an active IPO path or only historical process noise remains unresolved in retained evidence. | 低 | |
| CM001 | Atom’s relevant market is neutral-atom quantum computing sold as cloud access, on-prem systems, and application co-design inside the broader quantum-computing hardware and services market, not the full quantum-technology stack. | 高 | SM001, SM003, SM019 |
| CM002 | Included spend for Atom therefore centers on quantum processing access, on-prem deployment, hybrid HPC integration, and algorithm-development services tied to neutral-atom systems. | 中 | SM003, SM018, SM019, SM022 |
| CM003 | Excluded spend includes quantum sensing, quantum networking or QKD, and most post-quantum cryptography migration budgets because those categories do not require buying Atom hardware. | 中 | SM007, SM010, SM029 |
| CM004 | The status-quo substitutes for Atom’s target jobs are classical HPC, AI-accelerated simulation, classical optimization solvers, and quantum-safe software migration rather than quantum hardware. | 中 | SM003, SM014, SM029 |
| CM005 | Atom’s neutral-atom platform is specifically framed around simulation and optimization workloads because AWS and Atom both highlight physics, chemistry, materials, and combinatorial optimization as native fits. | 高 | SM018, SM019, SM026 |
| CM006 | McKinsey’s 2026 monitor says quantum-computing company revenue exceeded $1 billion in 2025 and could reach $3.2 billion to $4.4 billion by 2028. | 中 | SM001 |
| CM007 | Across McKinsey’s 2025 and 2026 monitors, public market lenses place quantum-computing revenue around $0.65 billion to $1.4 billion in 2024-2025 and around $43 billion to $72 billion by 2035, with substantial upside only after commercialization matures. | 高 | SM001, SM002 |
| CM008 | QED-C’s 2026 quantum-computing forecast estimates a $1.4 billion market in 2025 growing roughly 30 percent annually to more than $3 billion by 2028. | 中 | SM004 |
| CM009 | QED-C’s state-of-industry report separately pegs 2025 market size at $1.9 billion with 30 percent average annual growth, underscoring that even closely related industry datasets do not perfectly align. | 中 | SM005 |
| CM010 | BCG’s most conservative provider-revenue lens still sees a $1 billion to $2 billion quantum-computing provider market by 2030 and a $90 billion to $170 billion hardware-and-software provider market by 2040. | 中 | SM003 |
| CM011 | BCG also projects $450 billion to $850 billion of long-run economic value by 2040, which is much larger than provider revenue and therefore should be treated as an outer value envelope rather than spend that vendors can directly capture. | 中 | SM003 |
| CM012 | The public market-size spread is already meaningful before 2030: McKinsey and QED-C cluster near $1.1 billion to $1.9 billion for 2025, while long-term value pools jump into tens of billions only if fault-tolerant adoption arrives. | 中 | SM001, SM004, SM005, SM003 |
| CM013 | QED-C says on-premises systems are projected to become the largest quantum-computing segment by 2028, which is especially relevant for Atom because its strongest public lighthouse customer is an on-prem national-lab deployment. | 中 | SM004, SM016, SM035 |
| CM014 | Atom-specific SAM is narrower than generic quantum TAM because its evidence-backed early use cases concentrate in simulation, materials, chemistry, optimization, and hybrid HPC rather than the full set of quantum-software or cryptography spend. | 中 | SM004, SM018, SM019, SM024, SM026 |
| CM015 | The most defensible evidence-constrained SOM lens for Atom is not broad enterprise IT but named sovereign, national-lab, and research programs that already buy access, equipment, or co-design engagements. | 中 | SM009, SM016, SM022, SM032, SM034 |
| CM016 | BCG estimates that public orders of quantum computers already support more than half of the market, making sovereign demand the clearest near-term demand base for Atom-like hardware vendors. | 中 | SM003 |
| CM017 | QED-C simultaneously describes dependence on government funding as a structural vulnerability and says 54 percent of respondents see some probability of a quantum winter by 2031, so public funding is both a driver and a fragility. | 中 | SM004, SM005 |
| CM018 | The UK alone committed £2.5 billion over ten years to quantum technologies and explicitly names life sciences, finance, aerospace, defence, and HPC-adjacent capabilities as domestic user sectors, showing how sovereign programs shape buyer development before commercial ROI is settled. | 中 | SM007 |
| CM019 | The U.S. National Quantum Initiative reauthorization summary extends the federal program through 2034 while authorizing NIST funding, new testbeds, workforce hubs, and application research, reinforcing that the market is still being scaffolded by public institutions. | 中 | SM008 |
| CM020 | DOE’s Quantum Genesis initiative aims to stand up a fault-tolerant, scientifically relevant quantum capability by 2028 and to integrate it with national supercomputing infrastructure, further validating government and HPC centers as the first serious buyers. | 中 | SM011 |
| CM021 | NSA’s CNSA 2.0 guidance says organizations should plan, prepare, and budget now for quantum-resistant algorithms, so post-quantum migration creates quantum-related security budgets even before fault-tolerant hardware is broadly useful. | 中 | SM010 |
| CM022 | Atom’s 2026 readiness report says government and defense are expected to lead commercialization over the next three years. | 中 | SM012 |
| CM023 | The same Atom 2026 survey places large enterprises second and pharmaceuticals and life sciences third at 11 percent, while financial services ranks last at 5 percent for near-term commercialization. | 中 | SM012 |
| CM024 | Atom’s 2025 readiness survey says organizations prioritize cost-effectiveness, access to cutting-edge technology, and cloud availability when deciding how to buy quantum capability. | 中 | SM013 |
| CM025 | Atom’s cloud path is real rather than hypothetical: AWS and Atom say Aquila has been available on Amazon Braket since 2022, and Atom reported almost 1,000 machine-hours of usage with availability expanded to more than 100 hours per week by late 2023. | 高 | SM019, SM020 |
| CM026 | Government and national-lab adoption is also concrete: NERSC’s 2026 open call offers Aquila and Gemini access for DOE-aligned work in materials science, chemistry, high-energy physics, and HPC-integrated workflows. | 高 | SM009, SM022 |
| CM027 | AIST awarded Atom a 6.5 billion JPY contract, roughly $41 million, to place a neutral-atom system beside the ABCI-Q supercomputer, making sovereign on-prem procurement the strongest public Atom buyer proof. | 中 | SM016, SM035 |
| CM028 | The later AIST MOU expands that relationship into a hybrid cloud-and-HPC environment and explicitly mentions industrialization, external users, and supply-chain work, showing a path from flagship installation to broader ecosystem formation. | 中 | SM017, SM035 |
| CM029 | AWS and Atom now publicly target Libra on Braket in 2028 for chemistry, high-energy physics, and materials simulation, which implies Atom’s next buyer step is hybrid cloud usage for scientifically relevant workloads rather than only isolated research demonstrations. | 高 | SM018, SM032 |
| CM030 | AWS also describes current Atom users pushing quantum reservoir computing, high-energy-physics simulation, and financial-optimization algorithms, while the Braket hardware page highlights graph optimization, protein design, traffic coordination, and network problems. | 高 | SM018, SM019 |
| CM031 | Atom’s buyer-development motion in pharma and biology is visible through its drug-discovery partnership with Quantum Intelligence Corp. and its participation in Wellcome Leap Quantum for Bio projects. | 中 | SM024, SM025 |
| CM032 | Atom’s optimization motion is visible through Kipu, where both parties position neutral atoms for large-scale optimization in logistics, portfolio optimization, pharmaceutical research, and telecom. | 中 | SM026 |
| CM033 | Enterprise and government co-design are themselves part of the product: BCG X and Deloitte both frame Atom as a problem-first co-development partner for governments, HPC centers, life sciences, materials, logistics, and financial-services clients. | 中 | SM015, SM033 |
| CM034 | Atom’s ICSC program widens access for universities, national laboratories, and businesses in Italy via premium cloud access and mentoring, which shows that the adoption path often begins with subsidized access and training before hardware ownership. | 中 | SM034 |
| CM035 | The finance segment is active but still pre-production: Data Center Knowledge says JPMorgan’s London quantum-AI platform is for research into portfolio optimization, quantum machine learning, and algorithms in a secure enterprise setting, while practical business value remains an open question. | 中 | SM027 |
| CM036 | The Quantum Insider’s 2026 banks survey says more than 15 global banks have research programs across portfolio optimization, risk modeling, derivative pricing, fraud detection, and post-quantum migration, but no bank has deployed production-ready quantum systems for live operations. | 中 | SM028 |
| CM037 | Atom’s 2026 survey says 62 percent of organizations actively factor sovereignty into procurement decisions, so regional sourcing and domestic-control concerns now shape adoption alongside raw performance. | 中 | SM012 |
| CM038 | Atom’s 2026 survey also says 37 percent cite skilled-worker shortages as a major obstacle and only 13 percent have scaled quantum applications productively even though 56 percent are evaluating or piloting. | 中 | SM014 |
| CM039 | QED-C echoes that talent remains inadequate and the supply chain is still custom, fragile, and strategically sensitive, especially for photonics, control electronics, and other specialized components. | 中 | SM005 |
| CM040 | BCG argues that quantum computing currently provides no tangible commercial or scientific advantage over classical systems, because fidelity and circuit depth remain limiting while GPUs, algorithms, and AI keep raising the classical baseline. | 中 | SM003 |
| CM041 | Moody’s reaches a similarly skeptical conclusion, saying commercial value is delayed because the hardware is not ready yet and that media coverage has tended to be overblown. | 中 | SM029 |
| CM042 | IBM’s Advantage Tracker argues that credible quantum advantage will emerge through iterative community validation rather than a single press release, because classical methods can overturn apparently strong quantum runtime leads within months. | 中 | SM031 |
| CM043 | HPCwire’s coverage of D-Wave shows how quickly headline “quantum supremacy” claims can be met by classical rebuttals, reinforcing that technical announcements do not automatically clear buyer trust hurdles. | 中 | SM030 |
| CM044 | BCG says quantum-computing provider revenue in the NISQ era could be only $100 million to $500 million per year for analog and hybrid simulation use cases even while long-run upside remains large, which is why near-term valuation depends on disciplined channel selection rather than TAM slogans. | 中 | SM003 |
| CM045 | BCG’s priority industries for error-corrected value are technology, chemicals and agriculture, pharmaceuticals, defense and space, and financial institutions, followed by the public sector. | 中 | SM003 |
| CM046 | BCG also says quantum computing is roughly 100,000 times more expensive per hour than classical computing today and that corporate buyers usually want one-year break-even, with three- to five-year payback acceptable only in some cases. | 中 | SM003 |
| CM047 | McKinsey’s 2026 monitor says the largest use-case value pools by 2035 sit in energy and materials, pharmaceuticals, and finance, which broadly matches Atom’s simulation-first positioning even if public Atom customer proof is still heaviest in government and HPC. | 高 | SM001, SM002, SM016 |
| CM048 | McKinsey’s finance deep dive values finance use cases at roughly $400 billion to $600 billion by 2035, but frames them around optimization, risk modeling, and cryptography security rather than immediate production deployment. | 中 | SM001 |
| CM049 | Atom and AWS claim cloud-accessible fault-tolerant workflows can begin in 2028 through Libra, with chemistry, materials, optimization, and other scaling-limited workloads as the initial targets. | 高 | SM018, SM032 |
| CM050 | BCG still places full-scale fault tolerance after 2040, so Atom’s 2028 roadmap is a material outlier rather than a consensus market assumption. | 中 | SM003 |
