Building the Future of Data Infra on the Back of Open Source

Highlights

• ClickHouse maintains a monthly release cycle that has produced over 120 releases and nearly 3,000 contributors since 2016.

• ClickHouse Cloud separates stateless compute from durable S3 object storage to enable dynamic autoscaling and pay-as-you-go pricing.

• More than 50 percent of ClickHouse revenue originates outside the United States across 38 cloud regions.

• Anthropic, OpenAI, and major AI labs utilize ClickHouse under the hood for large-scale analytical and agentic workloads.

• ClickHouse launches on the Vercel Marketplace with one-click provisioning, a 30-day free trial, and $300 in free credits.

Timeline

1. 00:00 - 02:30

   Open Source Origins and Growth Velocity

   Enterprise adoption prompted the creation of a commercial company in 2021 to govern the open-source project and build commercial products. A rapid release cadence drove continuous community engagement and rapid deployment of ClickHouse Cloud within a year of company formation.

   • Alexey Milavidov initially developed ClickHouse in 2009 before open-sourcing the project in 2016.
   • Major technology enterprises including Cloudflare, Uber, Lyft, and Spotify adopted the database engine early on.
   • Monthly open-source releases accumulated over 100,000 pull requests from nearly 3,000 contributors.
2. 02:30 - 06:20

   Cloud Architecture and Global Deployment Models

   Architectural separation removes the scaling limitations of local-storage clusters. Flexible deployment models include fully managed cloud, customer-account deployments, and air-gapped private environments. Providing a functionally equivalent open-source version serves as an insurance policy against vendor lock-in while functioning as a marketing channel.

   • Traditional open-source deployments use a shared-nothing architecture with local SSDs that require peak-load over-provisioning.
   • ClickHouse Cloud decouples compute from S3 object storage to eliminate rigidity and reduce infrastructure costs.
   • Deployments span three major cloud providers across 38 regions globally to meet local residency and latency requirements.
3. 06:20 - 12:20

   Platform Expansion and Control Plane Uniformity

   Scaling a collection of data management products requires hiding configuration complexity behind a unified platform. Comprehensive infrastructure management incorporates identity management, encryption key handling, regulatory compliance certifications, and cost management tools designed for both human users and automated agents.

   • Product offerings expanded from core analytics to real-time observability, data warehousing, transactional Postgres, and AI engineering.
   • Seven corporate acquisitions integrated diverse technologies into a unified platform managed via Kubernetes.
   • Control plane standardization delivers consistent user interfaces, single sign-on security, and full API and command-line interface parity.
4. 12:20 - 16:35

   AI Workloads and Vercel Marketplace Integration

   AI models and LLMs frequently recommend ClickHouse based on public documentation, code repositories, and forum discussions. Four conference keynote companies—Cerebris, TypeSafe AI, Shopify, and Vercel—rely on ClickHouse to handle massive data scales despite running distinct business models.

   • LangFuse, an open-source AI observability platform built on ClickHouse, integrates natively with the Vercel AI SDK.
   • Agentic workflows collapse data silos by combining application data, traces, and analytics into unified queries.
   • ClickHouse joins the Vercel Marketplace with managed invoicing, trials, and access to all cloud capabilities.