初创公司尽调
尽调报告 Analytics database infrastructure Series D private company 2026-10-10

ClickHouse

将云端采用转化为企业规模、并以高估值运行的开源数据库公司。

ClickHouse 依然是开源分析领域强劲的增长故事,但 150 亿美元的 D 轮估值已经透支持续兑现的预期,而利润率、留存和客户集中度仍缺乏透明度。

覆盖范围与披露说明

公开证据更新至 2026-10-10;ARR、客户规模和 2026 年 1 月融资信息为最新数据,但利润率、留存率和客户集中度仍未披露。

封面要素

最新融资 01
400 USD M [CO004]
员工人数 05
430-623 employees [CO009]

公司概况

ClickHouse 是一家在特拉华州注册的私营数据库公司,通过 ClickHouse Cloud 及相邻企业服务将开源列式分析引擎商业化。公司把广泛的开发者采用漏斗与托管云、BYOC 和支持服务结合起来,面向分析、可观测性和 AI 基础设施团队。

官网
clickhouse.com
创始人
Aaron Katz, Alexey Milovidov, Yury Izrailevsky
创立地点
San Francisco Bay Area, California, United States
总部
San Francisco Bay Area, California, United States
产品
开源列式分析数据库,加上用于实时分析、可观测性和 AI 工作负载的托管 ClickHouse Cloud。
客户
企业和高增长软件公司的数据工程、基础设施、分析及 AI 原生应用团队。
商业模式
按用量计费的云消费、企业与 BYOC 订阅,以及围绕数据摄取和支持的托管增值服务。
阶段
Series D private company
融资情况
公司于 2026 年 1 月完成由 Dragoneer 领投的 $400 million Series D,投后估值 $15 billion,累计融资约 $1.15 billion。
[CO001, CO002, CO003, CO004, CO006, CO007, CO008, CO009]

执行摘要

主要优势

  • 开源采用持续转化为经验证的云业务规模和客户增长。
  • ClickHouse 资本实力雄厚,面向 AI 工作负载的产品覆盖也在扩大。
  • ClickHouse 在数据、基础设施和 AI 团队中拥有较强的品牌认知度。

主要风险

  • 毛利率、留存率和客户集中度仍未披露。
  • 150 亿美元估值几乎不给执行失误或估值倍数压缩留下空间。
  • 分布式系统和数据摄取的复杂性可能拖慢部署,并加重支持负担。

未决问题

  • 分 cohort 的 ARR 质量、留存率和毛利率数据仍未披露。
  • 从公开来源仍看不到客户集中度和头部客户敞口。
  • 开源使用转化为可持续付费云收入的经济性,公开层面尚未得到证明。

目录

Chapter 01

01公司概览

1.1 公司身份与历史基础

ClickHouse, Inc. 是一家商业化开源基础设施供应商,提供高性能实时分析、数据仓库和可观测性能力。底层技术于 2009 年起源于 Yandex,由 Alexey Milovidov 负责工程,最初为 Yandex.Metrica 处理 PB 级遥测数据,查询响应时间低于一秒。该技术证明相较传统关系型、行式系统具有巨大性能优势后,核心数据库引擎于 2016 年 6 月以宽松的 Apache 2.0 开源许可证发布。2021 年 9 月,公司在旧金山湾区成立 ClickHouse, Inc.,成为独立的特拉华州公司,并正式将技术商业化。如今,公司主要通过 ClickHouse Cloud 变现;这是一项全托管、无服务器数据库服务,面向多租户分析应用和机器学习工作负载。[CO001, CO002, CO006, CO007, CO008, CO009]

KPI 快照表
指标数值 / 状态截至日期置信度尽调缺口
最新估值$15 billion2026-01高私人二级市场定价折扣未经核实
累计融资额$1.15 billion2026-01中信贷额度是否全部提款未知
收入运行率$250M+2026-05高GAAP 收入确认及 ARR 定义未经核实
付费云客户4,000+2026-05高ACV 分布及客户层级拆分未披露
全球员工数430-623 名员工2026中外包人员确切数量及地区分布未披露
总部所在地美国加州旧金山湾区2026高无(已核实为 Delaware 公司)

指标汇总自公司官方公告、投资报告及经核实的高管披露;私营公司的经营数据仍未经审计。

[CO002, CO004, CO006, CO007, CO008, CO009]

1.2 高管团队与组织治理

ClickHouse 的管理团队既有深厚的开源数据库工程背景,也有成熟的企业级市场执行能力。联合创始人兼首席执行官 Aaron Katz 在公司成立时加入,曾在 Salesforce 和 Elastic 担任高级商业领导职务,拥有丰富的规模化经验。联合创始人兼首席技术官 Alexey Milovidov 继续掌舵技术方向;他创建了原始数据库引擎,并持续负责开源社区贡献和架构创新。联合创始人兼总裁 Yury Izrailevsky 负责运营及基础设施扩张,此前曾任 Google 工程副总裁和 Netflix 云系统副总裁。治理由早期投资机构的董事会席位支持,但 Milovidov 的架构主导权和 Katz 的企业分销策略仍形成明显的关键人依赖。[CO001, CO003, CO015]

领导层与创始人表
人员职务背景职能匹配关键人物依赖
Aaron Katz联合创始人兼 CEO曾任 Salesforce 执行副总裁及 Elastic 首席营收官企业级市场进入与商业规模化领导力高(推动企业商业战略及高管招聘)
Alexey Milovidov联合创始人兼 CTO2009 年起在 Yandex 创建 ClickHouse核心数据库架构与开源项目方向关键(该引擎的首席架构师及技术愿景负责人)
Yury Izrailevsky联合创始人兼总裁曾任 Google 工程副总裁及 Netflix 云业务副总裁云基础设施工程与大规模分布式系统高(负责云服务可靠性及运营规模化)

标示核心高管创始人;关键人物依赖反映其专业技术知识及运营规模化经验。

[CO001, CO003]

1.3 资本结构、融资轮次与所有权

ClickHouse 已成为数据基础设施领域资本最充裕的公司之一,自 2021 年剥离以来累计获得约 $1.15 billion 的资本承诺。公司于 2021 年 9 月完成由 Benchmark 和 Index Ventures 领投的 $50 million Series A,随后在 2021 年 10 月完成由 Coatue Management 和 Altimeter Capital 共同领投的 $250 million Series B,投后估值 $2 billion。2022 年 12 月,Thrive Capital 领投了一轮未披露金额的 Series B 扩展融资。2025 年 5 月,Khosla Ventures 领投 $350 million Series C,同时配套 $100 million 信贷额度,增长进一步加速。最近,公司于 2026 年 1 月完成由 Dragoneer Investment Group 领投的 $400 million Series D,估值 $15 billion,从而建立充足现金储备,加速拓展企业人工智能和分析工作负载产品。[CO004, CO005, CO006, CO011, CO015]

利益相关方或投资者图谱
利益相关方职能轮次 / 持股战略重要性尽调事项
Dragoneer Investment Group领投方Series D($400M)支持 AI 基础设施扩张的后期增长伙伴确认董事会席位及按比例跟投权
Khosla Ventures领投方Series C($350M)企业 AI 生态及战略市场定位的主导伙伴审查投票控制约定
Coatue Management联合领投方Series B($250M)支持商业化的主要科技跨界投资者核实其是否继续参与 Series D
Altimeter Capital联合领投方Series B($250M)聚焦云基础设施规模化的公私市场跨界机构评估其 IPO 前定位及顾问角色
Index Ventures & Benchmark联合领投方Series A($50M)指导 Yandex 商业剥离项目的基础风险投资伙伴检查早期清算优先权层级
Nebius Group(原名 Yandex N.V.)创始股东剥离股权(据报 Series D 前约 28%)历史母公司及原始 IP 转让伙伴确认监管合规性及二级市场流动性限制

代表已披露的领投方及主要机构投资者;完整股权结构表和少数天使持仓仍属私密信息。

[CO004, CO005, CO006, CO015]

1.4 商业规模、里程碑轨迹与尽调风险

从运营表现看,ClickHouse 增长势头极强:截至 2026 年 5 月,年化运行率收入超过 $250 million,付费客户组织超过 4,000 家。这一表现反映了企业云的快速采用,以及公司在生成式 AI 遥测和基础设施日志领域的战略定位与高净扩张率。2026 年 1 月,公司收购 Langfuse,扩大软件组合,把原生 LLM 应用追踪和评估直接纳入分析生态。不过,尽调仍需重点审视若干风险:列式存储的数据摄取流程可能脆弱且工程负担重;高并发分析查询可能耗尽数据库处理能力;AI 基础设施毛利率承压和早期 AI 应用切换成本低,也会持续带来风险,需结合私有单位经济数据验证。[CO004, CO007, CO008, CO010, CO012, CO013, CO014]

FO001: 公司里程碑时间线

ClickHouse 从 Yandex 的开源原型发展为数十亿美元级企业云平台的里程碑进程。

里程碑日期和轮次细节汇编自公司公告和监管披露。

开源项目诞生 [CO001]

开源发布 [CO001]

商业化设立与 A 轮 [CO001, CO002]

B 轮与 $2B 估值 [CO002, CO006]

B 轮扩展 [CO011]

C 轮与信贷额度 [CO005, CO006]

D 轮与 Langfuse 收购 [CO004, CO010]

年化运行率收入超过 $250M [CO007, CO008]

[CO001, CO002, CO004, CO005, CO007, CO008, CO010, CO011]

1.5 展示项

Chapter 02

02市场分析

2.1 市场边界与工作负载分类

ClickHouse 主要处于实时分析数据库管理系统(OLAP)市场,为跨数十亿行数据、要求亚秒级响应的结构化和半结构化分析查询提供服务。可触达的软件支出包括高性能列式存储引擎、向量化查询执行层、实时流式摄取管道和交互式 API 服务数据库。传统行式事务数据库(OLTP)和通用 ACID 应用后端明确不属于这一核心边界;不过,ClickHouse 最近与 Ubicloud 推出原生托管 Postgres 服务,借助变更数据捕获简化事务数据与分析数据同步。与此同时,ClickHouse 通过开源 ClickStack 框架不断切入相邻的云可观测性和日志分析市场,替代那些处理结构化遥测时计算和存储开销过高的传统搜索技术。收购 Langfuse 后,市场边界还延伸至 AI 可观测性和智能体基础设施,使 ClickHouse 成为下一代软件系统中 LLM 评估与追踪遥测的底层分析数据平面。[CM007, CM009, CM010, CM011, CM012]

市场定义表
类别 / 细分市场纳入支出排除支出主要买方 / 付款方战略相关性
实时分析型 DBMS(OLAP)列式存储引擎、实时摄取管道、亚秒级查询执行及高并发 API 查询服务行式事务型运营数据库、事务型 ACID 处理及通用关系型应用后端工程副总裁、数据平台负责人、基础设施副总裁ClickHouse 可实现 10x-100x 查询加速,并支撑内部及面向用户的分析功能的核心市场
云可观测性与日志分析结构化及半结构化日志存储、分布式追踪、指标聚合及 APM 遥测存储专有全栈 APM 代理许可、合成监控套件及前端会话回放工具DevOps 总监、站点可靠性工程副总裁(SRE)、首席安全官高容量邻近市场:ClickHouse ClickStack 可替代昂贵的专有搜索型日志集群
云数据仓库与 Lakehouse数据湖查询加速、交互式 SQL 数据集市、Iceberg / Delta Lake 读取路径及外部 BI 服务层批处理 ETL / ELT 管道、企业 ERP 事务整合及多租户数据治理套件首席数据官(CDO)、企业分析负责人共存层:ClickHouse 将对延迟敏感的查询从 Snowflake、BigQuery 等批处理仓库中卸载出来
AI 与智能体基础设施遥测LLM 输入 / 输出可观测性、追踪评估、向量搜索索引及实时智能体工作流执行日志前沿基础模型预训练算力、LLM token 推理托管及智能体应用开发框架AI 负责人、首席 ML 工程师、AI 应用副总裁Langfuse 收购及 Anthropic、OpenAI 广泛采用后形成的高增长新兴层

界定实时分析型 DBMS 支出的可服务范围,并与相邻的事务型 OLTP 及批处理数据仓库层区分开来。

[CM007, CM009, CM010, CM011, CM012]

2.2 可触达市场规模与渗透视角

估算 ClickHouse 的总可触达市场需要同时看运营和经济维度,因为独立分析机构不会单独发布纯实时列式数据库引擎的收入。按广义技术画像口径,6sense 统计 ClickHouse 在 2026 年覆盖全球 4,770 个客户域名,估计市场份额为 2.83%,在更广义的大数据分析类别中排名 #10。按地域看,采用高度集中于北美:美国占已识别全球部署的 1,625 个客户域名或 50.34%,其次是印度(13.69%)和英国(7.78%)。按商业变现口径,ClickHouse Cloud 快速扩张,付费商业账户从 2025 年 5 月的超过 2,000 个增至 2026 年 1 月的超过 3,000 个,同时商业 ARR 超过 $350 million,按同比口径增长 250%。按开源口径,全球调查显示 72% 的部署会将外部可视化层连接到 ClickHouse,说明可触达市场既包括底层基础设施存储引擎,也包括更广泛的分析查询服务生态。[CM001, CM002, CM003, CM004, CM006, CM007, CM014]

TAM / SAM / SOM 或规模测算视角表
规模测算视角 / 细分市场发布方 / 来源年份地理范围估算值增长 / CAGR方法论置信度分析局限
全球大数据分析软件6sense 技术画像2026全球(50.3% 美国)跟踪 48 个竞争平台未披露自动化网络技术画像爬虫,识别覆盖 40,000+ 项技术的生产级企业部署中衡量的是部署域名数量(4,770 个 ClickHouse 域名,占比 2.83%),而非直接美元支出
托管商业云数据库(ClickHouse Cloud)ClickHouse 公司披露2026全球(36 个超大规模云区域)$350M+ ARR 运行率ARR 同比增长 250%+20VC 上的高管财务披露及 Series D 融资公告,覆盖 3,000+ 个付费云账户高反映商业云收入运行率而非市场总支出;不包括自托管开源集群
开源分析社区覆盖面ClickHouse《State of ClickHouse》调查2025全球(40 个国家)1,200 个受访组织未披露分析数据库架构、BI 工具及部署环境的全球用户调查中受访者为自选样本;72% 使用独立 BI 层,38% 表示现有 BI 存在瓶颈
前沿 AI 工作负载遥测客户群组ClickHouse 管理层披露2026全球占商业总收入 <12%使用量增长 100x高管披露的 AI 原生客户集中度,包括 Anthropic、OpenAI 和 Decagon中未披露每个 AI 客户的具体美元支出;高查询吞吐量并不直接对应成比例的美元利润率

将 6sense 的第三方技术画像跟踪与商业云披露结合;纯实时列式 OLAP 的完整独立美元 TAM 数据仍不可得。

[CM001, CM002, CM003, CM004, CM005, CM006, CM007, CM014]
FM001: 市场规模视角

从大数据分析技术画像到核心实时 OLAP 部署,分层展示受约束的市场机会。

各层宽度仅表示层级,不代表数值比例。

金字塔层级代表从广泛技术画像到付费云客户和大规模生产集群的运营与市场规模视角;美元金额在披露时代表年化运行率。

[CM001, CM003, CM004, CM006, CM011, CM013]

2.3 买方画像、工作流与预算归属

ClickHouse 的采用者可分为四类买方,体现其架构在开发者主导的开源采用和企业自上而下采购之间的灵活性。主要商业买方通常是工程副总裁、基础设施负责人或首席技术官,他们掌管中央工程算力预算,需要为面向客户的应用或内部分析功能提供亚秒级查询性能。在可观测性场景,经济买方转为站点可靠性工程(SRE)和 DevOps 负责人,目标是压低来自传统专有供应商、不断上涨的日志存储账单。快速崛起的买方群体还包括 AI 负责人和机器学习工程负责人,他们用 ClickHouse 摄取并分析数百万条智能体追踪和 LLM 生成日志。在企业账户中,开发者和数据工程师自下而上的采用是主要催化剂:团队先部署开源 ClickHouse Local 或自托管集群,解决紧迫的延迟瓶颈;随后迁移到 ClickHouse Cloud,以免去集群运维和基础设施配置的负担。[CM003, CM005, CM006, CM009, CM011]

细分市场 / 买方图谱
客户细分目标买方主要用户经济付款方核心工作流预算负责人采用触发因素
高规模科技与数字平台工程副总裁、基础设施负责人后端开发者、数据工程师中央工程组织面向用户的分析、广告技术事件归因、欺诈监测及亚秒级外部报告工程副总裁传统行存储或 Postgres 在并发下失效;云数据仓库查询延迟超过 5-10 秒
企业可观测性与 SRE 团队站点可靠性工程副总裁、DevOps 总监DevOps 工程师、SRE、系统架构师IT 运营与基础设施预算高吞吐量日志索引、分布式追踪查询及 Kubernetes 集群遥测存储SRE 副总裁 / IT 运营随日志量每年增长 50%+,Elasticsearch、Splunk 或 Datadog 带来的许可及存储账单难以承受
AI 原生创新者与模型开发者AI 负责人、首席技术官ML 工程师、AI 应用开发者研发 / AI 平台预算智能体执行追踪评估、RAG 向量检索、通过 Langfuse 记录 LLM 交互日志及基准测试评分AI 负责人 / CTO智能体应用循环要求对数十亿条历史提示词和输出提供亚毫秒级查询响应
中端市场产品与业务分析分析总监、产品负责人数据分析师、产品经理、增长团队业务单元分析预算使用现代 BI 构建交互式自助仪表板、产品使用漏斗查询及客户群组留存计算分析负责人数据团队需要在 100M+ 条事件记录上快速运行 SQL,而无需将数据提取到专有 BI 引擎内存中

源自商业客户披露、案例研究及 4,700+ 个部署的技术画像和企业画像分布。

[CM006, CM007, CM008, CM009, CM011]

2.4 增长催化剂、采用约束与尽调现实

ClickHouse 的市场扩张由现代应用架构的结构性变化推动:机器生成遥测数据呈指数增长,实时面向用户的分析兴起,能够高频生成分析查询的自主 AI 智能体出现。不过,采用约束和尽调现实会压低扩张速度。首先,企业账户渗透要面对 Databricks、Azure Databricks 和 Microsoft Azure Synapse 等传统云平台;它们合计占已跟踪类别市场份额的 44% 以上,并受益于与企业签订的云厂商承诺协议。其次,虽然开源社区采用广泛,但 38% 的受访企业用户表示传统 BI 工具无法充分发挥 ClickHouse 的列式执行引擎,需要专门工具或定制工程,这会拖慢非技术用户实现价值的速度。最后,ClickHouse Cloud 的长期单位经济仍存在关键尽调缺口:群组毛利率、客户获取回收指标以及高管评论之外的净收入留存持久性,均未得到经审计的公开财务披露验证。[CM001, CM005, CM008, CM013, CM014]

2.5 展示项

Chapter 03

03竞争格局

3.1 竞争格局与市场分类

ClickHouse 所处的分析数据市场高度分化,包含专用实时在线分析处理(OLAP)引擎、企业云数据仓库和嵌入式执行环境。竞争的核心不是覆盖所有数据库工作负载的正面功能对等,而是不同的性能边界和架构取舍。ClickHouse 专为在海量扁平或反规范化数据集上进行高吞吐分析查询而设计,利用向量化 MergeTree 存储引擎,在数十亿条记录上执行亚秒级扫描。因此,替代方案分为五类:云数据仓库(Snowflake、Google BigQuery、Databricks SQL、Amazon Redshift)、实时流式 OLAP 系统(Apache Druid、Apache Pinot)、传统事务关系型引擎(PostgreSQL、MySQL)、专用时序平台(TimescaleDB)和嵌入式分析引擎(DuckDB)。传统企业数据仓库擅长复杂多表连接、临时商业智能、治理和集中式存储计算分离,但在高并发实时事件摄取下往往延迟更高、计算成本更高。相反,Druid 和 Pinot 等流式 OLAP 竞品专注于从 Kafka 等消息总线低延迟摄取,但过去需要复杂的多节点分布式微服务架构。与此同时,DuckDB 等单节点和边缘分析引擎越来越多地承接开发者工作站和本地管道的工作负载,省去了分布式开销。[CP001, CP002, CP007, CP008, CP013]

全面竞争格局与分类
竞争对手 / 系统架构类别部署模式核心工作负载主要差异化相对 ClickHouse 的关键局限
Snowflake云数据仓库托管 SaaS企业 BI 与数据共享多集群共享数据,计算与存储解耦实时事件流查询延迟更高
Databricks SQLLakehouse 平台托管 PaaS统一 ETL、Lakehouse 与 AI / ML基于 Delta Lake 开放格式的 Spark / Photon 引擎亚秒级服务的运营延迟更高
Google BigQuery无服务器云仓库托管式无服务器即席 SQL 分析与 ML无服务器多租户执行,零基础设施频繁流式查询的查询扫描计费不可预测
Amazon Redshift云数据仓库托管 PaaS企业数据仓库与 AWS 集成深度集成 AWS 生态,RA3 存储解耦集群容量调整复杂,摄取速度较慢
Apache Druid实时 OLAP自托管 OSS / 云时间序列与事件仪表板采用位图索引的实时流式摄取多组件分布式架构复杂
Apache Pinot实时 OLAP自托管 OSS / 云面向用户的实时分析高并发下的低延迟筛选聚合即席 SQL 表达能力及 join 支持有限
DuckDB嵌入式进程内 OLAP本地库 / 进程内本地数据科学与边缘分析进程内零依赖嵌入式向量化引擎单节点内存受限;不具备分布式多节点集群
PostgreSQL关系型 DBMS(OLTP)自托管 OSS / 云事务后端与应用状态完整 ACID 合规性、丰富的 SQL 标准及可扩展性行存储在 TB / PB 级列式扫描上存在瓶颈

分类及技术对比画像综合自独立行业基准报告与供应商文档。

[CP001, CP002, CP007, CP008, CP011]

3.2 竞争者画像与核心差异

在企业部署中,ClickHouse 面对的主要竞争者取决于架构层级和部署模式。在企业云数据仓库层,Snowflake 是最主要的全托管替代方案。Snowflake 采用多集群共享数据架构,将计算仓库与中央对象存储完全解耦,提供自动弹性扩展、稳健的工作负载隔离和顺畅的数据共享。该无服务器模式大幅降低企业数据团队的运维负担,因此在运营简洁性比亚秒级延迟更重要的集中式 BI、高管报告和跨职能企业数据汇总场景中,Snowflake 成为默认选择。Databricks 以统一 Lakehouse 架构挑战 ClickHouse,在开放的 Apache Parquet 和 Delta Lake 存储格式之上结合 Spark 与 Photon 执行引擎。Databricks 面向希望在单一数据湖上整合端到端数据管道、流式 ETL 和机器学习模型训练的组织,而 ClickHouse 专注于分析服务和可观测性。在实时事件分析领域,Apache Druid 和 Apache Pinot 是直接的架构类比,优化不可变事件流索引和实时仪表盘生成。两者都使用带位图索引的分段列式存储,但管理分布式协调器、代理和历史节点需要较强的运营能力。在开发者一端,DuckDB 已成为本地、单服务器和嵌入式分析工作流的替代方案,在应用进程内部直接执行向量化查询,不需要外部守护进程管理或网络延迟。[CP002, CP003, CP009, CP010, CP011]

功能与能力矩阵:ClickHouse 对比主要替代方案
评估维度ClickHouseSnowflakeDatabricks SQLApache DruidDuckDB
向量化查询执行原生向量化(SIMD)向量化(专有)向量化(Photon C++)列式批扫描原生向量化(C++)
存储架构MergeTree 列式分片微分区(云)Delta Lake / Parquet不可变分段单文件列式
事务模型非 ACID 快照隔离完整 ACID 多语句完整 ACID(Delta Lake)仅追加分段不可变性完整 ACID(MVCC)
点查询效率低(稀疏主索引)中(裁剪筛选器)中(Liquid Clustering)中(倒排索引)高(支持 ART 索引)
分布式 Join有限 / 更适合扁平化完整分布式哈希 Join优化的 Shuffle Join有限(Lookup Join)快速本地哈希 Join
运营部署OSS 集群或托管云全托管 SaaS客户 VPC 中的托管 PaaS分布式多守护进程 OSS零运维嵌入式库

评估维度来自架构技术规格及独立对比基准分析。

[CP003, CP004, CP005, CP006, CP009, CP010]
FP001: 竞争定位:运营负担与实时分析延迟

按序映射分析平台,比较运营管理负担与实时亚秒级查询性能。

条目运营负担(低至高) (1–9)实时查询延迟(批处理至亚秒级) (4–9)说明
1. ClickHouse Cloud 4 9 Fully managed cloud delivery with sub-second query latency on billion-row datasets.
2. ClickHouse OSS 8 9 High operational complexity managing sharded clusters, delivering industry-leading raw scan speed.
3. Snowflake 2 5 Minimal operational overhead with serverless compute, but higher latency on continuous streaming queries.
4. Google BigQuery 2 4 Serverless managed model optimized for ad-hoc batch SQL rather than high-frequency real-time serving.
5. Apache Druid(OSS) 9 8 Excellent low-latency event indexing, but requires maintaining complex multi-daemon distributed architecture.
6. DuckDB 1 7 Zero operational footprint embedded in application memory, bound by single-node compute limits.