| CM051 | Public sources do not isolate a neutral-atom-specific SAM or publish Atom pricing curves by workload, so any Atom-specific TAM/SAM/SOM model must remain evidence-constrained and directional. | 中 | SM003, SM004, SM005 |
| CM052 | No retained public source verifies a named Atom finance customer running production workflows as of 2026-07-05, so finance should be treated as a strategic target segment rather than a demonstrated current revenue base. | 中 | SM027, SM028, SM012 |
| CM053 | QED-C says simulation is the dominant near-term application cluster, with computational chemistry at 26 percent and materials science at 22 percent of identified use cases, which supports Atom’s simulation-heavy vertical focus. | 中 | SM004 |
| CM054 | Atom’s surveys and partner announcements consistently describe the market as moving from curiosity-driven experimentation to proof-driven procurement, meaning buyers increasingly want benchmarks, integration plans, and co-design pathways before committing capital. | 高 | SM012, SM014, SM015, SM033 |
| CM055 | Neutral-atom technical credibility is stronger than for many private peers because Harvard, Atom, MIT, and collaborators demonstrated logical-qubit and high-fidelity milestones in 2023, but that progress still does not eliminate commercialization risk. | 高 | SM036, SM037, SM029 |
| CP001 | Atom's buyer-facing competitive set spans direct neutral-atom peers, broader universal quantum hardware platforms, annealing substitutes, cloud brokers, and hybrid-HPC status quo alternatives. | 中 | SP001, SP005, SP024, SP027, SP029, SP030 |
| CP002 | Atom offers three distinct access paths in 2026: Amazon Braket, premium direct access, and on-prem neutral-atom deployments integrated with HPC environments. | 高 | SP001, SP002, SP005 |
| CP003 | Aquila is a 256-qubit neutral-atom system positioned for simulation, optimization, and machine-learning workloads and available more than 100 hours per week on Amazon Braket. | 中 | SP001 |
| CP004 | Atom's public roadmap emphasizes logical-qubit capability, room-temperature deployment, and enterprise or government applications rather than only a larger physical-qubit headline. | 高 | SP003, SP004 |
| CP005 | Atom reported more than $230 million of financing in February 2025 and tied the round to both strategic investors and commercial progress with customers such as AIST. | 中 | SP004 |
| CP006 | IBM Quantum is the main incumbent benchmark because it combines the broadest published hardware fleet, uptime metrics, and a full access ladder from free usage to dedicated on-prem systems. | 高 | SP007, SP008 |
| CP007 | IBM publicly lists Open, Pay-As-You-Go, Flex, Premium, and On-Prem plans, making it the most transparent quantum hardware packaging surface in this retained set. | 中 | SP007 |
| CP008 | IBM's published pricing ladder starts at free Open access, then $96 per minute PAYG, $72 per minute Flex, and $48 per minute Premium, while on-prem pricing remains quote-led. | 中 | SP007 |
| CP009 | Google's Willow hardware is not publicly available in 2026 and instead is offered only to a select cohort of research partners through an early access program. | 高 | SP009, SP011 |
| CP010 | Google's open-source Cirq framework broadens developer familiarity with its stack even while broad commercial hardware access remains unavailable. | 中 | SP010 |
| CP011 | Willow gives Google a major technical and ecosystem threat position, but its commercial pressure remains lower than IBM's because access is still proposal-gated rather than catalog-like. | 高 | SP009, SP011 |
| CP012 | IonQ offers direct cloud access, reservations, and major-SDK compatibility including Braket, Q#, Cirq, Qiskit, and tket integrations. | 中 | SP012 |
| CP013 | IonQ Forte Enterprise is marketed as a rack-based, data-center-deployable 36-qubit trapped-ion system, giving IonQ one of the clearest on-prem enterprise offers in the field. | 中 | SP014 |
| CP014 | IonQ reported $64.7 million of Q1 2026 revenue and 755% year-on-year growth, showing commercial scale beyond a purely technical roadmap story. | 中 | SP015 |
| CP015 | IonQ's 2026 roadmap targets 100-256+ physical qubits and 12 logical qubits, making trapped-ion competition directly relevant to Atom's own logical-qubit positioning. | 高 | SP013, SP015 |
| CP016 | Quantinuum combines direct subscriptions, Microsoft Azure distribution, cloud access, and on-prem Helios availability rather than relying on a single commercial route. | 高 | SP016, SP017, SP019 |
| CP017 | Helios is marketed as a 98-physical-qubit trapped-ion platform with cloud and on-prem availability plus named enterprise collaborators, placing Quantinuum among the strongest procurement rivals to Atom. | 高 | SP016, SP017 |
| CP018 | pytket can import external circuit formats and run on a range of devices and simulators, reducing compiler-layer lock-in for Quantinuum users. | 中 | SP018 |
| CP019 | Azure Quantum publishes Quantinuum H2 subscription prices of $125,000 per month for Standard and $175,000 per month for Premium, giving Quantinuum one of the few explicit enterprise price signals in quantum hardware. | 高 | SP019, SP016 |
| CP020 | Azure also publishes IonQ token pricing and Rigetti time-based billing, showing that a cloud broker can normalize comparison shopping across otherwise dissimilar hardware vendors. | 中 | SP019 |
| CP021 | PsiQuantum is a long-horizon strategic threat rather than a broad commercial compute option today because its public story is centered on utility-scale photonic infrastructure and government-backed buildouts. | 高 | SP020, SP021 |
| CP022 | Nature corroborates PsiQuantum's claim that its photonic platform is being built as a manufacturable, foundry-style architecture rather than a lab-only prototype. | 高 | SP020, SP021 |
| CP023 | Rigetti competes as a full-stack superconducting platform with in-house Fab-1 manufacturing and less-than-one-millisecond hybrid connectivity between classical infrastructure and its QPUs. | 中 | SP022 |
| CP024 | Rigetti's Novera is an immediately shippable 9-qubit on-prem testbed with 24/7 system access and deep stack control, making it relevant for buyers who prioritize hands-on integration. | 中 | SP023 |
| CP025 | D-Wave is a real substitute because it sells production-ready annealing systems today while also keeping a gate-model research path alive. | 高 | SP024, SP025 |
| CP026 | D-Wave Leap offers 99.9% uptime, subsecond responses, and hybrid solvers for up to two million variables, which is a stronger near-term operational offer than most universal gate-model vendors provide. | 中 | SP025 |
| CP027 | D-Wave's January 2026 $20 million FAU system agreement shows that on-prem quantum contracts are real and can anchor regional government and workforce ecosystems. | 高 | SP024, SP026 |
| CP028 | AWS Braket lists multiple vendors on one procurement and execution surface, which lowers initial buyer commitment to any one modality and makes comparison shopping easier. | 高 | SP005, SP006 |
| CP029 | Amazon Braket uses comparable task, shot, and reservation concepts across QPUs, and Atom Aquila is one of the few private quantum hardware offers with fully public rates. | 高 | SP006, SP001 |
| CP030 | Atom's public Braket rates are $0.30 per task, $0.01 per shot, and $2,500 per reserved hour. | 高 | SP006, SP001 |
| CP031 | Atom Computing shows that Atom's direct neutral-atom peer set is real, not hypothetical, with 1,200+ fully connected qubits and a $300 million-plus raise announced in June 2026. | 中 | SP029 |
| CP032 | Pasqal also crowds the direct peer set by marketing cloud, on-prem, Google Cloud, and Azure access while citing expected financing of at least €340 million and 25+ commercial use cases. | 中 | SP030 |
| CP033 | NVIDIA frames useful quantum as accelerated quantum supercomputing, which strengthens the status-quo substitute of hybrid AI/HPC workflows before buyers commit to one hardware vendor. | 中 | SP027 |
| CP034 | Cloud marketplaces and open SDKs keep software-side switching costs moderate because buyers can multi-home across Braket, Azure, Cirq, pytket, IonQ integrations, and other brokered routes. | 高 | SP005, SP010, SP012, SP018, SP019 |
| CP035 | Lock-in still exists at the hardware and workload layer because neutral atoms, trapped ions, superconductors, photonics, and annealing expose different connectivity, timing, and algorithm-fit trade-offs. | 中 | SP001, SP011, SP013, SP020, SP024, SP028 |
| CP036 | Atom's clearest durable edge is not generic software lock-in but a bundle of neutral-atom analog simulation, room-temperature and HPC deployment, and a live logical-qubit roadmap. | 高 | SP001, SP002, SP003 |
| CP037 | Atom's channel disadvantage versus IBM, IonQ, and Quantinuum is that those rivals expose broader enterprise procurement paths or stronger public commercialization signals today. | 中 | SP007, SP015, SP016, SP017, SP019 |
| CP038 | Trust and procurement posture in 2026 favor vendors with published uptime, explicit plans or prices, on-prem options, or named enterprise deployments—not just large qubit counts. | 中 | SP007, SP008, SP017, SP023, SP025, SP026 |
| CP039 | IEEE Spectrum's 2026 framing is that the first error-corrected machines represent scientific advantage rather than broad commercial advantage, which should bound Atom's moat claims. | 中 | SP028 |
| CP040 | IEEE Spectrum also highlights a neutral-atom trade-off: strong scalability and qubit maneuverability, but slower operations than superconducting systems. | 中 | SP028, SP011 |
| CP041 | Atom's risk register is dominated by cloud-broker commoditization, better-capitalized incumbents, direct neutral-atom crowding, and the chance that hybrid classical workflows satisfy customer needs earlier. | 中 | SP019, SP027, SP028, SP029, SP030 |
| CP042 | Atom's public pricing transparency is stronger than Google, PsiQuantum, and most direct private hardware peers, but weaker than IBM's fully published plan ladder. | 中 | SP006, SP007, SP009, SP019, SP020 |
| CP043 | On-prem and dedicated deployment paths can raise switching costs once integrated into a customer's security and HPC environment, which benefits Atom but is increasingly table stakes across major rivals. | 高 | SP002, SP007, SP014, SP016, SP023, SP024 |
| CP044 | IBM and Google remain the most credible likely entrant or displacement threats because they can improve hardware, software, and customer education quickly even when public access is selective or contract-heavy. | 中 | SP008, SP009, SP010, SP011 |
| CP045 | The direct neutral-atom peer set is now crowded enough that Atom's differentiation durability depends on faster enterprise proof and workload wins, not on modality choice alone. | 中 | SP028, SP029, SP030 |
| CI001 | Atom’s clearest public revenue surfaces are on-premises system delivery, logical-qubit system integration, and high-touch partner-backed deployments rather than broad self-service cloud usage. | 高 | SI001, SI002, SI008, SI009, SI010 |
| CI002 | The June 2026 funding release says Atom is using fresh capital to expand on-premises systems for government, enterprise, and research customers worldwide. | 高 | SI002, SI003 |
| CI003 | QuNorth is the strongest public evidence that Atom has monetizable hardware-delivery revenue rather than only research collaborations. | 高 | SI009, SI010, SI011, SI012, SI013 |
| CI004 | Public retained sources do not disclose Atom’s current revenue or ARR. | 高 | SI001, SI002, SI005 |
| CI005 | Public retained sources do not disclose Atom’s pricing, discounting, contract duration, or standard commercial terms. | 高 | SI001, SI002, SI008 |
| CI006 | Atom announced more than $300 million of cumulative funding as of 2026-06-16. | 高 | SI002, SI003 |
| CI007 | The announced Series C round was $100 million led by Third Point Ventures with participation from DCVC and Cisco Investments. | 高 | SI002, SI003 |
| CI008 | The same announcement referenced a signed Letter of Intent with the U.S. Department of Commerce for another $100 million. | 高 | SI002, SI003 |
| CI009 | DCVC described the Department of Commerce support as a $100 million equity stake in Atom. | 中 | SI004 |
| CI010 | Forge reports a July 2026 Series D-1 valuation of $2.06 billion and about $89.07 million of financing. | 中 | SI005 |
| CI011 | Forge’s funding table implies that public market-data visibility into Atom’s private financing is stronger than visibility into Atom’s operating economics. | 中 | SI005 |
| CI012 | The QuNorth materials describe Magne as a 50-logical-qubit system funded through an €80 million Danish initiative, but they do not disclose Atom’s recognized revenue from the program. | 高 | SI009, SI010, SI011, SI012, SI013 |
