Skip to Content
Artificial intelligence

Tesla Is Building Its Own Custom AI Chips

December 8, 2017

At a party on the fringe of the year’s biggest AI conference, Elon Musk put an end to rumors about his automaker’s ambitions for AI hardware.

According to people invited to the event, Musk said he “wanted to make it clear that Tesla is serious about AI, both on the software and hardware fronts.” He added, “We are developing custom AI hardware chips.” He was joined on stage by Jim Keller, an engineer involved with developing silicon at AMD and Apple who joined Tesla last year, as well as Andrej Karpathy, Tesla’s director of AI, and Shivon Zilis, also at Tesla.

Rumors have swirled about the possibility of Tesla’s building its own machine-learning chips since we explained last year that the firm had split with its original AI hardware supplier, MobilEye, after an accident involving Tesla’s AutoPilot technology. Now we know it’s true.

The event was orchestrated to help Tesla recruit engineers during the Neural Information Processing Systems (NIPS) conference. As well as talking about AI hardware, Musk also discussed topics including automated driving, artificial general intelligence, and brain-computer interfaces.

Predictably, he made some bold claims. He suggested that artificial general intelligence, the grand challenge of the field, would be achieved in just five or 10 years. For all the academic progress that’s been made in the past few years, it’s hard to imagine that anyone involved in the field would take that claim seriously. He also made typically bold remarks about when autonomous driving would arrive. 

Unsurprisingly, the claims raised eyebrows on Twitter among some tech-savvy attendees.

Deep Dive

Artificial intelligence

Why Meta’s latest large language model survived only three days online

Galactica was supposed to help scientists. Instead, it mindlessly spat out biased and incorrect nonsense.

A bot that watched 70,000 hours of Minecraft could unlock AI’s next big thing

Online videos are a vast and untapped source of training data—and OpenAI says it has a new way to use it.

Google’s new AI can hear a snippet of song—and then keep on playing

The technique, called AudioLM, generates naturalistic sounds without the need for human annotation.

Responsible AI has a burnout problem

Companies say they want ethical AI. But those working in the field say that ambition comes at their expense.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.