NVIDIA launched the Nemotron Coalition on March 16 at GTC in San Jose, uniting eight AI labs to co-develop open frontier models on DGX Cloud. The founding members include Black Forest Labs, Cursor, LangChain, Mistral AI, Perplexity, Reflection AI (led by former OpenAI CTO Mira Murati), Sarvam, and Thinking Machines Lab.

What Happened

Jensen Huang announced the Nemotron Coalition during the GTC keynote, describing it as a collaborative effort to build open-source frontier AI models that any developer or company can use. The first project is a base model co-developed by Mistral AI and NVIDIA, trained on NVIDIA DGX Cloud. Once complete, it will be shared as open-source and underpin the NVIDIA Nemotron 4 model family.

Each coalition member contributes domain expertise, proprietary evaluation benchmarks, and specialized data for post-training. In return, members get access to frontier-scale training runs that would otherwise be prohibitively expensive for any single organization. Training on the first collaborative model is currently underway, though NVIDIA has not announced a specific release timeline.

Why It Matters

The open vs. proprietary AI debate has defined the industry for the past two years. The Nemotron Coalition is NVIDIA's answer: fund open frontier models at the same scale as proprietary ones by splitting the cost across multiple labs. As Jensen Huang put it, "Open models are the lifeblood of innovation."

The member list signals what NVIDIA thinks matters. Mistral AI brings open-source model expertise. Cursor represents the AI coding market. Black Forest Labs contributes image and video generation. Perplexity adds search and retrieval. LangChain covers the agent and tooling layer. Reflection AI, led by Mira Murati, brings frontier research experience from OpenAI. Sarvam focuses on multilingual AI for underserved markets. Thinking Machines Lab adds research depth.

Mistral CEO Arthur Mensch framed the strategic argument: "Open frontier models are how AI becomes a true platform." By pooling resources on DGX Cloud, smaller labs can participate in training runs at a scale that previously required billions in capital, leveling the playing field against closed-model providers.

For developers already using NVIDIA's Nemotron 3 Super model, the coalition means the next generation of Nemotron models will benefit from diverse training data and evaluation expertise across eight different AI domains.

Key Details

  • Founding members: Black Forest Labs, Cursor, LangChain, Mistral AI, Perplexity, Reflection AI, Sarvam, Thinking Machines Lab.
  • First project: Base model co-developed by Mistral AI and NVIDIA on DGX Cloud.
  • License: Open-source, underpinning the Nemotron 4 family.
  • Member contributions: Domain data, evaluation benchmarks, and post-training expertise.
  • Member benefit: Access to frontier-scale training infrastructure on DGX Cloud.
  • Status: Training underway, no public release date announced.

What to Do Next

Developers and teams building on open-source AI models should follow the coalition's progress for updates on the first collaborative model release. Those already invested in the Nemotron ecosystem can expect direct improvements to the model family. For broader context on NVIDIA's expanding open model strategy, the SiliconAngle analysis covers how the coalition fits into NVIDIA's wider GTC announcements.