| CI013 | The June 2026 funding release says Atom plans to use capital for higher-qubit-count systems, software, control systems, error correction, deployments, and team growth. | 高 | SI002, SI003 |
| CI014 | Those stated use-of-funds categories indicate a hardware-heavy cost structure with significant R&D, integration, and deployment spending. | 高 | SI002, SI003 |
| CI015 | DARPA Stage B participation implies continued engineering spend on prototypes, risk burn-down, and utility-scale development planning before broad commercialization. | 中 | SI006, SI014 |
| CI016 | Atom’s public commercialization path appears concentrated in government, research, and sovereign ecosystem programs rather than broad recurring enterprise demand. | 高 | SI002, SI009, SI010, SI017 |
| CI017 | The absence of public CAC, payback, NRR, churn, or sales-cycle disclosure means Atom cannot yet be judged with mature software efficiency metrics. | 高 | SI001, SI002, SI005 |
| CI018 | IonQ’s 2025 Form 10-K shows that even a more mature public quantum peer still reports significant operating losses and ongoing capital needs. | 高 | SI019, SI022 |
| CI019 | Rigetti’s 2025 Form 10-K similarly reinforces that quantum-hardware businesses remain capital intensive and commercially early. | 中 | SI020 |
| CI020 | D-Wave’s 2025 Form 10-K shows that public quantum companies can have real revenue and still face heavy operating and financing pressure. | 中 | SI021 |
| CI021 | Compared with public peers, Atom’s private status leaves the market with far less visibility into revenue mix, margin profile, and cash runway. | 中 | SI005, SI019, SI020, SI021 |
| CI022 | IonQ and Rigetti market-cap pages show that investors currently ascribe multi-billion-dollar values to public quantum narratives, even while economics remain imperfectly proven. | 中 | SI023, SI024 |
| CI023 | Quantinuum’s June 2026 IPO coverage shows that public markets remain open to scaled quantum stories with better disclosed revenue than Atom currently provides. | 中 | SI025 |
| CI024 | Sacra’s PsiQuantum profile reinforces that private quantum capital still flows toward long-duration hardware programs before broad revenue disclosure. | 中 | SI026 |
| CI025 | Public retained sources do not disclose Atom’s cash balance, monthly burn, runway, gross margin, or working-capital profile. | 高 | SI001, SI002, SI005 |
| CI026 | The Department of Commerce commitment and Series C materially improve Atom’s financing posture even though final capital availability still depends on private terms. | 高 | SI002, SI003, SI004 |
| CI027 | Because Atom is still investing ahead of broad revenue disclosure, financing dependency remains a live thesis variable rather than a solved issue. | 高 | SI002, SI005, SI018 |
| CI028 | The strongest positive read on Atom’s financial story is that capital raised and customer proof are both real, which lowers immediate existential financing risk. | 高 | SI002, SI004, SI009, SI010, SI011 |
| CI029 | The strongest negative read is that Atom’s visible economics may still be dominated by a small number of flagship system programs rather than recurring diversified demand. | 高 | SI009, SI010, SI011, SI018 |
| CI030 | Public evidence is not strong enough to support a clean revenue multiple or margin-based valuation framework for Atom. | 中 | SI004, SI005, SI018 |
| CI031 | The QuNorth contract is strategically important but insufficient on its own to prove repeatable revenue quality. | 高 | SI009, SI010, SI011, SI012 |
| CI032 | The company’s public messaging stresses product, control systems, and error correction investment more than software-like unit economics. | 高 | SI002, SI003 |
| CI033 | Independent skeptical commentary on quantum timelines remains relevant because long commercialization cycles can magnify burn and financing risk. | 中 | SI018 |
| CI034 | Whether Atom already generates meaningful software, services, or maintenance revenue alongside hardware remains unresolved in retained sources. | 低 | |
| CI035 | Whether the Department of Commerce capital closes on the same timeline and economics implied by current public commentary remains unresolved. | 低 | |
| CI036 | Whether Magne-class deployments can scale into a broader customer base without heavy custom work remains unresolved. | 低 | |
| CE001 | Atom markets the AC1000 system and the logical-qubit era as the centerpiece of its current product story. | 中 | SE001 |
| CE002 | Atom positions its systems for enterprise, government, and research customers that need on-premises quantum deployment. | 高 | SE002, SE003 |
| CE003 | The company says its on-premises systems provide customers with logical-qubit capabilities for complex applications. | 高 | SE002, SE007 |
| CE004 | Atom’s core architecture uses arrays of optically trapped neutral atoms as qubits. | 高 | SE001, SE002 |
| CE005 | The 2023 milestone announcement described a 1,225-site atomic array populated with about 1,180 qubits. | 中 | SE004 |
| CE006 | Atom said the 2023 platform crossed the 1,000-qubit threshold for a universal gate-based system. | 中 | SE004 |
| CE007 | The 2023 announcement also highlighted 40-second coherence and mid-circuit measurement as enabling capabilities for fault tolerance. | 中 | SE004 |
| CE008 | Microsoft and Atom reported a commercial machine that entangled 24 logical qubits. | 高 | SE006, SE015 |
| CE009 | The same Microsoft-Atom system demonstrated computation on 28 logical qubits with error detection and correction. | 高 | SE006, SE015 |
| CE010 | The arXiv technical paper describes fault-tolerant quantum computation on a 256-qubit neutral-atom processor. | 中 | SE015 |
| CE011 | The arXiv paper says full connectivity is enabled by atom movement and that error sources can be converted into detectable atom loss. | 中 | SE015 |
| CE012 | Microsoft says Atom recently achieved 99.6% two-qubit gate fidelity on a commercial neutral-atom system. | 中 | SE006 |
| CE013 | Microsoft’s qubit-virtualization system is designed to sit on top of Atom’s neutral-atom hardware to create logical qubits. | 高 | SE006, SE025 |
| CE014 | The commercial Microsoft-Atom package also integrates Azure Elements, AI, and HPC workflows with the quantum hardware. | 高 | SE006, SE025, SE026 |
| CE015 | DARPA selected Atom for Stage B of QBI to test a utility-scale development plan and its associated risks. | 高 | SE005, SE013, SE014 |
| CE016 | QBI participation gives Atom an external benchmark for whether its technical roadmap can plausibly scale toward utility. | 高 | SE013, SE014 |
| CE017 | QuNorth’s Magne system is described as a 50-logical-qubit, 1,225-physical-qubit Level 2 system. | 高 | SE007, SE008, SE009, SE010 |
| CE018 | Independent and customer-side sources agree that Atom will deliver the quantum hardware for Magne while Microsoft contributes software and cloud connectivity. | 高 | SE007, SE008, SE009, SE010, SE011 |
| CE019 | QuNorth documents chemistry, materials, biology, and optimization as target use cases for the Atom-Microsoft system. | 高 | SE007, SE008, SE010 |
| CE020 | Atom’s homepage and funding release both emphasize on-premises deployment rather than open public-cloud self-service as the current commercialization center of gravity. | 高 | SE001, SE002 |
| CE021 | The Nu Quantum partnership shows Atom extending its product narrative into quantum networking and utility-scale systems integration. | 中 | SE012 |
| CE022 | NERSC’s 2026 neutral-atom proposal call shows that neutral-atom workflow integration with HPC buyers is becoming operational rather than purely speculative. | 中 | SE018 |
| CE023 | Relative to IBM and Google, Atom’s public product story centers on physical-scale neutral-atom hardware plus logical-qubit partnerships rather than an established cloud product catalog. | 中 | SE019, SE020, SE021, SE022 |
| CE024 | Relative to Quantinuum and QuEra, Atom’s public differentiation is the combination of neutral-atom scale and a Microsoft-backed commercial logical-qubit machine. | 中 | SE006, SE023, SE024 |
| CE025 | The Magne description frames the Atom-Microsoft offering as a complete full-stack quantum computer including hardware, algorithms, software, operating system, compiler, control electronics, AI agents, system governance, and cloud connectivity. | 中 | SE009 |
| CE026 | The Microsoft blog says the commercial package was available to order in late 2024 with delivery in 2025. | 中 | SE006 |
| CE027 | Public retained sources do not disclose Atom’s list pricing, realized pricing, or contract SLA structure. | 高 | SE001, SE002, SE006 |
| CE028 | Public retained sources do not disclose SOC 2, ISO 27001, uptime targets, or a public reliability dashboard for Atom systems. | 高 | SE001, SE002 |
| CE029 | Independent skeptical commentary still implies that technical milestones alone do not prove near-term broad commercial utility. | 中 | SE016, SE017 |
| CE030 | Atom’s current product evidence is strongest on architecture and flagship deployments and weaker on repeatable enterprise operations metrics. | 高 | SE001, SE002, SE007, SE016 |
| CE031 | Microsoft Research materials position logical qubits and quantum error correction as core to useful quantum computing, which aligns with Atom’s hardware story. | 中 | SE025 |
| CE032 | Google, IBM, Quantinuum, and QuEra product pages show that Atom competes in a field where multiple modalities claim roadmap credibility, raising the bar for differentiation. | 中 | SE019, SE021, SE023, SE024, SE027, SE028, SE029 |
| CE033 | The product module set visible in retained sources includes AC1000 hardware, Microsoft logical-qubit integration, QuNorth system deployment, and utility-scale networking work with Nu Quantum. | 高 | SE001, SE006, SE007, SE012 |
| CE034 | The strongest public use-case surfaces for Atom are chemistry, materials science, drug-discovery-adjacent workloads, and complex optimization. | 高 | SE002, SE007, SE008 |
| CE035 | DARPA, Microsoft, QuNorth, and NERSC collectively show a product strategy aimed first at high-touch research and sovereign deployments before broader enterprise scale. | 高 | SE005, SE006, SE007, SE018 |
| CE036 | Whether Atom’s public product evidence is enough to support a durable enterprise software layer remains unresolved in retained sources. | 低 | |
| CE037 | Whether Atom can move from flagship sovereign deployments to a broad installed base without disclosing pricing and support metrics remains unresolved. | 低 | |
| CU001 | Atom’s strongest named customer proof is QuNorth, the Nordic initiative that will acquire and operate the Magne system. | 高 | SU006, SU007, SU008, SU009, SU010 |
| CU002 | Atom said in June 2026 that it sold its first commercial on-premises quantum computer to QuNorth in 2025. | 高 | SU002, SU003, SU008 |
| CU003 | QuNorth is backed by EIFO and the Novo Nordisk Foundation with roughly €80 million of initial funding. | 高 | SU006, SU008, SU009, SU010 |
| CU004 | The Magne program is designed to give Nordic researchers and industries priority access to logical-qubit quantum computing. | 高 | SU006, SU008, SU009 |
| CU005 | Public use cases for the Atom-Microsoft-QuNorth system include chemistry, materials science, biological systems, and optimization. | 中 | SU006, SU007, SU008 |
| CU006 | Microsoft and Atom present the commercial system as serving both academia and industry rather than one narrow vertical. | 高 | SU005, SU006 |
| CU007 | DARPA Stage B and the broader QBI program make the U.S. government a critical stakeholder and quasi-customer for Atom’s roadmap validation. | 高 | SU004, SU012, SU013 |
| CU008 | NERSC’s 2026 neutral-atom proposal call shows that national-lab style buyers are building workflows around neutral-atom systems now. | 中 | SU015 |
| CU009 | U.S. and U.K. quantum-policy materials show that sovereign and public-research buyers remain early anchors for quantum-computing demand. | 高 | SU018, SU019, SU020 |
| CU010 | The NSA’s CNSA 2.0 guidance reinforces why governments and security-sensitive institutions are preparing for quantum capability now. | 中 | SU021 |
| CU011 | Atom’s customer base is most visible today in sovereign, research, and high-performance-computing channels rather than broad enterprise adoption. | 高 | SU002, SU006, SU012, SU015 |
| CU012 | The funding release names materials science, pharmaceuticals, energy, and logistics as commercial application areas driving demand for Atom systems. | 高 | SU002, SU003, SU028, SU029 |
| CU013 | Independent market sources continue to identify government, defense, pharma, and advanced-industrial research as the most credible early quantum buyers. | 高 | SU016, SU017, SU022, SU023, SU026, SU027 |