坐标依据架构要求和基准延迟画像进行标准化序位定位,并非实测毫秒级遥测数据。

[CP001, CP002, CP006, CP010, CP011, CP013]

3.3 架构取舍、护城河持久性与反向证据

将 ClickHouse 与替代平台比较,可以看到决定客户留存和工作负载扩张的结构性限制。第一,ClickHouse 的设计严格优先保障 OLAP 扫描效率,而非在线事务处理(OLTP)能力。平台不提供完整的 ACID 事务保证、行级锁定,也不擅长任意单条记录的更新和删除。ClickHouse 虽支持后台 Mutations 和 Lightweight Deletes,但变更是异步、非原子且资源密集型的,后台合并时必须重写整个数据分片。第二,相比 PostgreSQL 或 MySQL 等行式事务数据库,单主键记录的点查询仍效率低下,因为 ClickHouse 依赖稀疏索引,定位的是包含 8,192 行的粒度,而非单条记录。第三,分布式多表连接是公认的性能瓶颈;要获得最佳执行效果,摄取前需把关系型模式反规范化或摊平为宽表,这会增加外部管道复杂度。第四,自管 ClickHouse 集群需要较强的运营能力,尤其是对分片 MergeTree 节点进行数据再平衡,以及通过 ClickHouse Keeper 或 ZooKeeper 管理共识状态。ClickHouse Cloud 虽然减轻了基础设施管理负担,并能在主要超大规模云厂商上自动扩展,但厂商锁定和专有云管理层会让开源纯粹主义者感到受限。此外,Snowflake 和 Google BigQuery 等企业巨头拥有根深蒂固的采购协议、广泛的合作伙伴生态和集成式合规框架;ClickHouse 若要从部门级可观测性与遥测部署进入核心企业数据仓库支出,就会遇到强大的分销壁垒。[CP004, CP005, CP006, CP012, CP014, CP015, CP016]

打包方式、商业模式与运维权衡
平台许可证 / 交付方式定价模式运维复杂度客户权衡
ClickHouseApache 2.0 OSS / 云 SaaS基础设施资源消耗 / 云 CU自管集群高;Cloud 低扁平表速度极快,但需要手动分片和调优连接
Snowflake专有商业 SaaS按秒消耗计算积分 + 按 TB 计存储极低(无服务器抽象)企业治理无缝,但流式工作负载会快速消耗积分
Databricks商业 PaaS + 开放标准Databricks Units(DBU)+ 底层云 VM中等(Workspace 和集群策略管理)统一湖仓生态,但纯分析服务有额外开销
Apache DruidApache 2.0 OSS / Imply Cloud自管计算实例或 Imply SaaS极高(Broker、Historical、MiddleManager、ZooKeeper)支持实时流索引,但多节点运维繁重
DuckDBMIT 开源 / MotherDuck Cloud本地执行免费 / MotherDuck 无服务器用量本地嵌入式为零;云扩展低本地查询即时完成,但受单节点规模限制

定价机制反映公开价目结构;不含企业批量折扣和协商后的合同承诺。

[CP007, CP008, CP012, CP014, CP015]

3.4 展示项

Chapter 04

04财务

4.1 收入来源与变现架构

ClickHouse 通过 ClickHouse Cloud 将开源列式数据库技术商业化;这项全托管云服务在 AWS、GCP 和 Microsoft Azure 上提供无服务器和专用部署模式。核心 ClickHouse 引擎继续采用 Apache 2.0 许可,以维持广泛的开发者采用漏斗;企业收入则来自活跃计算单元和托管云对象存储的按用量计费。除闲置时可缩容至零的多租户无服务器集群外,ClickHouse 还提供专用集群和 Bring-Your-Own-Cloud(BYOC)部署模式,面向受监管的企业环境。在 BYOC 模式下,专有控制平面负责编排,数据库集群直接运行在客户管理的虚拟私有云中。其他收入来源包括通过 ClickPipes 提供的托管流式摄取连接器、原生变更数据捕获同步和企业支持协议,构成多层次商业组合。[CI001, CI002, CI014]

收入流表
收入流变现机制计费单位当前状态 / 跑率收入质量尽调要求
ClickHouse Cloud(无服务器与专用版)按用量收取计算和存储费用计算单元小时($0.22-$0.75/CU-hr)+ 存储($25.30-$50/TB-mo)核心增长引擎;ARR 于 2026 年 5 月突破 $250M,2026 年 8 月估算为 $350M高(经常性、由扩张驱动)客户群留存、净扩张率和计算利用率
Enterprise BYOC(自带云)带专用 VPC 控制平面的年度订阅按配置规模计的年度合同金额(ACV)面向受监管客户的活跃企业部署模式高(年度合同经常性收入)企业合同期限、折扣安排和承诺最低支出
托管摄取与附加组件(ClickPipes / CDC)托管 Kafka/S3/DB 连接器的用量附加费摄取数据量 / 连接器小时数已纳入 Cloud 各档位的商业功能中等(与 Cloud 摄取量挂钩)托管连接器计算的毛利率与公有云出口成本
开源支持与咨询(历史)年度支持 SLA 和商业许可证按节点计的年度许可证 / 集群费用相比 Cloud 已不再优先;开源仍是 Apache 2.0 漏斗中等(非按用量计费的服务 / 支持)纯支持收入与 Cloud 订阅收入的构成比例

数据反映已报道的云变现结构和分析师估算(Sacra、Improvado)。无服务器用量与承诺型企业 BYOC 合同之间的实际收入构成尚未公开披露。

[CI001, CI004, CI014]

4.2 定价机制、产品打包与 TCO 取舍

ClickHouse Cloud 从两个主要维度将基础设施消费变现:计算单元小时数和托管存储容量。公开计算价格从基础开发层级的每计算单元小时 $0.22,到高并发生产工作负载的每计算单元小时 $0.75;云对象存储按每 TB 每月 $25.30 至 $50.00 计费。与自托管部署进行总拥有成本基准比较,显示潜在客户需要权衡显著的运营取舍。运营自有 ClickHouse 集群,每月需要为原始云计算、高可用复制存储和协调节点支付 $2,435 至 $30,720,另需占用全职数据库管理员工程能力的 10% 至 20%。此外,公开价目页面经常遗漏跨区域数据出口、灾难恢复快照和跨区域私有网络等次要成本驱动因素,因此必须严格规划管道架构,避免账单意外上升。[CI002, CI006, CI012]

定价 / 变现表
部署档位公开定价 / 单位指标合同 / 计费模式实际定价与公开价对比折扣与未知项来源
ClickHouse Cloud Basic / Development每计算单元小时 $0.22;存储 $25.30/TB-month;入门价 $67/mo按需信用卡支付或自助计费标准公开价,含 $300 注册额度自助用量无企业折扣;流失率未披露Improvado TCO Guide / ClickHouse Official
ClickHouse Cloud Scale / Production每计算单元小时 $0.39-$0.75;存储 $35-$50/TB-month按用量月度开票或按承诺额度扣款实际费率随持续并发和自动暂停而变动分级用量和承诺额度扣款会降低有效单价Improvado TCO Guide / Sacra
企业专用 / BYOC定制企业合同($100k+/yr 规模承诺)年度经常性订阅 + 超额用量 / VPC 许可证计算承诺采用协商后的企业折扣私有定价;多年承诺折扣和支持打包尚未核实Improvado TCO Guide / Sacra
自托管开源(对比)$0 软件许可证(Apache 2.0);基础设施 $2,435-$30,720/mo硬件 / 云 IaaS 基础设施 + 内部工程开销软件费为零;基础设施 + 10-20% FTE DBA 运维开销硬件配置、备份复制和 Keeper 仲裁成本转由客户承担Improvado TCO Guide / ClickHouse Official

公开价取自 Improvado 2026 年定价基准和 ClickHouse 文档。实际折扣安排、平均合同金额和最低承诺门槛均属非公开信息。

[CI002, CI006, CI012]

4.3 单位经济、成本结构与利润率驱动因素

ClickHouse Cloud 构建在 AWS、Google Cloud 和 Azure 等公有超大规模云基础设施之上,因此销售成本(COGS)主要由批发计算、对象存储和网络带宽费用构成。列式架构和向量化执行相较行式数据库带来更高压缩率和计算效率,但公开披露仍未证明毛利率表现。ClickHouse 在十多个国家拥有全球分布式员工,但公司不披露合并毛利、研发投入或客户获取成本。关键软件单位经济指标——尤其是毛利率、净收入留存(NRR)和年度客户回本速度——仍属于未披露的私有数据。因此,要评估长期软件盈利能力,仍需独立验证托管基础设施成本转嫁和客户扩张动态。[CI009, CI010, CI011]

单位经济表
财务指标已披露或估算值置信度分析意义尽调要求
年度经常性收入(ARR)$350M(2026 年 8 月估算);$250M(2026 年 5 月里程碑)中体现 IPO 准备度所需的商业规模和增长速度审阅经审计财务报表、订阅 ARR 定义和 LTM 实现收入
客户数4,000+(2026 年 5 月);3,000+(2026 年 1 月)高体现 ClickHouse Cloud 的客户采用速度按合同档位拆分(自助服务与 ARR 超过 $100k 的企业客户)
毛利率(%)低直接反映云基础设施相对于 AWS/GCP 托管成本的效率要求提供公有云计算 / 存储与客户支持之间的 COGS 拆分
净收入留存率(NRR)低决定有机增长持续性的关键 SaaS / 用量扩张指标要求提供客户群层面的净扩张曲线及各客户群的毛流失率

ClickHouse 完全未披露毛利率和净收入留存率。投资判断前,尽调必须建立经审计的 COGS 和客户群留存数据。

[CI004, CI007, CI010]

4.4 收入模式桥接与变现转化

从开源社区使用转向付费云消费,是一座运营上的转化桥梁。免费的开源部署先带来开发者采用和工作负载概念验证;随着数据量和运营复杂度上升,ClickHouse 再把它们转化为付费云消费。客户支出会根据活跃查询执行和存储留存直接转化为总账单。但云交付成本会立即通过底层超大规模云计算实例、托管存储层和跨区域出口流量吞噬总账单的相当一部分。其他服务交付成本包括自动备份监控、托管摄取基础设施和持续云支持运营。对于企业投资判断,理解扣除托管 COGS 和持续平台再投资后的剩余贡献利润,是评估规模化运营杠杆的关键。[CI001, CI011, CI012]

FI001: 收入模型桥接图

通过模型化的扣减项,展示从云端毛账单到经营贡献利润率的示意性单位经济转化。

条目数值
云端毛账单 100
云计算 IaaS(AWS/GCP/Azure) -28
对象存储与跨区域出口流量 -7
托管支持与云运维 -5
估算毛利 60
销售与 GTM 再投资 -35
经营贡献利润率 25

示意性单位经济桥接图反映模型化的 60% 云端毛利率结构和托管成本扣减;公司未披露经审计的毛利率和 COGS。

[CI001, CI011, CI012]

4.5 资本充足性、现金续航与投资判断

ClickHouse 凭借 2026 年 1 月融资方案和现有信贷额度建立了较强的资本充足性,短期扩张准备充分。本轮融资使机构轮次累计股权融资超过 $1 billion;此外,公司在 Series C 期间从 Goldman Sachs 和 Stifel 获得 $100 million 信贷额度。公司商业势头出色:客户数从 2026 年 1 月的 3,000 家增至 2026 年 5 月的超过 4,000 家;ARR 则从 2025 年底的 $160 million,增至公司报告的 2026 年 5 月 $250 million,以及估算的 2026 年 8 月 $350 million。然而,以 $15 billion 估值计算,公司的隐含收入倍数介于估算远期 ARR 的 43x 和历史交易日收入的 90x 以上。由于缺少经审计的毛利率披露和客户群留存数据,尽管商业势头极佳,投资判断仍承担较高估值风险。[CI003, CI004, CI005, CI007, CI008, CI010, CI013]

4.6 展示项

Chapter 05

05产品与技术

5.1 产品定义与核心分析工作流

ClickHouse 为结构化遥测、日志、指标和面向用户的仪表盘提供高速在线分析处理。产品组合包含三种主要部署形态:开源自管 ClickHouse 数据库引擎、零依赖的 ClickHouse Local 命令行工具,以及全托管无服务器平台 ClickHouse Cloud。在实时遥测和临时分析工作流中,ClickHouse 通过压缩列式数据、直接查询原始不可变事件流,替代传统行式数据库的瓶颈。这省去了复杂的批量预聚合管道,让企业每秒摄取数百万个事件,同时在数十亿行数据集上保持亚秒级查询响应。[CE004, CE005, CE006, CE009, CE010]

产品模块 / 资产矩阵
模块 / 产品线目标用户 / 工作负载成熟度 / 状态核心差异化尽调缺口
ClickHouse 开源服务器自托管分析集群的数据工程师和基础设施团队生产环境(成熟)向量化 SIMD C++ 引擎、丰富的表引擎、本地磁盘缓存自管运维复杂,集群再平衡开销高
ClickHouse Cloud需要无服务器自动扩缩容 OLAP 的企业数据团队生产环境(商业 GA)计算与共享对象存储(S3/GCS/Azure)解耦,支持自动扩缩容私有多租户 noisy neighbor 隔离基准仍未披露
ClickHouse Local通过 CLI 分析本地数据文件、无需部署服务器的开发者生产环境(成熟)零依赖嵌入式分析引擎,可替代 awk/grep/sqlite受限于单主机内存和核心规模,无法分布式协同
ClickHouse for PostgreSQL 与 LangfuseAI/ML 应用开发者和 Postgres 运营数据库用户积极扩张(2026)原生 PG wire 集成和开源 LLM 可观测性平台长期变现模式及其与核心 OLAP 引擎的技术集成深度

产品组合状态以 2026 年 10 月的商业 GA 和开源可用性为准;私有云多租户遥测数据仍未披露。

[CE004, CE005, CE009, CE010]
工作流 / 用例表
用户任务 / 目标传统 / 替代工作流ClickHouse 方案可量化收益技术限制
实时遥测和即席日志探索行式 RDBMS 或 Elasticsearch 集群,索引开销很高采用列式压缩和稀疏主索引的 ClickHouse MergeTree可在数十亿行上以亚秒级扫描查询,存储压缩率达 3x 至 5x单行点变更会造成显著写放大和合并开销
面向用户的分析和实时仪表盘预聚合数据集市或批量计算的缓存层直接在原始流式事件表上执行即席 SQL省去批量预聚合管道,将数据延迟降至毫秒级复杂的多表分布式连接需要大量内存和手动共置
LLM 应用追踪和评估碎片化的第三方可观测性 SaaS 工具或彼此割裂的向量存储由 ClickHouse 列式分析存储支撑的 Langfuse 平台在 PB 级事件规模上统一端到端追踪分析和评估LLM 追踪与核心数据库模式的集成仍处于早期采用阶段
运营数据库分析卸载将事务数据抽取到远程数据湖的 ETL 管道用 ClickHouse for PostgreSQL 与事务数据并行处理分析无需 ETL,直接对运营数据库进行实时查询卸载无法为运营写入路径提供完整事务 ACID 语义

