AMD Expanding Into Tesla Model 3 and Model Y

One constant theme throughout AMD’s recent resurgence into high-performance computing has been the messaging around the scalability of its platform. Building a processor that can scale both from single digit watts all the way up to big water cooled compute servers is no easy task, but also combining multiple types of processors into a single chip to also scale just adds layers of difficulty. AMD were keen to point this out at its recent CES presentation, stating that the RDNA2 graphics architecture is immensely scalable, from mobile to notebook to desktop to server, but also through to embedded, industrial, and automotive. It’s that last part I asked CEO Dr. Lisa Su about.

Last year it was announced, and subsequently confirmed through model numbers, that the Tesla infotainment systems in the Model X and Model S are using AMD’s embedded platform to drive the display and graphics in those vehicles. Our understanding is that the first versions of that silicon in those vehicles are based on Zen plus Vega, so I asked Dr Su about what she meant by RDNA2 being in automotive solutions. Beyond that, I also asked about the AMD and Tesla relationship.

Dr Su reaffirmed that RDNA2 was ever-prevalent in the ecosystem, from consoles to PCs, but she also mentioned the Samsung [partnership] in the mobile space. She stated that Tesla is always pushing the envelope and that [AMD] appreciates that they’ve chosen Ryzen and Radeon in vehicles like the Model S and Model X. She went on to say that they’ve also started with the Model 3 and Model Y, adopting [AMD] technologies for their infotainment solutions. There was no explicit detailing about the depth of the relationship or the extent of the agreements between the two, but it seems clear that four of Tesla’s major vehicles using AMD are a sizeable win for the company.

From an outside perspective, it’s interesting just how, where, and which embedded technologies are used in different markets. We hear about so few (AMD plays big in gambling machines, for example) because of the nature of those markets and how accessible they are to the public. At one stage AMD showcased me around their showroom in the Santa Clara HQ that had a number of these implementations, even going back as far as the old G-series embedded silicon, given that the silicon has to be supported for 10-15 years. I wonder if AMD has updated that showroom – I’m going to have to go visit again soon.

*AMD after the interview with Dr Su clarified that Tesla using Ryzen embedded + Navi (RDNA2) in Model S and X. They just started shipping Model 3 and Y (higher volume vehicles) with Ryzen embedded.