Nvidia is developing a new family of open-source AI models named Nemotron 4.
The largest model version will contain at least one trillion parameters.
This initiative positions the company to compete directly with top global open-source alternatives.
The project signals a strategic expansion from chip manufacturing into the AI software ecosystem.
Nvidia expects these models to increase enterprise demand for its GPU hardware.
Project insiders report the model could be ready by late fall 2026.