性能收益来自面向结构化分析工作负载的列式向量化执行;事务和变更工作负载仍受限制。

[CE006, CE009, CE010, CE013, CE014]

5.2 向量化列式架构与云端解耦

ClickHouse 用 C++ 实现为单一静态链接二进制文件,不依赖强制运行时组件或第三方协调软件。其查询处理架构将 SQL 查询解析为优化的逻辑和物理执行树,借助 SIMD 硬件向量化和 CPU 缓存局部性评估列块。持久化存储依赖 MergeTree 引擎系列,把数据划分为不可变分片,并使用带 8,192 行粒度的稀疏主索引进行快速范围过滤。在 ClickHouse Cloud 中,计算执行与持久化存储解耦:无状态工作节点负责查询执行和本地 NVMe 缓存,底层数据集存放在 AWS S3 或 Google Cloud Storage 等共享云对象存储中,由 ClickHouse Keeper 协调。[CE001, CE002, CE003, CE007, CE008, CE011]

技术 / 运营架构表
架构层角色与功能职责底层技术 / 依赖运维与扩展风险
客户端连接与摄取通过原生 wire、HTTP REST、MySQL/Postgres 协议接收查询和流式批量摄取静态链接 C++ 网络栈、Kafka/Pulsar 流式连接器未池化客户端流量突发时连接池耗尽
查询解析器与执行引擎解析 SQL、执行 AST 优化,并在列块上进行向量化处理利用 CPU 缓存和 SIMD 指令集的向量化查询引擎宽表未分区连接或高基数 GROUP BY 可能导致查询因内存不足失败
存储引擎与分区将列式数据片段组织在磁盘上,并使用稀疏主索引(每个粒度 8,192 行)采用 LZ4/ZSTD 压缩和后台数据片段合并的 MergeTree 引擎系列频繁小批量插入时写放大和文件系统 I/O 抖动
云解耦存储与共识将无状态计算节点与持久化云对象存储解耦共享云对象存储(S3/GCS/Azure Blob)由 ClickHouse Keeper 协调多云区域间共享存储元数据延迟和尾延迟波动

架构角色反映 ClickHouse VLDB 2024 学术架构和 ClickHouse Cloud 解耦存储模型。

[CE001, CE002, CE003, CE007, CE008, CE011]
FE001: 产品架构图

ClickHouse 多层技术架构覆盖采集协议、向量化计算、MergeTree 存储和云端解耦。

客户端采集与协议接口
原生 TCP 线协议、HTTP REST API、MySQL 和 PostgreSQL 兼容线协议,以及流式采集连接器(Kafka、Pulsar)。
模块
  • 原生二进制 TCP 协议
  • HTTP REST 接口
  • MySQL / PostgreSQL 线协议
  • 流式连接器(Kafka、对象存储)
查询处理与优化层
SQL 解析、抽象语法树(AST)生成、基于规则和基于成本的查询优化,以及逻辑计划转换。
模块
  • SQL 解析器与 AST 构建器
  • 逻辑与物理查询优化器
  • 分布式查询协调器
向量化计算执行引擎
面向列的向量化执行按数据块处理数据,优化 CPU 指令流水线和硬件 SIMD 寄存器。
模块
  • 向量化算子流水线
  • SIMD 加速表达式
  • 多线程工作线程执行
MergeTree 列式存储引擎
物理列式存储格式按不可变分片组织,配备 8,192 行粒度的稀疏主索引和后台异步压缩。
模块
  • MergeTree 引擎核心
  • 稀疏主索引
  • LZ4 / ZSTD 列式压缩
  • 后台异步分片合并
云端解耦存储与元数据层
无状态计算工作节点在共享对象存储(S3、GCS、Azure Blob)之上解耦,由 ClickHouse Keeper 协调分布式共识。
模块
  • 共享对象存储抽象层
  • ClickHouse Keeper(Raft 共识)
  • 本地 NVMe 缓存与预取
安全、合规与多租户管理层
基于角色的访问控制(RBAC)、TLS 加密、查询配额、资源治理和持续集群监控遥测。
模块
  • 基于角色的访问控制(RBAC)
  • TLS/mTLS 网络加密
  • 查询资源配额与限制
  • 审计日志与系统遥测
[CE001, CE002, CE003, CE007, CE011, CE012]

5.3 技术差异化与生态扩张

ClickHouse 的分析引擎通过激进的向量化数据处理、稀疏索引架构和灵活的线协议兼容性实现差异化。ClickHouse 不维护会拖慢高吞吐流式写入的重型 B-tree 索引,而是使用指向压缩数据粒度的轻量主键,最大化顺序磁盘吞吐。客户端应用可通过原生 TCP 二进制协议、HTTP REST 接口,或标准 MySQL 和 PostgreSQL 线协议兼容性与 ClickHouse 交互。2026 年,ClickHouse 收购开源 LLM 可观测性平台 Langfuse,捕获非确定性 AI 应用追踪;同时推出 ClickHouse for PostgreSQL,将列式 OLAP 处理直接带入既有运营数据库拓扑。[CE003, CE005, CE009, CE010, CE011]

5.4 运营约束、多租户与技术尽调缺口

ClickHouse 擅长追加型分析扫描,但架构设计也带来明确的运营边界和技术取舍。高频单行更新和变更会产生显著写放大,因为不可变分片必须通过后台合并异步重写。宽型分布式连接和高基数分组还需要大量内存余量;超过内存上限时,查询可能被终止。ClickHouse 纳入企业安全控制,包括基于角色的访问控制、TLS 网络加密、查询配额和审计日志。不过,针对 ClickHouse Cloud 在共享存储争用下的多租户邻居干扰隔离和尾延迟保证,尚无独立基准数据,仍是重要技术尽调领域。[CE012, CE013, CE014]

5.5 展示项

Chapter 06

06客户

6.1 客户细分与买方画像

ClickHouse 的商业客户群已从早期的开源数据库爱好者,显著转向企业基础设施团队和高增长 AI 应用构建者。过去,软件工程师和数据库管理员看重其原始分析速度与列式压缩;如今,首席技术官、数据工程负责人和负责实时运营工作流的基础设施负责人越来越多地采购 ClickHouse。客户群可分为三类主要商业群体:需要亚秒级遥测和智能体推理响应的 AI 原生前沿模型开发者;整合传统数据仓库的企业云现代化团队;以及把实时面向客户分析嵌入 SaaS 平台的产品工程团队。此外,公司在拉丁美洲、亚太和中东扩张后,采购动态各不相同:在部分地区,通过区域经销商和系统集成商间接采购的收入最高占商业收入的 25%。[CU001, CU004, CU005, CU006, CU013, CU015]

客户分层表
客户分群买方 / 用户 / 付费方主要用例已观察到的规模与工作负载收入与战略价值尽调缺口
AI 与前沿模型构建者基础设施副总裁 / AI 工程师 / CTO实时 AI agent 分析、提示词评估、模型遥测和向量检索数十亿条日志和亚秒级 agent 评估查询;支撑 Anthropic Claude 4 基础设施战略灯塔客户带来高用量,也增强现代基础设施的可信度客户层面的 ARR 集中度和毛利率情况仍未披露
企业云现代化客户数据负责人 / 首席数据官 / 采购实时数据仓库、运营仪表盘和传统 OLAP 卸载每分钟数千个并发查询处理数千亿个事件(Tesla、Capital One)通过年度多年期 ClickHouse Cloud 企业承诺推动 ARR 核心增长传统迁移客户群的续约率和净扩张指标未披露
云原生与高速增长初创公司创始人 / 平台工程师 / 产品负责人产品分析、金融交易监控和面向外部客户的指标高速摄取事件且延迟要求严格(Polymarket、Lovable、Decagon、Airwallex)从免费档位快速扩展到多席位云档位用量计费下的早期初创公司倒闭风险和用量波动
开源自管采用者软件开发者 / 数据库管理员 / 社区本地部署和自托管 OLAP、日志聚合及内部分析实验社区部署累计超过 50,000 个 GitHub stars 和数亿次下载转化为托管 ClickHouse Cloud 的主要漏斗入口免费开源用户转为付费云账户的转化率未公开报告

客户分群综合自 C/D 轮客户披露、合作伙伴生态公告和开发者社区指标。规模反映公开记录的客户工作负载;财务贡献反映公司定性披露。

[CU001, CU004, CU008, CU009, CU010, CU013, CU015]

6.2 采用轨迹、云端速度与增长指标

ClickHouse Cloud 是 ClickHouse, Inc. 的核心商业引擎,过去 18 个月客户账户增长迅速。公司官方披露显示,托管云服务的客户数从 2025 年 5 月的超过 2,000 家,增至 2026 年 1 月、伴随 $400m Series D 融资的超过 3,000 家,并在 2026 年 5 月下旬达到超过 4,000 家。换算下来,2026 年初每季度净增约 1,000 个客户账户。与此同时,ARR 在 2026 年初同比增长超过 250%,到 2026 年 5 月超过 $250 million。采用速度由庞大的开源社区推动:社区拥有超过 50,000 个 GitHub 星标和数亿次终身下载,为 ClickHouse Cloud 的按消费计费基础设施提供高效的自然转化入口。[CU001, CU002, CU003, CU013, CU014]

客户增长 / 采用轨迹表
指标已披露值截至日期来源归属置信度战略含义缺失分母 / 缺口
ClickHouse Cloud 客户数超过 2,000 个客户2025-05-29ClickHouse C 轮公告高证明初步商业 traction 以及从开源社区向云服务的转化自助信用卡账户与企业承诺之间的拆分未披露
ClickHouse Cloud 客户数超过 3,000 个客户2026-01-16ClickHouse D 轮公告高采用加速迅猛,约 8 个月增加 1,000 个付费账户平均合同金额(ACV)和账户档位分布未说明
ClickHouse Cloud 客户数超过 4,000 个客户2026-05-27ClickHouse Open House / 渠道负责人访谈高延续在三个月内净增 1,000 个客户账户的速度客户总流失率和 logo 留存率仍未说明
年度经常性收入(ARR)超过 $250 million2026-05-27CRN 对全球合作伙伴副总裁的访谈高商业规模可观,以 $15B 估值基准验证企业变现能力贡献利润率、专业服务占比和经常性客户群持续性未披露

所有增长指标均直接来自公司在官方融资公告和经核实的高管访谈中的披露。缺失分母反映公司没有 SEC 级财务报告。

[CU001, CU002, CU003, CU014]
FU001: 客户旅程图

客户从开源评估、生产云部署到合作伙伴赋能的企业扩张的推进路径。

采用阶段反映 ClickHouse Cloud 部署和 House Mates 合作伙伴生态中观察到的标准商业化开源变现模式。

[CU004, CU005, CU007, CU013, CU015]

6.3 具名生产验证与工作负载演进

前沿 AI 创新者和《财富》500 强企业的关键生产部署验证了采用质量。最值得注意的是,Anthropic 确认 ClickHouse 在开发和发布旗舰 Claude 4 AI 模型中发挥了重要作用,利用 ClickHouse 的高吞吐架构分析海量模型训练和遥测数据。Tesla 在高可用分析环境中保持 ClickHouse 生产部署,并认可其速度和耐久性。在金融服务和全球商业领域,Capital One、Airwallex 和 Mercado Libre 依靠 ClickHouse 进行高并发欺诈检测、运营监控和实时交易处理。Sony、Lyft、Modal、Polymarket、Decagon 和 Lovable 等其他现代软件公司也在生产环境运行 ClickHouse,用于日志分析、乘客数据探索和实时客户仪表盘。这些已验证的参考客户表明,ClickHouse 已从部门试验走向主要运营基础设施。[CU008, CU009, CU010, CU015]

具名客户证明表
客户分群部署 / 用例生产环境还是试点已记录结果技术或商业限制
Anthropic前沿 AI 与 LLM 系统模型开发和遥测的底层分析引擎生产环境对 Claude 4 模型基础设施的开发和部署发挥关键作用具体查询量、硬件配置和年度支出未披露
Tesla企业 / 自动驾驶技术实时遥测分析和高吞吐量数据处理生产环境在关键任务系统中实现高可用性、查询速度和数据持久性总数据规模和集成架构仍属保密信息
Capital One金融服务企业事务和分析事件流处理生产环境扩大云分析部署,用于高并发欺诈和交易监控具体部署范围和多云基础设施依赖未披露
ModalAI 云基础设施摄取高吞吐日志和事件流的实时仪表盘生产环境通过 Hex 连接器,每分钟摄取超过 1 million 个事件,覆盖数千亿条日志依赖第三方 BI 连接器的性能和维护
Lyft交通 / 市场平台覆盖多个城市和服务区域的实时出行与司机工时分析生产环境在显著节省基础设施成本的同时,大幅提升查询性能自行管理配置的运维负担与托管云服务之间的取舍

客户部署反映了官方公告、高管访谈和案例研究中披露的已验证生产用例。商业支出水平和合同条款未公开。

[CU008, CU009, CU010, CU015]

