Confluent on Wednesday launched new features that help turn the vendor’s longstanding streaming data platform into a streamhouse for agents and other AI applications.
First introduced in September by a working group of streaming data specialists that includes Aiven, Confluent and Redpanda, a streamhouse combines streaming data platforms with data lakehouse architectures to feed AI agents with continuous, real-time data so AI agents can act on relevant business context at machine speed. Without streaming data, agents lack the up-to-the-minute context to carry out processes such as supply chain management and fraud detection that rely on real-time information.
Among the new tools aimed at turning Confluent’s platform into a streamhouse is Lightning Tables, a feature within Confluent’s Query suite that feeds agents with live, governed tables so the agents have a real time view of an enterprise’s business.
Query is one of Confluent’s five core capabilities that, when combined with Stream to capture and move business events between systems, Connect to deliver event data, Process to prepare data and Govern to classify and secure data, completes the pipeline required to serve agents with current, governed business logic.
In addition to Lightning Tables, which is now generally available on AWS, Confluent also introduced hybrid connectors in early access to help enterprises connection data stored across cloud and on-premises environments.
Stewart Bond, an analyst at IDC, called the new features “moderately” significant additions; Lightning Tables is the most valuable for Confluent customers, which now include IBM users after the longtime tech giant acquired Confluent in December 2025.
“Lightning Tables stand out because they change the read path,” he said. “But speed and ease of use alone will not advance the real-time imperative.”
Instead, ensuring that data can be trusted is what will make stream processing more effective, Bond continued, noting that an IDC survey earlier this year showed that data trust and quality rather than latency are the top technical barrier to real-time data initiatives.
“That gap is sharpest in exactly the regulated, legacy-heavy verticals — financial services, healthcare, government/telco — where IBM’s installed base concentrates. Financial services also show the highest share of event brokers with no formal governance at all, despite the heaviest regulatory pressure,” he said.
Building a streamhouse
To feed agents with real-time context, Lightning Tables joins three pre-existing capabilities within Query: a real-time context engine to give agents access to governed business logic through Model Context Protocol servers, Snapshot Queries to support analysis over both live and historical data to determine patterns over time, and Tableflow to make streaming data available in open table formats for lakehouses.
Sean Falconer, Confluent’s vice president of product for AI products and strategy, noted that for years, customers could capture and process streaming data but had to copy the data and move it to another platform to query it. Lightning Tables is designed to complete Confluent’s query capabilities, eliminating the need to build pipelines and move data.
“The main driver was how customers were using us,” Falconer told TechTarget. “Now, customers get a built-in way to run fast queries, one plane of governance, fewer systems to manage and keep in sync, and they can hand data freshness over to us.”
In addition, broad market trends influenced Confluent’s decision to build Lightning Tables, he continued.
“Companies are increasingly able to have software that runs the business, rather than just helping a human run it,” Falconer said. “For people to trust those systems, the AI has to understand the current truth of the business. … Lightning Tables gives teams a familiar way to query the latest state instantly, through SQL and MCP.”
Lightning Tables stand out because they change the read path. But speed and ease of use alone will not advance the real-time imperative.
Stewart Bond
Analyst, IDC
While the new capabilities in Query complete Confluent’s five-pillar model of a streamhouse, Bond noted that there remain gaps in the vendor’s effort to provide agents, applications and human users with the same, governed view of their organization.
To deliver that consistent context, Confluent needs to add capabilities such as open semantic standards and ontologies that address shared definitions, agent governance including lineage and more widespread hybrid connections.
“[The new features] complete Confluent’s own five-pillar model [of a streamhouse], but the same governed view for every agent is a higher bar,” Bond said.
However, he noted while Confluent’s evolution toward building a streamhouse for AI is ongoing, the vendor’s streaming data capabilities are the most complete across brokering, stream processing, data serving and context and governance with Microsoft perhaps its closest rival.
“Microsoft is closest on event-to-semantics-to-action with Fabric Real-Time Intelligence and Fabric IQ,” Bond said. “What [Confluent] lacks in coverage, it makes up for in scale … via the IBM acquisition.”
Next steps to building a streamhouse
With Lightning Tables now generally available on AWS, Confluent plans to improve the new feature’s performance at scale, according to Falconer.
“We’ll keep pushing on how much data Lightning Tables can handle and on overall query throughput, so customers can run larger workloads and serve more concurrent applications and agents from the same data,” he said.
In addition, as Confluent builds up its streamhouse, it will focus on expanding capabilities to all public clouds.
Bond, meanwhile, suggested that Confluent provide evidence of the efficacy of Lighting Tables and other new features it’s touting as part of a pipeline for fueling agents and other AI tools. In addition, incorporating some of IBM’s existing capabilities could better enable Confluent to connect agents with relevant context.
“It would be nice to see Confluent publish benchmarks and reference architectures for Lightning Tables,” Bond said. “Confluent also needs to leverage the semantic and ontology advancements in the watsonx portfolio to provide real-time context capabilities.”
Eric Avidon is a senior reporter for Informa TechTarget and a journalist with more than three decades of experience. He covers analytics and data management.
