Engineers building self-driving cars and artists creating virtual worlds are the target consumers for Nvidia’s CES 2022 news. The first half of the company’s virtual press conference was all about gaming—the people playing the games and the people building them. Jeff Fisher, senior vice president, GeForce, announced over 160 new GeForce RTX and Studio laptops and an RTX 3050 graphics card. Ali Kani, vice president and general manager, automotive, described the company’s automotive advancements in the second half.
Fisher shared all the stats driving the gaming market that’s worth almost $3 billion: esports, streaming channels, the popularity of watching online gaming content and world-building games such as Roblox.
Fisher said the company has made a massive investment in its software stack to attract gamers as well as the creators inventing the games.
The company’s cloud gaming service GForce was also central to the product news. NVIDIA announced a partnership with Samsung that will make the gaming platform available on smart televisions. The service is already available on LG 2021 Web OS Smart TVs.
Fisher also announced a partnership with AT&T to make GForce available on the telecom’s 5G network.
The GeForce RTX 3080 Ti and RTX 3070TI laptop GPUs will be available Feb. 1 in laptops from Alienware, ASUS, MSI and Razer with more following in the coming months. The new and updated GeForce laptops range from 14-inch ultraportables starting at $799 to 17-inch models.
Fisher also shared an update about the Max-Q design approach the company uses to build laptops. These updates include:
- A CPU optimizer to allow the transfer of more power to the GPU
- Rapid core scaling that can sense real-time demands of an application and use only the cores required at that time
- Battery boost 2.0 that uses artificial intelligence to improve power usage and increase battery life
Omniverse software for world builders
NVIDIA wants gamers and game makers as customers and is building software and hardware to attract those customers. The company announced that its Omniverse software will now be free for individual NVIDIA Studio creators using GeForce RTX and NVIDIA RTX GPUs. Artists and designers use the real-time 3D design platform to create 3D assets and scenes with rendering, physics and other tools. NVIDIA reports that the software has been downloaded by almost 100,000 designers since its open beta launch a year ago.
At the CES 2022 press conference, Fisher announced new platform developments including:
- Nucleus Cloud with “one-click-to-collaborate” simple sharing of large Omniverse 3D scenes.
- New support for the Omniverse ecosystem provided by leading 3D marketplaces and digital asset libraries such as Omniverse-ready assets for creators, all based on Universal Scene Description (USD) format from TurboSquid by Shutterstock, CGTrader, Sketchfab and Twinbru.
- Omniverse Machinima with new, free characters, objects and environments from leading game titles which allows creators to remix and recreate their own game cinematics with these assets by dragging and dropping them into their scenes.
- Omniverse Audio2Face, an AI-enabled app that instantly animates a 3D face with just an audio track, now offers blendshape support and direct export to Epic’s MetaHuman Creator app.
NVIDIA’s virtual assistant for your car
There was some crossover from the gaming portion of the press conference into the automotive segment. Kani showed off the talents of a virtual assistant built with NVIDIA’s Omniverse platform as he described the company’s DRIVE Concierge service. The virtual assistant is designed to work with the company’s DRIVE Chauffeur and uses real-time conversational AI to complete various tasks. In a demo video, the driver asks the assistant to open the sunroof, check messages and make dinner reservations. The Concierge mispronounces miso when describing the chef’s speciality at the recommended restaurant but is able to suggest a white wine to go with the meal.
Kani also described how customers are using the eighth generation of the company’s DRIVE Hyperion platform, which has been adopted by automakers, truck makers, Tier 1 automotive manufacturers and robotaxi companies.
“We have some partners that just buy our chips and core operating system while developing their own software applications while other partners rely on us across the entire stack from self-driving software to training AI models in the cloud,” Kani said.
DRIVE Hyperion is a reference architecture for the development, testing and validation of autonomous vehicles. The platform integrates a sensor suite that includes 12 exterior cameras, three interior cameras, nine radars, 12 ultrasonics and one front-facing lidar, plus one lidar for ground truth data collection. DRIVE Hyperion features the full software stack for autonomous driving as well as driver monitoring and visualization which can be updated over-the-air.
Kani said the company is working with many new electric car makers, including many in China.
“These electric cars will get better and better over time with each over-the-air update,” he said.
“Such companies can benefit from new business models that are software driven rather than selling a car and seeing it degrade once it’s driven out of the dealership.”
NVIDIA also is working with TuSimple, an autonomous trucking company that is developing a freight network in partnership with UPS, Navistar and Penske.
Kani said developing autonomous-vehicle technology is one of the biggest challenges for AI. NVIDIA is building Drive Sim Replicator to get around this difficulty. The platform uses synthetic data to help AV technology analyze and understand challenging environments, such as driving at night and navigating construction zones.
“With synthetic data, engineers can train AI models even before any data has been collected,” Kani said. “It’s accurate, low cost and fills in gaps for data not easily found in the real world.”