6.4 持久性、渠道合作伙伴依赖与摄取摩擦

尽管 ClickHouse 收入增长迅速,客户留存经济性和实际部署摩擦仍存在关键尽调不确定性。ClickHouse 不披露客户群层面的净收入留存(NRR)、毛收入留存(GRR)或客户 logo 流失率,使外部无法独立验证收入持久性。与此同时,企业扩张暴露出运营依赖。在 2026 年 5 月正式推出 House Mates 伙伴计划之前,ClickHouse 没有正式的系统集成商和经销商计划,国际客户缺少本地实施支持,公司还必须支持超过 200 个未经审查的开源集成。分析领域的独立从业者进一步指出,数据摄取管道脆弱、自定义连接器维护以及未优化的 BI 工具查询,是客户部署在实现生产价值前反复卡住的技术瓶颈。[CU004, CU005, CU006, CU007, CU011, CU012, CU014]

6.5 展示项

Chapter 07

07风险

7.1 监管、法律与公司治理风险

ClickHouse 的核心采用 Apache 2.0 许可的开源模式,既能广泛全球分发,又避免严格的 copyleft 义务。公司于 2021 年 9 月在特拉华州成立,起源于俄罗斯的 Yandex,最初得到 Index Ventures 和 Benchmark Capital 等创投机构支持。拆分为独立美国实体和荷兰子公司后,公司治理边界清晰。公开法律文件和开源审计日志未显示存在威胁核心数据库引擎的未决专利主张、商业秘密诉讼或结构性知识产权争议。监管合规风险主要涉及跨境数据传输法规、区域数据主权要求,以及国际部署区域的 SOC 2 / ISO 认证。此外,Stifel 和 Goldman Sachs 领衔提供的 $100 million 创投债务额度附带资本结构契约,需要持续监测流动性;近期股权融资缓解了契约压力。[CR001, CR009, CR010, CR014]

监管 / 法律风险登记表
风险类别 / 项目司法辖区状态可能性严重程度缓释措施剩余风险尽调路径
公司设立与 Yandex IP 渊源美国 / 特拉华州历史问题 / 已解决低中2021 年 9 月在特拉华州设立,采用独立法人结构并完成风险投资融资Apache 2.0 开源许可证下的剩余知识产权主张风险较低查阅原始贡献者许可协议和知识产权转让文件
开源商业化与许可全球 / 多司法辖区活跃 / 已治理低低核心数据库引擎按 Apache 2.0 许可证分发,云端专属管理软件为专有软件风险极小;开源核心允许分叉,但托管云控制平面构成商业护城河查阅代码审计日志和第三方依赖许可扫描报告
风险债务信贷额度契约美国活跃 / 监控中中中与 Stifel 和 Goldman Sachs 签订 $100M 额度,由 $400M D 轮现金储备提供后盾偿债义务取决于流动性契约和经营现金消耗核查信贷额度协议条款、最低流动性契约和利率息差
跨境数据隐私与云合规美国 / 欧盟 / 全球活跃 / 已落实合规中中AWS、GCP 和 Azure 提供跨区域云可用性,并支持区域数据驻留跨境分析数据传输的监管变化要求持续认证确认 SOC 2 Type II、ISO 27001 和 GDPR 数据处理附录文件

截至 2026 年 10 月,评估公开的公司设立、开源许可治理和信贷额度条款。

[CR001, CR009, CR010, CR011]

7.2 运营、架构与可靠性故障模式

分布式分析数据库在 PB 级规模下的运营可靠性面临严峻工程挑战。ClickHouse Cloud 通过在独立服务和共享仓库之间设置软运营护栏,缓解资源争用,包括限制数据库、表、活跃分片和副本数量。为维持共享元数据目录的稳定性,仓库对共享同一 ClickHouse Keeper 集群的副本总数设置 50 个的软上限。协调故障仍是关键失效模式:当 ZooKeeper 或 Keeper 会话断开,或纪元事务计数器溢出时,领导者重新选举会使表副本进入只读状态,打断摄取管道。在应用查询层,Temporal Cloud 等企业客户的生产影子测试发现了细微的数字异常,包括 64 位整数累加器溢出;在实现查询级类型转换前,该问题曾导致账单总额为负。自动拉取请求审查工具和 clickhouse-diagnostics 等诊断工具,为调查内存耗尽、副本延迟和写入限流提供必要遥测。[CR002, CR003, CR004, CR005, CR006, CR007, CR008]

运营 / 质量 / 安全风险登记表
故障模式可能性严重程度缓释成熟度剩余风险未解决缺口
协调会话丢失与副本只读锁中高已建立ZooKeeper 或 Keeper 重连和协调期间出现暂时性写锁监测多区域集群中的会话断开频率
Keeper 纪元计数器回绕低高改进中非计划的主节点重新选举导致暂时无法写入相较旧版 ZooKeeper 的 32 位计数器限制,强化 Keeper Raft 共识
计费聚合中的 64 位整数累加器溢出低高已解决极端数据量下应用层符号回绕,产生负的聚合值查询类型转换规范,以及客户端 SQL 中自动校验整数宽度
表分区与活跃分片爆炸(TOO_MANY_PARTS)中中已建立后台合并跟不上高频写入时触发插入限流诊断工具(clickhouse-diagnostics)及每个服务 10,000 个分片的软上限

综合生产事故复盘和诊断工具中记录的运营故障模式与架构限制。

[CR002, CR004, CR005, CR006, CR007, CR008]
FR001: 风险热力图

按可能性、严重程度和缓解成熟度评估主要风险领域的矩阵。

严重程度和可能性评分是对运营、财务及监管证据的定性综合判断。

[CR002, CR003, CR005, CR007, CR012, CR013]

7.3 伙伴依赖、财务杠杆与尽调路径

ClickHouse Cloud 在架构上解耦,但在运营上依赖 Amazon Web Services、Google Cloud Platform 和 Microsoft Azure 的底层超大规模云基础设施。云对象存储或虚拟化层中断会直接损害查询延迟和数据摄取 SLA。战略合作也带来运营依赖,尤其是与 Ubicloud 的企业集成——为混合事务与分析工作负载提供托管 PostgreSQL——以及对 Langfuse 的商业整合,用于 AI 可观测性。从财务角度看,ClickHouse 于 2026 年 1 月通过 Series D 股权融资获得 $400 million,现金续航充足,但长期现金流可持续性取决于企业合同留存。由于客户集中度、毛收入留存(GRR)和大客户账单分布未在公开记录中披露,机构投资者必须在保密尽调中索取经验证的披露明细,以评估单租户暴露。[CR009, CR010, CR011, CR012, CR013, CR014]

合作伙伴 / 依赖风险登记表
依赖项交易对手方角色集中度失效场景严重程度缓释措施剩余风险
公有云基础设施AWS / GCP / Azure为 ClickHouse Cloud 提供计算、网络和对象存储托管关键区域云中断或对象存储延迟飙升,导致查询 SLA 恶化高多云架构支持跨多个主要超大规模云厂商部署中
托管事务引擎Ubicloud为统一事务 / 分析栈提供高性能 PostgreSQL 引擎高CDC 复制管道服务降级或延迟,扰乱混合分析中采用原生 CDC 复制和解耦分析引擎的联合架构中
风险债务融资Stifel & Goldman Sachs$100M 信贷额度提供方中等发生极端运营失调时违反契约或触发提前偿还中$400M D 轮带来的充足股权现金缓冲低
开源 LLM 生态Langfuse 社区开源开发者采用与追踪 SDK 集成中等生态分化,或竞争者分叉开源追踪库中深度集成 ClickStack,并获得公司官方赞助低

评估关键外部平台、云基础设施和企业合作伙伴依赖。

[CR009, CR010, CR011, CR012]

7.4 展示项

Chapter 08

08估值

8.1 投资论点、立场与建议

ClickHouse 是现代数据基础设施中最具吸引力的运营增长案例之一,核心驱动力是明确的开源采用和云端加速。投资论点的核心在于:ClickHouse 的列式数据库架构相较传统系统可带来最高 100x 的查询速度优势,从而支持高并发分析和 AI 智能体工作负载。商业势头也很突出:ClickHouse Cloud 服务超过 3,000 个客户,ARR 同比增长超过 250%。Meta、Tesla 和 Cursor 等领先科技创新者已在企业规模部署该平台。 但这种出色的运营表现,完全抵不过激进的私有市场估值。ClickHouse 的 $15 billion Series D 融资对应 $160 million 已确认收入的 93.8x 收入倍数,远高于 Datadog(24.0x)和 Dynatrace(8.0x)等成熟公开可比公司。此外,缺少经审计的客户群透明度时,IPO 前私募股份天然存在二级市场流动性不足和资本损失风险。因此,我们给出明确建议:以高风险评级和估值偏高立场继续跟踪;在投入资本前,投资者应等待倍数正常化或完整的利润率披露。 [CV001, CV003, CV005, CV006, CV007, CV010, CV014]

建议汇总表
指标 / 维度状态 / 数值基准 / 背景决策含义
建议跟踪SaaS 基础设施中位数:买入 / 跟踪持续监测;在倍数压缩或财务披露增加前暂缓配置股权
信心等级中历史报告评级:中对运营增长势头的高信心被对客户群利润率和单位经济性了解有限所抵消
风险评级高企业基础设施中位数:中受 93.8x 倍数压缩风险、二级市场流动性不足和竞争性超大规模云厂商反击驱动
估值立场偏高上市同业平均:5.1x 至 24.0x EV/Revenue当前 $15B 估值已计入零失误执行和长期云收入激进转化
目标回报周期3 至 5 年后期风险投资的标准持有期假设倍数向 25x EV/Sales 正常化,需保持 70% 以上 CAGR 才能实现 2.0x MOIC

数据反映 $15B 的 D 轮定价和 2026 年 1 月指标;基准来自 Multiples.vc 和 UpMarket 二级市场披露的同业交易水平。

[CV001, CV003, CV005, CV006, CV014]
FV001: 建议逻辑

将运营验证点、产品扩张、估值倍数与下行风险串联起来,推导出继续跟踪的建议。

节点 / 连接节点 / 起点终点说明
节点 1 运营规模 [node-scale]ClickHouse Cloud 已有超过 3,000 个客户,ARR 同比增长 >250%,并获得标杆企业采用。
节点 2 AI 平台扩张 [node-expansion]收购 Langfuse 布局 LLM 可观测性,并与 Ubicloud 推出原生 Postgres 服务。
节点 3 极高估值倍数 [node-valuation]D 轮估值 $15B,按已记录收入 $160M 计算,对应 93.8x 收入倍数。
节点 4 与公开市场可比公司的脱节 [node-comps]公开市场龙头 Datadog、Dynatrace 和 Elastic 的交易倍数为收入的 5x 至 24x,意味着估值大幅压缩风险严重。
节点 5 私募市场流动性不足 [node-illiquidity]二级市场股份存在结构性流动性不足,没有有保障的退出时间表,并且面临全部损失风险。
节点 6 投资建议:跟踪 [node-recommendation]出色的运营和强劲的产品需求被过高估值抵消;等待估值倍数重置或利润率变得明朗。
连接 1 运营规模 [node-scale]AI 平台扩张 [node-expansion]促成
连接 2 AI 平台扩张 [node-expansion]极高估值倍数 [node-valuation]已计入定价
连接 3 极高估值倍数 [node-valuation]与公开市场可比公司的脱节 [node-comps]对比
连接 4 与公开市场可比公司的脱节 [node-comps]私募市场流动性不足 [node-illiquidity]叠加加剧
连接 5 私募市场流动性不足 [node-illiquidity]投资建议:跟踪 [node-recommendation]决定
[CV001, CV003, CV005, CV006, CV007, CV008, CV009, CV014]

8.2 当前融资背景与同行估值倍数

ClickHouse 的融资轨迹显示,资本供给前所未有地充裕,估值也在加速扩张。截至 2026 年,公司通过七笔融资交易累计募集了超过 $1.05 billion 的股权和债务资本。2025 年 5 月,Khosla Ventures 领投 $350 million 的 Series C,估值达到 $6.3 billion;同时 Stifel 和 Goldman Sachs 提供了 $100 million 的信贷额度。不到八个月后,Dragoneer 领投 $400 million Series D,估值升至 $15 billion。 这轮快速抬升使 ClickHouse 的隐含收入倍数从 Series C 的 63.5x 膨胀至 Series D 的 93.8x。相比之下,成熟的上市分析与可观测性公司即便已证明具备 GAAP 盈利能力或强劲自由现金流,估值仍大幅折价。公开市场龙头的 EV / 收入倍数分布在 1.5x 至 24.0x:Datadog 为 24.0x,Dynatrace 为 8.0x,Elastic 为 5.1x,老牌供应商 Teradata 为 1.5x。即使在高增长软件供应商中的前四分位,93.8x 也已按近乎完美的执行定价,让私人市场投资者高度暴露于宏观重估或下轮融资修正风险。 [CV001, CV002, CV003, CV004, CV005, CV011]

正 / 反论点表
支柱核心投资论点反论点 / 关键风险论点失效指标
市场与产品实时列式分析的事实标准,相较面向行的 DBMS 具备 100 倍速度优势开源核心支持自托管却无法实现商业变现;云端转化可能停滞云 ARR 增长降至同比 80% 以下,或开源社区出现分叉
客户采用覆盖 3,000 多家客户并快速扩张,其中包括 Meta、Tesla、Capital One 及领先 AI 开发商企业客户集中度和净收入留存率完全未披露失去知名企业背书,或企业客户品牌流失加剧
AI 扩张通过收购 Langfuse 和与 Ubicloud 联合打造统一 Postgres,战略性拓展 LLM 可观测性可观测性市场竞争极其激烈,已有 Datadog、Dynatrace 等成熟厂商Langfuse 集成受阻,或无法向数据库客户交叉销售可观测性
估值与回报定义品类的数据平台,有望从数十亿美元级现代数据栈中占据可观份额93.8x 入场倍数没有安全边际,使投资者暴露于严重估值重置风险下一轮定价融资或 IPO 的企业倍数低于 30x,且收入未加速

