Redpanda Agentic Data Plane (ADP) now in limited availability

Securely integrate AI agents and MCP servers, all in real time

February 18, 2026
Last modified on
TL;DR Takeaways:
No items found.
Learn more at Redpanda University

2026: the year AI gets to work

Over the past year, there has been a revolution in the enterprise adoption of agentic systems. In early 2025, executives still had significant doubts, capped off by the MIT NANDA paper, which cited that only 5% of enterprise AI projects were in production. This became widely misquoted as “95% of generative AI pilots are failing,” or “fail to deliver ROI.” Yet by the dawn of 2026 the evidence was quite the opposite: Google reported in September, “74% of executives report achieving ROI their first year.” In fact, 39% of executives had already deployed more than 10 agents across their enterprises.

Even highly-skeptical developers could see marked gains in their own productivity. What was dismissed as a distraction in 2020 is now ready-for-use in 2026. As CoreWeave’s CEO Michael Intrator stated, 2026 is “The Year AI Gets to Work.

Now, everyone from engineering to marketing wants AI-based tools in production. They want them to have direct read-write access to enterprise data, which further elevates the stakes. It requires running at scale. Reliably. Securely.

This is why the enterprise agentic AI market is struggling while consumer AI flourishes. Building an agent is remarkably easy. Running one safely with access to sensitive corporate systems and data remains hard. Redpanda is focused on bringing control to the connectivity layer. Rather than configuring access policies individually at each data source, enterprises need a central point through which all agent interactions flow: an agentic data plane.

That is why we built Redpanda Agentic Data Plane (ADP): so you can achieve your scalability goals and time-to-market with agentic systems, while meeting your ROI and staying within your token budgets.

"While individual developers will love using our platform, Redpanda ADP is designed to impress your boss, and your boss’s boss."

Watch our Founder and CEO, Alexander Gallego, talk to NYSE about what we’re delivering with Redpanda ADP.

What is Redpanda ADP?

We announced Redpanda ADP back in October of last year. You can read more about our vision in the related article. It’s a unified, governed access layer that connects all your data systems and mediates every agentic interaction.

Redpanda ADP safely exposes enterprise data to AI agents. It combines:

  • A low-latency streaming layer for events and human-in-the-loop (HITL) workflows
  • A fully-managed, global policy, governance, and observability layer that enforces access, records intent, and enables replayable audits
  • 300+ high-quality connectors to bring context to the models
  • A distributed, Apache Iceberg™-native query engine for real-time context

How can I get my hands on Redpanda ADP?

Redpanda ADP is now available for approved Redpanda Design Partners. Design Partners can work with the Redpanda team to develop solutions built on the Redpanda Bring Your Own Cloud (BYOC) offering on AWS. Other cloud providers (GCP, Azure) will be available soon, as will Self-Managed and Serverless deployments.

To apply to be a Redpanda Design Partner, contact your Redpanda representative, or, if you aren’t yet a customer, reach out to our team.

What’s included in Redpanda ADP?

Redpanda ADP builds on the core of Redpanda’s Kafka®-compatible data streaming service, with a number of unique elements:

  • AI Gateway
  • Remote MCP
  • AI Agents
  • Authentication & authorization
  • Observability
Learn more about what’s included in Redpanda ADP in this on-demand webinar: Putting AI to Work: Governing Enterprise Agents in Production

AI Gateway

Redpanda AI Gateway is a unified access layer that sits between your applications and the LLM or MCP services they use. It provides centralized routing, policy enforcement, cost management, and observability for all your AI traffic. Users can also specify token budgets and spending limits and minimize token consumption through deferred tool loading. Using AI Gateway, users gain control, auditing, and flexibility for complex agentic workflows.

Redpanda AI Gateway also provides a central point to aggregate and manage MCP servers, both local and remote, through an admin-controlled MCP Server Registry. Configuration can be easily set up via YAML, enabling developers to create and deploy agents via text prompts. This enables centralized governance and security for MCP servers. Its JavaScript-based orchestrator reduces round trips needed in multi-step workflows.

AI Agents

Redpanda ADP includes AI agents that can be equipped with tools and given instructions declaratively. They use the industry-standard A2A protocol to accept input and integrate with authorization and observability services. We also provide built-in triggers in Redpanda Connect to hook into pipelines and workflows, with over 300 connectors already available.

Authentication and authorization

Redpanda ADP is built around unified authentication based on OpenID Connect (OIDC). Every resource requires authentication, which can be a user or a service account. Overall, Redpanda ADP has fine-grained access control policies for authorization that can be managed at the organization level or the data plane level.

Observability

Redpanda ADP connects to LLMs via the OpenTelemetry Protocol (OTLP), enabling easy collection and analysis of agent transcripts, including comprehensive metrics, logs, and traces of every agent interaction. Not to be confused with audio or video transcripts, agent transcripts provide a holistic record of an agent’s behavior, allowing for both post-facto analysis as well as continuous online performance evaluation and adjustment.

AI agent transcripts are easy to view and analyze within Redpanda Console

Transcript data can be viewed in our Console to debug and understand what actions your agent took. In addition, you can add your own traces using the new OTLP endpoint in Redpanda Connect. Because Kafka-compatible data streaming is included, you can also easily send all of the traces to an observability platform or to Apache Iceberg tables with Redpanda Connect as well.

Let’s talk!

This is just the first set of deliverables for Redpanda Agentic Data Plane. For more details, check out our Redpanda ADP documentation.

We’d love to hear what you’re building with AI. Contact us to let us know how Redpanda can help you create the next generation of data-intensive, AI-fueled applications.

Also, we’re working on tons more for both Redpanda ADP and our streaming service, so subscribe to our monthly newsletter to get notified as new capabilities are delivered!

Handy links

No items found.
Subscribe to our monthly newsletter

Related articles

View all posts
David Yu
,
,
&
Feb 5, 2026

Redpanda Terraform provider: manage Connect pipelines, PrivateLink, and more

We're going "beyond the cluster" to standardize streaming for modern enterprises

Read more
Text Link
Towfiqa Yasmeen
,
Mike Broberg
,
&
Feb 3, 2026

Redpanda Serverless now Generally Available

Zero-ops simplicity meets enterprise-grade security to unlock production-ready data streaming for builders

Read more
Text Link
Towfiqa Yasmeen
,
Mike Broberg
,
&
Nov 25, 2025

What's new in Redpanda Cloud: A simpler Serverless now on GCP, & more

Putting governed agents to work, plus a refreshed onboarding UX

Read more
Text Link
PANDA MAIL

Stay in the loop

Subscribe to our VIP (very important panda) mailing list to pounce on the latest blogs, surprise announcements, and community events!
Opt out anytime.