NVIDIA has formed 'Nemotron Coalition': What will it do?
What's the story
NVIDIA has announced the formation of a coalition with leading artificial intelligence (AI) labs across the globe, including India's Sarvam. The group, dubbed the Nemotron Coalition, was launched at NVIDIA's annual GPU Technology Conference in San Jose, California. Other members of this pioneering alliance include Thinking Machines Lab founded by former OpenAI CTO Mira Murati, Cursor, Mistral AI, Perplexity, Reflection AI, Black Forest Labs, and LangChain.
Coalition goals
The coalition's 1st project is a base model
The Nemotron Coalition aims to fast-track AI model development, broaden intelligence beyond individual models, and reinforce an open ecosystem. It also hopes to make model development more efficient for organizations. The coalition's first project is a base model co-developed by Mistral AI and NVIDIA. This collaboration will leverage the AI expertise and technology of both companies for further development.
Member input
Each member will contribute their unique expertise
Each member of the Nemotron Coalition will bring their unique expertise to collaboratively build a shared model. This includes Sarvam's expertise in sovereign language AI development, Mistral AI's capabilities in building efficient customizable models, and Cursor's real-world performance requirements and evaluation datasets. Other contributions include a data collaboration with Thinking Machines Lab, multimodal capabilities from Black Forest Labs, high-performance AI systems from Perplexity, open systems from Reflection AI, and specialization in enabling reliable tool use and long-horizon reasoning by LangChain.
AI for all
Sarvam co-founder's take on the initiative
Sarvam Co-founder Pratyush Kumar emphasized that "AI reaches its full potential when it works in every language and for every community." He added, "Open models make this possible by giving builders the freedom to adapt frontier capabilities to real-world needs." Last month, Sarvam had introduced two large language models with 30 billion and 105 billion parameters at the India AI Impact Summit.