Run agents securely and efficiently with the next generation AgentCore runtime
AI agents now run multi-step workloads for hours and act on behalf of users, raising the bar for secure, scalable agent compute. Learn how the next generation of Amazon Bedrock AgentCore runtime delivers hardware-isolated sessions with faster starts, more compute, usage-based billing, and isolation that now extends to your data.
As AI agents grow in complexity and autonomy, securing their operations at scale presents new challenges. Traditional approaches to isolation, such as pre-provisioned environments or on-demand environments with cold-start latency, struggle to balance performance, cost, and efficiency at production levels. Only 15% of organizations have successfully scaled orchestrated, multi-agent systems, and 70% express concerns about trust and governance.
The next generation of agent infrastructure must deliver robust isolation, rapid startup times, elastic scaling, and usage-based economics. Amazon Bedrock AgentCore's new runtime addresses these demands by providing a managed compute layer that creates a secure, hardware-isolated Firecracker micro virtual machine for each agent session.
This update enables sessions to start up to 15 times faster, scale to four times more compute, and run up to 68% more efficiently while maintaining full isolation. Teams can now scale memory and compute within isolated sessions to match workload demands, eliminating cold-start delays and unnecessary resource consumption. The new runtime has already demonstrated performance improvements and cost savings for several organizations, streamlining the deployment and scaling of AI agents in production environments.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
