Third-party cyber evaluations involving OpenAI models
openai.com·13h ago
Cloud & Infrastructure
AWS Blog·May 18·5 min read
AWS Blog·May 18·5 min readAWS Weekly Roundup: AWS Transform at 1 year, Claude Platform on AWS, EC2 M3 Ultra Mac instances, and more (May 18, 2026)
TL;DR
Enterprise applications often require costly, manual code modernization across multiple platforms (.NET, mainframe, VMware), creating bottlenecks at scale.
✦ Why It Matters
Engineers can now automate large-scale code modernization using custom rules, reducing manual refactoring effort and accelerating legacy system upgrades.
Key Takeaways
- Evaluate AWS Transform Custom for automating language version upgrades in legacy .NET and mainframe codebases.
- Define custom transformation rules specific to your architecture patterns to extend AWS-managed templates.
- Plan modernization projects using agentic AI to reduce manual code review and refactoring cycles.
- Assess compatibility with VMware, mainframe, and .NET workloads in your enterprise portfolio.
Full Summary
AWS Transform, launched one year prior, addresses the enterprise challenge of modernizing legacy applications across heterogeneous platforms (.NET, mainframe, VMware workloads) without manual rewriting.
The service employs agentic AI—software agents capable of autonomous reasoning and multi-step task execution—to identify, plan, and execute code transformations at enterprise scale. At re:Invent 2025, AWS introduced AWS Transform Custom, extending the platform to support organization-specific transformation rules alongside AWS-managed templates.
This hybrid approach allows teams to encode domain knowledge and proprietary patterns into the transformation engine, reducing manual intervention. The capability addresses a critical gap: enterprises need to upgrade language versions and migrate architectures but lack tooling that scales beyond small codebases.
By automating these transformations, organizations reduce engineering overhead and accelerate time-to-modernization for mission-critical systems.