Briski ties Nemotron interest to U.S. origin and data inspectability
Nvidia’s three new models, released Thursday as part of its Nemotron family, are aimed at helping companies build a variety of AI agents, the company said. Kari Briski, Nvidia’s vice president of generative AI software for enterprise, said the company is committed to regularly releasing newer, better open models.
“This is a message to developers and our partners and enterprises that they can depend on us,” Briski said.
One of the three models, dubbed Nemotron 3.5 Super VL, is designed to help businesses build what Nvidia describes as “always-on” AI agents and is faster for long-running agentic tasks, the company said. Briski said a distinguishing feature is that it can handle images and video in addition to text.
The model has been customized by several companies using their own data, including Atlassian, CrowdStrike, Dassault Systèmes, Salesforce and Synopsys. Briski said the companies are post-training Nvidia’s model because they can get better results from a customized model than from a top-of-the-line frontier model.
Separately, Meta Platforms said Thursday it was releasing a new open-weight AI model and announced a collaboration with Apple to run it locally on Apple’s high-end Mac computers. Taken together, the Nvidia and Meta releases add to the roster of American-made AI options, though it remains too soon to tell whether the new models will entice developers away from cheaper open models from China.
Briski said Nvidia has seen greater interest in its Nemotron models because they are U.S.-made and because their underlying data can be inspected. Public-sector organizations and financial institutions are particularly wary of using Chinese-made models, she said. Companies have also grown more comfortable with customizing, or fine-tuning, open models, according to Briski.