Nvidia targets performance parity with top open-source AI models for Nemotron 4

Nvidia isn’t just selling the shovels in the AI gold rush anymore. It wants to dig, too. The company announced the Nemotron Coalition at GTC on March 16, a collaborative effort with leading AI laboratories designed to produce open, frontier-level foundation models. The first output: a base model co-developed with Mistral AI on Nvidia’s DGX Cloud platform, which will be open-sourced and serve as the bedrock for the upcoming Nemotron 4 family. What the coalition actually looks like The founding members read like a curated guest list for an AI power dinner. Mistral AI, the Paris-based startup that has become synonymous with efficient open models, is the lead co-developer. Black Forest Labs, the team behind the Flux image generation models, brings multimodal expertise. LangChain, the developer tooling company that has become infrastructure for thousands of AI applications, rounds out the initial roster. The first model built through this partnership will run on DGX Cloud, Nvidia’s managed infrastructure service that gives partners access to clusters of its latest GPUs without needing to build and maintain their own data centers. Jensen Huang, Nvidia’s CEO, framed the effort in characte...