Elon Musk stated on Wednesday evening that Tesla and SpaceX AI (xAI) plan to continue their substantial procurement of Nvidia chips. This move highlights Musk's demand for powerful computing capabilities in advancing the development of autonomous vehicles, humanoid robots, and AI systems.
Notably, this is the first time Musk has publicly mentioned "SpaceX AI" since SpaceX's acquisition of xAI last month through an all-stock transaction.
The increase in chip demand is closely related to Tesla's next-generation self-developed processor, "AI5," which is designed to support its autonomous driving projects and is intended to run the company's proprietary systems, including the "Full Self-Driving" software.
Musk elaborated on the X platform, stating that the AI5 processor will not only be used for training tasks in data centers but is primarily aimed at edge AI applications for the Optimus robot and Robotaxi. He also revealed that Tesla expects to widely release an update for "Full Self-Driving" (supervised mode) in the coming weeks. Prior to this, he mentioned last Saturday that Tesla's chip manufacturing project, "Terafab," would launch within seven days.
Musk Pushes Tesla to Secure More Chip Supply to Support Autonomous Driving and Robotics Development
Musk's latest remarks align with his comments last year at Tesla's annual shareholder meeting regarding chip supply challenges. At that time, he indicated that Tesla might collaborate with Intel to seek more manufacturing solutions.
He noted, "We might work with Intel. We haven't signed any agreements yet, but discussions with Intel are valuable."
Although no collaboration agreements have been announced, Musk's intentions are clear: Tesla is actively exploring all possible avenues to secure more chip supply.
Meanwhile, Musk also revealed that Tesla is collaborating with Taiwan's TSMC and South Korea's Samsung. Nevertheless, he still believes the company may struggle to obtain sufficient chip production.
During last year's shareholder meeting, Musk stated, "Even if we forecast chip production based on the best-case scenario from suppliers, it still won't be enough." He then outlined the next steps: "I think we may have to build Tesla's Terafab. It's much larger than a Giga factory. I can't find any other way to achieve the chip production we need. So, I think we will likely have to build a massive chip factory. This must be done."
While the AI5 chip is primarily designed for vehicles, Musk also connects it to the Optimus robot and Robotaxi projects.
Nvidia Expands Automotive Business Landscape, Facing Challenges in China's Electric Vehicle Market
Nvidia has launched the Rubin platform, which includes six new chips aimed at powering AI supercomputers. Additionally, the company has released more efficient Inference Context Memory Storage and Spectrum-X Ethernet photonic switches.
At the same time, Nvidia is actively collaborating with Tesla's competitors, such as China's BYD and Geely, to expand its business scope.
Jensen Huang stated on Monday at the GPU Technology Conference in San Jose, California, "The 'ChatGPT moment' for autonomous vehicles has arrived."
These two Chinese automakers will adopt Nvidia's Drive Hyperion platform, which integrates chips and computing modules.