ClickHouse’s 50M ARR Milestone: What Changes for Data Teams?
By
Alexander Gran
Founder & Editor
ClickHouse announced at Open House 2026 on May 27, 2026 that its serverless cloud offering has crossed $250 million in annual run-rate revenue — more than triple compared to a year earlier — and that it has surpassed 4,000 customers after adding more than 1,000 net new customers since January 2026. The company also launched ClickHouse Agents, a fully managed agentic analytics service powered by Anthropic’s Claude.
For data teams evaluating analytics infrastructure or already running ClickHouse, here’s what the announcements mean in practice.
The growth context
ClickHouse raised a $400 million Series D in January 2026 when it had just crossed 3,000 customers. One quarter later, it’s at 4,000. The ARR tripled year-over-year. Recent customer additions include Capital One, Lovable, Decagon, Polymarket, and Airwallex; the existing customer base includes Anthropic, Meta, Cursor, Sony, Tesla, Memorial Sloan Kettering, Lyft, and Instacart.
For teams evaluating database vendors, this growth pattern indicates platform momentum and reduced vendor-risk. A database that’s tripling ARR with enterprise customers across financial services, AI companies, and consumer tech is not a narrow niche product — it’s being validated across varied, demanding workloads.
ClickHouse Agents
ClickHouse Agents is the most substantive product announcement from Open House 2026. It’s a no-code agent builder built directly into ClickHouse Cloud, powered by Claude. Out of the box, it includes:
- A chat interface for querying ClickHouse data in natural language
- A sandboxed code interpreter
- Shareable artifacts (exportable results and visualizations)
- Skills management and memory
- Multi-agent workflows
- Native integration with ClickHouse data and MCP-compatible third-party systems
- Integration with the AWS Agent Registry
The practical value: data teams with non-technical stakeholders can use ClickHouse Agents to run analytics queries without writing SQL. Engineers can build automated reporting agents that connect ClickHouse to other systems via MCP. Multi-agent workflows allow orchestrating multiple analytics tasks without building custom infrastructure.
This positions ClickHouse not just as a query-fast analytics database but as a platform for agentic analytics — where AI agents reason over data and return actionable results, rather than requiring humans to write and execute queries.
CostBench
ClickHouse launched CostBench, an open benchmark comparing ClickHouse Cloud, Snowflake, Databricks, BigQuery, and Redshift on cost-performance — applying each vendor’s actual compute billing to the same analytical workload. The claim: ClickHouse Cloud is the only system in the “Fast and Low-Cost” zone across datasets, with the nearest competitor at 23x worse on cost-performance.
Vendor-published benchmarks should always be read critically — they’re designed to favor the publisher. The value of CostBench is that it’s open and reproducible: the methodology is published, and you can run the benchmark against your own workload. For teams doing a thorough evaluation, running CostBench on a representative workload is more useful than accepting the headline number.
The benchmark is available at clickhouse.com/benchmarks.
Other product announcements
Managed Postgres in public beta: ClickHouse is offering managed Postgres with native integration to ClickHouse analytics. For AI applications that need both transactional state (Postgres) and high-throughput analytics (ClickHouse), this reduces the infrastructure complexity of running both databases separately.
Full-text search generally available: Previously, full-text search required workarounds. General availability makes ClickHouse more directly useful for observability and AI-grounding workloads that need to search log text or unstructured content.
Agentic onboarding: New users can go from sign-up to first production query with AI-assisted schema setup, without manual configuration.
House Mates partner program: ClickHouse’s first formal partner ecosystem with 25+ technology partners and 35+ services partners, including dbt Labs, Fivetran, Sigma, and Temporal.
Who should pay attention
Data engineering and analytics teams already evaluating or running ClickHouse: the agent and Postgres integration announcements are directly relevant to your roadmap. Review current ClickHouse Cloud pricing after the launch period to account for any changes that follow from the cost-performance claims.
Teams on Snowflake, BigQuery, or Redshift evaluating alternatives: CostBench is worth running on your actual workload if cost is a primary concern. Don’t take the 23x headline at face value — run the benchmark on your specific usage pattern.
Teams not running analytics at the scale where database infrastructure matters: this announcement is vendor news, not a workflow change. You don’t need to take action.