What happened
On April 28, 2026, OpenAI and AWS announced a strategic expansion bringing OpenAI's GPT-5.5 models, Codex coding agent, and Amazon Bedrock Managed Agents (powered by OpenAI) to AWS customers. The announcement follows a revised Microsoft partnership that ended OpenAI's Azure exclusivity. More than 4 million people use Codex weekly; enterprises can now deploy it and production-ready AI agents within AWS environments with native IAM, PrivateLink, and CloudTrail integration.
Why it matters
This fundamentally reshapes agentic AI deployment options for enterprises. Organizations can now access frontier models and autonomous coding agents within existing AWS security perimeters, procurement workflows, and cloud commitments. For security teams, it means agentic workloads will increasingly run inside trusted cloud environments rather than via external SaaS endpoints — shifting the control plane but also expanding the attack surface within cloud tenants.
Applicability
Relevant for any organization with AWS commitments evaluating agentic AI for software development, document workflows, or multi-step business processes. Security and cloud architecture teams should assess how Bedrock Managed Agents interact with existing IAM policies, service principals, and data governance controls.