Company eliminates 90% of streaming costs while bridging data silos and enabling production-ready AI agents

StreamNative, the cloud-native data streaming company, made a series of announcements at the Data Streaming Summit in San Francisco that solve the three-part infrastructure crisis blocking enterprise Artificial Intelligence (AI) adoption at scale. The company unveiled Ursa lakehouse-native streaming across all deployment models and introduced Orca – an Event-Driven Agent Engine for production AI agents, delivering the industry’s first unified platform for real-time AI at enterprise scale.

The Real-Time AI Infrastructure Crisis

The explosion of real-time AI is creating unprecedented demands on data streaming infrastructure that traditional architectures simply cannot meet. Organizations deploying AI agents, real-time analytics and autonomous systems face three critical bottlenecks that are fundamentally incompatible with legacy streaming designs.

• First, streaming costs scale exponentially with AI workloads as legacy systems with leader-based replication and separate ETL pipelines can’t handle AI’s massive data throughput demands economically, often creating cost structures that become prohibitive at scale.
• Second, streaming and analytics systems remain disconnected, creating data silos that force AI agents to work with stale data snapshots instead of live context. This prevents AI from accessing the real-time information streams it needs to make informed decisions.
• Third, while agent frameworks proliferate in research labs and demos, production environments lack the event-driven runtime, persistent memory and observability that autonomous systems require to operate reliably at enterprise scale.

StreamNative’s announcements today deliver the industry’s first complete solution to this crisis with two products that redesign how enterprises approach real-time AI.

Marketing Technology News: MarTech Interview with Lee McCance, Chief Product Officer @ Adverity

Ursa Everywhere – Lakehouse-Native Streaming Across All Deployments
StreamNative is announcing that Ursa, its lakehouse-native streaming engine, is now available across all deployment models, including Serverless, Dedicated and Bring-Your-Own-Cloud (BYOC) environments. Ursa’s architecture eliminates separate Extract, Transform, and Load (ETL) pipelines by writing streaming data directly to open table formats like Apache Iceberg and Delta Lake in cloud object storage. This approach delivers up to 90-95% cost reduction compared to traditional streaming systems while maintaining full Kafka API compatibility.

The lakehouse-native design bridges the streaming-analytics divide by ensuring events become immediately queryable tables in data lakehouses without additional engineering overhead. Organizations can now stream once and immediately access that data through SQL, analytics tools and AI systems without building separate batch processing pipelines.

Marketing Technology News: What is a Full Stack Marketer; What MarTech Matters Most to Full Stack Marketers?

Orca Agent Engine Enters Private Preview
StreamNative also announced the Private Preview of Orca Agent Engine, the industry’s first event-driven runtime designed specifically for production AI agents. Orca transforms AI agents from stateless, request-response functions into always-on, event-driven services with persistent streaming memory and built-in observability. The platform is framework-agnostic, allowing organizations to bring existing agents built with OpenAI Agents SDK, Google Agent Development Kit (ADK) or custom Python code without modification. The support for LangGraph and other agent frameworks is on the company’s roadmap and forthcoming.

Orca leverages Pulsar or Kafka as a shared event bus to enable agent collaboration and tool discovery via the Model Context Protocol (MCP), creating an “agent mesh” where autonomous systems can coordinate actions and share context in real time. Enterprise-grade governance provides full audit trails of agent decisions and actions, addressing the transparency and accountability requirements for production AI deployments.

“Unify Data. Accelerate Insights. Autonomous Actions. That’s the bar for the next generation of data systems,” said Sijie Guo, CEO and co-founder of StreamNative. “Today’s announcements show we’re not tackling streaming, analytics, and AI in isolation—we’re delivering a single, unified foundation that removes the boundaries between them.”