| CU014 | Public retained sources do not disclose Atom’s total active customer count. | 高 | SU001, SU002 |
| CU015 | Public retained sources do not disclose Atom’s NRR, GRR, churn, renewal rates, or contract length. | 高 | SU001, SU002, SU003 |
| CU016 | Public retained sources do not quantify customer concentration, even though the named customer proof is highly concentrated. | 高 | SU002, SU006, SU008 |
| CU017 | The strongest durability proxy is progression from strategic partnership to actual customer deployment, not disclosed renewal cohorts. | 高 | SU005, SU006, SU008 |
| CU018 | Because Magne is a flagship sovereign-backed project, Atom’s public customer evidence currently overweights strategic importance relative to breadth. | 高 | SU006, SU008, SU009 |
| CU019 | The Atom-Microsoft partnership acts as both a delivery channel and a customer-acquisition dependency. | 高 | SU005, SU006, SU007 |
| CU020 | The Department of Commerce and DARPA relationships indicate that government-aligned demand and policy support may be as important as commercial demand in the near term. | 高 | SU003, SU004, SU012, SU013 |
| CU021 | The customer journey for Atom appears to run from strategic ecosystem qualification to flagship deployment rather than from self-serve trial to seat expansion. | 高 | SU005, SU006, SU015, SU025 |
| CU022 | QuEra’s 2026 “proof” framing and broader market commentary imply that quantum vendors increasingly need named deployment evidence to win buyers. | 中 | SU016, SU017, SU025 |
| CU023 | The UK strategy and U.S. reauthorization language both support a public-sector buyer journey that starts with capability-building and testbeds before scaled procurement. | 高 | SU018, SU019 |
| CU024 | Research and Markets and BCG both imply that enterprise quantum demand is real but still mostly pre-mainstream, reinforcing why Atom’s broad customer ramp is likely to be gradual. | 中 | SU022, SU023, SU024, SU026, SU027 |
| CU025 | The strongest expansion upside for Atom is land-and-expand from sovereign or research anchors into adjacent enterprise use cases in chemistry, materials, and optimization. | 高 | SU006, SU007, SU016, SU017 |
| CU026 | The main concentration risk is that one or two flagship programs could dominate public perception before broad recurring demand exists. | 高 | SU006, SU008, SU015 |
| CU027 | The funding release and Microsoft materials suggest Atom is targeting both scientific discovery and commercial industrial users, but named production enterprise users are still sparse. | 高 | SU002, SU005, SU006 |
| CU028 | The Magne timeline points to first operational tasks around early 2027, so some customer adoption evidence is still future-tense rather than realized usage. | 中 | SU006, SU007 |
| CU029 | Independent skeptical commentary remains relevant because long utility timelines can slow enterprise conversion even when flagship customers exist. | 中 | SU014 |
| CU030 | Public evidence is currently strong enough to prove one flagship commercial customer and several government-aligned demand signals, but not enough to prove a diversified installed base. | 高 | SU002, SU006, SU012, SU015 |
| CU031 | Whether QuNorth becomes a repeatable template for additional sovereign or enterprise customers remains unresolved. | 低 | |
| CU032 | Whether any current pilots or partner programs already generate recurring support or services revenue remains unresolved in retained public sources. | 低 | |
| CU033 | Whether Atom’s eventual customer base skews more toward governments or enterprise industry remains unresolved because the market is still early. | 低 | SU016, SU017, SU022, SU023, SU030, SU031 |
| CU034 | The customer chapter should treat Microsoft more as channel and commercialization partner than as end customer. | 高 | SU005, SU006, SU007 |
| CU035 | Policy-backed demand likely lowers near-term buyer-risk perception for Atom but can also raise procurement friction and concentration risk. | 高 | SU012, SU013, SU018, SU019 |
| CR001 | Atom announced more than $230 million of 2025 financing and said $60 million remained contingent on a prerequisite funding condition. | 高 | SR001, SR015 |
| CR002 | TechCrunch reported that Atom's 2025 financing was a convertible note rather than a priced equity round and that management did not disclose the expected valuation or timing of the next equity round. | 中 | SR015 |
| CR003 | Atom entered DARPA's Quantum Benchmarking Initiative in Stage A, a program aimed at testing whether a commercially useful fault-tolerant quantum computer can be built within roughly a decade. | 高 | SR002, SR011 |
| CR004 | After Stage A, DARPA selected Atom for Stage B and Atom said the program can provide up to $15 million over 12 months before later independent verification stages. | 高 | SR003, SR011 |
| CR005 | Atom's public hiring page shows the company is simultaneously trying to design and deliver systems, support global partners, and commercialize neutral-atom platforms, which is evidence of a broad execution load rather than a narrow lab agenda. | 中 | SR004 |
| CR006 | Atom's current FTQC narrative asks investors and customers to underwrite Libra on Amazon Braket in 2028 and a gigaquop-class follow-on system in 2028-2029, so valuation upside remains highly sensitive to roadmap timing. | 高 | SR005, SR006, SR017, SR039 |
| CR007 | Atom's January 2024 roadmap had already targeted 100 logical qubits and more than 10,000 physical qubits in 2026, underscoring that the company has long operated on an unusually aggressive schedule. | 中 | SR036 |
| CR008 | Aquila is currently distributed through Amazon Braket, which makes AWS a major public access channel for Atom's most visible commercial product. | 高 | SR007, SR009, SR038 |
| CR009 | Amazon Braket charges per-shot, per-task, or hourly reservation fees and bills associated AWS services separately, which means Atom's cloud adoption curve is partly mediated by Amazon platform economics rather than by Atom alone. | 高 | SR008, SR007 |
| CR010 | NVIDIA and Atom publicly frame quantum error-correction decoding as a key bottleneck, implying that Atom's FTQC execution depends not only on the QPU but also on external accelerated-computing performance. | 高 | SR010, SR017 |
| CR011 | The named public customer and validation set in this chapter is concentrated in government, HPC, and sovereign programs—DARPA, AIST, NERSC, NQCC, and AWS-linked access—rather than in a broad disclosed enterprise install base. | 高 | SR011, SR012, SR013, SR014, SR035, SR006 |
| CR012 | NERSC's 2026 QCAN call offered up to six projects using Atom systems and positioned the work as research access, which is meaningful proof but not the same thing as diversified production demand. | 中 | SR012 |
| CR013 | NQCC first awarded Atom a multimillion-pound testbed contract and later described Atom's project as an error-corrected neutral-atom testbed, making UK public-program execution a real proof point but also a visible delivery obligation. | 高 | SR013, SR014 |
| CR014 | Atom announced a 6.5 billion JPY AIST contract, roughly $41 million, to deliver an on-prem neutral-atom system alongside the NVIDIA-powered ABCI-Q supercomputer, making one named lighthouse deployment unusually important to the commercial narrative. | 中 | SR035 |
| CR015 | Because public evidence still does not reveal note-conversion mechanics, preferences, or a next-round valuation anchor, the 2025 financing should be read as runway support rather than as clean market-price validation. | 高 | SR001, SR015 |
| CR016 | U.S. export controls that took effect in September 2024 cover quantum computers, related equipment, components, materials, software, and technology. | 高 | SR016, SR017, SR018, SR028 |
| CR017 | The BIS rule also created a general license with annual reporting for certain quantum deemed exports and reexports involving foreign nationals from D:1 and D:5 countries, so cross-border hiring and collaboration can create material compliance overhead. | 高 | SR016, SR017, SR027, SR028 |
| CR018 | The same BIS rule added worldwide national-security and regional-stability controls for newly listed quantum ECCNs and explicitly notes the CFIUS critical-technology context, linking export compliance and foreign-investment sensitivity. | 高 | SR017, SR023, SR024 |
| CR019 | OMB M-23-02 was issued to implement NSM-10, which means U.S. government interest in quantum is tied to a live cryptographic-migration program rather than to generic long-term curiosity. | 高 | SR019, SR033 |
| CR020 | CISA says critical-infrastructure and government network owners and operators should transition toward post-quantum cryptography, raising the procurement bar for vendors that want to serve security-sensitive accounts. | 高 | SR020, SR021, SR032 |
| CR021 | White House, Skadden, and K&L Gates materials say covered contractors will be required to meet certain federal cybersecurity standards and vulnerability-disclosure policies by the end of 2030, creating a future contractor-compliance burden around government-linked deals. | 高 | SR025, SR026, SR032 |
| CR022 | NIST SP 800-171 Rev. 3 sets the baseline for protecting controlled unclassified information in nonfederal systems, which is directly relevant if Atom touches government data or federally controlled environments through on-prem or hybrid deployments. | 高 | SR022, SR025 |
| CR023 | The UK National Security and Investment Act allows government scrutiny of acquisitions in sensitive sectors, so Atom's UK public-program activity sits inside a sovereignty-screening environment even without a disclosed Atom-specific review. | 高 | SR023, SR013 |
| CR024 | NSA says it does not recommend QKD or QC for National Security Systems unless important limitations are overcome, which constrains overbroad quantum-security marketing to federal buyers. | 高 | SR020, SR034 |
| CR025 | PACER, CourtListener, and USPTO search are the obvious public tools for litigation and patent screening, but this chapter did not retain a company-specific cleared dispute record from those systems, so legal comfort still depends on direct diligence rather than on absence-by-default. | 高 | SR029, SR030, SR031 |
| CR026 | Independent legal analyses describe commercial quantum as increasingly shaped by export controls, government contracting, IP, and cross-border research restrictions even before any Atom-specific enforcement record appears. | 高 | SR024, SR027 |
| CR027 | Atom markets secure, direct, supported premium access and public cloud access, but the retained sources do not yet surface third-party security certifications or uptime commitments, leaving enterprise assurance partly unproven in public. | 高 | SR007, SR008, SR038 |
| CR028 | Atom's business model now includes cloud access, on-prem delivery, and application co-design, so execution risk includes deployment, support, and program management rather than only lab physics. | 高 | SR004, SR035, SR038 |
| CR029 | Atom's public FTQC Founders Circle explicitly asks enterprises, HPC centers, and government programs to start multi-year application co-design before the relevant hardware is online, which increases expectation-management risk if milestones slip. | 高 | SR005, SR017, SR018 |
| CR030 | Atom's own 2026 market report says 46% of respondents expect flat quantum budgets in 2026, 44% expect increases, and 10% expect decreases, which is a proof-driven demand backdrop rather than a pure narrative boom. | 中 | SR037 |
| CR031 | The same report says government mandates and grant availability are the top budget driver for 28% of respondents and government and defense are the most likely near-term commercialization segment at 24%, reinforcing public-sector concentration risk. | 高 | SR037, SR011 |
| CR032 | The report also says 62% of organizations factor sovereignty into procurement decisions, which can complicate Atom's go-to-market in UK and other national-program contexts. | 高 | SR023, SR037 |
| CR033 | The same report says specialist workforce shortages are cited by 37% of respondents and are especially acute in QEC, a direct risk to Atom because its roadmap and NVIDIA collaboration both emphasize decoder-heavy fault tolerance. | 高 | SR004, SR010, SR037 |
| CR034 | Observer Research Foundation argues commercially relevant quantum computers are likely at least a decade away and warns hype can create a financial bubble, providing an adverse external lens on aggressive sector roadmaps. | 中 | SR040 |
