NVIDIA has acquired SchedMD, the primary developer of Slurm, an open-source workload manager widely used to schedule and allocate compute resources across high-performance computing and AI clusters. NVIDIA said it will continue to develop and distribute Slurm as vendor-neutral open-source software, maintaining support across a range of hardware and software environments. Financial terms of the deal were not disclosed.
Slurm plays a central role in managing large-scale parallel workloads and is deployed on more than half of the top supercomputing systems globally, making it core infrastructure for both traditional HPC and modern AI model training and inference. By bringing SchedMD in-house, NVIDIA is deepening its involvement in the software layer that orchestrates how compute resources are consumed, while committing to preserve Slurm’s open-source governance and broad ecosystem support. The acquisition reflects NVIDIA’s growing focus on owning more of the AI and accelerated-computing stack beyond hardware, particularly in areas that determine efficiency, scalability, and utilization across increasingly complex and heterogeneous clusters.