Qualcomm and Nvidia are unequivocally two shining stars in the field of future automotive technology. They have left all the traditional carmakers behind in the development of smart car technologies. Given the current mercurial competition in the automotive industry, one of the right directions for carmakers is to join the platforms of Qualcomm and/or Nvidia for the development of their own EV models because they are unlikely to keep up with these two companies by drawing on their own R&D resources alone. First, let's look at how the two tech colossuses have been developing smart car platforms…
- Qualcomm Snapdragon has various digital open application-ready platforms for networking, intelligent cockpits, ADAS, and even fully autonomous driving, integrating multiple CPUs into an SoC (a system on a chip) on an ARM architecture to collaborate with carmakers on the development of future automotive technology. The iterative and incremental development in recent years has led Qualcomm Snapdragon to attract more than 20 big traditional carmakers such as Volkswagen, BMW, Hyundai, BYD, Stellantis, Renault, Honda, Volvo, Ferrari, and Mercedes-Benz to join the partnership, with R&D capital claiming to total as huge as US$30 billion.
- Nvidia started with image recognition technology and is now feeling at home in the domain of self-driving technology. Orin and Atlan AI self-driving chips are world-renowned, and the recently unveiled DRIVE Thor reached a peak of 2,000 TOPS, eight times the computing power of Orin, in just three years, allowing Nvidia to stay far ahead of all tech companies worldwide. The Orin chip has been widely used in new car models of many carmakers around the world, including Mercedes-Benz, Volvo, Hyundai, Audi, BYD, Volkswagen, as well as new startups such as the NXL trifecta (NIO, XPeng, and Li Auto) and Lucid.
If we compare computing power, the Snapdragon Ride 5-nm single chip reaches 360 TOPS and is the current leader in mass production for use in vehicles, followed by the Orin 7-nm chip with 254 TOPS. Qualcomm is slightly ahead of Nvidia for the time being. Nvidia announced last year that the Atlan chip, which reaches 1,000 TOPS, would not be ready for mass production until 2025. This schedule would slow down Nvidia in the race of autonomous driving and intelligent cockpit technologies. In this September, therefore, Nvidia moved up the mass production date of its more powerful DRIVE Thor chip straight to 2025, which means that Atlan will no longer be instrumental in the future development of self-driving technology, and that DRIVE Thor will directly face the challenge of new chips from Qualcomm and other tech giants. The ADAS, automatic parking, DMS, digital dashboard, infotainment system and other functions in a car need to be controlled by different carputers at present, but they will all be effectively integrated to save costs when the 2,000 TOPS chip arrives. Judging from this trend, will Nvidia overtake Qualcomm to take the lead in self-driving technology? Not necessarily. BMW and Volkswagen have both signed long-term partnerships with Qualcomm for self-driving system development in the past year (Mercedes-Benz is partnering with Nvidia). Within the next three years, Qualcomm will surely launch a heavyweight self-driving AI chip to take on DRIVE Thor.
Does a chip with computing power reaching 2,000 TOPS mean that fully autonomous vehicles will be introduced soon? It is a question that needs to be answered by the global leaders in self-driving technology: Waymo, Baidu, Cruise, and AutoX. Waymo is currently working with Samsung; Baidu is using Orin while developing its own chip; Cruise has left the partnership with Nvidia and started developing its own chip; and AutoX is using both Qualcomm and Nvidia chips to build its computing platform. So, it seems that the two self-driving AI chipmakers have more room for cooperation with carmakers. Two camps will form, self-driving chip giants and self-driving tech giants, and compete in the application to vehicles on another level of business.