Teaching Everyone to Fish for Tokens

Nvidia is investing in nearly open-source language models like Nemotron, releasing training data and code to foster a broad ecosystem where many can build AI models, increasing demand for Nvidia's inference chips. Open-source model recipes remain resource-intensive but accessible, allowing companies to create and improve models collaboratively. This approach aims to democratize AI development and expand Nvidia's market for AI hardware.