评估纳入 ClickHouse 发布的运营指标、二级交易观察和分析师同业倍数。

[CV001, CV003, CV005, CV007, CV008, CV009, CV010]

8.3 乐观、基准与悲观情景预测

评估三年持有期的投资结果,需要把收入增速、利润率结构和退出倍数压缩纳入模型。UpMarket 和 Sacra 的第三方估算预计 ClickHouse 到 2026 年 5 月的年化收入将达到 $250 million,但回报画像仍高度取决于入场纪律和退出倍数环境。 在乐观情景下,ClickHouse 的 ARR 年增 120%,达到 $1.2 billion;随着 AI 查询量激增,Langfuse 可观测性也有效交叉销售,毛利率达到 75%。即使倍数从 93.8x 压缩至 30.0x,公司估值仍达到 $36.0 billion(2.4x MOIC,约 34% IRR)。在基准情景下,ARR 以 75% CAGR 增至 $650 million,毛利率为 68%;但倍数压缩至 22.0x 后,退出估值仅为 $14.3 billion(0.95x MOIC,约 -2% IRR),说明即便经营增长亮眼,回报也会完全被侵蚀。在悲观情景下,ARR 增速降至 40% CAGR(ARR 为 $350 million),倍数正常化至 12.0x,估值降为 $4.2 billion,造成 72% 的资本减值(-35% IRR)。 [CV001, CV003, CV005, CV012, CV013]

乐观 / 基准 / 悲观情景表
情景运营假设3 年 ARR 与利润率画像退出倍数与估值IRR / MOIC 回报概率信号
乐观情景ARR 按 120% CAGR 增长至 $1.2B;云端毛利率达到 75%;Langfuse 和 Postgres 变现成功ARR:$1,200M;毛利率:75%;NRR:140%退出倍数:30x ARR;隐含估值:$36.0BMOIC:2.4x;隐含 IRR:3 年约 34%AI 工作负载查询快速增长,企业从 Snowflake 迁移
基准情景ARR 按 75% CAGR 复合增长至 $650M;毛利率稳定在 68%;云端转化平稳ARR:$650M;毛利率:68%;NRR:125%退出倍数:22x ARR;隐含估值:$14.3BMOIC:0.95x;隐含 IRR:-2%(保本)稳定的开源转化抵消倍数从 93.8x 压缩至 22x 的影响
悲观情景ARR 增长放缓至 40% CAGR,达到 $350M;超大规模云厂商激烈打价格战;毛利率降至 60% 以下ARR:$350M;毛利率:58%;NRR:105%退出倍数:12x ARR;隐含估值:$4.2BMOIC:0.28x;隐含 IRR:-35%(资本严重受损)超大规模云厂商的托管产品压低价格,宏观环境下企业软件预算收紧

情景建模采用 2026 年入场、3 年投资持有期;退出倍数反映 Multiples.vc 可比公司的正常化公开软件估值。

[CV001, CV003, CV005, CV012, CV013]

8.4 战略执行、退出准备度与尽调重点

ClickHouse 正在主动拓展平台覆盖面,以支撑其数十亿美元的市值。2026 年 1 月收购 Langfuse 后,ClickHouse 进入生产级 LLM 可观测性的核心位置,处理非确定性 AI 追踪与评估问题。同时,公司与 Ubicloud 合作推出企业级托管 Postgres 服务,为开发者提供原生 CDC 同步以及统一的事务与分析查询能力。 不过,要达到 IPO 前的投资准备状态,仍需补齐关键的私域尽调缺口。投资者必须评估二级市场流动性不足,以及连续超大轮融资可能带来的清算优先权悬置。 直接投资股权前,重点尽调包括:(1)审计 ClickHouse Cloud 各部署的客户群组级净收入留存率(NRR)和毛利率走势;(2)审查 AI 实验室及头部企业客户的客户集中度;(3)将基础设施单位经济性与 AWS、Google Cloud、Snowflake 的竞争性托管产品对比测试。经核实的数据尚未到手前,继续跟踪该资产是更审慎的受托选择。 [CV006, CV008, CV009, CV010, CV014]

8.5 图表

免责声明

本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开,任何投资决策前都应向管理层询问,并通过一手文件核实。

