Nvidia is fully committed to the development of the open-source model Nemotron 4, driving chip demand while directly facing competition from customers.

Nvidia is intensifying its investment in the open-source space by dedicating resources to develop a significant new model. The company hopes to drive hardware demand through this model, but this move also means it may compete with its own customers and partners. Multiple individuals involved in the Nemotron project revealed that Nvidia plans to create the largest foundational model in the Nemotron 4 series, with performance targets aligning with the worlds leading open-source large models. The previous generation's flagship model had 570 authors, and employees noted that the team involved in developing Nemotron 4 will further expand. One former employee stated, "At this stage, everyone hopes to be involved." Nvidia has recently launched several open-source models, with this development building on that foundation. On Tuesday, Nvidia released the Nemotron 3.5 Lightning lightweight model, which focuses on efficient and high-speed operation of intelligent agents. The company also introduced free model routing software to help enterprises quickly build model scheduling toolssuch tools can allocate different AI tasks to the most suitable and cost-effective models. Several Nemotron project employees disclosed that the flagship version of Nemotron 4 will have at least a trillion parameters. The parameter count refers to the adjustable units in the model's learning process, and this scale is approximately double that of Nvidia's current flagship model, Nemotron 3 Ultra. Even with a trillion parameters, this model is still smaller than the top open-source large models in the United States; nonetheless, Nvidia places significant importance on model compression techniques, and with these, even smaller models are expected to achieve better performance.