AWS AI Updates: May 13, 2026

1. AWS Security Agent adds full repository code review

AWS. Security Agent now performs context-aware analysis across an entire codebase, reasoning about architecture, trust boundaries, and data flows to surface systemic vulnerabilities that pattern matchers miss. When issues land, it generates remediation tied to a specific file and line, and AWS claims the agent can build working exploits at speeds traditional tools cannot match. The capability is in preview at no extra cost in all regions where Security Agent runs. Source

2. SageMaker Feature Store supports Python SDK V3 with Lake Formation and Iceberg

AWS. Feature Store now works through SageMaker Python SDK v3.8.0+, exposing Lake Formation column- and row-level access controls on feature groups and Apache Iceberg table property configuration for offline storage. The combination lets teams enforce governance and tune Iceberg query performance from one SDK rather than juggling separate tools. Available in all regions where Feature Store runs. Source

3. SageMaker Unified Studio ships in-product tutorials and release notes

AWS. SMUS now ships guided tutorials covering SQL, data analysis, visual ETL, and ML model training, plus a What’s New panel surfacing recent feature additions directly inside the product. The release also adds automatic light and dark mode matching to system preference. Available in all IAM-based SMUS regions. Source