The Xavier has over 9 billion transistors with a custom 8-core CPU, a 512-core Volta GPU, an 8K HDR video processor, a deep-learning accelerator and new computer-vision accelerators. NVIDIA says the SoC can perform 30 trillion operations per second using only 30 watts of power. NVIDIA says that's 15 times more efficient than the previous architecture. Important for EVs where everything that pulls from the battery reduces vehicle range.
All those huge numbers mean that the Xavier can crunch more sensor and vehicle data for the AI systems that will power self-driving vehicles. That should make OEMs and ride-hailing companies working on autonomous cars very happy. After all, artificial intelligence is already processor intensive. Add the non-stop stream of data coming from sensors and cars are maxing out the capabilities of current computers.
Two Xavier processors will sit on the recently announced Pegasus AI computing platform (which is capable of level 5 autonomy, according to NVIDIA). The setup will also include two NVIDIA GPUs. Shove the whole system into a car and we're that much closer to truly self-driving vehicles.
Click here to catch up on the latest news from CES 2018.