Real-world adoption stories from the CNCF ecosystem. Each case study highlights how organizations use HAMi to improve GPU utilization and scale AI infrastructure.
Building a unified heterogeneous AI compute scheduling platform on Kubernetes with HAMi for topology-aware scheduling and fine-grained accelerator sharing.
100% hardware pool utilization with topology-aware scheduling
Orchestrating 1,000+ A100 GPUs for GenAI features serving 200M+ global users with HAMi GPU sharing and KEDA autoscaling.
2Ã fewer GPUs with HAMi GPU sharing, handling 700% traffic spikes
Improving GPU utilization for autonomous driving workloads with HAMi-based GPU virtualization on Kubernetes.
10Ã GPU utilization improvement in CI pipelines
Scaling machine learning infrastructure with HAMi-based GPU virtualization on Kubernetes.
3x improvement in platform GPU utilization
Building a flexible GPU cloud with HAMi to increase utilization and improve delivery speed.
>80% average GPU utilization after vGPU adoption
Building a heterogeneous AI virtualization pooling solution (Effective GPU) with HAMi.
Up to 57% GPU savings for production and test clusters
Improving AI inference orchestration with HAMi in education-focused workloads.
90% of GPU infrastructure optimized using HAMi
