Mon, 16 Nov 2026

Confluent adds Lightning Tables as enterprises seek live data for AI decision-making

Confluent has introduced the Streamhouse, a data architecture designed to give AI agents, applications and analytics a continuously updated, governed view of a business’s current state. The launch adds Query to Confluent’s Data Streaming Platform, completing its five pillars of Stream, Connect, Process, Govern and Query.

The company says traditional data lakes, warehouses and lakehouses were designed mainly for reporting and batch processing. As AI agents take on more operational work, they may act on outdated or conflicting information if systems update at different times—a problem Confluent argues becomes more consequential when decisions trigger downstream workflows.

What a Streamhouse does

A Streamhouse continuously captures, processes, governs and serves current business context across agents, applications and analytics. It is intended to work alongside existing warehouses and lakehouses rather than requiring enterprises to consolidate all data first.

Confluent’s example is an airline managing a cancelled flight. Flight operations, scheduling and reservations systems may each hold a different view of the same disruption; an autonomous rebooking agent could therefore assemble an itinerary that appears valid in individual systems but fails in practice. With a Streamhouse, updates on aircraft location, departure time and seat availability feed a shared, current operational picture before an itinerary is offered.

Greg Taylor

“Across APAC, businesses are scaling digital services exponentially, yet customers and operations increasingly expect experiences to respond in the moment,” said Greg Taylor, senior vice-president and general manager for Asia-Pacific at Confluent.

“A Streamhouse gives teams a continuously updated, governed foundation for that context, so applications and agents can work from the same current picture.” Greg Taylor

Lightning Tables go live

The new Lightning Tables capability is now generally available on AWS. It is a read-only serving layer that stores and materialises Apache Kafka data into tables, keeping them current as changes occur and enabling low-latency SQL and API queries.

In Confluent’s internal testing, the company says it takes approximately 100 milliseconds from new data creation to serving, with query latency of 10–100 milliseconds and throughput of more than 1,000 queries per second. It says the service can reduce the need for separate ETL pipelines, databases and caches, while applying schemas, lineage and access policies from capture through to serving.

The platform also includes Tableflow, which makes streaming data available to lakehouses and warehouses in open table formats, and a Real-Time Context Engine that serves current business context to AI agents through the Model Context Protocol.

Asia-Pacific use cases

For Asian enterprises, the architecture could be relevant where operations depend on rapidly changing data across distributed systems: fraud detection and real-time payments in financial services; inventory visibility and demand forecasting in retail; predictive maintenance and supply-chain optimisation in manufacturing; and network monitoring and 5G service optimisation in telecommunications.

Confluent says its platform supports hybrid and multi-cloud environments, allowing organisations to retain data where it resides rather than creating another centralised silo. It also highlights enterprise governance, security and compliance features, including schema management, encryption, lineage and access controls.

Shaun Clowes

“The next era of enterprise AI will be won by the agents that know what’s happening in the business the moment they act,” said Shaun Clowes, general manager for Confluent and Data at IBM. “With Lightning Tables, our Data Streaming Platform gives each agent and application a live, governed view of the business.”

Accenture is among the early users. Its chief data strategist, Thomas Mahaffey, said moving to a Streamhouse architecture powered by Confluent had reduced the company’s time to insight from months to days and cut reporting times across its operations in more than 120 countries.

Related Stories

MORE STORIES

Subscribe