It's been announced that Tesla, increasingly focusing on self-driving AI technology, has partnered with AMD to develop a dedicated processor that's specifically created to make artificial intelligence systems within Tesla's self-driving cars run with more efficiency.
A more power-efficient purpose-built chip could help Tesla get closer to delivering totally autonomous driving. He views the new chip as a move to address specific computing tasks for the vehicle while an Nvidia chip would still handle general goal computing. The chip is being manufactured by GlobalFoundries, and AMD spin-off, which has a wafer supply agreement in place with AMD through 2020.
Steves thinks the chip being developed in the Tesla-AMD partnership will focus on specific computing tasks, while the general processing will still be handled by Nvidia's chips. Tesla CEO Elon Musk promised this year that capability will be available to consumers in 2019.
In extending trade the shares of AMD were up to 2.2%.
On Wednesday, at GTC 2017, GlobalFoundries CEO Sanjay Jha cited the American electric carmaker as one of the companies who chose to work with fabricators, though he did not elaborate. Tesla and AMD did not immediately return Business Insider's request for comment.
Interestingly, Tesla's head of AI, Jim Keller, is an ex-AMD CPU ar- chitect having designed Zen (used in Ryzen and EPYC), and has surrounded himself at Tesla with several key AMD players, which may have been the initial reason Tesla is aligning themselves with AMD. Chris Lattner, Tesla's vice president of Autopilot software, left the company in June after just six months at the helm. As it seems now, they have partnered up with AMD to do so. Since Keller's hiring, Tesla has also brought on other big talents from AMD like director Ganesh Venkataramanan, Bill McGee, a principal hardware engineer, and Dan Bailey, a system circuit design lead.