Nvidia releases Nemotron 3.5 Lightning, its first open-source model
Nvidia has released Nemotron 3.5 Lightning, its first open-source AI model, allowing companies to download, modify, and use it at no cost. The lightweight model is designed to run on a single GPU on a PC and is tailored for autonomous AI agents. The release follows CEO Jensen Huang's debut on X defending open-source models and comes amid a U.S. policy debate over Chinese models like Kimi K3. For Nvidia, open source drives demand for its GPUs. The company also launched NeMo Switchyard, a tool to select the most cost-effective model for tasks. Notably, Nvidia used distillation to give the model capabilities similar to larger Nemotron versions, a technique that has raised national security concerns when used by Chinese developers.
Nvidia releases Nemotron 3.5 Lightning, its first open-source model
The new Nemotron offering is free for companies to download, use and modify without getting permission or paying Nvidia. Nvidia is releasing Nemotron 3.5 Lightning, which the company says is "lightweight" and can run on a single graphics processing unit on a PC. The chipmaker said it was developed particularly for agents, or AI programs that can run autonomously in the background.
Key takeaway
Nvidia's first open-source model, Nemotron 3.5 Lightning, is free, lightweight, and agent-focused, positioning Nvidia to sell more GPUs by lowering AI costs and countering proprietary rivals.
What happened
Nvidia released Nemotron 3.5 Lightning, its first open-source model, which is free for companies to download, use and modify without paying Nvidia. The chipmaker says the model is lightweight and can run on a single GPU on a PC, and it was designed for autonomous AI agents.
The release comes less than three weeks after CEO Jensen Huang posted on X for the first time to defend open-source AI models, urging the U.S. government to support open models while avoiding premature restrictions. Nvidia also launched NeMo Switchyard, a tool to select the most appropriate AI model for a task, and said it used distillation to give the model similar capabilities to its larger Nemotron models.
Evidence
Nvidia released Nemotron 3.5 Lightning as its first open-source model, free for companies to download, use and modify.
CNBC · attributed
The new Nemotron offering is free for companies to download, use and modify without getting permission or paying Nvidia.
Why it matters
This strategic release could intensify competition in the AI model market, potentially lowering prices and increasing adoption, while also raising questions about the use of distillation and its implications for AI governance and national security.
Limits and uncertainties
The article does not provide independent performance benchmarks for Nemotron 3.5 Lightning, so its actual capabilities relative to other models are unclear.
The use of distillation, while acknowledged, may face regulatory scrutiny given concerns about IP theft.
Practical implications
Developers can now access a free, lightweight model for building AI agents, potentially reducing costs.
Organizations might use NeMo Switchyard to optimize model selection, but they should verify performance against their specific use cases and consider the hardware requirements.
What to watch
Watch for adoption by companies like CrowdStrike and Harvey, and for any regulatory actions regarding distillation in open-source models.
