Nvidia Releases New Open Model

Good morning. In the great “open” versus “closed” debate sweeping Silicon Valley, Nvidia has emerged as among the strongest supporters of open artificial intelligence. The chip giant continued down that path on Tuesday, with the release of a new open model called Nemotron 3.5 Lightning.

Kari Briski, Nvidia’s vice president of generative AI software for enterprise, told the Wall Street Journal Leadership Institute that Lightning contains the knowledge of more powerful models, but is packaged into a much smaller size.

Nvidia also announced Nemo Switchyard, what it calls an “open-source model routing library”—essentially a tool that can automatically select models for supporting different AI agent tasks.

Both are key in helping enterprises, including Nvidia, save on AI token costs and protect their proprietary data, Briski said.

Why? Smaller models like Lightning are generally cheaper to run than larger, more powerful ones, and model routers know when to choose smaller models for simpler tasks. “You don’t always have to send those routine tasks to the bigger model, but the router will continuously balance quality, latency, and cost,” Briski said.

Then there’s the need to keep corporate intellectual property safe. For many companies, especially those in specialty domains like cybersecurity and material science, there’s the concern that AI providers could become competitors, Briski said.

“They really need to think about, ‘Who are they sending their IP to?’ Because they’ll just end up taking on their market,” she said.

Open models are customizable, allowing companies to self-train those models with their own data.

For Nvidia, those same reasons hold true. The chip company initially built Switchyard to address its own need to keep AI token cost under control, and it wanted to control its own IP, Briski said.

“We are now able to demonstrate the cost savings in both efficiency and accuracy. That’s something I think a lot of people are really keen on right now with the amount of tokens being spent,” she added.

Nvidia’s announcement comes a day after Meta Platforms released its latest open model, Muse Glimmer, which the company touted as “small enough to run on a Mac or PC with a single consumer GPU.”

This September 14–15, technology leaders will gather in New York City for the WSJ Technology Council Summit to explore how enterprise AI is moving from experimentation to measurable business value. Join the Technology Council and be part of the conversations shaping the future of leadership, as executives tackle AI deployment, cybersecurity, evolving technology policy, enterprise transformation and the strategies driving the next generation of business innovation.

Request an Invitation

Follow Isabelle Bousquette on LinkedIn, Instagram, X, and TikTok for more behind the scenes on her tech and AI coverage, and lately, her contributions to the WSJ Leadership Institute’s new Executive Resilience series, where she’s profiling America’s top execs about their fitness and wellness habits.

Follow Belle Lin on LinkedIn and X for her latest reporting on enterprise technology and AI.

Steven Rosenbush is chief of the enterprise technology bureau at the WSJ Leadership Institute. He also has a column. You can follow him on LinkedIn.

Tom Loftus is the editor of The Morning Download. He suggests following Isabelle, Belle and Steve on their various social channels. But if you insist, here’s his LinkedIn.