| CR035 | When Atom's own 2024 and 2026 roadmaps are read against DARPA's 2033 utility-scale benchmark and ORF's skepticism, the company looks credible but still fundamentally milestone-driven rather than de-risked. | 高 | SR003, SR011, SR036, SR040 |
| CR036 | Delay or reprioritization at AWS, NVIDIA, DARPA, AIST, NERSC, or NQCC would hit validation, revenue proof, and roadmap credibility at the same time because those counterparties sit on overlapping critical paths. | 高 | SR006, SR010, SR012, SR013, SR014, SR035 |
| CR037 | The 2025 financing materially extends Atom's runway for manufacturing and FTQC development, but the company still appears financing-dependent because public sources do not disclose cash on hand, burn, or a revenue concentration bridge. | 高 | SR001, SR015, SR037 |
| CR038 | Atom's financing announcement says capital will accelerate development and production of large-scale fault-tolerant quantum computers, which confirms that the business still absorbs significant up-front investment before broad commercial scale is proven. | 高 | SR001, SR036 |
| CR039 | Public customer proof shows Atom can win cloud and on-prem programs, but the retained evidence still does not demonstrate broad recurring renewals beyond the lighthouse set. | 高 | SR007, SR012, SR013, SR014, SR035 |
| CR040 | Atom's public roadmap continues to extend beyond Libra into a later gigaquop system, so investors are underwriting not one but multiple generations of timely execution. | 高 | SR005, SR006, SR017, SR039 |
| CR041 | The strongest current mitigants—DARPA stage-gates, AWS distribution, NVIDIA decoding support, and sovereign or HPC deployments—are real, but each is also a dependency node rather than a fully independent risk offset. | 高 | SR003, SR006, SR010, SR012, SR013, SR014, SR035 |
| CR042 | AIST, NERSC, and NQCC prove Atom has more operating reality than many private peers, yet they also bias public proof toward institutions with long procurement cycles and policy goals instead of broad enterprise repetition. | 高 | SR012, SR013, SR014, SR035, SR037 |
| CR043 | What exact ECCNs, product classifications, license applications, or deemed-export logs does Atom use for Aquila, Gemini, Libra, software, and associated know-how? | 低 | |
| CR044 | What cash balance, monthly burn, and milestone-to-milestone runway bridge can management prove through Libra and the later gigaquop program? | 低 | |
| CR045 | What share of revenue, backlog, or pipeline depends on AIST and other government or HPC accounts versus repeat commercial customers? | 低 | |
| CR046 | What third-party security certifications, uptime SLAs, or government-contractor compliance attestations can Atom show for sensitive deployments? | 低 | |
| CR047 | Has any core Atom IP been challenged, licensed restrictively, or encumbered in a way not visible from public patent and docket-search tools? | 低 | |
| CV001 | Forge Global lists Atom Computing at a $2.06 billion Series D-1 valuation in July 2026. | 中 | SV001 |
| CV002 | Forge records a July 8, 2026 Series D-1 financing of about $89.07 million at $3.92 per share. | 中 | SV001 |
| CV003 | Forge flags Atom with IPO Mentioned, Confidential Filing, and S-1 Filed status markers. | 中 | SV001 |
| CV004 | Atom said in June 2026 that it had raised more than $300 million in total, including a $100 million Series C and a planned $100 million Department of Commerce commitment. | 中 | SV002 |
| CV005 | The June 2026 Series C was led by Third Point Ventures with participation from DCVC and Cisco Investments. | 中 | SV002 |
| CV006 | DCVC said the Department of Commerce support included a $100 million equity stake in Atom Computing. | 中 | SV003 |
| CV007 | NIST said the Department of Commerce signed nine letters of intent totaling $2.013 billion to accelerate U.S. quantum leadership. | 中 | SV031 |
| CV008 | Atom disclosed that it sold its first commercial on-premises quantum computer to QuNorth in 2025. | 中 | SV002 |
| CV009 | Independent QuNorth coverage says the Magne system is a 50-logical-qubit program for Nordic users and is expected to be operational by early 2027. | 高 | SV008, SV009, SV010 |
| CV010 | Microsoft and Atom announced a commercially available machine that created 24 entangled logical qubits and computed on 28 logical qubits. | 中 | SV007 |
| CV011 | DARPA selected Atom for QBI Stage B, and the program is explicitly designed to test whether utility-scale quantum computing can be achieved by 2033. | 高 | SV004, SV005, SV006 |
| CV012 | Public sources reviewed for this chapter do not disclose Atom’s revenue, ARR, gross margin, or recurring-revenue mix. | 中 | SV001, SV002, SV008 |
| CV013 | Because Atom’s first named sale is recent and its operating economics remain undisclosed, a revenue-multiple framework is not supportable as the main valuation method today. | 中 | SV002, SV011, SV030 |
| CV014 | A milestone-and-strategic-value framework fits Atom better than a current-fundamentals multiple because the market is underwriting technical proof, sovereign demand, and platform scarcity. | 中 | SV001, SV007, SV011, SV030 |
| CV015 | Atom’s $2.06 billion valuation against more than $300 million raised implies roughly a 6.9x price-to-raised ratio. | 中 | SV001, SV002 |
| CV016 | Forge’s funding-history data implies about a 2.9x markup from the $714.4 million Series C-1 valuation in June 2026 to the $2.06 billion D-1 mark in July 2026. | 中 | SV001 |
| CV017 | At a $2.06 billion entry valuation, a 2x gross outcome requires about a $4.1 billion exit and a 3x gross outcome requires about a $6.2 billion exit before dilution. | 中 | SV001 |
| CV018 | IonQ reported $64.7 million of Q1 2026 revenue and raised its full-year 2026 revenue guidance to $260 million to $270 million. | 高 | SV011, SV012 |
| CV019 | Stock Analysis listed IonQ’s market cap at about $12.98 billion on July 17, 2026. | 中 | SV014 |
| CV020 | IonQ’s public valuation is therefore roughly 6.3x Atom’s current private mark while also carrying visible revenue and backlog. | 中 | SV001, SV011, SV014 |
| CV021 | Reuters reported through U.S. News that Quantinuum targeted a valuation of up to $14.3 billion in its June 2026 IPO. | 中 | SV015 |
| CV022 | Reuters reported through CNBC that Quantinuum ultimately raised about $1.68 billion in its IPO. | 中 | SV016 |
| CV023 | Quantinuum’s own product pages show a more mature commercial surface, including both cloud access and on-premises offerings for its trapped-ion systems. | 中 | SV017, SV018 |
| CV024 | Reuters reported through U.S. News that PsiQuantum raised $1 billion at a $7 billion valuation in September 2025. | 中 | SV019 |
| CV025 | PsiQuantum’s official technology page continues to emphasize utility-scale manufacturing rather than near-term broad revenue disclosure. | 中 | SV020 |
| CV026 | Stock Analysis listed D-Wave Quantum’s market cap at about $6.20 billion on July 17, 2026. | 中 | SV021 |
| CV027 | Stock Analysis listed Rigetti Computing’s market cap at about $4.69 billion on July 17, 2026. | 中 | SV022 |
| CV028 | QuEra disclosed a financing of more than $230 million in February 2025 and AWS says it aims to bring QuEra fault-tolerant systems to the cloud starting in 2028. | 中 | SV023, SV024 |
| CV029 | Oratomic announced a $300 million Series A in July 2026, showing strong investor appetite for neutral-atom challengers. | 中 | SV025, SV027 |
| CV030 | The free-source Oratomic pages reviewed here do not publish a precise post-money valuation, so Atom-to-Oratomic pricing comparisons are directional rather than precise. | 中 | SV025, SV026, SV027 |
| CV031 | Crunchbase said public pure-play quantum companies were collectively valued above $36 billion in 2026 even as startup funding slowed from 2025 peaks. | 中 | SV028 |
| CV032 | S&P Global said listed quantum stocks rallied after the U.S. $2 billion support program even though sector revenues remain limited. | 中 | SV029 |
| CV033 | Lambda Finance argues that public quantum pure-plays trade more like options on breakthroughs than like software businesses. | 中 | SV030 |
| CV034 | Lambda Finance warns that quantum pure-play position sizing should respect 50% to 70% drawdown risk. | 中 | SV030 |
| CV035 | Government support reduces near-term financing risk and validates policy importance, but it does not prove diversified customer demand or durable recurring revenue. | 中 | SV003, SV031, SV009 |
| CV036 | Forge’s confidential-filing signal creates IPO optionality, but the retained public evidence still lacks a public S-1, a roadshow range, or audited Atom financial statements. | 中 | SV001 |
| CV037 | The most credible IPO path is to deliver QuNorth, add another named customer, and show visible revenue progression before testing public markets. | 中 | SV001, SV009, SV015, SV018 |
| CV038 | A bull case requires repeat sovereign or enterprise system sales, successful logical-qubit commercialization, and market receptivity closer to IonQ or Quantinuum than to weaker public comps. | 中 | SV007, SV015, SV018, SV019 |
| CV039 | A base case is that Atom remains strategically important and raises again around or modestly above the current mark after QuNorth and additional government-backed milestones. | 中 | SV001, SV009, SV031 |
| CV040 | A bear case is that revenue stays concentrated, QuNorth slips, or public-comp sentiment weakens enough to produce a flat round or down round. | 中 | SV009, SV029, SV030 |
| CV041 | Atom’s $2.06 billion mark sits well below Quantinuum, IonQ, and PsiQuantum, but above many earlier-stage private neutral-atom financings and therefore places it in the upper middle of the sector’s private valuation stack. | 中 | SV001, SV015, SV019, SV023 |
| CV042 | That relative positioning makes the current valuation plausible as a scarcity premium, but stretched on disclosed fundamentals because Atom has only first-sale evidence and no public run-rate revenue. | 中 | SV001, SV002, SV008, SV030 |
| CV043 | The recommendation is research-more rather than pass because Atom combines real technical and policy signal with still-limited fundamental disclosure. | 中 | SV001, SV002, SV007, SV009, SV031 |
| CV044 | Confidence should be medium because the valuation mark is concrete via Forge COI data, but revenue quality, cap-table terms, and customer breadth are still private. | 中 | SV001, SV002, SV026 |
| CV045 | Risk rating should remain high because commercialization timing, capital intensity, and quantum multiple volatility can all impair private-mark durability. | 中 | SV006, SV029, SV030, SV033 |
| CV046 | Valuation stance is stretched rather than outright absurd because Atom already has a first commercial sale, logical-qubit commercialization proof, and unusually strong government support. | 中 | SV001, SV007, SV009, SV031 |
| CV047 | The highest-value diligence asks are the D-1 term sheet, investor roster, preference stack, QuNorth contract economics, and evidence of a second named paid deployment. | 中 | SV001, SV009, SV031 |
| CV048 | A thesis-break trigger would be any future round below the D-1 mark or with punitive preference terms, especially if Microsoft and DARPA credibility still fails to broaden paying customer adoption. | 中 | SV001, SV007, SV011 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Atom Computing | Home - Atom Computing | Atom Computing builds highly scalable, gate-based quantum computers with arrays of optically-trapped neutral atoms. |
| SO002 | Atom Computing | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | Atom Computing has raised over $300 million to date, including a $100 million Series C round led by Third Point Ventures and a planned $100 million from the U.S. Department of Commerce. |
| SO003 | PR Newswire | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SO004 | DCVC | A bold and timely bet: the U.S. Government invests in quantum… | The U.S. Department of Commerce’s $2 billion investment across nine quantum companies materially involves four DCVC portfolio companies, and it includes a $100 million equity stake in Atom Computing. |
| SO005 | Forge Global | Atom Computing IPO: Investment Opportunities & Pre-IPO Valuations - Forge | Post-Money Valuation represents the estimated valuation based on company-submitted Certificates of Incorporations (COIs). |
| SO006 | PR Newswire | Quantum Startup Atom Computing First to Exceed 1,000 Qubits | Atom Computing announced it has created a 1,225-site atomic array, currently populated with 1,180 qubits. |
| SO007 | PR Newswire | Atom Computing selected by DARPA for the next stage of exploring near-term utility-scale quantum computing with neutral atoms | DARPA announced that Atom Computing has been selected for Stage B of the QBI program. |
