Volvo expands NVIDIA partnership for its next generation of driving systems
Volvo and NVIDIA are expanding their partnership to combine vehicle safety engineering with shared sensing, computing and driving software. The technology is intended for a future Volvo, with the first vehicle expected within this decade. Existing electric cars already use NVIDIA chips. The companies describe broader automated-driving ambitions, including a gradual later move toward eyes-off features, without a deployment date for that stage.
Artificial Intelligence··Morning
An expanded partnership targets future vehicles
Volvo Cars and computing company NVIDIA expanded their automated-driving partnership on October 15. Volvo will combine its vehicle safety engineering with NVIDIA’s Hyperion sensing and computing architecture, Alpamayo driving software and Halos safety tools. The companies intend the system to support journeys across highways, urban streets and parking areas in future vehicles. Their announcement describes development goals for a next-generation Volvo.[1], [2]
Volvo software development head Alwin Bakkenes said the first vehicle would debut within this decade and was not arriving in the very near future. The companies have not named that model or disclosed the financial value of the arrangement. Bakkenes described moving from vehicle-line projects involving multiple suppliers toward a more integrated technology stack, intended for battery-electric, plug-in hybrid and gasoline vehicles. A specific on-sale or deployment date was not provided.[1]
Parallel software paths combine different sensing methods
NVIDIA automotive chief Ali Kani described two different software paths running simultaneously. One uses Alpamayo’s AI models, datasets, simulation and development tools. The other combines more conventional software with machine learning and camera and radar perception. He said the arrangement uses several sensing methods. His description of the safety benefits is a company claim about the planned approach.[1]
Hyperion’s general platform design combines cameras, radar, lidar, ultrasonic and cabin sensors with shared computing. That description does not specify the future Volvo’s final sensor arrangement. Halos covers vehicle architecture, system software, driving algorithms and safety evaluation. NVIDIA describes Alpamayo models producing a proposed driving trajectory and an explanation trace, which the company says can help with safety validation.[2]
Existing chips precede a longer automation roadmap
The partnership began in 2017. Volvo’s existing battery-electric vehicles already use NVIDIA Orin chips to process sensor information. Volvo and its software subsidiary Zenseact have also invested in NVIDIA systems for cloud model training. The expanded arrangement does not activate new capabilities in current cars. Volvo describes a gradual later introduction of eyes-off and mind-off driving, without stating when those features will arrive or that the first new vehicle will launch with them.[1], [2]
