Nvidia lays out its own vision for the ‘omniverse’ at its AI conference

The tech giant unveiled its simulation engine this week, which aims to make it easy for businesses to create digital twins.
article cover


· 3 min read

Stay up to date on emerging tech

Drones, automation, AI, and more. The technologies that will shape the future of business, all in one newsletter.

For Nvidia, metaverse buzz is misplaced—it’s all about the omniverse.

The tech giant centered its annual AI conference, GTC, on the idea of “digital twins” for virtually everyone and everything. The term essentially refers to simulation software—in goes the real-world data, out come the performance predictions.

“A constant theme you’ll see [is] how Omniverse is used to simulate digital twins of warehouses, plants and factories, of physical and biological systems, the 5G edge, robots, self-driving cars, and even avatars,” CEO Jensen Huang said in his keynote.

Emerging Tech Brew caught up with Ian Buck, VP and GM of accelerated computing at Nvidia, about Omniverse—or, as he put it, a “simulation engine for...creating and connecting physically accurate virtual worlds.”

Let’s get omni

Here are a few applications Nvidia showed off for Omniverse...

Avatars: Nvidia’s most metaverse-y announcement? A way to use Omniverse to create interactive avatars, combining the company’s natural language processing, computer vision, and recommendation algorithms to create ready-made reproductions for a variety of conversations.

Climate change: The company will create a digital twin of Earth, dubbed “E2,” in Omniverse. In order to power the simulation, real-world climate-change data will be run through large-scale supercomputers using Modulus, Nvidia’s new AI framework that specializes in physics-related machine-learning models.

  • The company is also partnering with Lockheed Martin and forest services on the state and federal level, in order to roll out what it bills as the world’s first AI lab for wildfire prevention and response.

Synthetic data: The more AI models we build, and the bigger they get, the more training data they need. One of the many obstacles to this? Data privacy, especially if the model is healthcare-focused, or needs another type of sensitive data in order to learn.

  • To help circumvent that bottleneck, Nvidia announced Replicator, an Omniverse–based engine that can generate large amounts of synthetic data.

5G: The company is also teaming up with Ericsson, the Swedish telecom giant, to use radio-signal data to build digital twins of cities—and simulate how well 5G will cover different areas, especially around potential obstacles like vegetation and thick walls.

  • “We’re planning where to deploy our 5G infrastructure,” Buck told us.

Looking ahead

Over the next two years, Buck predicted, we’ll see companies apply this tech to their own vertical markets—especially ahead of the deployment of self-driving vehicles and autonomous robots.

“It really offers a platform for people to develop and invent autonomous vehicles, robots, for all sorts of different applications,” Buck said, adding that the technology offers a way to “get it right in the virtual world before they ever need to deploy it in the real world.”

Stay up to date on emerging tech

Drones, automation, AI, and more. The technologies that will shape the future of business, all in one newsletter.