SOURCE ARCHIVE · ORIGINAL LINKS INCLUDED
AWS customers can now access three new AI models via Amazon SageMaker JumpStart: GLM-5.2 FP8, NVIDIA Nemotron-Nano-12B-v2, and GLM-OCR, expanding AWS’s foundation model portfolio.
AWS Identity and Access Management introduced account access manager to simplify assigning IAM roles to workforce users via AWS IAM Identity Center, replacing two previous access management approaches.
Amazon EC2 has launched application status checks to detect application-level issues on instances, complementing existing system checks and enabling automated recovery via Auto Scaling.
AWS Elastic Disaster Recovery (DRS) now maintains UEFI boot mode for recovered Linux servers, eliminating post-recovery configuration steps and ensuring consistency with source environments.
AWS has launched EC2 R8i and R8i-flex instances in the Europe (Milan) region, featuring custom Intel Xeon 6 processors with improved performance and memory bandwidth for memory-intensive workloads.
AWS Parallel Computing Service (PCS) now meets FedRAMP, SOC, ISO, CSA STAR, and PCI compliance standards, enabling regulated organizations to run sensitive HPC workloads on AWS while adhering to governance requirements.
Amazon ECS now allows fractional GPU scheduling on EC2 G6f instances, enabling smaller GPU partitions for workloads like AI inference and rendering to reduce costs.
AWS Lambda console now supports console-to-IDE integration for Kiro and Cursor, expanding cloud-to-local development workflows for serverless applications.
AWS has launched Amazon EC2 G7 instances in the Europe (Spain) Region, featuring NVIDIA RTX PRO 4500 Blackwell GPUs for enhanced AI inference, graphics, and data analytics workloads.
AWS has introduced runtime instances for Amazon Bedrock AgentCore, enabling AI agents to run on custom EC2 instances with managed infrastructure, supporting sessions up to 14 days.
Amazon ElastiCache now offers Graviton4-based M8g, R8g, and C8gn nodes for Valkey and Memcached, delivering up to 47% higher throughput and 43% lower latency compared to Graviton3-based nodes.
AWS Transform now automates post-launch actions in migration workflows, reducing manual configuration errors and accelerating multi-account server migrations through predefined or custom SSM documents.
AWS Security Agent now supports penetration testing for applications using email-based MFA, expanding automated testing coverage beyond TOTP methods.
Amazon WorkSpaces Applications now provides enhanced observability metrics in CloudWatch at no extra cost, improving IT teams' ability to monitor and troubleshoot application streaming workloads.
Amazon WorkSpaces now provides enhanced observability metrics in CloudWatch at no extra cost, offering deeper visibility into virtual desktop performance and session health across all supported AWS Regions.
Amazon QuickSight introduces multi-dataset topics, allowing users to model relationships across multiple datasets without pre-joining data, simplifying dashboard creation and natural-language queries.
AWS introduces temporal policies and rate limiting in Amazon Bedrock AgentCore to enhance control over AI agent workflows and traffic management.
AWS Lambda introduces scalable network bandwidth for functions outside a VPC, increasing throughput from 625 Mbps to 3,000 Mbps based on memory configuration.
AWS Marketplace now offers AI Insights in the pricing section of product listings, explaining pricing structures in plain language to help buyers make informed decisions before purchasing.
AWS has launched general availability of vector search for Amazon DynamoDB, enabling real-time indexing and querying of vector embeddings at scale with single-digit millisecond latency.
AWS Security Hub Extended now includes Supply Chain Security as its 10th category, adding Chainguard and Socket as curated partners to detect malicious open-source dependencies before deployment.
AWS has expanded availability of Amazon EC2 C8g instances to four new regions, featuring AWS Graviton4 processors with up to 30% better performance than Graviton3-based instances.
Amazon S3 Vectors is now accessible in the AWS European Sovereign Cloud (Germany) Region, expanding vector storage capabilities for AI applications within the EU.
AWS has launched Amazon EC2 Storage Optimized I8g instances in Europe (Paris) and Asia Pacific (Jakarta) regions, featuring AWS Graviton4 processors and enhanced storage performance for I/O-intensive workloads.
Amazon RDS for SQL Server now offers Developer Edition in 13 additional AWS regions, enabling free development and testing with full Enterprise Edition features.
AWS has made its Transform continuous modernization service generally available, enabling engineering teams to analyze and address technical debt across code repositories at scale.
AWS Lambda has raised the maximum event pollers for Provisioned Mode in Amazon SQS event source mappings from 2,000 to 10,000, enabling up to 100,000 concurrent invocations per mapping.
AWS announced serverless full fine-tuning support in Amazon SageMaker for over 25 open-source models, enabling deeper customization beyond parameter-efficient methods.
Amazon ECR now allows image layers up to 200 GB via Docker push, simplifying storage for large datasets and models, while SDK/CLI uploads remain capped at 50 GB.
AWS Transform now supports offline schema conversion for SQL Server to Aurora PostgreSQL, allowing enterprises to modernize legacy .NET applications without a live database connection.