Solo.io Extends Support for AI Agents Running on Kubernetes Clusters
TL;DR — Key Takeaways
-
AI agents don’t have to keep Kubernetes pods busy while doing nothing. Agent Substrate separates stateful sessions from active compute, allowing idle agents to be suspended and restored in under 500 milliseconds.
-
Kubernetes is getting a new class of workload — and new scaling headaches. Unpredictable agent activity can rapidly consume cluster capacity, making efficient resource allocation essential.
-
Solo.io wants to make agentic AI more manageable and secure. Agent Substrate, kagent, agentgateway and agentshield combine orchestration, isolation, identity controls and security capabilities for enterprise deployments.
Solo.io this week at the KubeCon + CloudNativeCon North America 2026 conference announced that it is making enterprise editions of Agent Substrate, a framework that makes it possible to run long-running agent workloads at scale on Kubernetes clusters more efficiently, and kagent, a tool for building and managing AI agents, generally available.
Keith Babo, chief product officer for Solo.io, said IT teams are under significant pressure to find ways to safely deploy AI agents at scale. The Agent Substrate, jointly developed by Google and Solo.io, makes it possible to decouple stateful agent sessions and identities from active compute in a way that creates a snapshot of idle agents that is then stored and, when needed, restored in under 500 milliseconds to reduce costs by consuming infrastructure resources more efficiently.
The Agent Substrate framework has been donated to the Cloud Native Computing Foundation (CNCF), which should result in it being included in distributions of Kubernetes in 2027. However, for IT teams that need to deploy AI agents at scale sooner, Solo.io is making an edition available that it will support.
It’s not clear how quickly IT teams are moving to build and deploy AI agents on Kubernetes clusters, but many of them are looking to deploy what amounts to a new class of workloads on a platform they have already deployed. Otherwise, they will wind up deploying one more platform to support a specific class of workloads that they will then have to separately manage.
In the meantime, Solo.io is also now providing a kagent tool to enable IT teams to build and deploy AI agents that can now invoke the Agent Substrate. At the core of Agent Substrate is a runtime, known as the actor, that provides a sandbox that isolates execution and a gateway that governs how it is reached and what it can access. State, identity, and access controls stay with the actor to make isolation and governance integral to every session.
An agentgateway developed by Solo.io can then be used to enforce identity and policy controls on ingress and egress, while an agentshield tool detects attack paths to protect the environment.
The fundamental challenge IT teams will encounter is that agentic AI workloads are not predictable. As a result, IT teams can suddenly discover that a wave of AI agents are suddenly locking up every pod available on a Kubernetes cluster.
There are, of course, still other issues that need to be resolved to fully optimize Kubernetes for AI agents, including revamping its control plane to enable Kubernetes clusters to handle all the calls to application programming interfaces (APIs) that will be made by AI agents that will soon be deployed in the thousands by enterprise IT teams.
One way or another, AI agents will find their way onto Kubernetes clusters. The only thing that remains to be seen is just how many will be deployed in a way that consumes a limited amount of infrastructure resources as efficiently and cost-effectively as possible.
Frequently Asked Questions
What is Solo.io's Agent Substrate?
Agent Substrate is a framework jointly developed by Solo.io and Google that enables stateful AI agent sessions to be separated from active compute resources. It allows idle agents to be suspended and restored quickly, helping organizations reduce infrastructure consumption while maintaining session state and identity.
How does kagent help organizations deploy AI agents on Kubernetes?
Kagent provides tools for building, deploying and managing AI agents within Kubernetes environments. Its integration with Agent Substrate enables IT teams to run agents using a framework designed for persistent state, workload isolation and more efficient resource utilization.
Why are AI agents challenging traditional Kubernetes infrastructure?
AI agent workloads are unpredictable, stateful and potentially resource-intensive. Large numbers of agents can consume available pods and generate significant API traffic. Organizations need better mechanisms for managing idle sessions, isolating execution and controlling resource consumption without compromising security.
