Overview
- SchedMD’s sale closes with undisclosed terms; the 2010-founded firm employs about 40 people and serves customers such as CoreWeave and the Barcelona Supercomputing Center.
- Slurm is used on more than half of the top 10 and top 100 systems on the TOP500 list and is considered critical for scheduling GPU resources in generative AI training and inference.
- Nvidia will continue open-source support, training and development for Slurm and says it will speed access to new systems while supporting heterogeneous hardware environments.
- The announcement arrived alongside Nvidia’s Nemotron 3 open-model family, which includes Nano for targeted tasks, Super for multi‑agent applications and Ultra for complex workloads.
- Industry analysts characterize the deal as strengthening Nvidia’s grip on AI infrastructure and deepening its software advantage built around CUDA.