证据索引

结论
编号陈述可信度来源
CO001 ClickHouse began development in 2009 as an open-source columnar analytics database project at Yandex. 高 SO001, SO009
CO002 ClickHouse, Inc. was incorporated as a Delaware corporation in September 2021 with headquarters in the San Francisco Bay Area. 高 SO001, SO005, SO009
CO003 Aaron Katz serves as Co-Founder and Chief Executive Officer, Alexey Milovidov serves as Co-Founder and Chief Technology Officer, and Yury Izrailevsky serves as Co-Founder and President. 高 SO003, SO005
CO004 ClickHouse raised $400 million in Series D financing in January 2026 led by Dragoneer Investment Group at a post-money valuation of approximately $15 billion. 高 SO003, SO005, SO008
CO005 ClickHouse closed a $350 million Series C financing round in May 2025 led by Khosla Ventures at a reported $6.35 billion valuation. 高 SO004, SO008
CO006 ClickHouse has raised approximately $1.15 billion in total capital across six equity rounds and a $100 million credit facility. 中 SO005, SO008
CO007 ClickHouse Cloud annualized run-rate revenue crossed $250 million in May 2026, more than tripling year-over-year. 高 SO002, SO005
CO008 ClickHouse Cloud served more than 4,000 paying customer organizations as of May 2026, expanding from 3,000 in January 2026. 高 SO002, SO005
CO009 ClickHouse employed between 430 and 623 people across global engineering and go-to-market teams in 2026. 中 SO005
CO010 ClickHouse completed the acquisition of LLM observability platform Langfuse in January 2026 alongside its Series D financing. 高 SO003, SO008
CO011 In December 2022, ClickHouse closed an undisclosed Series B extension with Thrive Capital to fund serverless database architecture. 中 SO008
CO012 AI infrastructure companies face gross margin compression below traditional 70% software levels, with switching costs remaining low across agentic applications. 中 SO006
CO013 ClickHouse's AI customer cohort accounts for less than 12% of total revenue to mitigate customer concentration risk. 中 SO006
CO014 Data ingestion into ClickHouse is complex, and high-concurrency real-time dashboard queries against unoptimized tables can saturate database CPU capacity. 中 SO007
CO015 Major institutional shareholders include Dragoneer, Khosla Ventures, Index Ventures, Benchmark, Coatue, Altimeter, and Nebius Group. 高 SO003, SO008
CM001 Technographic tracking by 6sense records 4,770 customer domains using ClickHouse for big data analytics globally in 2026, representing an estimated 2.83% market share and a #10 category ranking. 中 SM005
CM002 The United States represents the largest geographic concentration of ClickHouse customer domains tracked by 6sense, accounting for 1,625 accounts or 50.34% of identified deployments. 中 SM005
CM003 ClickHouse Cloud expanded its paying commercial customer base from over 2,000 customers announced in May 2025 to more than 3,000 customers by January 2026. 高 SM006, SM007
CM004 ClickHouse commercial ARR expanded at over 250% year-over-year entering 2026, surpassing $350 million in annual recurring revenue following a 300% growth trajectory in 2025. 高 SM004, SM006, SM007
CM005 ClickHouse maintains a net dollar retention rate exceeding 200% alongside gross revenue retention above 99% across its commercial customer base. 中 SM004
CM006 Workloads from AI-native startup accounts such as Anthropic, OpenAI, Harvey, Sierra, and Decagon represent under 12% of ClickHouse commercial revenue despite rapid ingestion volume growth. 中 SM004
CM007 A survey of 1,200 ClickHouse users conducted by ClickHouse Inc. found that 72% of deployments rely on a separate BI or visualization layer for operational dashboarding. 中 SM001
CM008 Thirty-eight percent of surveyed ClickHouse users report that their existing BI tooling fails to fully exploit the columnar database engine's real-time query capabilities. 中 SM001
CM009 ClickHouse acquired open-source LLM observability project Langfuse in January 2026 to incorporate application-layer AI telemetry and evaluation data into its core platform. 中 SM006
CM010 ClickHouse and Ubicloud co-developed an enterprise-grade managed Postgres service backed by NVMe storage and native CDC to unify transactional OLTP and analytical OLAP workloads. 中 SM006
CM011 Extreme-scale production deployments of ClickHouse include Tesla's internal Comet observability platform ingesting 1 billion events per second and Mercado Libre reducing trace query latency by 50x. 中 SM003, SM004
CM012 Alternative managed ClickHouse deployment options and ecosystem partners in 2026 include specialized real-time data platforms such as Tinybird alongside ClickHouse Cloud and self-hosted open-source clusters. 中 SM002, SM008
CM013 Incumbent cloud data platforms Databricks (17.99%), Azure Databricks (17.60%), and Microsoft Azure Synapse (9.01%) command over 44% of tracked big data analytics market share, creating severe competitive density for independent database vendors. 中 SM005
CM014 Comprehensive third-party market sizing specifically isolating pure real-time columnar analytical database revenue from general cloud data warehousing and big data platforms remains unavailable from independent analyst firms. 中 SM001, SM005
CP001 ClickHouse competed primarily as an open-source columnar database management system engineered specifically for real-time analytical processing on large datasets. 中 SP007, SP008
CP002 The competitive landscape for ClickHouse spans five primary categories: cloud data warehouses, direct real-time OLAP engines, transactional databases, time-series engines, and embedded analytics. 中 SP007
CP003 Snowflake provides a fully managed multi-cluster shared-data warehouse separating compute and storage, eliminating cluster administration compared to self-managed ClickHouse. 中 SP007, SP008
CP004 ClickHouse lacks full ACID transaction guarantees and is not designed for transactional workloads with frequent single-row updates and deletes. 中 SP007
CP005 ClickHouse utilizes sparse indexing with granules of 8,192 rows, rendering single-row point lookups significantly less efficient than traditional row-oriented relational databases. 中 SP007
CP006 Complex multi-table joins in ClickHouse face performance degradation, frequently requiring data denormalization and table flattening before ingestion. 中 SP007
CP007 ClickHouse is distributed under the open-source Apache 2.0 license, while offering ClickHouse Cloud as a managed serverless cloud service across major hyperscalers. 中 SP002, SP007
CP008 ClickHouse Cloud benchmarks demonstrate performance and cost advantages against Snowflake, Databricks, BigQuery, and Redshift on large-scale analytical workloads. 中 SP002
CP009 Databricks SQL provides an integrated Lakehouse architecture powered by Spark and Photon over Delta Lake, targeting unified ETL, BI, and machine learning. 中 SP007
CP010 Apache Druid and Apache Pinot represent direct real-time OLAP competitors optimized for streaming event ingestion, but involve complex multi-service distributed architectures. 中 SP007
CP011 DuckDB serves as an embedded in-process OLAP engine executing columnar queries directly inside host applications, challenging ClickHouse on single-node workflows. 中 SP007, SP008
CP012 Operating self-managed ClickHouse clusters entails operational complexity around sharding, replication with ClickHouse Keeper, and manual data rebalancing. 中 SP007
CP013 ClickHouse reported crossing $250 million in annualized revenue run rate in May 2026, tripling its business from the prior year ahead of planned IPO readiness. 中 SP001
CP014 ClickHouse relies on asynchronous background Mutations and Lightweight Deletes for modifications, which rewrite entire data parts and incur I/O overhead. 中 SP007
CP015 ClickHouse’s commercial growth is driven by enabling net-new real-time and AI workloads rather than solely replacing incumbent enterprise data warehouses. 中 SP004
CP016 ClickHouse raised a $400 million Series D financing round led by Dragoneer in January 2026, expanding its capitalization to fund product broadening and AI infrastructure. 中 SP004
CI001 ClickHouse Cloud operates a serverless and dedicated consumption revenue model charging customers for active compute unit-hours and data storage. 中 SI001
CI002 ClickHouse Cloud published compute pricing ranges from $0.22 to $0.75 per compute unit-hour alongside storage pricing of $25.30 to $50 per terabyte-month. 中 SI002
CI003 ClickHouse entered 2026 with a well-capitalized balance sheet and an active credit facility that reduced near-term financing pressure. 高 SI003, SI005
CI004 Sacra estimates that ClickHouse reached $350 million in annual recurring revenue in August 2026 following a company-disclosed $250 million annualized milestone in May 2026. 中 SI001
CI005 Multiples.vc estimates ClickHouse's revenue multiple at approximately 93.8x based on a $15 billion valuation and $160 million in recorded round-date revenue as of January 2026. 中 SI004
CI006 Operating a self-hosted ClickHouse deployment requires an estimated $2,435 to $30,720 per month in infrastructure along with 10% to 20% full-time engineering overhead. 中 SI002
CI007 ClickHouse Cloud served more than 3,000 customers as of January 2026, expanding past 4,000 reported customer accounts by May 2026. 高 SI003, SI007
CI008 ClickHouse secured a $100 million credit facility led by Stifel and Goldman Sachs alongside its $350 million Series C equity financing in May 2025. 高 SI006, SI001
CI009 ClickHouse maintains a distributed workforce spanning employees across more than 10 countries without public disclosure of total personnel payroll or operating expenses. 中 SI008
CI010 Critical software unit economics metrics including gross margin percentage and net revenue retention remain undisclosed in public disclosures and third-party reports. 中 SI001
CI011 ClickHouse Cloud's gross margin is heavily influenced by underlying public cloud infrastructure costs across AWS, GCP, and Azure including compute and egress charges. 中 SI002
CI012 Official ClickHouse Cloud list prices omit secondary operational costs such as data egress fees, high-availability replication multipliers, and backup storage expenses. 中 SI002
CI013 ClickHouse's lifetime equity financing exceeds $1 billion across priced investment rounds through January 2026. 中 SI001, SI007
CI014 Enterprise customers adopting Bring-Your-Own-Cloud deployments pay for proprietary control plane orchestration while running database instances directly inside customer-managed virtual private clouds. 中 SI001
CE001 ClickHouse is a columnar database management system engineered for high-speed online analytical processing and real-time data analytics. 高 SE001, SE004
CE002 ClickHouse is implemented in C++ as a standalone statically-linked binary that executes queries without mandatory external runtimes or operational dependencies. 高 SE001, SE005
CE003 The query engine evaluates analytical queries using vectorized execution over column data chunks, utilizing CPU caches and SIMD instructions for sub-second aggregations. 高 SE001, SE006
CE004 ClickHouse core product offerings encompass open-source self-managed ClickHouse, ClickHouse Local CLI utility, and the fully managed ClickHouse Cloud service. 高 SE001, SE002
CE005 ClickHouse provides native binary wire protocols, HTTP REST endpoints, and compatibility layers for MySQL and PostgreSQL client connectivity. 高 SE001, SE004
CE006 ClickHouse for analytics workloads replaces row-oriented query bottlenecks by compressing columnar data and scanning billions of rows in under one second. 高 SE006, SE002
CE007 ClickHouse Cloud architecture decouples compute execution from shared object storage layers such as AWS S3, Google Cloud Storage, and Azure Blob Storage. 高 SE001, SE008
CE008 In ClickHouse Cloud deployments, stateless worker nodes execute queries while shared object storage provides durability and independent horizontal scaling. 中 SE008
CE009 In January 2026, ClickHouse acquired open-source LLM observability platform Langfuse to extend its real-time analytics stack into non-deterministic AI application tracing. 中 SE003
CE010 ClickHouse launched ClickHouse for PostgreSQL to bring analytical columnar processing directly alongside operational Postgres database environments. 中 SE003
CE011 The MergeTree engine family uses sparse primary indexes with 8,192-row granules instead of traditional B-trees, optimizing sequential scans over random point updates. 高 SE001, SE007
CE012 ClickHouse includes integrated security capabilities comprising role-based access control, TLS encryption, quota management, and continuous monitoring. 中 SE001
CE013 High-frequency single-row mutations and point updates cause heavy write amplification and merge overhead because data parts must be rewritten asynchronously. 中 SE007, SE006
CE014 Memory headroom and concurrency ceilings enforce strict query limits, causing unoptimized joins and wide groupings to fail with out-of-memory errors if unpartitioned. 中 SE007
CU001 ClickHouse Cloud expanded its paying customer base from over 2,000 accounts in May 2025 to more than 3,000 in January 2026, reaching over 4,000 customers by late May 2026. 高 SU001, SU003, SU006, SU007
CU002 ClickHouse reported annual recurring revenue exceeding $250 million by May 2026, following year-over-year ARR expansion of more than 250 percent reported in January 2026. 高 SU001, SU006
CU003 ClickHouse closed a $400 million Series D financing round led by Dragoneer Investment Group in January 2026 at a private valuation of approximately $15 billion. 高 SU005, SU006
CU004 ClickHouse launched its inaugural House Mates partner program in May 2026 with an initial roster of over 60 ISVs, consulting firms, systems integrators, and resellers. 高 SU001, SU002, SU003
CU005 Partner channels account for roughly 25 percent of ClickHouse commercial revenue in the Asia-Pacific region and 10 to 15 percent in other international geographies. 中 SU001
CU006 ClickHouse lacked formal reseller processes and regional systems integration support programs prior to establishing the House Mates framework in mid-2026. 中 SU001, SU002, SU003
CU007 More than 200 software integrations were created by the open-source community in the wild without official support or validation before ClickHouse created a dedicated partner organization. 高 SU001, SU002, SU003
CU008 Anthropic deployed ClickHouse as key analytical infrastructure supporting the development and deployment of its Claude 4 AI model. 高 SU007, SU008
CU009 Tesla, Capital One, Sony, Meta, and Mercado Libre maintain verified production deployments of ClickHouse across real-time analytics and data warehousing workloads. 高 SU003, SU006, SU007
CU010 High-growth software innovators including Lovable, Decagon, Polymarket, Modal, and Airwallex adopted or expanded ClickHouse Cloud deployments in late 2025 and early 2026. 高 SU003, SU006
CU011 Data ingestion pipeline brittleness and custom API connector maintenance represent a primary technical point of failure where real-time marketing analytics deployments on ClickHouse stall. 中 SU004
CU012 Traditional BI tools generate row-oriented SQL with unoptimized JOINs and subqueries that degrade ClickHouse analytical query performance in the absence of a pre-aggregated semantic layer. 中 SU004
CU013 The open-source ClickHouse project has accumulated hundreds of millions of downloads and more than 50,000 GitHub stars, acting as the primary organic top-of-funnel customer discovery vehicle. 高 SU003, SU008
CU014 ClickHouse provides no public disclosure regarding cohort-level net revenue retention (NRR), gross revenue retention (GRR), or customer logo churn rates. 中 SU001, SU006
CU015 ClickHouse's customer mix has broadened from open-source technical adopters toward mainstream enterprise organizations implementing real-time agentic AI and observability applications. 高 SU001, SU002, SU006
CR001 ClickHouse was originally developed at Yandex before ClickHouse, Inc. was incorporated in Delaware in September 2021 with backing from venture investors to commercialize the open-source Apache 2.0 database. 高 SR008, SR009
CR002 ClickHouse Cloud enforces default soft guardrails including caps of 1,000 databases, 5,000 tables, 50,000 partitions, 10,000 parts, and 1,000 concurrent queries per replica. 中 SR001
CR003 ClickHouse Cloud warehouses impose a soft ceiling of 50 combined replicas per warehouse to protect the shared ClickHouse Keeper coordination ensemble from stability degradation. 中 SR001
CR004 ClickHouse server failure modes include insert rejections, memory exhaustion, read-only replicas, and disk saturation requiring inspection of system tables and background merge logs. 高 SR001, SR002
CR005 ZooKeeper and Keeper coordination interruptions force replica reconnection and reconciliation cycles during which replicated tables can become temporarily read-only for writes. 中 SR003
CR006 Epoch counter rollover in coordination engines precipitates unscheduled leader re-elections and transient write unavailability across active database clusters. 中 SR003
CR007 In production shadowing at Temporal Cloud, 64-bit integer accumulators in ClickHouse summed large account storage totals into negative values due to integer overflow. 中 SR004
CR008 Validating ClickHouse against a legacy warehouse required six weeks of continuous query shadowing because floating-point summation differences and timing variations masked true logic mismatches. 中 SR004
CR009 ClickHouse completed a $400 million Series D financing round in January 2026 led by Dragoneer, expanding capital resources alongside the acquisition of Langfuse and launch of a managed Postgres service. 中 SR006
CR010 ClickHouse secured a $100 million credit facility led by Stifel and Goldman Sachs in May 2025 to supplement Series C equity funding. 中 SR007
CR011 ClickHouse Cloud relies on hyperscaler object storage and managed compute infrastructure across AWS, Microsoft Azure, and Google Cloud Platform for cloud-hosted deployments. 高 SR001, SR006
CR012 ClickHouse partners with third-party infrastructure providers including Ubicloud for managed Postgres and Stifel and Goldman Sachs for credit facilities, introducing partner operational dependencies. 高 SR006, SR007
CR013 Automated pull request analysis and linting tools evaluate ClickHouse codebase modifications for safety, compliance, and correctness prior to merge. 中 SR005
CR014 ClickHouse Cloud enterprise customer revenue concentration remains undisclosed publicly, creating uncertainty regarding top-account churn exposure. 低
CV001 ClickHouse's private-market pricing remains a stretched entry point relative to public software peers. 高 SV002, SV003, SV008
CV002 ClickHouse has raised over $1.05 billion in cumulative equity and debt capital across its financing history through 2026. 高 SV005, SV007
CV003 ClickHouse's $15 billion post-money valuation implies an estimated 93.8x revenue multiple against recorded revenue of $160 million as of January 2026. 中 SV008
CV004 ClickHouse's revenue multiple expanded from 63.5x at its $6.3 billion Series C valuation in May 2025 to 93.8x at its Series D in January 2026. 中 SV008
CV005 Publicly listed analytics and monitoring comparables trade at substantially lower revenue multiples, including Datadog at 24.0x, Dynatrace at 8.0x, Elastic at 5.1x, and Teradata at 1.5x. 中 SV008
CV006 Secondary pre-IPO market transactions in private shares carry severe illiquidity, lack guaranteed exit timelines, and present risks of total capital loss. 中 SV009
CV007 ClickHouse reported serving over 3,000 customers on ClickHouse Cloud with annual recurring revenue growing more than 250 percent year over year at the time of its Series D. 高 SV002, SV004
CV008 ClickHouse acquired open-source LLM observability platform Langfuse in January 2026 to expand into production AI application monitoring and evaluation. 高 SV002, SV003
