Nvidia Unveils Nemotron 3, Acquires SchedMD, and Launches MGX

Nvidia launches Nemotron 3 AI models, acquires SchedMD to keep Slurm open-source, and unveils MGX architecture for flexible, efficient AI data centers, strengthening its position in open-source AI and high-performance computing.

Nvidia has made significant moves in the AI and high-performance computing sectors with the launch of its third-generation Nemotron large-language models and the acquisition of SchedMD, the developer of the widely used Slurm workload management system. The Nemotron 3 Nano model is available now, with larger versions expected in 2026, offering improved efficiency and performance for writing, coding, and complex tasks. Nvidia’s open-source approach, including transparent training data and security tools, aims to build trust with enterprise and government users, especially as other providers shift strategies and Chinese models face bans in the U.S. The SchedMD acquisition ensures Slurm remains open-source and vendor-neutral, supporting diverse computing environments and accelerating development for next-generation AI workloads. Additionally, Nvidia’s MGX modular architecture is set to revolutionize data centers with flexible, efficient designs, supporting advanced AI-driven workloads and offering enhanced serviceability, security, and high-performance networking. These initiatives reinforce Nvidia’s commitment to open-source AI, robust infrastructure, and innovation in accelerated computing.