Nvidia Wants to Fix the Shortage of Strong American Open AI Models
Nvidia is expanding its open-weight Nemotron family — including Super and ~500B-parameter Ultra variants — and plans to release weights, training data and RL environments for customisation.
Nvidia wants to close what it sees as a gap in strong, openly licensed American AI models — and is using its Nemotron family to do it.
The short version
- Nvidia is expanding Nemotron, its line of open-weight models, with larger Super and Ultra variants.
- The Ultra tier is around 500 billion parameters, aimed at complex agentic tasks.
- Nvidia plans to release weights, training data and reinforcement-learning environments for customisation.
Why it matters
Open, inspectable models give developers an alternative to closed APIs — and put pressure on commercial labs to justify their value.
Summary by Nerd News Network. Read the full article at The Register via the links above and below.