Something has shifted in the way we think about AI agents. Today, the conversation is about production. How do you give an agent real-time context? How do you orchestrate tool calls reliably at scale?
We are thrilled to announce the official agenda for the Data Streaming Summit 2026, where the data and agent communities unite!
This year's summit is a 2-day event designed for builders and operators:
- Day 1: Dive into our interactive Hackathon, where you can get hands-on and build real-time agentic systems.
- Day 2: The main Conference, featuring a jam-packed schedule of technical deep dives, real-world case studies, and forward-looking discussions.
The full session agenda is now live---explore the complete schedule and speaker details on our website.
🚀 The Data+Agent Hackathon
We'll open the summit on Wednesday, October 7, with a full day dedicated to building. The Data+Agent Hackathon gives developers an opportunity to experiment with what becomes possible when real-time data meets AI agents.
Participants can explore:
- Real-time context
- Event-driven agent workflows
- Streaming infrastructure
- Memory and orchestration
- Agent tooling and more
Join us for an immersive experience with access to a dedicated StreamNative environment, including Kafka clusters and agent workspaces. Our engineers will provide expert guidance to help you build, run, and showcase your creations. Space is limited, so secure your spot for a day of real-time building.
🛤️ Track Summaries
Keynotes: Join us for the 2026 keynote, where we'll trace the path from real-time data to trusted context to agent action. Expect major announcements across open-source streaming, plus governed lakehouse advances with next-generation open table formats and deeper catalog integration alongside Databricks, and new capabilities with RisingWave that turn streaming topics into queryable, trusted context. OpenAI takes the stage for a keynote segment on building infrastructure in the age of agents, and we'll introduce a new engine and complete lifecycle for building, running, observing, and governing agents in production. Come see what it takes to build infrastructure where systems can understand, reason, and act.
Streaming Infrastructure: This is the foundational layer. Dive deep into the architecture, internals, and operations of next-generation engines that run the modern data stack. Sessions in this track will explore how to resolve data skew, vectorizing Apache Flink, achieving key order without compromise in Apache Pulsar 5.0, and mastering real-time state over messy, high-volume event streams.
Agent Runtime & Governance: Explore the application layer that makes AI agents safe and reliable in production. This track covers the runtimes, orchestration frameworks, and governance patterns necessary for autonomous systems. Discover insights into running thousands of browsers as a streaming data plane, building machine-speed defenses, optimizing compute efficiency, and governing agents safely when they query sensitive customer data.
Lakehouse & Real-Time Pipelines: Data in motion meets data at rest. Discover how to build durable, queryable context pipelines at scale. Expect technical deep dives into streaming massive workloads into a lakehouse, beating the concurrency tax on Iceberg tables, declarative data movement, and enabling hybrid ML inference for both streaming and batch workflows.
Context & Memory for Agents: Your agent is only as good as the pipeline behind it. This bridge track focuses on giving your agents the memory they need to act autonomously. Sessions will unpack event-driven memory for agent swarms, the mechanics of forking shared logs so agents can safely operate on live data, and how to codify your analytics knowledge into composable skills.
🔥 Session Highlights
We have an incredible lineup of talks this year, featuring sessions from leading AI pioneers and infrastructure innovators across the industry, including OpenAI, NVIDIA, Together, Harvey, Meta, Airbnb, Pinterest, Adobe and more.
As an independent highlight reel, here is a sneak peek at what some of the biggest tech companies in the world are bringing to the stage:
- OpenAI is bringing heavy-hitting insights across multiple tracks, discussing everything from building their in-house data agent to exploring the AI-native Flink experience and testing nondeterministic changes in your agent harness.
- Meta will explore how the agent runtime fundamentally functions as a stream processor and share practical strategies for applied AI compute efficiency.
- NVIDIA will dive into the cutting-edge streaming infrastructure required for vision-language reasoning across real-time and batch workflows.
- Airbnb will unpack how they are achieving declarative data movement at scale by unifying streaming and offline-to-online pipelines.
- Pinterest is set to reveal their next-generation DB ingestion process---from CDC and Iceberg to agentic onboarding---and will even share how they get away with "lying" to Kafka using MirrorMaker.
- Snowflake will expose the hidden complexities of real-time AI pipelines, shedding light on the pitfalls no one talks about until they break.
- SK Telecom will showcase how they handle carrier-scale, real-time decisioning by blending streaming data with AI agents.
🎟️ Secure Your Spot!
You don't want to miss the defining event for the infrastructure layer of the AI era.
One Important Registration Note:
The Hackathon (Day 1) is absolutely FREE, but space is limited! You must reserve a free Hackathon ticket through the registration page to secure your spot and participate. Plus, all participants who submit their work during the Hackathon will receive free admission to the Conference on Day 2.
👉 Register for the Data Streaming Summit 2026 Here!
Whether you're a longtime member of the Pulsar and Kafka communities, a platform engineer integrating streaming with AI, or an agent builder trying to take your system from demo to production---this is your conference. We'll see you in San Francisco!
