We are thrilled to announce that Canadian customers can now leverage advanced foundation models such as Anthropic’s Claude Sonnet 4.5 and Claude Haiku 4.5 via Amazon Bedrock, utilizing cross-Region inference (CRIS). This significant enhancement enables organizations across Canada to access the latest AI technologies, thereby accelerating their AI initiatives.\n\nAmazon SageMaker HyperPod clusters, integrated with Amazon Elastic Kubernetes Service (EKS), now facilitate the creation and management of interactive development environments like JupyterLab and Visual Studio Code. This development simplifies the machine learning lifecycle by providing managed environments tailored for data scientists. A new feature called Amazon SageMaker Spaces allows developers to create self-contained environments for running notebooks, maximizing GPU investments by running both interactive and training jobs on the same infrastructure.\n\nTo get started, administrators can install the Spaces add-on through the SageMaker AI console, utilizing either a quick or custom installation approach. This allows for seamless setup of environments where data scientists can create and connect to their unique workspaces. Users can access these workspaces via a web UI or connect directly from their local Visual Studio Code setup using secure SSH-over-SSM tunneling.\n\nIt’s essential for organizations to manage their resources wisely. Recommendations for best practices include enforcing user access controls through EKS Access Entries, creating templates for standardized environments, and utilizing fractional GPU support for better resource allocation efficiency. Additionally, enabling automatic shutdown of idle workspaces can optimize compute resource usage.\n\nBy integrating these capabilities, teams can enhance their productivity, focus on developing AI models, and ensure that they are working within secure, managed environments. As Canadian organizations embrace these advancements, they can look forward to a more streamlined pathway to AI innovation.