Well, it seems Tesla's plans to go all-in with its custom silicon have become much more aggressive, as Musk says the automaker intends to compete with NVIDIA's leading offerings.
Tesla's CEO Shows Intentions To Produce "Highest Volume" Chips in the World, By Providing Vehicle & DC Compute
Musk has been hyping up Tesla's chipmaking ambitions for quite some time now, especially after the approval of his mega-trillion-dollar pay package. The billionaire initially announced plans to develop a 'TeraFab', after saying that Samsung/TSMC cannot fulfill Tesla's incoming chip demands, and now, at a post on X, not only did Tesla's CEO show plans to scale up to four generations ahead of the firm's custom chips, but the Dojo3 supercomputer is now back on track as well. Musk also gave us insight into the AI5 silicon's performance, which we'll also discuss.
Tesla announced a few months ago that it was shutting down its Dojo supercomputer project, given the automaker's reliance on hardware from companies like NVIDIA. It was claimed that the decision came after key figures around the Dojo project had left Tesla, which ultimately forced the company to abandon it. However, according to a new announcement by Musk, Dojo3 is back in action, indicating that Tesla is in dire need of compute power, which is why it is doubling down on custom silicon ambitions.
If you are curious about what the Dojo3 supercomputer could look like, there is little information available, but based on Musk's post, we could expect an AI5-based cluster, given Tesla's intentions to have a unified silicon onboard for automobiles and Optimus. Tesla's CEO also mentions AI5 as the "highest volume" chip, implying that the company plans to make its chip and cluster configurations much more extensive than in previous projects.
Musk also commented on how the custom silicon lineup would evolve moving ahead, claiming that Tesla plans to scale up to AI9 chips, with a "nine-month" product cadence, which is quite similar to how NVIDIA operates. The reason why Musk has decided to go "all in" with chipmaking is that Tesla intends to churn out a cost structure advantage when the FSD tech stack becomes mainstream, and by controlling the chip supply chain itself, Tesla could fine-tune silicon to however the automaker intends to, giving them an edge over competitors.
Regarding AI5 performance, Tesla's CEO mentioned that the company is targeting Hopper-class performance with a single-chip configuration and that, with two dies onboard, the performance rivals Blackwell. Elon claims that the pricing of the AI5 chip is "peanuts", which again justifies the argument of how Tesla plans to come up with a cost-effective silicon ecosystem. Musk is known to be an 'optimistic' personality, but dealing with semiconductors is an art that takes "decades to master," if not less, given the several aspects to consider.
Could Tesla be the next major chipmaker? Well, Musk's comments indeed indicate that the automaker is working toward this milestone, but execution would be pivotal here, since Tesla would need to nail all stages of chip manufacturing, from design to verification and silicon stability.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
