Skip to content
AI Radar by Wagner Solution
CS
Menu
worth noting Open-source

Google releases Agent Substrate for running AI agents in isolation

only one source so far

Google has released Agent Substrate, an open-source runtime for Kubernetes that runs agents in isolated microVMs. According to the company, it allows ten times as many agents to run on a single server.

Google has released Agent Substrate as an open-source tool, a runtime for Kubernetes designed to run AI agents in isolated microVMs.

According to Google, this runtime allows ten times as many agents to run on a single server. The company also says that pausing and resuming took less than 500 milliseconds in a test with 200 000 agents. These are results reported by the company, rather than independent measurements.

What changed

Why it matters

Teams running AI agents gain an open runtime for isolating their execution. The reported increase in server capacity may affect the amount of infrastructure needed; however, the actual impact on costs depends on the results of each deployment.

Relevant practical impact

What this means

01

For a business

For companies running large numbers of AI agents, Agent Substrate offers a way to change server utilization while maintaining isolation between individual agents. The claimed capacity gains are so far supported by measurements from Google.

Development
What to decide Before deciding whether to deploy it, verify the number of agents per server and the time needed to pause and resume them using your own workload.
More business impacts →
Agent Substrate Google Kubernetes microVM

Check the original

Event sources

only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.

1
The Decoder (daily AI news) independent context · first detected Google's Agent Compute Engine gives every Gemini agent its own hardware-isolated micro virtual machine