StarTree Targets Next-Generation Data Lakehouses With Apache Iceberg Support

By adding support for the Apache Iceberg data table format to its real-time data analytics platform, the company looks to simplify the use of lakehouse-stored data for customer-facing applications and agentic AI systems.

StarTree has added support for Apache Iceberg to its StarTree Cloud, bringing that platform’s real-time analytics capabilities to data lakehouses for a range of customer-facing and AI applications.

With the native Iceberg support StarTree Cloud can directly access and analyze data in lakehouses that support the Apache Iceberg data table format, eliminating the need to move, replicate or transform the data for use by other query engines with proprietary data formats.

That allows businesses and organizations to more easily leverage the huge volumes of data stored within data lakes for applications – both internal and external-facing – that require sub-second query speeds, high levels of concurrency and frequently updated information.

“A big part of this for us is responding to the demands of the market and aligning with all that momentum around open [data] table formats,” said Chad Meley, StarTree senior vice president of marketing, in an interview with CRN. But more broadly, he said the move is a way of “opening the data lakehouse to the outside world.”

StarTree, headquartered in Mountain View, Calif., bases its real-time data analytics software on the open-source Apache Pinot online analytical (OLAP) datastore and analytics technology that provides high-speed analytics on large datasets with low-latency and high-concurrency (the ability to simultaneously handle a large number of queries – as many as tens of thousands of queries per second).

Because of those capabilities, StarTree is particularly focused on providing its platform as a back-end analytical processing engine for customer-facing data products, operational intelligence applications and – more recently – AI agents. Use cases include customer-facing analytics applications such as self-service financial services and merchant online applications, gaming performance trackers, and self-service portals for logistics, transportation and delivery companies.

Until now businesses and organizations largely have been hesitant to tap into data lakehouses to support such external applications, Meley said. That’s because of the costs and complexities associated with moving and preparing data to work with the data analytics systems that support those applications.

Pinot has its own proprietary data file format. StarTree has decoupled that from its query engine and its high-performance indexing, intelligent materialized views, localized data caching and data serving capabilities. That allows StarTree to add Iceberg support and make it possible for StarTree to work directly with lakehouse data in the Iceberg format, said Chinmay Soman, head of product at StarTree, in the CRN interview.

“The beauty of that is you get all the query optimization of Pinot in this model,” Soman said of the added Iceberg support. “The goal is not to be yet another [data] warehouse query engine. We still focus on the serving side of these use cases. The things that Pinot was built for can be extended for Iceberg and you can get a very fast serving layer on top of Iceberg using these Pinot attributes.”

“We do see customers making Iceberg as a [single] source of truth for all data,” Soman added. “There is a trend today [of] people moving data from Iceberg [-based systems] into all kinds of things to do processing.”

Soman said StarTree is now working with early adopter customers to use the new Iceberg-based capabilities with general availability rolled out through the rest of 2025.

StarTree’s work around Apache Iceberg closely follows the company’s announcement in April that its platform supports Model Context Protocol, allowing customers to integrate StarTree’s data analytics capabilities with AI applications and agents that support the MCP standard.