Nvidia’s Nemotron 3.5 Just Launched, But A Bigger 1 Trillion Parameter Model Is Reportedly Already In The Works

• The Information said the model speculated to be under development could put Nvidia in more direct competition with customers such as OpenAI and Microsoft.
• It added that the push is towards Nvidia building an ecosystem of open models optimized for its GPUs to support broader demand for its AI infrastructure.
• Nvidia's cloud-compute budget for Nemotron development is capped at $7 billion through fiscal 2028, according to The Information.

Nvidia (NVDA) on Tuesday released its Nemotron 3.5 Lightning AI model while reportedly developing a much larger Nemotron 4 model that could put the chipmaker in more direct competition with some of its biggest customers.

According to a report by The Information, the flagship Nemotron 4 model is expected to exceed 1 trillion parameters, roughly twice the size of Nvidia's current largest model, Nemotron 3 Ultra, which was released in June. In comparison, the Nemotron 3.5 Lightning has around 30 billion parameters.

NVDA stock edged 0.6% in morning trade amid a muted broader market. The shares also saw a slew of bullish price hikes from Wall Street following the announcement of a $500 billion financing deal with six top financiers to fund artificial intelligence (AI) infrastructure.

What Happens With OpenAI, Microsoft, SpaceX And Others?

The development would mark a deeper push into AI models for Nvidia, whose chips power many of the frontier models being developed by customers including OpenAI, Microsoft (MSFT) and SpaceX (SPCX). 

Some of those companies are also developing their own AI chips, creating a potential conflict for Nvidia if its models compete directly with their offerings. According to a report by the South China Morning Post, some of China’s top AI models are still being trained on Nvidia chips, despite Beijing’s push to go local.

The Information’s report said Nvidia has not set a release date for Nemotron 4 yet, though two people familiar with the project said it could arrive as soon as this fall.

Nvidia Releases Nemotron 3.5 Lightning

Nvidia's Nemotron 3.5 Lightning is a smaller model designed to run AI agents efficiently. The company said the model can deliver up to four times faster output and complete agentic tasks 30% faster than comparable models, while allowing companies to customize it with their own data, tools and workflows.

The company also released free model-routing software that allows businesses to direct tasks to different models based on their cost and capabilities.

Nemotron 4 Could Put Nvidia Closer To Its Customers

According to The Information, Nvidia's push into open models is partly aimed at creating a broader selection of high-quality models optimized for its hardware, which could help drive demand for Nvidia's GPUs even if the company ends up competing with some of the customers buying those chips.

It added that tension is particularly notable with OpenAI, one of the companies that Nvidia is heavily invested in. The company has thus far put in $30 billion towards the startup, alongside an infrastructure partnership announced in September 2025 that could involve up to $100 billion to deploy at least 10 gigawatts of Nvidia systems for OpenAI's next-generation infrastructure.

Nvidia has also backed other open-source AI companies including Reflection AI and Thinking Machines Lab. The company has also organized a "Nemotron Coalition" that includes Mistral, Cursor, Cognition and Prime Intellect.

The Information reported that more than 570 authors contributed to Nvidia's most recent major model paper and that even more employees are expected to work on Nemotron 4. It also added that Nvidia's cloud-compute budget for Nemotron development is capped at $7 billion through fiscal 2028.

Read also: IBM Lands $240M Together AI Deal To Build Nvidia-Powered Neocloud Cluster

For updates and corrections, email newsroom[at]stocktwits[dot]com.