NVIDIA's new Nemotron coalition aims to build open-source AI models
NVIDIA just kicked off the Nemotron Coalition at its GTC 2026 event in San Jose, California, bringing together eight leading AI groups, including Mistral AI, LangChain, and Perplexity, to cocreate powerful open-source AI models on NVIDIA's DGX Cloud.
First project is a base model for the Nemotron 4 family
The coalition's first project is a base model for the new Nemotron 4 family, built by NVIDIA and Mistral AI.
Each member brings something unique, like Sarvam's sovereign AI tech, Cursor's coding benchmarks, LangChain's agent reasoning (which is already super popular), and Black Forest Labs' multimodal tools.
More developers get to build cool stuff across different industries
Nemotron wants to make advanced AI more accessible by creating open models that anyone can use or adapt, not just big tech companies.
This could mean more developers get to build cool stuff across different industries using NVIDIA's cloud power.
The launch was highlighted by a keynote from NVIDIA CEO Jensen Huang.