ClickHouse deepens Microsoft Fabric tie-up with OneLake

ClickHouse has expanded its collaboration with Microsoft, adding a native ClickHouse workload to Microsoft Fabric and extending OneLake interoperability.

The companies outlined four elements in the broader tie-up: a public preview of the ClickHouse workload for Microsoft Fabric, general availability of OneLake read support, a public preview of OneLake write support, and the listing of ClickHouse Bring Your Own Cloud in Microsoft Marketplace.

The new Fabric workload brings the ClickHouse query engine into the Microsoft Fabric environment, allowing users to query data in OneLake from within Fabric instead of moving to a separate service.

The workload is available in public preview through the Fabric workload hub and is aimed at data teams running interactive analytics, dashboards, and observability workloads on data stored in OneLake.

Fabric integration

Microsoft Fabric is positioned as a unified data platform, with OneLake serving as a single storage layer across the environment. Adding ClickHouse as a native workload gives Microsoft customers another analytical engine that can run against the same data estate.

Support for reading Apache Iceberg tables from OneLake through OneLake Table APIs is now generally available. This allows users to query Iceberg tables managed in OneLake directly from ClickHouse without duplicating data or carrying out a separate ingestion step.

The arrangement lets organisations keep OneLake as the main governed data store while using ClickHouse for query processing. Iceberg has become a widely used open table format for lakehouse deployments, making interoperability a key issue for software suppliers working across cloud analytics systems.

Support for writing Apache Iceberg tables back to OneLake is also now in public preview. This will let users write results into OneLake as governed tables that can then be read by other Fabric services, including Power BI and notebook environments.

According to ClickHouse, the write function uses OneLake Table APIs with credential vending. In practical terms, this extends the relationship beyond read access and allows ClickHouse to fit more directly into a data workflow that starts and ends in Microsoft's lakehouse layer.

Marketplace route

A separate part of the announcement concerns deployment. ClickHouse Bring Your Own Cloud, or BYOC, is now available to Azure customers through Microsoft Marketplace.

Under this model, customers run ClickHouse Cloud inside their own Azure tenant rather than on infrastructure operated outside their environment. They retain control over identity and access management policies, network settings, and encryption configuration, while the vendor handles provisioning, upgrades, monitoring, and scaling.

The arrangement is likely to appeal to organisations with data residency, compliance, or sovereignty requirements that want managed database services without giving up control of infrastructure boundaries. It also gives Microsoft customers a procurement route through existing commercial arrangements.

Purchases through Microsoft Marketplace can count towards Microsoft Azure Consumption Commitments. That may matter for large corporate buyers seeking to align software purchases with broader cloud spending targets.

Competitive context

The expanded relationship reflects the growing importance of open table formats and shared storage layers in cloud data platforms. Suppliers increasingly compete on query engines, management models, and integration paths rather than on closed data silos.

For ClickHouse, deeper ties with Microsoft place the company more firmly inside a major enterprise analytics ecosystem. The company has built its business around an open-source analytical database used for real-time analytics, observability, and data-intensive applications.

Microsoft, for its part, has been building out Fabric as a unified platform for analytics, business intelligence, and data engineering. Bringing external engines closer to OneLake strengthens that model by giving users more choice in how they process data while staying within a common governance framework.

Yury Izrailevsky, President, Product & Engineering at ClickHouse, described the rationale for the move in operational terms.

"Expanding our collaboration with Microsoft brings ClickHouse performance and deployment flexibility to where our customers' data increasingly lives," said Yury Izrailevsky, President, Product & Engineering, ClickHouse.

He added: "Our engineering teams have worked closely together so that these capabilities are seamlessly integrated into the Microsoft data platform rather than layered on top of it, and joint customers no longer have to trade off between the platforms they've standardised on."

Microsoft also framed the move around keeping data in OneLake while widening the range of analytical engines available to customers.

"We've long believed that unifying data in OneLake using open, standards-based formats will enable customers to innovate faster," said Dipti Borkar, Vice President, Microsoft IQ and OneLake, Microsoft.

She added: "With the ClickHouse workload available in Fabric, running on OneLake, joint customers have more best-in-class analytical engines to choose from to deliver analytics and AI, without moving or copying data between platforms. This collaboration is a strong example of our partner ecosystem extending what teams can do on a single, unified data estate."