Teradata and Microsoft Erase a Costly Step in Enterprise AI

SAN DIEGO, CA – September 02, 2026 – For decades, a fundamental tax on innovation has been the costly and complex process of moving data before it can be analyzed. Today, Teradata and Microsoft announced a landmark integration aimed at repealing that tax for enterprise AI workloads. The new partnership allows Teradata's Autonomous Knowledge Platform to run its high-performance AI and analytics directly on data stored within Microsoft OneLake, effectively eliminating the need for data duplication, migration, or brittle ETL (extract, transform, load) pipelines.

This move signals a significant shift in the data analytics landscape, where the gravity of data has long dictated the location of compute. By bringing its powerful analytical engine to the data's native residence, Teradata is making a bold statement about the future of AI architecture: intelligence should go to the data, not the other way around.

The End of the Data Shuffle

For any data engineer or AI team, the workflow has been painfully familiar. A promising dataset lives in one system, but the preferred analytics or AI tool lives in another. This predicament kicks off a resource-intensive cycle of building and maintaining pipelines to copy, synchronize, and govern duplicate datasets. The process introduces latency, creates potential security and governance gaps, and inflates storage costs—all before a single valuable insight is generated.

This integration between Teradata and Microsoft Fabric’s OneLake directly dismantles that paradigm. Built on the open Apache Iceberg table format, the collaboration enables Teradata to query OneLake tables in place. There is no data movement. Data engineers can reclaim time once spent on pipeline maintenance, while AI teams can accelerate feature engineering without waiting for data replication. Critically, governance policies and data lineage established in OneLake are natively respected, as cross-platform authentication is handled by Microsoft Entra ID. This ensures that security travels with the data, rather than being precariously rebuilt for each copy.

“Moving data to analyze it is a constraint that adds cost and complexity without adding value,” said Sumeet Arora, Chief Product Officer at Teradata. “Bringing Teradata enterprise AI directly to Microsoft OneLake means customers can run the workloads that matter, on the data where it already lives, inside the governance model they've already built. This is a concrete step in making the Autonomous Knowledge Platform work wherever enterprise data exists.”

A New Blueprint for Enterprise Intelligence

The implications of this partnership extend far beyond operational efficiency. It provides a strategic blueprint for businesses struggling to operationalize AI at scale. Microsoft Fabric was introduced as a unified, all-in-one analytics platform, with OneLake positioned as its central, tenant-wide data lake—a single source of truth for the AI era. By opening OneLake to a best-of-breed enterprise AI engine like Teradata's, Microsoft is making a powerful statement about its commitment to an open ecosystem over a walled garden.

This interoperability offers enterprises the flexibility to leverage their existing investments and expertise without being forced into a single-vendor architecture. For organizations already invested in Microsoft Fabric, Teradata becomes a powerful, accessible option for mission-critical AI that demands robust performance, workload management, and enterprise-grade service level agreements. For existing Teradata customers, it unlocks the vast ecosystem of Microsoft Fabric and Azure without the daunting prospect of a massive data migration project.

“Microsoft OneLake is built to be the unified data foundation for the AI era, connecting organizations with the platforms, tools, and data ecosystems they already trust,” noted Dipti Borkar, Vice President of Microsoft OneLake and ISV Ecosystem at Microsoft. “Teradata's integration through Apache Iceberg exemplifies the openness and interoperability at the heart of Microsoft Fabric, enabling customers to bring their data together and accelerate their AI transformation without the complexity or rearchitecting around a single vendor.”

The Unsung Hero: Open Standards Take Center Stage

The technical linchpin of this collaboration is Apache Iceberg. As an open-source table format, Iceberg brings the reliability and management features of a traditional data warehouse—such as ACID transactions, schema evolution, and time-travel queries—to the vast and flexible world of data lakes. Its adoption by major players is a key driver in the industry-wide shift toward the “data lakehouse” architecture and a direct challenge to the vendor lock-in of proprietary formats.

What makes the OneLake integration particularly elegant is its use of table format virtualization. OneLake can dynamically generate the necessary Iceberg metadata for tables natively stored in the Delta Lake format, and vice-versa. This on-the-fly conversion means organizations don't have to choose or maintain separate copies for different tools. An Iceberg-compatible engine like Teradata's can read a table, and a Delta-compatible engine can read the same table moments later, both accessing the same underlying data files. This commitment to open standards is what truly enables a flexible, future-proof data strategy, allowing businesses to adopt the best tools for the job without creating new data silos.

From Strategy to Reality

This announcement is not a distant roadmap item; read access to OneLake tables via Teradata is available now. The integration is a continuation of a multi-year strategic alliance between the two technology giants, which has included the integration of Teradata Vantage with Microsoft Azure and the launch of Teradata VantageCloud Lake on the Azure marketplace. The companies will be demonstrating their new joint capability at the European Microsoft Fabric + SQL Community Conference later this month.

While the technology promises a frictionless future, success still hinges on strategy. The integration removes a significant technical barrier, but it doesn't remove the need for disciplined data management. Organizations that treat OneLake as a dumping ground without a clear governance framework will simply create a more accessible data swamp. The true winners will be those who leverage this newfound interoperability to build a cohesive, governed, and intelligent data foundation, finally allowing them to focus on generating value from their data instead of just moving it around.