Amazon EKS Auto Mode reduces GPU management fees by up to 60%

Amazon Elastic Kubernetes Service (Amazon EKS) Auto Mode now offers significantly reduced management fees for GPU and accelerated instance types. Beginning July 1, 2026, G-series Auto Mode management fees are reduced by 35%, and P-series and AWS Trainium fees are reduced by 60%. These reductions apply automatically to all EKS Auto Mode clusters and no action is required from customers already using GPU instances with Auto Mode.
EKS Auto Mode simplifies Kubernetes operations by automatically provisioning and managing infrastructure for machine learning inference, fine-tuning, rendering, and batch processing workloads. It includes capabilities built for accelerated workloads: automatic parallel image pulling and unpacking on GPU instances with local NVMe storage, so large container and model images start faster, and accelerator-aware node repair that detects GPU hardware failures and automatically replaces unhealthy nodes. With today’s price reduction, customers can run GPU workloads on Auto Mode at lower management fees, making its fully managed infrastructure more cost-effective.
This pricing update is available in all AWS Regions where EKS Auto Mode is available. Amazon ECS is implementing identical management fee reductions for GPU instances on ECS Managed Instances. See the ECS What’s New post for details.
To get started with GPU workloads on EKS Auto Mode, see the EKS for AI/ML documentation. For the complete updated rate table, see Amazon EKS pricing.
Quelle: aws.amazon.com

Amazon EC2 C8ine instances are now available in AWS Europe (Frankfurt) region

Starting today, Amazon Elastic Compute Cloud (Amazon EC2) C8ine instances are available in the AWS Europe (Frankfurt) region. C8ine instances are powered by custom sixth generation Intel Xeon Scalable processors, available only on AWS. These instances feature the latest sixth generation AWS Nitro cards, delivering up to 43% higher performance compared to previous generation C6in instances.
C8ine instances offer up to 2.5 times higher packet performance per vCPU versus prior generation network optimized instances, providing up to 2x higher network throughput for traffic going through Internet gateways compared to existing C6in network optimized instances. C8ine instances are designed for security and network virtual appliances, including virtual firewalls, load balancers, and Telco 5G UPF workloads.
Amazon EC2 C8ine instances are available in US East (N. Virginia), US West (Oregon), Asia Pacific (Tokyo), and Europe (Frankfurt) regions. C8ine instances are available via Savings Plans and On-Demand instances. For more information, visit the Amazon EC2 C8i instance pages.
Quelle: aws.amazon.com

Amazon EMR Serverless now supports larger worker sizes to run more compute and memory intensive workloads

Amazon EMR Serverless now offers larger worker configurations of 32 vCPUs with up to 244 GB of memory, allowing you to run more compute and memory-intensive workloads. Previously, the largest worker configuration available on EMR Serverless was 16 vCPUs with up to 120 GB of memory. Larger workers can help you improve the runtime performance as well as cost profiles for your workloads.
For shuffle-heavy workloads, larger workers reduce inefficient data transfers between executors. For jobs with data skew, larger workers reduce the chances of out-of-memory failures. For jobs that need to cache data, larger workers allow holding more data in memory, boosting job performance. To take advantage of these benefits, we recommend using larger workers for your compute and memory-intensive Spark and Hive workloads.
To learn more about different worker configurations, please visit EMR Serverless documentation. Larger workers are available in all AWS Regions where EMR Serverless is available.
Quelle: aws.amazon.com

Amazon ECS Managed Instances reduces GPU management fees by up to 60%

Amazon Elastic Container Service (Amazon ECS) Managed Instances now offers significantly reduced management fees for GPU and accelerated instance types. Beginning July 1, 2026, G-series ECS management fees are reduced by 35%, and P-series and AWS Trainium fees are reduced by 60%. These reductions apply automatically and no action is required from customers already using GPU instances with ECS Managed Instances. With ECS Managed Instances, you get the application performance you want and the simplicity you need. Simply define your task requirements such as the number of vCPUs, memory size, and CPU architecture, and Amazon ECS automatically provisions, configures and operates most optimal EC2 instances within your AWS account using AWS-controlled access. You can also specify desired instance types, including GPU-accelerated, network-optimized, and burstable performance, to run your workloads on the instance families you prefer. ECS Managed Instances includes capabilities built specifically for accelerated workloads: GPU metrics (utilization, memory, and temperature) through Amazon CloudWatch Container Insights, and automatic health monitoring that detects GPU-specific hardware failures and replaces unhealthy instances to minimize workload disruption. With today’s pricing update, customers running GPU workloads on ECS Managed Instances can now benefit from fully managed infrastructure at lower management fees. This pricing update is available in all AWS Regions where ECS Managed Instances is available. For the complete updated rate table, see ECS Managed Instances pricing. Amazon EKS is implementing identical management fee reductions for GPU instances on EKS Auto Mode. See the EKS What’s New Post for details. To learn more about ECS Managed Instances, visit the feature page, documentation, and AWS News launch blog.
Quelle: aws.amazon.com

Amazon SageMaker Studio now integrates with Hugging Face for one-click model deployment and customization

Amazon SageMaker Studio now supports direct integration from Hugging Face, letting you go from discovering a model to working with it inside a fully configured Studio environment in a single click. Select any supported model on Hugging Face and choose “Customize on SageMaker AI” or “Deploy on SageMaker AI” to land directly on the corresponding workflow page with the model pre-loaded and ready to use.
Previously, getting from model discovery to a working environment required navigating the AWS Console to find SageMaker AI, configuring an environment, setting up IAM permissions for serverless model customization, and in many cases requesting GPU quota increases through Service Quotas before running a first job. Now, new customers complete a standard AWS sign-up and receive a SageMaker Studio environment created in seconds with pre-configured permissions for serverless model customization jobs including fine-tuning with custom reward functions for reinforcement learning, model evaluation, and deployment to SageMaker or Bedrock endpoints. Verified customers receive default GPU access to G5, G6, and G4dn instances across endpoint deployments, training jobs, and notebooks without requesting quota increases, and quota limit and utilization information is visible for each instance type directly inside the Studio environment. Returning customers signing in from Hugging Face or SageMaker product pages select their environment and land directly inside SageMaker Studio with the model ready to use.
This feature is available in all AWS Commercial Regions where Amazon SageMaker Studio is supported. To get started, visit any supported model on Hugging Face and select “Customize on SageMaker AI” or “Deploy on SageMaker AI,” or click Get Started from the SageMaker Studio page. To learn more, see Service quotas for Studio in the Amazon SageMaker documentation.
Quelle: aws.amazon.com