ClickHouse raises $400M as AI companies including Anthropic, OpenAI, and Tesla use it for real-time analytics

Summary

ClickHouse closed a $400 million funding round led by Dragoneer Investment Group in San Francisco, assembled over the holiday period without a formal process. Existing and new investors participated, though specific names and valuation were not disclosed. The raise arrived alongside the acquisition of Langfuse, a Berlin-based LLM observability platform, signaling a deliberate push into the agentic AI infrastructure stack.

Aaron Katz, who joined after 12 years at Salesforce, leads the company as CEO. ClickHouse's client roster now includes Anthropic, OpenAI, Tesla, Cursor, Lovable, and Ramp, covering a broad cross-section of both AI-native and fintech enterprise customers. Tesla is cited as the scale benchmark, ingesting one billion events per second into the database.

AI Discovery Is Driving Pipeline

One of the more telling data points is how Anthropic itself landed on ClickHouse. The company's team asked Claude what observability database to use and the model recommended its own vendor. A European fintech CEO separately reported that every LLM consulted for their real-time analytics build returned the same answer. This LLM-native word-of-mouth is becoming a legitimate enterprise sales channel, bypassing traditional analyst reports and vendor evaluations entirely.

The broader shift Katz describes is from batch-oriented analytics toward sub-second, real-time query responses. Ramp's spend dashboards are the consumer-facing example, where users now expect millisecond refresh times rather than the two-to-five-second waits that were acceptable under older architectures.

Engineering Roots and Open Source Leverage

ClickHouse originated inside Yandex, the Russian internet company, built by roughly a dozen engineers. Co-founder Alexei Milavidov, now based in Amsterdam, named the product as shorthand for Clickstream Data Warehouse and made the decision to open source it in 2016. That move created a frictionless adoption path that now accelerates enterprise deployment, with developers able to push the database to production without any vendor relationship.

The commercial layer sits on top as ClickHouse Cloud, a managed service running on AWS, Google Cloud, and Microsoft Azure, with an Alibaba partnership covering China. The pricing model is consumption-based, storage and compute, auto-scaling to handle the spiky workload profiles typical of analytical applications.

Global Footprint Is Already Substantial

The company employs people across 22 countries, with half of both its workforce and revenue outside North America. For Japan, ClickHouse structured a joint venture with Japan Cloud, formerly Sunbridge, the same firm that executed Salesforce's Japan go-to-market two decades ago. John Roos, former U.S. Ambassador to Japan and Salesforce board member, is among the investors through Geodesic Capital, providing relationship infrastructure in a market Katz describes as requiring deep localization that most foreign software companies underestimate.