Redpanda Data

🚀 REDPANDA Q1 HIGHLIGHTS We wrapped up Q1 filled with new product innovations, exciting partnerships, and record growth. 💥 In case you missed it, we rounded up the best highlights from this quarter, including: 📈 New ARR up 70% year-over-year 🎉 Launched the industry's first adaptable #streaming engine 🏆 Named one of the "Top 120 Agentic AI Companies" 🏢 Opened new offices in San Francisco and Austin Our pandas have been busy. But we're just getting started. 🤘 Read more: redpanda.com/press/record-q…

Most enterprises are deploying #AI agents without a plan for what those agents can access, what they can do, or how to prove it when something goes wrong. Sound familiar? 🤔 Redpanda CEO @emaxerrno sat down with @NeilCHughes from Tech Talks Network to talk about why #streamingdata is the foundation #agentic AI actually needs, and what happens when you skip that part. The conversation covers a lot of ground: #governance gaps, agent transcripts, explainability in regulated industries like banking and manufacturing, and why the Redpanda #AgenticDataPlane exists to close the gap between ambitious #agent deployments and production-ready ones. Also covered: Redpanda's work with @nvidia Vera, where benchmarks showed 5.5x lower latency and 73% higher throughput. Speed matters. But so does knowing what your agents did and why. 🔍 Listen to the full episode 🎧 techtalksnetwork.com/podcast/tech-t…

Catch us at Booth 18 at @APIdaysGlobal #NYC today & tomorrow! 🗽 ☀️ Drop by and chat with our expert pandas about real-time #datastreaming, #agenticAI, and what it actually takes to build systems that keep up with your agents. Plus, pick up some seriously good swag. 🎉 Event details: apidays.global/events/new-york #APIDaysNYC #FOSTNYC2026 #apidays

Back to the Future promised us flying cars by 2015. We got AI agents that hallucinate and stall in pilot forever instead. 🙃 2026 is the year that changes, but only for the teams who build the right foundation. Redpanda's latest blog breaks down 5 predictions on #agenticAI and #analytics, from how agents will actually get trusted with enterprise data to why real-time infrastructure is the new competitive moat. Spoiler: governance wins. ✅ Speed wins. ✅ Static dashboards don't. 🚫 See the full breakdown here: redpanda.com/blog/5-predict… Prefer to watch vs. read? Check out this on-demand tech talk: bit.ly/4uMDcpf

Building an AI agent is the easy part. Running one safely in production, with access to sensitive systems, real-time data, and full #governance? That's where most enterprises are still figuring things out. 🤔 Tomorrow, Redpanda CTO @takidau and Solutions Engineer Garrett Raska are hosting a session built specifically for IT leaders: Inside the Redpanda #AgenticDataPlane. You'll come away knowing: 👉 Where Redpanda ADP fits into your existing enterprise stack 👉 How policy enforcement and observability work at the infrastructure level 👉 What the real-world business impact of agent-driven systems looks like 🗓️ Tuesday, May 12 🕗 8am PDT / 11am EDT / 4pm BST Register here: redpanda.com/events/inside-…

An agent told "never delete the production database" will follow that rule until it doesn't. @takidau argues that's not just a model alignment problem. It's an HR problem. Tyler Akidau, CTO of @redpandadata spent 12 years on streaming systems at Google, wrote Streaming 101/102, and authored the O'Reilly Streaming Systems book. His new O'Reilly Radar piece, "Post-Human: We All Built Agents, Nobody Built HR," was the spark for an upcoming @chain_ofthought Podcast. His thesis: enterprise agentic AI has stalled in production because we keep trying to govern agents with identity, auth, and observability tools designed by humans for humans. But agents are a new kind of coworker. They're unpredictable in structurally novel ways, capable at machine speed, and directable to a fault. Prompt-level guardrails & CLAUDE.md guidance hold up until they don't. Some of what we covered: - Why inline guardrails collapse under prompt injection, and what out-of-band enforcement has to look like instead - The four pillars enterprises actually need: identity, authorization, observability, accountability (and why each builds on the last) - Why every agent instance needs a fresh, task-scoped "badge," and how service-account thinking breaks the moment agents start cloning themselves - The intersection problem: an agent acting on behalf of a human with full prod write access should still never get write permissions, no matter who delegated - Why recording every prompt, tool call, and response is the foundation everything else builds on (and why OpenTelemetry is the right place to start) - A teaser on the academic paper Tyler is co-writing with a psychologist on the human value systems agents structurally lack, and what that means for how we design them next Episode drops soon - follow @chain_ofthought to get it first📺 #AIAgents #AgenticAI #AIGovernance #Redpanda

