Nvidia Gives Away Key AI Compute Piece with Nemotron 3.5 Lightning

Nvidia has quietly given away a significant piece of its AI compute business by releasing Nemotron 3.5 Lightning, a free 30-billion-parameter open-weight model that runs on a single GPU.

This move is part of Nvidia's strategy to own the software layer that decides which model runs on their chips. By making this layer available for free, they can compete with other companies in the AI market and reinforce demand for their hardware.

Nemotron 3.5 Lightning uses a hybrid architecture combining Mamba-2 with attention and mixture-of-experts layers, supports a 1-million-token context window, and ships under the permissive OpenMDW-1.1 license.