Solo.io Releases Agent Substrate and kagent
The enterprise releases target infrastructure for running AI agents at scale, with OutSystems and Swissquote as launch customers.
Updated on Nov. 10, 2026 in Software
Solo.io announced general availability of enterprise releases of Agent Substrate and kagent. The infrastructure layer combines serverless scale, sandbox isolation and governed connectivity for agentic runtimes.
Why it matters
Dedicated pods for each agent session tie up resources while agents wait on models, tools or human input. Agent Substrate instead suspends idle sessions and resumes them on demand.
Agent Substrate 1.0 has more than 6,000 GitHub stars and over 100 contributors from more than 40 organizations. Solo released a sizing calculator on the kagent project site, but the source does not report independent benchmarks of its resource savings.
The players
Solo.io
Solo.io is a company that develops cloud-native networking and application infrastructure software, including the kagent and Agent Substrate projects.
OutSystems
OutSystems is a low-code software development platform company and one of the launch customers running this infrastructure in production.
Swissquote
Swissquote is a Swiss online bank and financial services firm and one of the launch customers running this infrastructure in production.
The details
Agent Substrate introduces the actor as the fundamental runtime unit for agents, pairing each with a sandbox that isolates execution and a gateway that governs access. Solo's enterprise distributions use agentgateway to enforce identity and policy controls on ingress and egress, with air-gapped deployment options available.
Timeline
November 10, 2026: Solo.io announced general availability of enterprise Agent Substrate and kagent.
The Tech Race
The release departs from the Kubernetes model of dedicating a pod to each workload, reorienting infrastructure around idle agent sessions instead. It positions Solo.io in the emerging contest to build runtimes for agentic AI systems.
Developers building agent applications can evaluate whether suspending idle sessions reduces the infrastructure they provision. Teams in regulated settings can consider the air-gapped deployment option for keeping agent workloads isolated.
The takeaway
Teams running many long-lived but mostly idle agent sessions may find that the actor model changes how they size infrastructure. Evaluating the sizing calculator against a realistic session profile is a practical first step.
Further reading
Explore more Software coverage for context on enterprise infrastructure tools.
More information
Estimate resource needs with the Agent Substrate sizing calculator on the kagent project site.
