Tesla's new A15 chip, with 40x improvement, boosts AI operations and memory for autonomous driving. Learn about production plans and future AI ecosystem.
© RusPhotoBank
Tesla has taken a significant step in advancing its own artificial intelligence technology. Elon Musk recently unveiled the A15 chip for the first time, confirming it is already operational, and shared plans for the next generation of solutions.
The A15 chip features a large central computing matrix and 12 memory modules along the edges. These components are designed to deliver high bandwidth and data processing capabilities. The memory used is supplied by SK hynix, underscoring Tesla's commitment to top-tier components.
This new chip will serve as an evolution of the HW4 platform and will be integrated into next-generation autonomous driving systems. According to Musk, the A15 offers up to a 40-fold improvement over its predecessor, with substantial gains in computational power and memory capacity. It is expected to handle up to 2,500 AI operations and feature approximately 144 GB of memory.
Production of the A15 is set to be handled by TSMC and Samsung, with mass production anticipated by late 2026 or early 2027. Looking ahead, Tesla is also considering shifting production to its own TeraFab facilities in the future.
In parallel, the company is already developing the next-generation A16 chip and advancing the Dojo3 supercomputer. This effort is part of Tesla's long-term strategy to build a comprehensive AI ecosystem—spanning from chips to data centers, all integrated into a unified infrastructure.