Nvidia is building Nemotron 4, a new family of open-source AI models whose largest version is expected to have at least 1 trillion parameters, The Information reported on Tuesday. The chip maker is positioning the models to compete with the best open-source systems available, a category that has drawn increasing attention as enterprises weigh the cost of proprietary AI against the risks and flexibility of open weights.
The company has not set a release date and has yet to finish training the models, though employees said the largest could be ready as early as late fall, according to the report. Nvidia pointed to previous remarks that it was working on Nemotron 4 but did not confirm the details.
Why open models are gaining ground
Open-source models have moved to the center of industry debate this year as AI bills balloon and cheap Chinese models approach the capabilities of top systems from Anthropic and OpenAI. A spate of recently disclosed hacks involving autonomous AI agents has added to the attention, especially because open models do not have curbs on cybersecurity use.
Nvidia is among the few major U.S. firms releasing open-source models. Last month it formed a coalition with other companies to develop and share tools for AI safety and cybersecurity, and it signed an open letter with tech heavyweights such as Microsoft backing open-weight models so that innovation does not drift overseas.
For IT and development teams, the distinction matters operationally: open-weight models can be deployed on internal infrastructure, fine-tuned for specific codebases, and audited directly, whereas closed models require sending data to external APIs and accept whatever guardrails the vendor imposes.
What Nvidia says about the project
Kari Briski, vice president of generative AI at Nvidia, said in an emailed statement: "Nvidia is investing in Nemotron because we believe every company and every country needs accessible frontier open models to strengthen safety and security, accelerate innovation, and provide a foundation they can rely on from one generation to the next."
Nvidia also released Nemotron 3.5 Lightning on Tuesday, an addition to its offerings aimed at code review, tool use, security alert monitoring, answering billing questions, and other tasks. The company released NeMo Switchyard, an open-source model-routing library designed to automatically direct AI tasks to the most suitable models.
Why this matters for IT and development professionals
The arrival of a 1-trillion-parameter open model from Nvidia would give development teams a serious alternative to proprietary systems for tasks like code review and security monitoring. The release of NeMo Switchyard also signals a shift toward practical infrastructure: instead of betting on a single model, teams can route different tasks to whichever model performs best, which is how many production AI systems are likely to operate. For professionals evaluating Generative AI and LLM tools, the practical question is whether open weights can match closed models on reliability and security - and Nvidia is betting they can. As the open-versus-closed debate plays out, developers who understand both sides will be better positioned to make deployment decisions for their own organizations, a skill set increasingly relevant to AI for IT & Development roles.
Your membership also unlocks:
