Nvidia developing Nemotron 4 open-source models, The Information reports

• The largest Nemotron 4 model is expected to have at least 1 trillion parameters, with final training still incomplete and a potential late fall release
• Nvidia is among few major U.S. firms releasing open-source models, a sector gaining traction as AI costs rise and Chinese models approach capabilities of leading American labs like Anthropic and OpenAI
• The company separately unveiled Nemotron 3.5 Lightning and NeMo Switchyard, an open-source routing library for directing AI tasks to suitable models

Nvidia Developing Nemotron 4 Open-Source AI Models to Compete Globally

Nvidia is developing Nemotron 4, a new family of open-source AI models designed to rival leading global systems, according to a report by The Information citing project insiders. The largest model in the family is expected to feature at least 1 trillion parameters, though no official release date has been set. Employees indicated the model could be ready as early as late fall, pending completion of final training.

This development positions Nvidia among the few major U.S. firms releasing open-source models, which have gained traction in 2026 as AI costs escalate and Chinese models approach capabilities of top American systems from Anthropic and OpenAI. Concerns have emerged around autonomous AI agents and cybersecurity risks, as open models typically lack usage restrictions.

Separately, Nvidia announced Nemotron 3.5 Lightning on Tuesday, targeting practical applications including code review, security monitoring, billing inquiries, and tool use. The company also released NeMo Switchyard, an open-source library that automatically routes AI tasks to optimal models.

In July, Nvidia joined other tech giants including Microsoft in signing an open letter supporting open-weight models to prevent innovation from shifting overseas. The company previously expanded its open-source model family late last year amid proliferating Chinese AI lab offerings.

Market Implications: The move signals Nvidia's strategic expansion beyond hardware into AI model development, potentially positioning the chip giant to capture value across the AI stack while addressing cost-conscious enterprises seeking alternatives to expensive proprietary systems. The open-source approach may also help Nvidia maintain competitive advantage against emerging international players.

Nvidia did not immediately respond to Reuters' request for comment.