Elon Musk Affirms Continued Nvidia Chip Purchases for Tesla and SpaceX AI
In a significant announcement, Elon Musk has confirmed that both SpaceX AI and Tesla plan to maintain large-scale purchases of Nvidia chips, even as the electric vehicle giant aggressively develops its own in-house AI silicon, known as AI5. This dual strategy highlights Musk's commitment to leveraging established technology while pushing forward with proprietary innovations.
Musk's Public Reference to SpaceX AI and Details on Tesla's AI5 Chip
In a post on X, Musk publicly referred to the combined entity of SpaceX and xAI as "SpaceX AI" for the first time, following last month's all-stock acquisition of his AI startup xAI by SpaceX. He elaborated on Tesla's AI5 chip, emphasizing its design for efficiency and edge computing applications.
"AI5 will punch far above its weight, because the entire Tesla AI software stack is designed to make maximally effective use of every circuit. We co-signed our AI software and hardware," Musk stated. He noted that while AI5 can be used for training in data centers, it is primarily optimized for AI edge compute in projects like Optimus and Robotaxi.
Musk also hinted at future advancements, adding, "There is still significant room for improvement. In the same half reticle and same process node, we think a single AI6 chip has the potential to match a dual SoC AI5."
Key Differences Between Nvidia and Tesla AI5 Chip Designs
The discussion was sparked by a post from former Apple and Rivian employee Phil Beisel, who explained the design distinctions between Nvidia's chips and Tesla's AI5. Beisel highlighted that Tesla's AI5 uses a half-reticle design, which enhances yield by allowing two chips per shot in lithography machines.
- Half-Reticle Design: Tesla's AI5 fits two chips per shot, effectively doubling yield and reducing fabrication requirements.
- Full-Reticle Design: Nvidia's Blackwell chip nearly fills a full reticle, making it a single-reticle design with different manufacturing implications.
Beisel further explained, "If Tesla hits its compute and efficiency targets with AI5 in this half-reticle format, it's almost like cutting fab requirements in half. And this has a big impact on Terafab, especially if it carries forward for AI6, AI7, etc."
Musk's Admiration for Nvidia and Ongoing Chip Orders
In response, Musk reiterated his admiration for Nvidia and its CEO Jensen Huang, stating, "I am a huge admirer of Nvidia and Jensen btw. That market cap is well-deserved. SpaceX AI and Tesla expect to continue ordering Nvidia chips at scale." This underscores the reliance of AI tech giants, including Musk's companies, on Nvidia's well-established chipmaking capabilities to meet processor demands while developing in-house solutions.
Imminent Launch of Tesla's Terafab Project
Adding to the developments, Musk announced on Saturday, March 14, that Tesla's Terafab project—a facility dedicated to manufacturing AI chips—is set to launch within seven days. If this timeline holds, Tesla's in-house chip production could be operational by Saturday, March 21, marking a critical step in reducing dependency on external suppliers and enhancing control over its AI infrastructure.
This move aligns with Tesla's broader strategy to integrate hardware and software seamlessly, driving innovation in autonomous vehicles and robotics. As the AI race intensifies, Musk's dual approach of partnering with industry leaders like Nvidia while advancing proprietary technology positions Tesla and SpaceX AI at the forefront of technological evolution.
