AWS Upgrades AgentCore Runtime for Faster Starts and More Compute for AI Agents
Source Amazon AWS Machine Learning blog

At a glance
AWS has upgraded AgentCore runtime, a managed environment for running AI agents. Sessions now start faster, can use more computing resources, and run more efficiently, the company said.
Key points
Each session runs in its own hardware-isolated virtual machine (MicroVM).
Session startup is up to 15 times faster, and available computing power is up to 4 times greater.
Each session can use up to 8 vCPUs and 32 GB of memory, and supports container images up to 10 GB.
Sessions start in under 2 seconds without pre-warmed resources, which reduces cold-start delays.
Why it matters
For agents to be used in real work, security, speed, and cost must be balanced, and this upgrade aims to strike that balance.
Original article
Run agents securely and efficiently with the next generation AgentCore runtime
Amazon AWS Machine Learning blog · Nov 13, 2026, 2:59 AM GMT+9 · aws.amazon.com
Read on Amazon AWS Machine Learning blogAI read the original article and summarized it in plain language. It may contain mistakes, so please check the original for accuracy. The article's copyright belongs to Amazon AWS Machine Learning blog.