Many AI workloads need only a fraction of a GPU. Giving each one an entire device can leave capacity idle; sharing GPUs lets platform teams serve more workloads with the infrastructure they already have.
Official announcementNVIDIA Developer Blog
Memory-Enforced GPU Sharing in Kubernetes with NVIDIA CUDA 13.4
From the official release
NVIDIA GPUs already support several approaches to sharing. Multi-instance GPU (MIG) splits a GPU into hardware-isolated instances that are transparent to the application, time-slicing lets several processes take turns without memory enforcement, and Multi-Process Service (MPS) lets them run concurrently in separate address spaces (sm-sharing).
CUDA 13.4 now enforces a memory boundary that is transparent to the application with a new time-sharing mode, alongside the compute sharing mode (sm-sharing) that already existed in MPS.
Here is what the CUDA release brings:
This is a short excerpt. Read the full announcement on the official source.
Continue on developer.nvidia.com → (opens in a new window)Related stories
Official announcement
Muse Is Now Available in Singapore: The World’s First Personal AI Agent Built for Everyone
Meta NewsroomFacebook
Following our global announcement of Muse , we are excited to share that Muse is now available in Singapore. https://about.fb.com/wp-content/uploads/2026/09/Introducing-Muse_Sizzle-Video.mp4 What Muse Can Do in Singapore Connectors are a big reason Muse can do so much. People in
Official announcement
Amazon ECS now supports GPU sharing with NVIDIA Multi-Process Service (MPS)
AWS What’s New
Amazon Elastic Container Service (Amazon ECS) now supports GPU sharing with NVIDIA Multi-Process Service (MPS), enabling you to run multiple Amazon ECS tasks concurrently on a single NVIDIA GPU. GPU sharing helps you increase GPU utilization and reduce infrastructure costs for wo
Official announcement
NVIDIA RTX Spark Powers Neural Rendering and AI Characters in Games
NVIDIA Developer Blog
NVIDIA RTX Spark is a new Windows PC platform for AI development, creation, and gaming in slim laptops and compact desktops. Powered by an NVIDIA Blackwell RTX... NVIDIA RTX Spark is a new Windows PC platform for AI development, creation, and gaming in slim laptops and compact de
Official announcement
SageMaker MLflow Apps now support MLflow 3.16
AWS What’s New
Amazon SageMaker AI MLflow Apps now support MLflow 3.16, bringing a redesigned tracing experience and faster experiment tracking to your generative AI and machine learning workflows. MLflow 3.16 improves debugging and analysis of generative AI applications and agents. The redesig
Official announcement
Snowflake and Oracle: Bringing the AI Data Cloud to OCI
Snowflake Blog
Connected data for trusted AI A business decision rarely depends on information from just one application. Financial results need context from sales activity. Inventory plans depend on supplier performance and customer demand. AI becomes more useful when it can work with those co
Official announcement
Automated Snowpipe Setup with Snowflake CoCo Is Now GA
Snowflake Blog
Load the same data into open Apache Iceberg tables Today, we’re announcing the general availability of automated Snowpipe setup with Snowflake CoCo , enabling teams to go from raw files arriving in cloud storage to a continuously loading standard Snowflake table or Snowflake-mana