While Nvidia has long been known for its high-performance graphics and AI chips, the company has also made a name for itself in software. Its proprietary CUDA platform, widely adopted by AI researchers and developers, has been a key factor in cementing its dominance. Now, by adding SchedMD to its portfolio, Nvidia is doubling down on software that enables efficient management of the massive computational demands of AI training and inference.
SchedMD, founded in 2010 in Livermore, California by Morris “Moe” Jette and Danny Auble, is best known for its Slurm software, an open-source system that schedules large computing jobs across data center servers. The software is free for developers and organizations to use, while SchedMD generates revenue by offering engineering support and maintenance services. Its client roster includes major players like cloud infrastructure provider CoreWeave and the Barcelona Supercomputing Center.
In a statement, Nvidia highlighted that Slurm is “part of the critical infrastructure needed for generative AI,” supporting developers in managing model training and inference on Nvidia hardware. The company confirmed that it would continue distributing SchedMD’s software under its open-source model, signaling a commitment to keeping the platform widely accessible. Financial terms of the acquisition were not disclosed.
The announcement comes alongside Nvidia’s unveiling of a new family of open-source AI models, which the company claims are faster, cheaper, and more efficient than previous offerings. These moves reflect Nvidia’s response to mounting competition from rival AI labs, including a wave of new open-source models emerging from China.
Investors responded positively, with Nvidia shares climbing 1.35% following the dual announcements. The acquisition positions Nvidia not just as a chipmaker, but as a more holistic AI ecosystem player, emphasizing the symbiotic role of hardware and open-source software in shaping the next generation of AI innovation.