Nvidia bulks up open source offerings with an acquisition and new open AI models 

Nvidia continues to grow its footprint successful unfastened root AI connected 2 fronts: an acquisition and a caller exemplary release.

The semiconductor elephantine announced Monday it acquired SchedMD, the starring developer of fashionable unfastened root workload absorption strategy Slurm. Nvidia said the institution volition proceed to run the program, which is designed for high-performance computing and AI, arsenic an unfastened source, vendor-neutral software.

Slurm was primitively launched successful 2002 and SchedMD was founded successful 2010 by the pb Slurm developers Morris Jette and Danny Auble. Auble is the existent CEO of SchedMD.

Terms of the woody weren’t disclosed. Nvidia declined to remark connected the quality beyond the company’s blog post.

Nvidia has been moving with SchedMD for much than a decennary and said successful its blog station the exertion is captious infrastructure for generative AI. The institution plans to support investing successful the exertion and “accelerate” its entree to antithetic systems.

The semiconductor institution besides released a caller household of unfastened AI models connected Monday. The institution claimed this radical of models, called Nvidia Nemotron 3, is the astir “efficient household of unfastened models” for gathering close AI agents.

This exemplary household includes the Nemotron 3 Nano, a tiny exemplary for targeted tasks, the Nemotron 3 Super, a exemplary built for multi-AI cause applications, and Nemotron 3 Ultra, built for much analyzable tasks.

“Open innovation is the instauration of AI progress,” Jensen Huang, laminitis and CEO of Nvidia, wrote successful the company’s property release. “With Nemotron, we’re transforming precocious AI into an unfastened level that gives developers the transparency and ratio they request to physique agentic systems astatine scale.”

In caller months, Nvidia has pushed to bolster its unfastened root and unfastened AI offerings.

Last week, the institution announced a caller unfastened reasoning imaginativeness connection model, Alpamayo-R1, which is focused connected autonomous driving research. The institution besides said astatine the clip it added much workflows and guides covering its Cosmos satellite models, which are unfastened root nether a permissive license, to assistance developers amended usage the models to make carnal AI.

The enactment is reflective of Nvidia’s stake that physical AI volition beryllium the adjacent frontier for its GPUs. Nvidia wants to beryllium the go-to supplier for the galore robotics — oregon self-driving conveyance — companies looking for the AI and bundle to make the brains down the technology.

Topics