New updates on AWS GovCloud (US) | Issue #10 (2026)

Anthropic Claude Haiku 5.5 is now available on Amazon Bedrock in AWS GovCloud (US). Haiku 5.5 is the fastest, most cost-efficient model in the Claude 5.5 family, and Anthropic reports it runs about 75% cheaper than Haiku 4.5 for most tasks. It is also the first Haiku with effort controls, so mission owners can tune the balance of cost and intelligence for each task. That profile fits the high-volume, latency-sensitive work missions run every day, from document classification and field extraction to summarization, live support and voice agents, and in-app assistants.

With Haiku joining Claude Sonnet 5.5 and Claude Opus 5.5 on Amazon Bedrock, mission owners now have the complete Anthropic Claude 5.5 family available in AWS GovCloud (US). That gives teams the flexibility to pair models by workload: a larger model like Opus 5.5 can plan the work and hand off well-defined coding, tool-use, and browser-automation subtasks to Haiku 5.5, making it practical to run many agents in parallel at a fraction of the cost. Amazon Bedrock delivers this full selection through one API with AWS-managed Guardrails, Knowledge Bases, and in-region data residency, all without leaving AWS GovCloud (US) or re-architecting. See how AI model choice brings Mission Advantage to AWS GovCloud (US). Learn more by reading the blog.

Claude Sonnet 5.5 and Opus 5.5 are also now available in Kiro on AWS GovCloud (US), powered by Amazon Bedrock. Developers get a similar model-pairing approach in their agentic IDE: larger models drive planning and complex refactoring while cost-efficient models handle fast, high-volume coding subtasks and tool calls. The result is a cost-efficient, spec-driven development loop that runs entirely within AWS GovCloud (US).

Getting Started with Amazon Managed Workflows for Apache Airflow (MWAA) Serverless in AWS GovCloud (US)

Amazon Managed Workflows for Apache Airflow (MWAA) Serverless is available in both AWS GovCloud (US) Regions, giving mission teams a fully managed way to orchestrate data pipelines, ETL jobs, and operational workflows without provisioning or maintaining Apache Airflow infrastructure. For organizations that coordinate work across services like Amazon S3, AWS Glue, Amazon EMR, and Amazon SageMaker, MWAA Serverless removes the undifferentiated heavy lifting of environment management so teams can focus on the workflow logic that drives their mission.

Amazon MWAA Serverless is accessible through the AWS CLI, the REST API, or programmatically using AWS SDKs, giving teams flexibility in how they create and manage workflows across both AWS GovCloud (US-East) and AWS GovCloud (US-West) Regions. The Amazon MWAA Serverless console is not available in AWS GovCloud (US). Built on Apache Airflow 3, it eliminates the need to provision, configure, or tune infrastructure. Workflows are defined using YAML-based DAG definitions and scheduled using schedule expressions within the workflow definition. Because Amazon MWAA Serverless operates without the Airflow web interface, workflow monitoring and management are handled through Amazon CloudWatch and AWS CloudTrail.

For security and isolation, each workflow runs with its own execution role and isolated compute environment, enabling workflow-specific IAM permissions for least-privilege access without requiring separate Airflow environments. Teams can choose to run tasks in their own Virtual Private Cloud (VPC), providing precise security controls at the workflow level. This architecture prevents cross-workflow interference and reduces security administration overhead while strengthening overall security posture in regulated environments.

AWS service operators from the Amazon Provider Package enable native integration with services like Amazon S3, AWS Glue, Amazon EMR, Amazon Redshift, and Amazon SageMaker. A built-in versioning system enables teams to manage workflow evolution and roll back to previous versions when needed. Amazon MWAA Serverless automatically provisions compute resources when a workflow runs and releases them when execution completes, with a usage-based pricing model that charges per task for its duration. There are no upfront costs and no minimum provisioning required, making it well suited for variable or unpredictable workloads, cost-sensitive development and production environments, and multi-team organizations that need workflow isolation for security or compliance.

Learn more:  What's New Post

• Contact us for help by way of email
• AWS GovCloud (US) Website
• AWS GovCloud (US) User Guide
• AWS GovCloud (US) What's New