Nvidia is repositioning itself as a software and services publisher, allowing it to regularly derive revenue from those who use its graphics processors.
“We are now entering a new phase, a new phase where we are thinking about software and a business model for software to be sold separately,” said Colette Kress, chief financial officer of Nvidia, at Morgan Stanley Technology, Media and Telecom. conference this week.
Things haven’t been easy on Nvidia’s hardware side, with semiconductor supply shortages, disgruntled gamers, and Arm’s failed acquisition. Oh well too bad. The software side will provide a more predictable revenue stream, Kress said.
Not only the hardware at the very beginning generates income, but the software throughout the life of car ownership
The CFO gave an example of how self-driving cars will generate recurring software revenue. Mercedes-Benz, from 2024, and Jaguar Land Rover, from 2025, will deliver autonomous vehicles with Nvidia computers and the DRIVE operating system.
“All of their fleets will be operated with Nvidia software, and we will have the ability to share that software with those OEMs and monetize it for the life of the cars on the road,” Kress said, adding that she expected the number of such cars to reach about 10 million.
Nvidia is working with Mercedes-Benz and Jaguar Land Rover to customize software for their vehicles. The GPU giant has been preaching the car-as-a-service model, in which automakers can activate car features, such as driver assistance, remotely through subscription services. Nvidia could get a cut of those services delivered through its on-board hardware.
“We have the ability for the material to exist for the lifetime inside the car [and] the software should be addressed every time with over-the-air updates. Not only does the hardware in the very beginning generate revenue, but the software over the lifetime of car ownership will,” Kress said.
This is just the beginning of these types of opportunities, Kress said, adding, “The pipelines are growing for this and we’ll be looking to provide metrics when the software is relevant.”
Nvidia also expects software revenue from the emerging metaverse opportunity, the plumbing for which it provides in the form of its graphics cards. The biz ultimately wants to tap into a market of millions of makers who will need tools to create avatars as well as the 3D worlds in which they will collaborate and communicate, Kress said. The chip titan already offers Omniverse, which is a metaverse creation suite where people can create interactive virtual universes for work and play. Kress wasn’t shy about putting a potential dollar value on this opportunity.
“So for the 40 million different creations that are out there, we can add about $1,000 per person per year… That’s a big piece,” she said.
Omniverse can also be used by engineers and manufacturers to create and simulate products and machines in collaborative virtual spaces. “We were able to do this by bringing to market both ray tracing, which provides real-time visual effects, but also bringing AI together and now simulating all of this with just the physics of the environments they create,” Kress mentioned.
To that end, Nvidia is working on the materials needed to design and simulate robots and cars in virtual reality workshops. It also allows creating interactive bots to respond to people’s queries in immersive 3D environments.
Creating this bot, creating this digital twin, could cost around $1,000 a year for each of these
“It could be used for service centers, call centers, because they use these bots to have a more interactive discussion with our customers. To create this bot, to create this digital twin, could cost around 1,000 $ per year for each of these robots.” spat Kress.
Nvidia CEO Jensen Huang is expected to make further announcements about Omniverse and related software offerings at the company’s upcoming GPU technology conference on March 22.
Nv is a complete company with tightly integrated hardware and software. US giant CUDA’s parallel computing suite isn’t exactly open to competitors, and Huang reiterated that he won’t open it, calling it a centerpiece of his software strategy. It is, however, theoretically open to third parties designing chips suitable for CUDA.
Nvidia is known for its software: from AI framework code to apps to development kits. You buy its hardware, the software comes with it. “We now have the ability to monetize that separately,” Kress said. ®