AWS Transform One-Year Milestone: AI-Powered Application Modernization at Scale

Amazon Web Services has reached a one-year milestone with AWS Transform, an agentic AI service designed to modernize enterprise applications. Launched in May 2025, the service focuses on modernizing .NET, Mainframe, and VMware workloads at scale.

Over the first 12 months of operation, the service has been used by thousands of customers to migrate hundreds of thousands of servers. According to AWS, the tool has processed more than 4.5 billion lines of code and saved over 1.6 million hours of manual effort.

The platform expanded its capabilities at re:Invent 2025 with the introduction of AWS Transform custom. This version allows organizations to use both AWS-managed transformations and custom transformations to modernize code. These tools enable users to analyze code bases, optimize performance, migrate frameworks, and upgrade language versions based on specific organizational requirements.

AWS also introduced automated testing functionality for mainframe workloads, alongside Reimagine capabilities and full-stack Windows modernization tools. To increase accessibility, AWS Transform agents are now integrated into Codex, Cursor, Claude, and Kiro. This includes the Kiro power toolkit, which is used for the creation of customized transformation agents.

AI Platform and Prompt Optimization

During the week of May 11, 2026, AWS announced the general availability of the Claude Platform on AWS. This integration provides direct access to the native Claude Platform experience from Anthropic, including its console, APIs, and early-access beta features.

Users can access these features through existing AWS accounts, eliminating the need to manage separate billing or tracking. AWS noted that the Claude Platform on AWS is operated by Anthropic, and customer data is processed outside of the AWS security boundary.

AWS released Advanced Prompt Optimization for Amazon Bedrock. This tool allows developers to optimize prompts for any model hosted on Bedrock. The feature supports simultaneous comparison of original prompts against optimized versions across up to five different models, facilitating performance improvements or migrations to new models.

Infrastructure and Hardware Updates

AWS has expanded its compute offerings with the launch of Amazon EC2 M3 Ultra Mac instances. These instances utilize Apple M3 Ultra Mac Studio hardware, which includes a 32-core Neural Engine, 60-core GPU, 28-core CPU, and 256GB of unified memory.

When compared to EC2 M4 Max Mac instances, the M3 Ultra instances provide 2x the Neural Engine cores and unified memory, 1.75x the CPU cores, and 1.5x the GPU cores. These specifications are intended to help Apple developers accelerate on-device ML workflows and run a higher number of Xcode simulators in parallel.

For data warehousing, AWS introduced Amazon Redshift RG instances powered by AWS Graviton. These instances are designed to run data lake and data warehouse workloads up to 2.4x faster than the previous generation RA3 instances. AWS claims these RG instances offer a 30% lower price per vCPU.

The RG instances feature a custom-built vectorized data lake query engine capable of processing Parquet and Apache Iceberg data directly on cluster nodes.

Security and Multicloud Connectivity

A new preview capability for the AWS Security Agent now enables full repository code scanning. This tool performs context-aware security analysis across an entire codebase. When the scanner identifies vulnerabilities, it generates code remediation in the form of specific fixes tied to the exact line and file.

This security feature is available at no additional cost to existing AWS Security Agent customers during the preview period.

AWS is also expanding its multicloud strategy via AWS Interconnect. The service now includes a preview of multicloud connectivity with Oracle Cloud Infrastructure (OCI). OCI is the latest cloud service provider to adopt the open specification used by AWS Interconnect.

This connectivity service is already generally available for Google Cloud, and AWS expects to bring support for Microsoft Azure later in 2026.

Research and Community Initiatives

AWS has invested $110 million into the Build on Trainium program to support AI research and education. The program provides university researchers with access to AWS Trainium, purpose-built AI chips.

Participating institutions include Carnegie Mellon, MIT, and UC Berkeley. All research conducted through the program is open source to ensure that improvements benefit the wider developer community.

Other ecosystem updates include the return of the Kiro Startups Credit program, which offers up to one year of Kiro Pro+ credits for qualifying organizations. AWS also announced the full schedule for AWS Community Days 2026, a series of volunteer-led events for peers and developers occurring in cities across every continent.