Recently, Tesla Korea has released job listings for AI chip design engineers, signaling an accelerated push for its in-house chip development plans. Tesla CEO Elon Musk stated on social platform X that the team is building the world's highest-volume AI chip architecture for the future, aiming for total production to surpass the current global scale of AI chips. Tesla is focusing on mass production readiness for its next-generation AI5 and AI6 chips, adopting an on-site R&D model with deep collaboration from Samsung and TSMC to enhance yield rates and efficiency in advanced manufacturing processes. As Robotaxi and Optimus evolve, computing power demands are surging, making in-house chip development capabilities a critical pillar. Tesla's recruitment approach reflects Musk's signature style, emphasizing practical experience and problem-solving skills over mere academic credentials. The choice to recruit chip designers in South Korea stems from the country's advanced manufacturing capacity, led by Samsung, and its status as a hub for HBM technology, enabling Tesla to participate in nearby prototype verification processes and shorten the gap between design and mass production. Tesla's in-house chips demonstrate significant performance improvements: the AI5's hardened block quantization and softmax functions are enhanced by 5x, memory capacity increased by 9x, raw computing power boosted by 10x, and overall efficiency improved by 50x. The AI5 and AI6 will respectively power the HW5 and HW6 platforms, serving as the computational cores for next-generation in-vehicle systems. Specifically, the AI5 will support transitional products and advanced driver-assistance features, while the AI6 targets full self-driving scenarios, providing foundational computing power for future Robotaxi and higher-level FSD applications. Tesla also plans to establish a complete chip supply chain in the U.S., including a PCB center in Texas and an FOPLP facility, to reduce reliance on external suppliers.
