AWS ML BlogWednesday · June 24, 2026FREEShared infrastructure, isolated tenants: Pool model multi-tenancy with Amazon Bedrock AgentCore
AWS published a blog post titled 'Shared infrastructure, isolated tenants: Pool model multi-tenancy with Amazon Bedrock AgentCore' on the AWS Machine Learning Blog. The post describes a multi-tenancy architecture for AI agents using Amazon Bedrock AgentCore. In the pool model, multiple tenants share the same underlying infrastructure, such as compute and model inference resources, while tenant isolation is achieved through logical separation mechanisms. The blog provides architectural guidance, including how to configure AgentCore to route requests to the correct tenant, manage tenant-specific data, and enforce access controls. The consequence of this approach is cost efficiency: by pooling resources, organizations can reduce operational overhead compared to dedicated infrastructure per tenant. The post is aimed at developers building multi-tenant AI agent applications and includes code examples and best practices for implementing the pattern.
Pool model multi-tenancy reduces costs by sharing infrastructure while keeping tenant data isolated.