Database company ClickHouse just crossed $250 million in annualized revenue run rate, tripling its business in a single year. Co-founder and president of product and technology Yury Izrailevsky shared the numbers with TechCrunch AI, and he isn’t slowing the forecast: he expects revenue to hit the high nine digits by the end of the year. For a company that’s under five years old, that’s a steep climb.
The quick version for busy readers
- Revenue: $250M annualized run rate, up 3x year over year.
- Valuation: $15 billion, set in January after a $400M Series D led by Dragoneer.
- The multiple: over 60x annualized revenue. That’s premium pricing, even by today’s standards.
- The signal: an IPO within the next few years, per Izrailevsky.
Why this matters
ClickHouse sells managed cloud hosting for its open-source database, the kind built to chew through the massive datasets AI agents generate. That positioning is the whole story here. As companies push agents into production, they need infrastructure that can query enormous volumes of data fast and cheap. ClickHouse is selling exactly that, and the customer list backs it up: Anthropic, Meta, Capital One, and Decagon are among its 4,000-plus customers.
What stands out is the pricing argument. Izrailevsky claims the paid managed service ends up cheaper for clients than running the free open-source version themselves. He called it “something that’s a little counterintuitive, but it also has been a big tailwind for us.” That counterintuitive pitch is doing real work on the revenue line.
The IPO context
ClickHouse isn’t moving in a vacuum. TechCrunch AI reports the company joins a small but growing group of startups signaling public-market plans, with the IPO window expected to swing open after SpaceX’s June debut, followed by anticipated listings from OpenAI and Anthropic later this year. After a long drought, the exits are lining up.
The hiring tells the same story. Last fall ClickHouse brought on Jimmy Sexton as CFO. He previously ran investor relations at Snowflake, one of ClickHouse’s direct competitors. Naming a CFO with that background is one of the clearest tells that a company is building toward public markets. You don’t hire that profile to stay private.
Where it came from
There’s a backstory worth knowing. The core technology was built inside Russian search giant Yandex 17 years ago, then spun out as an independent startup in 2021. So the engine is mature even if the company is young. That gap between battle-tested tech and a fresh corporate structure is part of why growth has been this fast.
ClickHouse has also been buying. It’s acquired six startups so far, including Langfuse, a tool that helps developers track and evaluate AI agent performance. Izrailevsky said the company plans to stay acquisitive, targeting “relatively young, but showing very promising technology” startups, usually open-source, that round out its core product.
My take
The 60x revenue multiple is the number to watch. It prices ClickHouse for perfection, which means the path to a public listing depends on holding this growth rate through a market that punishes any slip. Tripling revenue once is impressive. Doing it while justifying a $15 billion tag is a different test.
But the strategic logic is sound. ClickHouse sits underneath the AI agent boom rather than competing in the crowded model layer. Every company shipping agents needs somewhere to store and query the data those agents produce, and that demand isn’t tied to which model wins. Picking the infrastructure layer over the application layer is a smart place to stand right now.
What to watch next
- Whether revenue actually reaches the high nine digits by year-end, as forecast.
- More acquisitions, especially in open-source AI tooling like the Langfuse deal.
- Concrete IPO filings, likely after the SpaceX and OpenAI listings test investor appetite.
- Whether the “managed is cheaper than self-hosted” pitch keeps converting open-source users into paying customers.
The full breakdown, including Izrailevsky’s revenue forecast and the acquisition strategy, is available at the original source.