| SO008 | Microsoft Azure Quantum Blog | Microsoft and Atom Computing offer a commercial quantum machine with the largest number of entangled logical qubits on record | Microsoft and Atom Computing have made rapid progress in reliable quantum computing by creating and entangling 24 logical qubits. |
| SO009 | The Quantum Insider | Microsoft and Atom Computing Partner on Level 2 Quantum System for Nordic Users | The system will be built by Atom Computing and Microsoft, combining neutral atom hardware with advanced quantum software, and is expected to be operational by early 2027. |
| SO010 | Data Center Dynamics | Microsoft and Atom Computing to build "world’s most powerful quantum computer" in Denmark | |
| SO011 | Quantum Computing Report | Denmark's QuNorth to Acquire 50-Logical-Qubit Magne Quantum Computer from Atom Computing and Microsoft | Magne is characterized as a Level 2 quantum computer... specified to consist of 50 logical qubits and 1,225 physical neutral-atom qubits. |
| SO012 | Novo Nordisk Foundation | EIFO and the Novo Nordisk Foundation Acquire the World’s Most Powerful Quantum Computer | |
| SO013 | Novo Nordisk Foundation | New quantum computer with great potential to boost Nordic research and innovation | |
| SO014 | Accura | Atom Computing and Microsoft to deliver the world’s most powerful quantum computer to Danish quantum initiative | |
| SO015 | HPCwire | Atom Computing Raises $100M Series C to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SO016 | The Quantum Insider | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SO017 | DARPA | Stage B selection | |
| SO018 | DARPA | QBI | DARPA | |
| SO019 | arXiv | Fault-tolerant quantum computation with a neutral atom processor | We demonstrate the entanglement of 24 logical qubits encoded into 48 atoms ... and implement the Bernstein-Vazirani algorithm with up to 28 logical qubits encoded into 112 atoms. |
| SO020 | Forbes | Atom Computing Announces Record-Breaking 1,225-Qubit Quantum Computer | |
| SO021 | Observer Research Foundation | Quantum Computing: Separating Hype from Reality | Commercially relevant applications of quantum computing are likely at least a decade away, as current quantum computers are highly error-prone and can only perform simple calculations. |
| SO022 | Boston Consulting Group | The Long-Term Forecast for Quantum Computing Still Looks Bright | |
| SO023 | NERSC | NERSC Issues 2026 Call for Proposals for Neutral Atom-Based Quantum Computing | |
| SO024 | U.S. Senate Commerce Committee | National Quantum Initiative Reauthorization Act of 2026 – Section Summary | |
| SO025 | AIP FYI | DOE Launches ‘Quantum Genesis’ Initiative | |
| SM001 | McKinsey & Company | Quantum Technology Monitor 2026: A Commercial Tipping Point | |
| SM002 | McKinsey & Company | Quantentechnologie wird zum Milliardenmarkt | |
| SM003 | Boston Consulting Group | The Long-Term Forecast for Quantum Computing Still Looks Bright | Quantum computing today provides no tangible advantage over classical computing in either commercial or scientific applications. |
| SM004 | Quantum Economic Development Consortium | 2026 Market Forecast: Quantum Computing | |
| SM005 | Quantum Economic Development Consortium | State of the Global Quantum Industry 2026 | |
| SM006 | Research and Markets | Quantum Computing Market Report 2026 | |
| SM007 | UK Government | National Quantum Strategy (accessible webpage) | |
| SM008 | US Senate Commerce Committee | National Quantum Initiative Reauthorization Act of 2026 – Section Summary | |
| SM009 | NERSC | NERSC Issues 2026 Call for Proposals for Neutral Atom-Based Quantum Computing | |
| SM010 | National Security Agency | Announcing the Commercial National Security Algorithm Suite 2.0 | |
| SM011 | American Institute of Physics | DOE Launches ‘Quantum Genesis’ Initiative | |
| SM012 | Atom Computing | The Quantum Market Shifts from Hype to Proof | |
| SM013 | Atom Computing | Global Quantum Budgets Set to Surge by ~20% | |
| SM014 | Atom Computing | Study: Companies Demand Reliable Results Rather Than Visions When It Comes to Quantum Computing | |
| SM015 | Atom Computing | BCG X & Atom Computing Forge Quantum Partnership | |
| SM016 | Atom Computing | AIST Selects Atom’s Neutral-Atom Quantum Computer | |
| SM017 | Atom Computing | AIST and Atom Sign Memorandum | |
| SM018 | Amazon Web Services | AWS Deepens Strategic Collaboration with Atom to Bring Fault-Tolerant Quantum Computing to Amazon Braket | |
| SM019 | Amazon Web Services | Atom - Quantum Computing Hardware Provider - AWS Braket | |
| SM020 | Atom Computing | World’s Largest Publicly Available Quantum Computer on Amazon Braket | |
| SM021 | Atom Computing | DARPA Selects Atom for Stage B of Quantum Benchmarking Initiative | |
| SM022 | Atom Computing | Atom to Bring Expanded Quantum Capability to NERSC | |
| SM023 | Atom Computing | Atom and Pawsey Partner to Drive Innovation | |
| SM024 | Atom Computing | Atom Computing and Quantum Intelligence Corp Team Up to Accelerate Drug Discovery with Quantum Computing | |
| SM025 | Atom Computing | Quantum Bio | |
| SM026 | Atom Computing | Kipu Quantum and Atom Collaborate | |
| SM027 | Data Center Knowledge | JPMorgan, OQC, and AMD Plan Quantum AI Data Center for Finance | |
| SM028 | The Quantum Insider | Top Global Banks Exploring Quantum Technologies in 2026 | |
| SM029 | Moody’s | Decoding Quantum Hype: What Big Tech Is Announcing | The delay in getting commercial value on real-world applications from quantum computing can be attributed to one overarching fact: The hardware is not ready yet. |
| SM030 | HPCwire | D-Wave Reports Quantum Supremacy; Stirs Immediate Challenge (and Rebuttal) | |
| SM031 | IBM Quantum | Quantum Advantage Tracker: the race to advantage | |
| SM032 | Atom Computing | Atom Announces 2028 Fault-Tolerant Quantum Computer and Expanded Multi-Year Strategic Collaboration with AWS | |
| SM033 | Atom Computing | Atom Computing and Deloitte Form Alliance to Accelerate Enterprise Adoption of Neutral-Atom Quantum Computing | |
| SM034 | Atom Computing | ICSC & Atom Computing Launch Partnership | |
| SM035 | The Quantum Insider | AIST Selects Atom’s Neutral-Atom Quantum Computer to Be Installed Alongside NVIDIA-Powered ABCI-Q Supercomputer | |
| SM036 | Harvard Gazette | Harvard researchers create first logical quantum processor | |
| SM037 | Harvard Gazette | Self-correcting quantum computers within reach? | |
| SP001 | Atom | Aquila | 256-qubit Quantum Computer | Available via Amazon Braket or via Premium Access: Secure, direct, supported environment with priority bookings. |
| SP002 | Atom | On-Premises Quantum Computers | Atom | Deploy Atom’s neutral-atom quantum computers on-premise for secure, continuous access and seamless integration with your HPC. |
| SP003 | Atom | Our Quantum Roadmap | Our roadmap recognizes this fact and doesn’t project systems still under development, instead, it outlines the products we feel are scientifically viable based on published, peer-reviewed research. |
| SP004 | Atom | Atom Completes $230 M Financing | Atom Computing, the leader in neutral-atom quantum computing, today announced it has successfully completed a financing of more than $230 million. |
| SP005 | Amazon Web Services | Quantum computers | Amazon Braket | Atom quantum computers are based on Rydberg atom qubits, which utilize internal states of individual Rubidium atoms that are trapped and manipulated using laser beams. |
| SP006 | Amazon Web Services | Amazon Braket Pricing | Amazon Braket offers three pricing components for on-demand use of a quantum computer (QPU): a per-shot fee and a per-task fee or a single hourly reservation fee. |
| SP007 | IBM | IBM Quantum Computing | Products and services | Open Plan... Pay-As-You-Go Plan... Flex Plan... Premium Plan... On-Prem Plan. |
| SP008 | IBM | IBM Quantum Computing | Hardware and roadmap | Quantum computers (>100q): 30+ ... Available qubits: 2300+ ... Circuits ran: 3.9T+ ... Availability (% uptime): 97%. |
| SP009 | Google Quantum AI | Willow Early Access Program | Google Quantum AI | Selected applicants to the Willow Early Access Program gain exclusive access to this hardware—which is not yet available to the public. |
| SP010 | Google Quantum AI | Cirq | Google Quantum AI | Cirq is a Python software library for writing, manipulating, and optimizing quantum circuits, and then running them on quantum computers and quantum simulators. |
| SP011 | Meet Willow, our state-of-the-art quantum chip | As part of Google Research, our team has charted a long-term roadmap, and Willow moves us significantly along that path towards commercially relevant applications. | |
| SP012 | IonQ | Quantum Cloud Services - IonQ Quantum Cloud | The IonQ Quantum Cloud offers various access models to meet your needs. Choose between on-demand access to run workloads flexibly or reserve time on a QPU for large-scale scheduled workloads. |
| SP013 | IonQ | IonQ | Roadmap | 2026: 100-256+ physical qubits ... 12 Logical qubits. |
| SP014 | IonQ | IonQ Forte Enterprise: Quantum Computer for Data Centers | IonQ Forte Enterprise’s installation specs are designed to be met by the typical, modern data center. |
| SP015 | IonQ | IonQ Posts Q1 2026 Earnings with Record Revenue | The company reported Record GAAP Revenues of $64.7 Million, Representing 755% Year-On-Year Growth. |
| SP016 | Quantinuum | Our Trapped Ion Quantum Computers | Purchase a subscription directly with Quantinuum ... Purchase a subscription on Microsoft Azure with access to Quantinuum Systems. |
| SP017 | Quantinuum | Helios | Quantinuum's Quantum Computers | The Helios platform is now available to customers through Quantinuum’s cloud service and on-premises offering. |
| SP018 | Quantinuum | Quantinuum Documentation | Build circuits manually or import from other instruction formats (qiskit, OpenQASM etc). |
| SP019 | Microsoft Learn | Pricing Plans for Azure Quantum Providers - Azure Quantum | Quantinuum provides two subscription plans: Standard and Premium ... Standard Plan: USD125,000/Month ... Premium Plan: USD175,000/Month. |
| SP020 | PsiQuantum | Technology — PsiQuantum | PsiQuantum’s wafers are now built by the thousands, at the highest possible level of technical maturity — in a high-volume, commercial semiconductor foundry. |
| SP021 | Nature | A manufacturable platform for photonic quantum computing | |
| SP022 | Rigetti Computing | Building scalable, innovative quantum systems | Rigetti pioneered hybrid quantum-classical computation with its Quantum Cloud Services platform, which has evolved to support ultra-low latency connectivity—less than one millisecond—between a customer’s high-performance classical hardware and Rigetti QPUs. |
| SP023 | Rigetti Computing | Novera | The Novera QPU is available to ship immediately. Allow 4-6 weeks for delivery once your order has been confirmed and shipping logistics are finalized. |
| SP024 | D-Wave | Annealing & Gate-Model Quantum Computing Systems | D-Wave | Annealing quantum computing is available today for real-world optimization and hybrid applications. |
| SP025 | D-Wave | The Leap™ Quantum Cloud Service | D-Wave | With 99.9% uptime and availability, and QPUs with subsecond response times, the Leap service has helped D-Wave customers address hundreds of millions of business and research problems. |
| SP026 | D-Wave | Florida Atlantic University Signs $20M Agreement to Purchase Advantage2 Quantum Computer | The agreement represents a $20 million commitment from FAU, aiming to accelerate and solidify the state of Florida’s position as a leader in quantum computing. |
| SP027 | NVIDIA | Quantum Computing Solutions from NVIDIA | Turning QPUs into useful quantum computers means integrating them with state-of-the-art AI supercomputers. |
| SP028 | IEEE Spectrum | Neutral Atom Quantum Computing: 2026's Big Leap | If someone says quantum computers are commercially useful today, I say I want to have what they’re having. |
| SP029 | Atom Computing | Home - Atom Computing | 1,200+ Fully-Connected Qubits ... Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers. |
| SP030 | Pasqal | Home - Pasqal | Pasqal is entering a new phase of development with new financing expected of at least €340 million, in anticipation of its public listing. |
| SI001 | Atom Computing | Home - Atom Computing | Atom Computing builds highly scalable, gate-based quantum computers with arrays of optically-trapped neutral atoms. |
| SI002 | Atom Computing | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | Atom Computing has raised over $300 million to date, including a $100 million Series C round led by Third Point Ventures and a planned $100 million from the U.S. Department of Commerce. |
