NVIDIA bought SchedMD, the developer of Slurm, a popular open-source scheduler for HPC and AI workloads, to support innovation among researchers, developers, and enterprises, and bolster its open-source software ecosystem.
The firm plans to maintain Slurm as an open-source, hardware-agnostic tool, ensuring it remains broadly accessible to researchers and AI developers.
Nvidia will increase SchedMD’s access to new systems and boost its development
Nvidia is increasing its investment in open-source tools as part of a broader strategy to stay ahead in the rapidly growing artificial intelligence market. It has been working closely with SchedMD for over ten years, and now, with the acquisition, it will continue to invest in Slurm.
Slurm, which translates to Simple Linux Utility for Resource Management, has long been an essential part of supercomputing. Currently, it is used in over half of the TOP500 supercomputers worldwide, allowing complex parallel computations to be scheduled. It also allows resources to be allocated across thousands of CPUs and GPUs.
By integrating SchedMD, NVIDIA gains stewardship of this crucial piece of the HPC and AI software stack, connecting hardware acceleration (via NVIDIA’s Blackwell GPUs and InfiniBand networking) to sophisticated job scheduling and resource orchestration. Such integration will enhance performance for everything from training large language models to running mission-critical scientific simulations
Danny Auble, CEO of SchedMD, spoke about the buy, saying, “We’re thrilled to join forces with NVIDIA, as this acquisition is the ultimate validation of Slurm’s critical role in the world’s most demanding HPC and AI environments. NVIDIA’s deep expertise and investment in accelerated computing will enhance the development of Slurm — which will continue to be open source — to meet the demands of the next generation of AI and supercomputing.”
Nvidia also asserted, “Slurm, which is supported on the latest Nvidia hardware, is also part of the critical infrastructure needed for generative AI, used by foundation model developers and AI builders to manage model training and inference needs.”
The AI chipmaker is bound to expand SchedMD’s reach into new systems, enabling customers to manage workloads more efficiently across their entire infrastructure. Additionally, the integration will enable customers to coordinate workloads across different hardware types and software better while benefiting from ongoing Slurm innovations.
The company also plans to continue supporting Slurm with open-source software services and training for SchedMD’s broad customer base, which spans cloud, AI, manufacturing, and research organizations.
Nvidia announced new Nano models earlier this week
On Monday, Nvidia unveiled a new generation of open-source AI models, designed to be faster, more efficient, and more capable than their predecessors, responding to a surge in similar releases from China. It revealed its latest Nemotron models for use cases such as writing and software development, starting with the release of Nemotron 3 Nano. According to the chipmaker, the new Nano model cuts costs while improving accuracy on longer, more demanding workloads.
Meanwhile, Meta is reportedly weighing a shift toward closed-source models, potentially making Nvidia one of the most prominent U.S. providers of open-source AI. To date, several U.S. states and government agencies have banned Chinese AI systems over security concerns. Most of those bodies claimed that Chinese models are being used in the Asian country’s military and intelligence operations.
Nonetheless, Kari Briski, Vice President of Generative AI at Nvidia, emphasized that the company aims to provide users with a trustworthy model and is making training data and tools available for security testing and customization. Briski noted, “This is why we’re committed to it from a software engineering perspective.”
The company’s shares even rose 1.35% following the announcement of its open-source AI models.
If you're reading this, you’re already ahead. Stay there with our newsletter.
