Nvidia Expands Open Source AI with Acquisition and New Models
▼ Summary
– Nvidia acquired SchedMD, the developer of the open-source Slurm workload management system for high-performance computing and AI.
– Nvidia plans to continue operating Slurm as vendor-neutral, open-source software and will invest to accelerate its access.
– Nvidia also released a new family of open AI models called Nemotron 3, which it claims is highly efficient for building AI agents.
– The Nemotron 3 family includes three models (Nano, Super, Ultra) designed for tasks ranging from targeted to complex applications.
– This activity reflects Nvidia’s broader strategy to bolster its open-source AI offerings and become a key supplier for physical AI and robotics.
Nvidia is making significant moves to strengthen its position in the open source artificial intelligence ecosystem through a strategic acquisition and the launch of a new model family. The company’s dual announcements signal a deeper commitment to providing the foundational tools and infrastructure necessary for the next generation of AI development, particularly for complex agentic and physical systems.
The first major development is Nvidia’s acquisition of SchedMD, the primary developer behind the widely-used open source workload manager Slurm. This software is a critical component for managing high-performance computing and AI clusters. Originally launched over two decades ago, Slurm has become essential infrastructure for generative AI workloads. Nvidia plans to continue operating Slurm as a vendor-neutral, open source project while accelerating its development and integration across diverse computing systems. The company has collaborated with SchedMD for more than ten years, viewing the technology as indispensable for scaling AI operations.
Simultaneously, Nvidia introduced a new suite of open AI models named Nvidia Nemotron 3. The company positions this as the most efficient family of open models available for constructing precise AI agents. The family consists of three tailored variants: the Nemotron 3 Nano for focused, specific tasks; the Nemotron 3 Super, engineered for applications involving multiple interacting AI agents; and the Nemotron 3 Ultra, designed to handle the most intricate and demanding workloads. In a statement, Nvidia CEO Jensen Huang emphasized that “open innovation is the foundation of AI progress,” framing Nemotron as an effort to transform advanced AI into an open platform that offers developers the transparency and efficiency required for large-scale agentic system development.
These initiatives are part of a broader push by Nvidia to expand its open source and open AI portfolio. Just last week, the company unveiled Alpamayo-R1, an open reasoning vision-language model targeted at autonomous driving research. It also released additional workflows and guides for its Cosmos world models, which are open-sourced under a permissive license, to assist developers in building physical AI applications.
This flurry of activity underscores a strategic bet by the semiconductor leader. Nvidia is positioning itself as the essential supplier for the emerging field of physical AI, which encompasses robotics and autonomous vehicles. By providing both the powerful hardware and the critical open source software stack, from workload management with Slurm to specialized AI models like Nemotron, the company aims to be the go-to partner for organizations developing the intelligent “brains” behind next-generation autonomous machines.
(Source: TechCrunch)
