The Ninth Circuit vacated Amazon's injunction against Perplexity's Comet on August 4. Under the CFAA an AI agent is a tool, not a person, so the user is who accessed Amazon, and the fight moves to terms of service.
AWS added an optimized Step Functions integration for AgentCore harness calls, turning agent reasoning into a workflow Task with approval, retry, trace, and token controls.
AWS released the next generation of Amazon OpenSearch Serverless for agentic AI apps, with scale-to-zero, 20x faster autoscaling, and up to 60% lower costs.
AWS Transform agents now run through Kiro, Claude Code, Cursor, Codex, plugins, and MCP, with AWS claiming 4.5B+ processed lines and 1.6M+ saved hours.
OpenAI GPT-5.5, GPT-5.4, and Codex are generally available on Amazon Bedrock, shifting authentication, billing, and feature boundaries into AWS.
AWS AgentCore Optimization preview productizes agent quality loops with trace-based recommendations, batch evals, and A/B testing.
AWS documented Nova Act limits for browser agents, including 100 sequential steps, 30-minute sessions, prompt injection boundaries, and IAM resources.
AWS SageMaker now supports /openai/v1 endpoints, lowering the migration cost for OpenAI SDK, LangChain, Strands Agents, and AI gateways.
Sysdig says an LLM-driven attacker chained a marimo RCE into AWS secrets, SSH bastions, and an internal PostgreSQL dump.
AWS opened SageMaker AI model customization to coding agents through Skills, turning SFT, DPO, RLVR, evaluation, and deployment into reviewable notebook workflows.
Anthropic detailed the isolation design behind Claude Code and Claude Cowork. The numbers turn agent security from approval prompts into sandbox, VM, and egress policy.
AWS Agent Toolkit and the AWS MCP Server GA show how coding agents can reach cloud accounts through IAM, CloudWatch, and CloudTrail.