BREAKING: Announcements at Data Streaming Summit'25 by StreamNative

Breaking Announcements from Data Streaming Summit by StreamNative

I’m here live, and the energy is amazing. This quarter’s announcements land with a clear message: real-time data should no longer be expensive, siloed, or stuck at the insight stage. It should flow all the way through to governed analytics and AI-driven actions, without detours.

Let’s unpack what was announced and why it matters.

1. Ursa Everywhere: Lakehouse Streaming at Scale

Ursa has been one of the most talked-about innovations in streaming: it rewrites the economics of data by eliminating redundant replication and writing streams directly into open table formats like Apache Iceberg or Delta.

Now, Ursa’s storage layer is available as a lakehouse tier across every Classic Engine cluster—Serverless, Dedicated, or BYOC.

What this means:

• Producers and consumers still get the familiar Pulsar/Kafka experience.

• Data automatically lands in your cloud object storage in Iceberg/Delta, ready for analytics.

• You skip the cost of duplicate pipelines or connectors.

• Future upgrades to Ursa brokers become seamless—your data is already in the right place.

💡 Why it matters: This is a pragmatic path for enterprises under pressure to reduce total cost of ownership. You get the durability and openness of the lakehouse without a disruptive migration.

2. Governed Analytics with Unity Catalog

A key friction in streaming has been governance—real-time data often bypasses the policies, catalogs, and lineage systems that enterprises rely on.

With this release, streams landing in Iceberg can now be managed through Databricks Unity Catalog, turning them into governed, query-ready tables immediately.

What this means:

• Analysts and data scientists query fresh, governed streaming tables in the same notebooks and dashboards they already use.

• Access control and lineage apply the moment data arrives.

• Interoperability stays intact—you can still use Spark, Trino, Flink, or Snowflake.

💡 Why it matters: Governance no longer lags behind when data goes “real-time.” This removes a huge blocker for enterprises that have struggled to balance speed with compliance.

3. RBAC General Availability

As more workloads consolidate on a shared streaming backbone, security becomes non-negotiable.

Role-Based Access Control (RBAC) is now GA in StreamNative Cloud. That means:

• Centralized definitions of who can publish, subscribe, evolve schemas, or manage namespaces.

• Roles applied across clusters instead of scattered ACLs.

• Auditable, predictable rollouts that reflect real organizational structures.

💡 Why it matters: Enterprises can finally enforce least-privilege access across multi-tenant environments without duct tape. This is a big step toward secure consolidation.

4. Orca Private Preview: Event-Driven Agents Go Production

Perhaps the most forward-looking announcement: Orca, an event-driven Agent Engine, is now in private preview.

Unlike agent frameworks that stay in notebooks or prototypes, Orca provides a production-grade runtime:

• Agents subscribe to topics, process events, and emit outputs in real time.

• Persistent state and memory allow knowledge to accumulate across sessions.

• Observability and auditability are built-in—every decision and action is logged.

• Existing agents (Python, OpenAI, Google ADK) can run without heavy rewrites.

💡 Why it matters: This closes the loop between data and action. Enterprises can deploy autonomous agents that don’t just analyze streams—they act on them. Think anomaly detection triggering remediation, customer signals driving personalization, or operational alerts leading to automated triage.

The Bigger Picture

These aren’t four separate announcements—they form one coherent story:

• Stream once (Ursa).

• Store in open formats (Iceberg/Delta).

• Govern centrally (Unity Catalog + RBAC).

• Act in real time (Orca).

It’s the path from data → insights → action without crossing tool silos or spinning up parallel pipelines.

For enterprises, this matters because:

• TCO comes down (no duplicate infrastructure).

• Governance scales (no compliance blind spots in real-time).

• AI becomes actionable (agents move from prototypes to production).

My Take

Data streaming is moving from plumbing to platform. The industry doesn’t need more disconnected systems—it needs a backbone that is practical, open, and intelligent.

This launch shows how StreamNative is trying to be that backbone. If you’re struggling with streaming costs, governance gaps, or the jump from insight to action, these updates hit those pain points directly.

I’ll be sharing on-site conversations and deeper analysis over the next few days, but for now, one thing is clear:
Real-time AI just got a lot more real.

I’ll be going deeper into these topics here at the summit and sharing conversations with leaders driving these innovations. Stay tuned.

Great announcements by StreamNative Team!!!!

🔍 Stay Ahead in AI & Data! Join 137K+ Data & AI professionals who stay updated with the latest trends, insights, and innovations.

📢 Want to sponsor or support this newsletter? Reach out and let's collaborate! 🚀

Best,

Ravit Jain

Founder & Host of The Ravit Show