CV009 ClickHouse introduced an enterprise-grade managed Postgres service in partnership with Ubicloud to unify transactional and analytical database workloads for AI applications. 高 SV002, SV003
CV010 ClickHouse's column-oriented architecture utilizes vectorized query execution and compression to deliver up to 100x faster analytical query performance than traditional row-oriented databases. 中 SV006
CV011 In May 2025 ClickHouse raised $350 million in Series C financing led by Khosla Ventures alongside a $100 million conventional credit facility led by Stifel and Goldman Sachs. 高 SV005, SV007
CV012 Third-party pre-IPO estimates by UpMarket model ClickHouse's valuation at $15.3 billion while referencing Sacra estimates of annualized revenue reaching $250 million by May 2026. 中 SV009
CV013 A conservative baseline valuation model implies ClickHouse requires sustained 80 percent ARR compounding to compress its entry multiple toward 25x over a three-year holding horizon. 中 SV008, SV009
CV014 ClickHouse warrants an overall investment recommendation of track with high risk rating and stretched valuation stance given the divergence between operational momentum and extreme pricing. 中 SV008, SV009
来源
编号出版方标题引文
SO001 ClickHouse, Inc. Our Story Work on ClickHouse began in 2009 with the vision of being the fastest OLAP database on earth.
SO002 ClickHouse, Inc. ClickHouse tops $250M ARR and 4,000 customers, launches Claude-powered agents at Open House 2026 | ClickHouse Annual run-rate revenue more than tripled year-over-year; new CostBench benchmark shows ClickHouse Cloud at 23x better cost-performance than the nearest cloud data warehouse
SO003 ClickHouse, Inc. ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure ClickHouse, a leader in real-time analytics, data warehousing, observability, and AI/ML, announced today the close of its Series D financing, raising $400 million.
SO004 ClickHouse, Inc. ClickHouse raises $350 million Series C to power analytics for the AI era ClickHouse, Inc., a leader in real-time analytics, data warehousing, observability, and AI/ML, today announced it has raised $350 million in Series C financing.
SO005 Silicon Valley Investclub ClickHouse — Funding, Valuation & Leadership - Silicon Valley Investclub As of January 2026, ClickHouse has achieved a valuation of $15 billion and serves over 3,000 enterprise customers
SO006 BigGo Finance ClickHouse's Aaron Katz: AI Margins Must Improve, and the Real Threat Isn't What You Think ClickHouse CEO Aaron Katz offers a contrarian take on the AI infrastructure boom: the cycle is real and accelerating, but the winners will be defined by revenue durability, not gross-margin optics.
SO007 Improvado ClickHouse Analytics: Complete Guide for Marketers 2026 ClickHouse is fast. Getting data into ClickHouse is slow, brittle, and expensive. This is where most implementations stall.
SO008 Revenue Memo ClickHouse funding: Every round, investor, and valuation (2026) ClickHouse has raised about $1.15 billion since it became a company in September 2021. That covers six equity rounds, from a Series A that month to a $400 million Series D in January 2026, plus a $100 million credit facility.
SO009 Wikipedia ClickHouse ClickHouse, Inc. was incorporated in San Francisco, California, in September 2021 to commercialize the open-source ClickHouse database.
SM001 Basedash Best ClickHouse BI and dashboard tools (2026) | Basedash According to the ClickHouse 2025 State of ClickHouse Survey, 72% of ClickHouse users rely on a separate BI or visualization layer for dashboarding — and 38% report that their current BI tool doesn’t fully leverage ClickHouse’s real-time capabilities (ClickHouse Inc., “State of ClickHouse Survey,” 2025, 1,200 respondents across 40 countries).
SM002 Tinybird Best managed ClickHouse® services compared in 2026 | Tinybird Skip the infra work. Deploy your first ClickHouse® project now.
SM003 ClickHouse The future of ClickHouse Cloud (feat. Mercado Libre case study) | ClickHouse Engineering leaders at Mercado Libre share their migration story, achieving 100% trace sampling, and cutting query times by 50x.
SM004 BigGo Finance ClickHouse's Aaron Katz: AI Margins Must Improve, and the Real Threat Isn't Visible Yet — BigGo Finance Nearly every AI-native firm — Anthropic, OpenAI, Harvey, Sierra, Decagon — runs on ClickHouse, yet that entire AI cohort represents less than 12% of revenue.
SM005 6sense Clickhouse - Market Share, Competitor Insights in Big Data Analytics Around the world in 2026, over 4,737 companies have started using Clickhouse as big-data-analytics tool. Companies using Clickhouse for big-data-analytics are majorly from United States with 1,625 customers. 50.34% of Clickhouse customers are from the United States.
SM006 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure | ClickHouse The company now serves more than 3,000 customers on its fully managed service, ClickHouse Cloud, with ARR growing more than 250 percent year over year.
SM007 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era | ClickHouse The momentum behind ClickHouse is accelerating: the company grew over 300% during the past year and now serves over 2,000 customers across a range of industries from fintech and transportation to consumer and healthcare.
SM008 ClickHouse Fast Open-Source OLAP DBMS | ClickHouse Column-oriented databases are better suited to OLAP scenarios. They are at least 100x faster in processing most queries. ClickHouse uses all available system resources to their full potential to process each analytical query as fast as possible.
SP001 TechCrunch ClickHouse triples annualized revenue to $250M, charting a path toward an IPO Database provider ClickHouse has crossed $250 million in annualized revenue run rate, tripling its business from last year, Yury Izrailevsky, co-founder and president of product and technology, told TechCrunch.
SP002 ClickHouse ClickHouse Cloud - Fast, Efficient, Effortless ClickHouse Cloud: the fastest, most cost-efficient way to build real-time analytics, observability, and AI-powered data applications. Fully managed, with sub-second query performance at any scale.
SP003 SoftwareSuggest 10 Best ClickHouse Alternatives & Competitors in 2026 SoftwareSuggest uses a combination of user responses to the question, “Which other vendor did you consider in your evaluation?”. We also consider the total review volume for those vendors/products within the market.
SP004 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure ClickHouse, a leader in real-time analytics, data warehousing, observability, and AI/ML, announced today the close of its Series D financing, raising $400 million. The round was led by Dragoneer Investment Group
SP005 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era ClickHouse, Inc., a leader in real-time analytics, data warehousing, observability, and AI/ML, today announced it has raised $350 million in Series C financing.
SP006 ClickHouse ClickHouse: Fast Open-Source OLAP DBMS Powering agentic systems with millisecond queries at petabyte scale.
SP007 Flexera ClickHouse alternatives: 10 true competitors (2026) ClickHouse is primarily designed for OLAP (Online Analytical Processing), which means it excels at handling complex queries and large datasets. But, it’s not perfect for online transaction processing (OLTP) workloads.
SP008 Data Stack Hub 15 Best ClickHouse Alternatives & Competitors in 2026 Snowflake is a fully managed cloud-native data warehouse that enables organizations to analyze massive volumes of structured, semi-structured, and unstructured data without managing infrastructure.
SI001 Sacra ClickHouse revenue, valuation & funding Sacra estimates that ClickHouse hit $350M in annual recurring revenue (ARR) in August 2026, up from $160M at the end of 2025.
SI002 Improvado ClickHouse Pricing 2026: Cloud vs Self-Hosted TCO Guide ClickHouse pricing in 2026 follows two distinct models... charges $0.22-$0.75 per compute unit-hour... It also charges $25.30-$50/TB-month for storage... requires 10-20% FTE engineering overhead.
SI003 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure SAN FRANCISCO — January 16, 2026 — ClickHouse, a leader in real-time analytics, data warehousing, observability, and AI/ML, announced today the close of its Series D financing, raising $400 million... The company now serves more than 3,000 customers on its fully managed service, ClickHouse Cloud, with ARR growing more than 250 percent year over year.
SI004 Multiples ClickHouse Revenue, Valuation, Funding & Investors ClickHouse is currently valued at $15B, based on its Series D in January 2026... ClickHouse's latest recorded revenue is $160M, as of its Series D in January 2026. At a $15B valuation, that implies a 93.8x revenue multiple.
SI005 UpMarket ClickHouse Stock for Accredited Investors | Pre-IPO Shares In January 2026, ClickHouse raised $400 million in its Series D at a $15 billion post-money valuation, led by Dragoneer Investment Group
SI006 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era SAN FRANCISCO, CA — May 29, 2025 — ClickHouse, Inc., a leader in real-time analytics, data warehousing, observability, and AI/ML, today announced it has raised $350 million in Series C financing... In addition to this financing, ClickHouse has secured a $100 million credit facility led by Stifel and Goldman Sachs.
SI007 Silicon Valley Investclub ClickHouse — Funding, Valuation & Leadership As of January 2026, ClickHouse has achieved a valuation of $15 billion and serves over 3,000 enterprise customers including Meta, Tesla, Sony, Anthropic, Lyft, and Instacart.
SI008 ClickHouse Job Openings at ClickHouse | ClickHouse Careers ClickHouse is proud to have employees in over 10 countries. We believe that diverse and distributed workplaces working inclusively are an essential advantage.
SE001 ClickHouse Architecture overview - ClickHouse Documentation ClickHouse is built in C++ as a single, statically-linked binary without dependencies.
SE002 ClickHouse ClickHouse: Fast Open-Source OLAP DBMS Analyze billions of rows in real time with millisecond results.
SE003 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate product innovation, acquires Langfuse, and launches ClickHouse for PostgreSQL ClickHouse is thrilled to announce the acquisition of Langfuse, the open-source LLM observability platform.
SE004 Flexera ClickHouse architecture 101: A comprehensive overview (2026) ClickHouse is an open source, columnar database management system specifically designed for high-speed OLAP (Online Analytical Processing).
SE005 Instaclustr Understanding ClickHouse®: Products, architecture, tutorial and alternatives ClickHouse is an open source columnar database management system for online analytical processing (OLAP) and real-time data analytics.
SE006 Khimananda Oli ClickHouse for Analytics Workloads: 2026 Performance Guide ClickHouse for analytics workloads solves this specific bottleneck by storing data column-wise and compressing it aggressively, enabling billion-row scans in under a second.
SE007 ChistaDATA ClickHouse Performance Tuning and Optimization by ChistaDATA The sort key and partition decide how many granules a query must read; the query shape decides how much work is done per granule; MergeTree settings decide write amplification
SE008 Diacto ClickHouse Cloud Architecture: Components, Scalability, and Best Practices At its core, it separates management from execution, storage from compute, and, in advanced deployments, one workload’s compute from another.
SE009 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era ClickHouse, Inc., a leader in real-time analytics, data warehousing, observability, and AI/ML, today announced it has raised $350 million in Series C financing.
SU001 CRN ClickHouse Launches Inaugural Partner Program Amid Global Expansion, Changing Data Demands In An AI World The launch of House Mates, the new ClickHouse partner community and program, comes as enterprise customers rapidly adopt the company’s OLAP (online analytical processing) database for real-time analytics, AI, observability, data warehousing and other data-intensive workloads.
SU002 ChannelE2E ClickHouse’s partner bet: More go-to-market engine than a badge program When I came in, ClickHouse already had a lot of adoption, but customers were struggling from an integration perspective. There were about 200 integrations available, and many were built by the open-source community.
SU003 ClickHouse Introducing House Mates: the ClickHouse partner community and program | ClickHouse ClickHouse Cloud has grown to over 4,000 customers in just four years, and that growth is global. Companies like Meta, Cursor, Sony, Tesla, Capital One, Lovable, Decagon, Polymarket, and Airwallex are running ClickHouse in production across real-time analytics, observability, AI, data warehousing, and just about every workload where speed and scale matter.
SU004 Improvado ClickHouse Analytics: Complete Guide for Marketers 2026 ClickHouse is fast. Getting data into ClickHouse is slow, brittle, and expensive. This is where most implementations stall.
SU005 Dakota ClickHouse Raises $400 Million Series D as Demand for AI-Driven Analytics Accelerates ClickHouse has raised $400 million in a Series D funding round led by Dragoneer Investment Group, valuing the company at roughly $15 billion and reinforcing its position as one of the most important players in modern data infrastructure.
SU006 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure | ClickHouse The company now serves more than 3,000 customers on its fully managed service, ClickHouse Cloud, with ARR growing more than 250 percent year over year. Over the past three months, customers including Capital One, Lovable, Decagon, Polymarket, and Airwallex have adopted the platform or expanded existing deployments.
SU007 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era | ClickHouse The momentum behind ClickHouse is accelerating: the company grew over 300% during the past year and now serves over 2,000 customers across a range of industries from fintech and transportation to consumer and healthcare. New customers include Anthropic, Tesla, and Argentina’s Mercado Libre, among others.
SU008 ClickHouse Fast Open-Source OLAP DBMS | ClickHouse ClickHouse played an instrumental role in helping us develop and ship Claude 4.
SR001 ClickHouse Usage limits - ClickHouse Documentation To prevent this, ClickHouse Cloud enforces limits across several operational dimensions. Soft limits apply at three different scopes.
SR002 ClickHouse GitHub - ClickHouse/clickhouse-diagnostics: Clickhouse diagnostics tooling When a ClickHouse server misbehaves — inserts rejected, queries running out of memory, a replica read-only, a disk filling up — the answers are already in its system tables, its logs and the host it runs on.
SR003 Altinity ClickHouse® Keeper: The Long Road from “ZooKeeper in C++” to Something More During that time, tables may still be available for reads, but writes can fail because the replicated table is temporarily read-only. Worse, this is not a scheduled maintenance window.
SR004 Temporal Technologies Running out of runway: migrating Temporal Cloud's billing store to ClickHouse A couple of days into shadowing, the comparison began reporting negative usage totals from ClickHouse. Because usage totals can't be negative, the sign pointed immediately at an overflow — we didn't need the other system's answer to know this one was wrong.
SR005 SkillBundle review skill by clickhouse | SkillBundle
SR006 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure | ClickHouse SAN FRANCISCO — January 16, 2026 — ClickHouse, a leader in real-time analytics, data warehousing, observability, and AI/ML, announced today the close of its Series D financing, raising $400 million.
SR007 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era | ClickHouse In addition to this financing, ClickHouse has secured a $100 million credit facility led by Stifel and Goldman Sachs.
SR008 ClickHouse Fast Open-Source OLAP DBMS | ClickHouse Column-oriented databases are better suited to OLAP scenarios. They are at least 100x faster in processing most queries. ClickHouse uses all available system resources to their full potential to process each analytical query as fast as possible.
SR009 Wikipedia ClickHouse ClickHouse, Inc. was incorporated in San Francisco, California, in September 2021 to commercialize the open-source ClickHouse database.
SV001 FundedIQ ClickHouse: Funding, Investors & Team (Oct 2026) | FundedIQ ClickHouse has raised $400M across 3 funding rounds on record, the most recent a Series D of $400M announced in Jan 2026. That round is above the $5.5M median round size for Data and Analytics companies tracked on FundedIQ over the last two years.
SV002 ClickHouse ClickHouse raises $400M Series D led by Dragoneer to accelerate expansion across analytics and AI infrastructure The company now serves more than 3,000 customers on its fully managed service, ClickHouse Cloud, with ARR growing more than 250 percent year over year.
SV003 Business Wire ClickHouse Raises $400M Series D Led by Dragoneer to Accelerate Expansion Across Analytics and AI Infrastructure ClickHouse, a leader in real-time analytics, data warehousing, observability, and AI/ML, announced today the close of its Series D financing, raising $400 million. The round was led by Dragoneer Investment Group, with participation from Bessemer Venture Partners, GIC, Index Ventures, Khosla Ventures, Lightspeed Venture Partners, accounts advised by T. Rowe Price Associates, Inc., and WCM Investment Management.
SV004 The AI Insider ClickHouse Secures $400M Series D Led by Dragoneer to Accelerate Expansion Across Analytics and AI Infrastructure ClickHouse raised $400 million in a Series D led by Dragoneer, citing rapid growth with over 3,000 ClickHouse Cloud customers and ARR up more than 250% year over year.
SV005 ClickHouse ClickHouse raises $350 million Series C to power analytics for the AI era Today’s round follows earlier investments of over $300 million, bringing total funding to over $650 million. This combined funding will be used to scale product development, support global expansion, and deepen partnerships with customers and technology providers building the next wave of AI-native applications. In addition to this financing, ClickHouse has secured a $100 million credit facility led by Stifel and Goldman Sachs.
SV006 ClickHouse ClickHouse: Fast Open-Source OLAP DBMS Column-oriented databases are better suited to OLAP scenarios. They are at least 100x faster in processing most queries. ClickHouse uses all available system resources to their full potential to process each analytical query as fast as possible.
SV007 Tracxn 2026 Funding Rounds & List of Investors - ClickHouse Dragoneer Investment Group... located in San Francisco (United States), made their first investment in ClickHouse on Jan 16, 2026 in its Series D round.
SV008 Multiples ClickHouse Revenue, Valuation, Funding & Investors ClickHouse's latest recorded revenue is $160M, as of its Series D in January 2026. At a $15B valuation, that implies a 93.8x revenue multiple. Revenue grew from $100M in May 2025 to $160M in January 2026, while the revenue multiple moved from 63.5x to 93.8x.
SV009 UpMarket ClickHouse Stock for Accredited Investors | Pre-IPO Shares Pre-IPO investments carry significant risks. ClickHouse shares are illiquid, meaning there is no public market to sell them quickly. There is no guaranteed exit timeline or return. The investment is speculative in nature, and investors should be prepared for the possibility of total loss.