Are Multi-Agent Systems Are Just Distributed Systems in a Trench Coat? Yes. #aiagents #software #coding youtube.com/shorts/Dnr-TqP…

We know all about the need for speed, but what about the need for...high-latency storage? 🤔 Here's the deal: the demand for low-cost storage solutions (like #cloud object storage) is reintroducing high-latency I/O into systems designed around the assumption of low-latency storage. In Redpanda, this shift became a performance bottleneck in Cloud Topics. 😅 Our #engineers are honest folks, so Noah Watkins wrote down exactly what went wrong when high-latency storage entered the picture when developing Cloud Topics, and how the fix made the architecture stronger. 💪 👉 redpanda.com/blog/littles-l…

⭐ NEW WEEKLY DEMO SERIES: Inside the Agentic Data Plane Building #AI agents is getting easier. Running them in production? That's where it gets complicated. Join us on May 12th for a strategic session designed for IT leaders — focused on how the #AgenticDataPlane (ADP) gives your organization the #governance, control, and real-time data access needed to run agentic systems safely at scale. Here's what you'll walk away with: 🏗️ A clear picture of what ADP is and why it matters 🔒 How to run #agentic systems safely in production 📐 Where ADP fits into your existing #enterprise architecture 📈 The real-world business impact of #agent-driven systems Event details: 📅 Tuesday, May 12, 2026 🕗 8:00am PDT | 11:00am EDT | 4:00pm BST 💻 Virtual Sign up here: redpanda.com/events/inside-… More technical? We've got a session for your architects and builders too — May 19th: redpanda.com/events/inside-…

If you're building #AI agents for production, this one's for you. Our CEO @emaxerrno is on stage at the AI Agent Conference today introducing the #AgenticDataPlane: a unified systems layer that gives agents everything they need — streaming + historical data through a single interface, deterministic low-latency eventing, real-time context compute, and governance built into the runtime from day one. Agents are moving from demos to deployments. The infrastructure has to move with them. Learn more about the Agentic Data Plane: redpanda.com/agentic-data-p…

We're LIVE at the AI Agent Conference in #NYC! 🗽 Come find us at Booth 116 in the Americas Hall — the swag is ready and our experts are on hand to show you how to power autonomous, data-driven #agents in #enterprise environments. ✅ And don't miss @emaxerrno taking the stage in the Grand Ballroom today at PM ET — he'll be diving into Redpanda's #AgenticDataPlane and what it means for the future of agentic infrastructure. Stop by & say hi! 👋 agentconference.com

Most agent architectures today are brilliant in demos but fragile in production. 😬 In episode 2 of our "Hello, Agent!" podcast, @DominikTornow, Founder & CEO at @resonatehqio, shares why durable execution is becoming non-negotiable for production AI agents. From the mechanics of interruption-tolerant execution to practical insights on how engineers can build multi-agent systems that actually hold up—this is one ep you'll want to listen to. 👀 Tune in below 🎧 redpanda.com/hello-agent/bu…

Observability Is the Real Agentic AI Moat #observability #compliance #euaiact #agenticai youtube.com/shorts/TmR43o4…

🗞️ REDPANDA NEWSLETTER  The sun is out, lattes are iced, and our paws are as busy as ever. 🌞 From the industry’s first adaptable streaming engine to a peek under the hood of our Cloud Topics architecture, this month's newsletter is packed with cool #engineering updates and a few hot takes on enterprise #AI. 💥👀 Read it here👇 linkedin.com/pulse/redpanda…

"AI doesn't create knowledge gaps once. It widens them continuously." This @OReillyMedia Radar piece by Marc Millstone (Head of AI, Redpanda) is a radically honest take on #AI autonomy in engineering. When engineers stop building things they fully own, they can no longer extend, defend, or debug them under pressure. Most teams focus on velocity and never ask what they’re risking or giving away in the process. 👀 To help #engineering leaders calculate a smarter approach, Marc presents a four-quadrant model for AI autonomy and the real questions you should be asking before filling your repo with AI-authored code that nobody understands. A worthy read indeed👇oreilly.com/radar/dont-aut…