Tesla's new Full Self Driving (FSD) chip was unveiled at the company's Tesla Autonomy Day investor event. Manufactured by Samsung in Austin, TX, the custom chip, says the company, was built with autonomy and safety in mind and is currently shipping in new Model S, Model X, and Model 3 vehicles.
The 260-mm 2 chip features twin neural network arrays each capable of 36 trillion operations per second (TOPS) and has a power consumption of 72 W. Two of the chips will be installed on each of the company's FSD computer boards, delivering 144 TOPS for collecting data from a range of surround cameras, radars, and ultrasonics, and to power deep neural network algorithms.
With these chips, claims the company, all of its cars will have the hardware they need for "full self-driving" capability. For that, says Tesla founder and CEO Elon Musk, "All you need to do is improve the software."
"How could it be," says Musk, "that Tesla, who has never designed a chip before, would design the best chip in the world? But that is objectively what has occurred. Not best by a small margin, best by a big margin."
In a blog post responding to Tesla's claims, Nvidia notes several "inaccuracies" in Tesla’s Autonomy Day presentation.
"It's not useful to compare the performance of Tesla's two-chip Full Self Driving computer against Nvidia's single-chip [Xavier] driver assistance system," says Rob Csongor, VP and GM, Automotive at Nvidia. "Tesla's two-chip FSD computer at 144 TOPs would compare against the Nvidia DRIVE AGX Pegasus computer which runs at 320 TOPS for AI perception, localization, and path planning."
In addition, says Nvidia's Csongor, "while Xavier delivers 30 TOPS of processing, Tesla erroneously stated that it delivers 21 TOPS. Moreover, a system with a single Xavier processor is designed for assisted driving AutoPilot features - not full self-driving. Self-driving, as Tesla asserts, requires a good deal more compute."
The two companies,