| SI003 | PR Newswire | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SI004 | DCVC | A bold and timely bet: the U.S. Government invests in quantum… | The U.S. Department of Commerce’s $2 billion investment across nine quantum companies materially involves four DCVC portfolio companies, and it includes a $100 million equity stake in Atom Computing. |
| SI005 | Forge Global | Atom Computing IPO: Investment Opportunities & Pre-IPO Valuations - Forge | Post-Money Valuation represents the estimated valuation based on company-submitted Certificates of Incorporations (COIs). |
| SI006 | PR Newswire | Atom Computing selected by DARPA for the next stage of exploring near-term utility-scale quantum computing with neutral atoms | DARPA announced that Atom Computing has been selected for Stage B of the QBI program. |
| SI007 | Microsoft Azure Quantum Blog | Microsoft and Atom Computing offer a commercial quantum machine with the largest number of entangled logical qubits on record | Microsoft and Atom Computing have made rapid progress in reliable quantum computing by creating and entangling 24 logical qubits. |
| SI008 | The Quantum Insider | Microsoft and Atom Computing Partner on Level 2 Quantum System for Nordic Users | The system will be built by Atom Computing and Microsoft, combining neutral atom hardware with advanced quantum software, and is expected to be operational by early 2027. |
| SI009 | Data Center Dynamics | Microsoft and Atom Computing to build "world’s most powerful quantum computer" in Denmark | |
| SI010 | Quantum Computing Report | Denmark's QuNorth to Acquire 50-Logical-Qubit Magne Quantum Computer from Atom Computing and Microsoft | Magne is characterized as a Level 2 quantum computer... specified to consist of 50 logical qubits and 1,225 physical neutral-atom qubits. |
| SI011 | Novo Nordisk Foundation | EIFO and the Novo Nordisk Foundation Acquire the World’s Most Powerful Quantum Computer | |
| SI012 | Novo Nordisk Foundation | New quantum computer with great potential to boost Nordic research and innovation | |
| SI013 | Accura | Atom Computing and Microsoft to deliver the world’s most powerful quantum computer to Danish quantum initiative | |
| SI014 | HPCwire | Atom Computing Raises $100M Series C to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SI015 | The Quantum Insider | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SI016 | DARPA | Stage B selection | |
| SI017 | DARPA | QBI | DARPA | |
| SI018 | Observer Research Foundation | Quantum Computing: Separating Hype from Reality | Commercially relevant applications of quantum computing are likely at least a decade away, as current quantum computers are highly error-prone and can only perform simple calculations. |
| SI019 | Securities and Exchange Commission | IonQ Form 10-K for fiscal year 2025 | |
| SI020 | Securities and Exchange Commission | Rigetti Computing 2025 Form 10-K | |
| SI021 | Securities and Exchange Commission | D-Wave Quantum 2025 Form 10-K | |
| SI022 | Stock Analysis | IonQ, Inc. (IONQ) Financials & Income Statement | |
| SI023 | CompaniesMarketCap | IonQ (IONQ) - Market capitalization | |
| SI024 | CompaniesMarketCap | Rigetti Computing (RGTI) - Market capitalization | |
| SI025 | The Quantum Insider | Quantinuum to Debut on Nasdaq After Raising $1.68 Billion in IPO | |
| SI026 | Sacra | PsiQuantum valuation, funding & news | |
| SE001 | Atom Computing | Home - Atom Computing | Atom Computing builds highly scalable, gate-based quantum computers with arrays of optically-trapped neutral atoms. |
| SE002 | Atom Computing | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | Atom Computing has raised over $300 million to date, including a $100 million Series C round led by Third Point Ventures and a planned $100 million from the U.S. Department of Commerce. |
| SE003 | PR Newswire | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SE004 | PR Newswire | Quantum Startup Atom Computing First to Exceed 1,000 Qubits | Atom Computing announced it has created a 1,225-site atomic array, currently populated with 1,180 qubits. |
| SE005 | PR Newswire | Atom Computing selected by DARPA for the next stage of exploring near-term utility-scale quantum computing with neutral atoms | DARPA announced that Atom Computing has been selected for Stage B of the QBI program. |
| SE006 | Microsoft Azure Quantum Blog | Microsoft and Atom Computing offer a commercial quantum machine with the largest number of entangled logical qubits on record | Microsoft and Atom Computing have made rapid progress in reliable quantum computing by creating and entangling 24 logical qubits. |
| SE007 | The Quantum Insider | Microsoft and Atom Computing Partner on Level 2 Quantum System for Nordic Users | The system will be built by Atom Computing and Microsoft, combining neutral atom hardware with advanced quantum software, and is expected to be operational by early 2027. |
| SE008 | Data Center Dynamics | Microsoft and Atom Computing to build "world’s most powerful quantum computer" in Denmark | |
| SE009 | Quantum Computing Report | Denmark's QuNorth to Acquire 50-Logical-Qubit Magne Quantum Computer from Atom Computing and Microsoft | Magne is characterized as a Level 2 quantum computer... specified to consist of 50 logical qubits and 1,225 physical neutral-atom qubits. |
| SE010 | Novo Nordisk Foundation | EIFO and the Novo Nordisk Foundation Acquire the World’s Most Powerful Quantum Computer | |
| SE011 | Novo Nordisk Foundation | New quantum computer with great potential to boost Nordic research and innovation | |
| SE012 | Accura | Atom Computing and Microsoft to deliver the world’s most powerful quantum computer to Danish quantum initiative | |
| SE013 | DARPA | Stage B selection | |
| SE014 | DARPA | QBI | DARPA | |
| SE015 | arXiv | Fault-tolerant quantum computation with a neutral atom processor | We demonstrate the entanglement of 24 logical qubits encoded into 48 atoms ... and implement the Bernstein-Vazirani algorithm with up to 28 logical qubits encoded into 112 atoms. |
| SE016 | Observer Research Foundation | Quantum Computing: Separating Hype from Reality | Commercially relevant applications of quantum computing are likely at least a decade away, as current quantum computers are highly error-prone and can only perform simple calculations. |
| SE017 | Boston Consulting Group | The Long-Term Forecast for Quantum Computing Still Looks Bright | |
| SE018 | NERSC | NERSC Issues 2026 Call for Proposals for Neutral Atom-Based Quantum Computing | |
| SE019 | IBM | IBM Quantum Computing | Hardware and roadmap | |
| SE020 | IBM | IBM Quantum Computing | Products and services | |
| SE021 | Google Quantum AI | Quantum Computer | Google Quantum AI | |
| SE022 | Google Quantum AI | Willow Early Access Program | |
| SE023 | Quantinuum | Our Trapped Ion Quantum Computers | System Model H2 | |
| SE024 | QuEra | Aquila | 256-qubit Quantum Computer | |
| SE025 | Microsoft Research | Quantum Computing - Microsoft Research | |
| SE026 | Microsoft Azure | Azure Quantum Computing | Microsoft Azure | |
| SE027 | IBM | IBM lays out clear path to fault-tolerant quantum computing | |
| SE028 | Meet Willow, our state-of-the-art quantum chip | ||
| SE029 | Quantinuum | Quantinuum Nexus | Access the world’s most powerful quantum computing stack | |
| SU001 | Atom Computing | Home - Atom Computing | Atom Computing builds highly scalable, gate-based quantum computers with arrays of optically-trapped neutral atoms. |
| SU002 | Atom Computing | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | Atom Computing has raised over $300 million to date, including a $100 million Series C round led by Third Point Ventures and a planned $100 million from the U.S. Department of Commerce. |
| SU003 | PR Newswire | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SU004 | PR Newswire | Atom Computing selected by DARPA for the next stage of exploring near-term utility-scale quantum computing with neutral atoms | DARPA announced that Atom Computing has been selected for Stage B of the QBI program. |
| SU005 | Microsoft Azure Quantum Blog | Microsoft and Atom Computing offer a commercial quantum machine with the largest number of entangled logical qubits on record | Microsoft and Atom Computing have made rapid progress in reliable quantum computing by creating and entangling 24 logical qubits. |
| SU006 | The Quantum Insider | Microsoft and Atom Computing Partner on Level 2 Quantum System for Nordic Users | The system will be built by Atom Computing and Microsoft, combining neutral atom hardware with advanced quantum software, and is expected to be operational by early 2027. |
| SU007 | Data Center Dynamics | Microsoft and Atom Computing to build "world’s most powerful quantum computer" in Denmark | |
| SU008 | Quantum Computing Report | Denmark's QuNorth to Acquire 50-Logical-Qubit Magne Quantum Computer from Atom Computing and Microsoft | Magne is characterized as a Level 2 quantum computer... specified to consist of 50 logical qubits and 1,225 physical neutral-atom qubits. |
| SU009 | Novo Nordisk Foundation | EIFO and the Novo Nordisk Foundation Acquire the World’s Most Powerful Quantum Computer | |
| SU010 | Novo Nordisk Foundation | New quantum computer with great potential to boost Nordic research and innovation | |
| SU011 | Accura | Atom Computing and Microsoft to deliver the world’s most powerful quantum computer to Danish quantum initiative | |
| SU012 | DARPA | Stage B selection | |
| SU013 | DARPA | QBI | DARPA | |
| SU014 | Observer Research Foundation | Quantum Computing: Separating Hype from Reality | Commercially relevant applications of quantum computing are likely at least a decade away, as current quantum computers are highly error-prone and can only perform simple calculations. |
| SU015 | NERSC | NERSC Issues 2026 Call for Proposals for Neutral Atom-Based Quantum Computing | |
| SU016 | McKinsey & Company | Quantum Technology Monitor 2026: A Commercial Tipping Point | |
| SU017 | Quantum Economic Development Consortium | 2026 Market Forecast: Quantum Computing | |
| SU018 | UK Government | National Quantum Strategy (accessible webpage) | |
| SU019 | U.S. Senate Commerce Committee | National Quantum Initiative Reauthorization Act of 2026 – Section Summary | |
| SU020 | American Institute of Physics | DOE Launches ‘Quantum Genesis’ Initiative | |
| SU021 | National Security Agency | Announcing the Commercial National Security Algorithm Suite 2.0 | |
| SU022 | Boston Consulting Group | Quantum Computing On Track to Create Up to $850 Billion of Economic Value By 2040 | |
| SU023 | Research and Markets | Quantum Computing Market Report 2026 | |
| SU024 | BusinessWire / ResearchAndMarkets | Quantum Computing Market Report 2026-2046: Quantum Funding Rounds Surpass $50M Average as Commercial Viability Accelerates | |
| SU025 | QuEra Computing | The Quantum Market Shifts from Hype to Proof | |
| SU026 | Quantum Economic Development Consortium | State of the Global Quantum Industry 2026 | |
| SU027 | PostQuantum | McKinsey Quantum Monitor 2026: Tipping Point? | |
| SU028 | The Quantum Insider | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SU029 | HPCwire | Atom Computing Raises $100M Series C to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SU030 | Google Quantum AI | Willow Early Access Program | |
| SU031 | Quantinuum | Quantinuum Nexus | Access the world’s most powerful quantum computing stack | |
| SU032 | Capital.com | IonQ, Inc. (IONQ) Market Cap – July 2026 Update | |
| SU033 | Capital.com | Rigetti Computing Market Cap – July 2026 Update | |
| SU034 | Yahoo Finance / Reuters | Quantum AI startup SandboxAQ valued at $5.3 billion after $300 million fundraising | |
| SU035 | TechCrunch | Eric Schmidt's SandboxAQ aims for $5B valuation for its AI/quantum Google moonshot | |
| SU036 | Yahoo Finance / Reuters | Quantum computing startup QuEra closes $230 million funding round | |
| SR001 | Atom Computing | Atom Completes $230 M Financing | Of the $230M, $60 million will be received in the near future upon satisfying a prerequisite funding condition, currently in progress. |
| SR002 | Atom Computing | Atom Joins DARPA’s Quantum Benchmarking Initiative | DARPA QBI is a multi-stage program which aims to determine whether it is possible to build a commercially useful, fault-tolerant quantum computer within a decade. |
