Tesla Is Developing Their Own AI Chips (With a Little Help From AMD)

Remember Jim Keller?, he worked for AMD on the K8, K12 and ZEN microarchitecture and was responsible for the x86-64 instruction set and HyperTransport interconnect. Well, last year he jumped ship and joined Tesla as Vice President of Autopilot Hardware Engineering.

The fruits of his knowledge now surface, Elon Musk confirmed at the NIPS conference that Tesla is developing their own chips for artificial intelligence and, among other things, chip architect Jim Keller is working on this. 

Elon Musk stated "I want to make it clear that Tesla is serious about artificial intelligence, such as software and hardware, and we are developing our own AI chips," said the Tesla CEO. There he added that Jim Keller is working on the ai-hardware. It's interesting as Musk has been warning about the risks of artificial intelligence for a long time now. In September there were already rumors that Tesla would work on their own AI chips. The company would work with AMD on this. According to rumors, a team of fifty people led by Jim Keller is working on it. Musk did not say when the chips are expected to be ready.

Tesla is up against Nvidia chips, who is very fully committed to creating processors that can be used in autonomous cars (PX series). Such vehicles collect a huge amount of data via their sensors, which must be processed and analyzed swiftly.
 

 
Tesla already has over 1.3 billion miles of road data accumulated through its first generation Autopilot program and that is useful for its second generation Autopilot, but the Autopilot team still also needs some direct data feeding Tesla Vision’s neural net with what the new hardware can see.