Last week Snowflake launched Datastream — Kafka-compatible streaming built directly into the platform. To me that's a strong signal: the industry now agrees streaming and the lakehouse belong together, especially for AI agents that need to act on fresh data. That convergence is exactly the shift we've been building toward. We call it - #Lakestream. AI agents can't act on yesterday's data. They need the latest events — seconds old — sitting next to the full history that gives them meaning, governed enough to trust. That's real-time context, and it's what the classic two-system setup — streaming on one side, the lakehouse on the other, pipelines in between — was never designed to deliver. A Lakestream unifies the two, end to end on open standards. You produce and consume the live stream through the Kafka API, and you query that same data as an open Apache Iceberg table from any engine. One stream, written once — a low-latency event stream and an open table at the same time. The log that backs the stream is the table, one copy serving the present and the past at once. The lakehouse merged the lake and the warehouse. The #Lakestream does it one step earlier in the data's life: it merges the stream and the lakehouse — and keeps both sides open. We wrote up the full thesis 👇 streamnative.io/blog/what-is…

Say hello to UniConn (Universal Connectivity)! After adding support for the Kafka protocol with Ursa engine, we've gone a step further by bringing in support for Kafka Connect. With UniConn, you can now run Kafka Connect connectors natively on StreamNative Cloud, using the same runtime and scheduler as Pulsar IO connectors. It’s all about making data streaming easier, no matter where your data lives! Check it out: streamnative.io/blog/revolut…