| SR003 | PR Newswire / Atom Computing | DARPA Selects Atom for Stage B of Quantum Benchmarking Initiative (QBI) | Stage B narrows the field and provides up to $15M over 12 months to validate Atom's baseline R&D plan before independent hardware verification and validation in Stage C. |
| SR004 | Atom Computing | Careers at Atom | Join us to design and deliver groundbreaking neutral-atom computer systems, support global partners in applying them, and help bring quantum computing from the lab to the world. |
| SR005 | PR Newswire / Atom Computing | Atom Unveils Gigaquop-Class Fault-Tolerant Roadmap and Invites Organizations to Co-Design Quantum Applications | The company is inviting enterprises, HPC centers, and government programs to co-design applications for fault-tolerant quantum hardware before it comes online. |
| SR006 | Amazon Web Services | AWS Deepens Strategic Collaboration with Atom to Bring Fault-Tolerant Quantum Computing to Amazon Braket | Today, we are announcing an expanded strategic collaboration with Atom Computing to bring Libra ... to Amazon Braket customers. |
| SR007 | Amazon Web Services | Atom - Quantum Computing Hardware Provider - AWS Braket | Aquila is Atom’s first generation of quantum processing units (QPU) available on Amazon Braket. It operates up to 256 qubits in analog mode. |
| SR008 | Amazon Web Services | Amazon Braket Pricing | Amazon Braket offers three pricing components for on-demand use of a quantum computer (QPU): a per-shot fee and a per-task fee or a single hourly reservation fee. |
| SR009 | Amazon Web Services | Submit an analog program using Atom Aquila | This page provides a comprehensive documentation about the capabilities of the Aquila machine from Atom. |
| SR010 | NVIDIA | NVIDIA and Atom Decode Quantum Errors with AI | Decoding is computationally challenging and is one of the primary bottlenecks of QEC today. |
| SR011 | DARPA | QBI | DARPA | QBI is designed to rigorously verify and validate whether any quantum-computing approach can achieve utility-scale operation — meaning its computational value exceeds its cost. |
| SR012 | NERSC | NERSC Issues 2026 Call for Proposals for Neutral Atom-Based Quantum Computing | NERSC ... is seeking project proposals to conduct research using neutral atom quantum processors from Atom Computing, including Aquila (analog) and Gemini (gate-based) systems. |
| SR013 | NQCC | Science Minister Andrew Griffith announces the results of the £30m quantum computing testbed competition | Atom Computing, Exeter ... [is among] seven quantum hardware companies ... awarded multimillion-pound contracts to build ... quantum computing testbeds. |
| SR014 | NQCC | Quantum computing testbeds | Atom’s project is designed to develop a more robust, scalable, and user-friendly neutral-atom testbed, offering an error-corrected operation. |
| SR015 | TechCrunch | Google-backed Boston quantum startup Atom raises $230M debt round | Notably, the financing is not equity. It’s a convertible note ... Atom’s team ... declined to say when the next equity funding round would happen or what valuation it expected. |
| SR016 | Bureau of Industry and Security | Department of Commerce Implements Controls on Quantum Computing and Other Advanced Technologies Alongside International Partners | This IFR includes controls related to quantum computing, semiconductor manufacturing, and other advanced technologies. |
| SR017 | U.S. Government Publishing Office | Federal Register: Commerce Control List Additions and Revisions; Implementation of Controls on Advanced Technologies | BIS is implementing export controls on several semiconductor, quantum, and additive manufacturing items for national security and foreign policy reasons. |
| SR018 | National Quantum Initiative | Department of Commerce Releases Export Controls on Quantum Technologies | BIS is requesting public comments, particularly on deemed exports, by November 5, 2024. |
| SR019 | Office of Management and Budget | M-23-02: Migrating to Post-Quantum Cryptography | This memorandum provides direction for agencies to comply with National Security Memorandum 10 (NSM-10). |
| SR020 | Cybersecurity and Infrastructure Security Agency | Post-Quantum Cryptography Initiative | CISA's Post-Quantum Cryptography (PQC) Initiative will unify and drive efforts ... to support critical infrastructure and government network owners and operators during the transition to post-quantum cryptography. |
| SR021 | National Institute of Standards and Technology | Post-Quantum Cryptography | CSRC | NIST’s Post-Quantum Cryptography (PQC) project leads the national and global effort to secure electronic information against the future threat of quantum computers. |
| SR022 | National Institute of Standards and Technology | NIST SP 800-171 Rev. 3, Protecting Controlled Unclassified Information in Nonfederal Systems and Organizations | Protecting Controlled Unclassified Information in Nonfederal Systems and Organizations. |
| SR023 | UK Government | National Security and Investment Act: the 17 types of notifiable acquisitions | Businesses and investors are legally required to tell the government about acquisitions of certain entities in 17 sensitive areas of the economy. |
| SR024 | Global Legal Insights | Quantum Computing Laws and Regulations 2026 | USA | The law is rushing to catch up amidst a backdrop of heightening geopolitical tension that stands poised to shape the future of the field. |
| SR025 | Skadden, Arps, Slate, Meagher & Flom LLP | New Executive Orders and Government Strategy Advance US Quantum Innovation and Mandate Post-Quantum Cryptography Transition | The executive orders establish timelines and enforceable obligations for federal agencies and government contractors to implement NIST-approved post-quantum cryptography standards. |
| SR026 | K&L Gates | Preparing for Q-Day: New Executive Orders Address Quantum Innovation and Post-Quantum Cryptography | The Administration is pursuing a whole-of-government effort focused on deployment, commercialization, national-security applications, and cyber resilience. |
| SR027 | PostQuantum | The Border Around Quantum: Export Controls, Deemed Exports, and “Research as a Controlled Flow” | Research as a Controlled Flow: Impact on Labs, Startups, and Teams. |
| SR028 | APS News | US Puts Export Controls on Quantum Computers | Entities must be licensed to export key components, and they must disclose when certain foreign nationals are working on the technology in the U.S. |
| SR029 | United States Courts | Find a Case (PACER) | PACER allows anyone with an account to search and locate appellate, district, and bankruptcy court case and docket information. |
| SR030 | CourtListener | CourtListener | CourtListener is a free legal research website containing millions of legal opinions from federal and state courts. |
| SR031 | United States Patent and Trademark Office | Search for patents | Use the following resources to search for patents. |
| SR032 | The White House | Fact Sheet: President Donald J. Trump Secures the Nation Against Advanced Cryptographic Attacks | The Order directs the Federal Acquisition Regulatory Council to require covered contractors to meet certain Federal cybersecurity standards and vulnerability disclosure policies by the end of 2030. |
| SR033 | The White House | National Security Memorandum on Promoting United States Leadership in Quantum Computing While Mitigating Risks to Vulnerable Cryptographic Systems | This memorandum outlines my Administration's policies and initiatives related to quantum computing. It identifies key steps needed to maintain the Nation's competitive advantage in quantum information science. |
| SR034 | National Security Agency | Quantum Key Distribution (QKD) and Quantum Cryptography (QC) | NSA does not recommend the usage of quantum key distribution and quantum cryptography for securing the transmission of data in National Security Systems unless the limitations below are overcome. |
| SR035 | Atom Computing | AIST Selects Atom’s Neutral-Atom Quantum Computer | Atom Computing ... announced it has been awarded a 6.5 Billion JPY contract (approx. $41M USD) by Japan’s ... AIST. |
| SR036 | Atom Computing | Roadmap for Advanced Error-Corrected Quantum Computers | 2026: Introduction of a third-generation quantum error-corrected model with 100 logical qubits and over 10,000 physical qubits. |
| SR037 | Intelligent CIO North America | Atom Computing report shows quantum market shifting from hype to proof-driven investment | Organizations are moving away from hype-driven quantum spending and demanding measurable value, as sovereignty, government funding and specialist talent shortages reshape the market. |
| SR038 | Atom Computing | Aquila | 256-qubit Quantum Computer | Available via Amazon Braket or via Premium Access: Secure, direct, supported environment with priority bookings. |
| SR039 | Data Center Dynamics | AWS to host Atom's next generation quantum computer | By 2028, Amazon will make Atom’s Libra ... available to its customers. |
| SR040 | Observer Research Foundation | Quantum Computing: Separating Hype from Reality | Despite decades of research and investment, progress in quantum computing remains hindered by its Achilles' heel: the lack of practical applicability. |
| SV001 | Forge Global | Atom Computing IPO: Investment Opportunities & Pre-IPO Valuations - Forge | |
| SV002 | PR Newswire | Atom Computing Raises More Than $300 Million to Accelerate Deployment of Fault-Tolerant, Neutral-Atom Quantum Computers | |
| SV003 | DCVC | A bold and timely bet: the U.S. Government invests in quantum… | |
| SV004 | PR Newswire | Atom Computing selected by DARPA for the next stage of exploring near-term utility-scale quantum computing with neutral atoms | |
| SV005 | DARPA | Stage B selection | DARPA | |
| SV006 | DARPA | QBI | DARPA | |
| SV007 | Microsoft Azure Quantum Blog | Microsoft and Atom Computing offer a commercial quantum machine with the largest number of entangled logical qubits on record - Microsoft Azure Quantum Blog | |
| SV008 | Quantum Computing Report | Denmark's QuNorth to Acquire 50-Logical-Qubit Magne Quantum Computer from Atom Computing and Microsoft - Quantum Computing Report | |
| SV009 | The Quantum Insider | Microsoft and Atom Computing Partner on Level 2 Quantum System for Nordic Users | |
| SV010 | Data Center Dynamics | Microsoft and Atom Computing to build "world’s most powerful quantum computer" in Denmark | |
| SV011 | IonQ Investor Relations | IonQ Announces First Quarter 2026 Financial Results | |
| SV012 | IonQ | IonQ Posts Q1 2026 Earnings with Record Revenue | |
| SV013 | U.S. Securities and Exchange Commission | 10-K | |
| SV014 | Stock Analysis | IonQ, Inc. (IONQ) Market Cap & Net Worth | |
| SV015 | U.S. News / Reuters | Honeywell's Quantinuum Eyes $14.3 Billion Valuation in Upsized US IPO | |
| SV016 | CNBC / Reuters | Honeywell's Quantinuum raises $1.68 billion in U.S. IPO as quantum computing heats up, Reuters reports | |
| SV017 | Quantinuum | Helios | Quantinuum's Quantum Computers | |
| SV018 | Quantinuum | Our Trapped Ion Quantum Computers | |
| SV019 | U.S. News / Reuters | PsiQuantum Valued at $7 Billion in Latest Funding Round, Teams up With Nvidia | |
| SV020 | PsiQuantum | Technology — PsiQuantum | |
| SV021 | Stock Analysis | D-Wave Quantum (QBTS) Market Cap & Net Worth | |
| SV022 | Stock Analysis | Rigetti Computing (RGTI) Market Cap & Net Worth | |
| SV023 | QuEra | QuEra Completes $230 M Financing | |
| SV024 | Amazon Web Services | AWS Deepens Strategic Collaboration with QuEra to Bring Fault-Tolerant Quantum Computing to Amazon Braket | Amazon Web Services | |
| SV025 | TechCrunch | Oratomic raises $300M to build a viable quantum computer that needs only 20K qubits | TechCrunch | |
| SV026 | Caplight | Oratomic | Valuation, Funding Rounds & Stock Price | Caplight | |
| SV027 | Seedtable | Oratomic Raises 300.0M USD in Series A Funding | |
| SV028 | Crunchbase News | Sector Snapshot: Quantum Computing Startup Investment Slows In 2026 While Public Markets Hold Strong | |
| SV029 | S&P Global Market Intelligence | Quantum computing stocks rise as US stakes $2B on sector build-out | |
| SV030 | Lambda Finance | Quantum Computing Stocks 2026: Pure-Plays, Tech Giants, and Private Leaders - Lambda Finance | |
| SV031 | NIST | Department of Commerce Announces Letters of Intent With 9 Companies for $2 Billion to Accelerate U.S. Leadership in Quantum Computing | |
| SV032 | Quantum Computing Report | IonQ Reports on Its Q1 2026 Financial Results: Historic Growth and Blueprint for Fault Tolerance - Quantum Computing Report | |
| SV033 | U.S. Securities and Exchange Commission | qbts-20251231 |