Tesla Inc (NASDAQ:TSLA), it seems, is not satisfied with just making cars and batteries, it wants to bring more manufacturing in-house. The company now reportedly plans to make the Tesla AI chip optimized for self-driving applications.
Tesla “serious” about AI
On Thursday, at the machine learning conference NIPS, CEO Elon Musk was heard talking about the Tesla AI chip, according to a report from The Register.
At the event, Musk told attendees that the company is “developing specialized AI hardware that we think will be the best in the world.”
Relying On Old-Fashioned Stock Picking, Lee Ainslie Reports His “Strongest Quarter” Ever
Lee Ainslie's Maverick Fund USA enjoyed its "strongest quarter in the fund's history" during the three months to the end of June. According to a copy of the firm's second-quarter letter to investors, which ValueWalk has been able to review, Maverick Fund USA gained 18% in the second quarter. Following this performance, the fund was Read More
Further, the CEO said that they are serious about AI, on both the hardware and software side.
“We are developing custom AI hardware chips,” Musk said, according to The Register.
In September, a report from CNBC claimed that Tesla already had over 50 employees working on the project, including names like Jim Keller, who joined the EV firm in January 2016 as Vice President of Autopilot Hardware Engineering. At the time, CNBC reported that Tesla is running tests on the samples it received of the first implementation.
Prior to Tesla, Keller has also worked at Apple and AMD. Keller, who joined Apple in 2008 after the iPhone maker acquired the Palo Alto Semiconductor, played a major role in Apple’s A4 and A5 iPhone chips, and at AMD, he worked on the Athlon K7 and K8. According to CNBC, other AMD veterans working with Tesla are principal hardware engineer Bill McGee, system circuit design lead Dan Bailey, and director Ganesh Venkataramanan.
What’s the need for the Tesla AI chip?
The Tesla AI chip is part of Tesla’s strategy to lower its dependence on other companies. But, the company is not doing it solely by itself, rather its chip would be based on AMD’s tech, CNBC said. Presently, Nvidia graphics processing units are used to power Tesla Autopilot self-driving hardware. Previously, the EV maker was using chips from Mobileye. Nvidia GPUs are widely used for AI projects, but they are not tailored for a specific use. Hence, Tesla might have felt the need to develop a custom chip for self-driving.
The Tesla AI chip would push the company’s effort to deliver fully autonomous driving. Musk has already promised to make it available to the consumers by 2019. Making custom chips is becoming a trend nowadays as companies realize the importance of artificial intelligence. Apple and Huawei are already using custom chips in their smartphones, while Google is using them to power its cloud AI services. Even drones and consumer cameras are using custom vision chips, notes The Verge.
At the event, Musk also talked about the future of driverless cars and AI. The Tesla CEO said that about “half of new cars built ten years from now will be autonomous.”
Further, he said that AGI (artificial general intelligence) would arrive in seven to eight years.
Musk, who on several occasions has raised a voice against the dangers of over-dependence on AI, said, “But I think there are a lot of ways that AI can be useful short of being god-like.”
In pre-market trading today, Tesla shares were up almost 1%.