Heterogeneous Management
Manage and schedule GPU, NPU, MLU, and other accelerators in one workflow.
HAMi is an open-source, cloud-native GPU virtualization middleware that brings sharing, isolation and scheduling of heterogeneous accelerators to AI workloads on Kubernetes.


Virtualization âĸ Sharing âĸ Isolation âĸ Scheduling






HAMi is a Sandbox project of the Cloud Native Computing Foundation (CNCF), listed in both the CNCF Landscape and the CNAI Landscape.
Manage and schedule GPU, NPU, MLU, and other accelerators in one workflow.
Slice memory and compute precisely with hard isolation at runtime.
Use binpack, spread, and topology-aware policies for better placement.
Work with Kubernetes APIs, DRA, and CDI for easier adoption.
Control memory and core quotas for fairer and more stable sharing.
Provide consistent metrics and visibility across device vendors.
From request to isolation, HAMi turns GPU slicing and heterogeneous scheduling into usable Kubernetes runtime paths.
nvidia.com/gpu+gpumem/gpucoresCompare traditional whole-GPU allocation with HAMi GPU sharing under the same workloads.
Broad accelerator ecosystem across vendors. See docs for full support matrix.
View full supported devices list âThe organizations below are evaluating or using HAMi in production environments.
Submit your organization through the contributor guide process.
See submission instructions âHAMi is advanced by contributors from the community and industry. These organizations actively participate in project development and ecosystem collaboration.
A live snapshot of HAMi community growth and open-source momentum.