Image: Tesla
Elon Musk has revealed new details about Tesla’s upcoming AI5 and AI6 chips, claiming they could set a new benchmark for automotive and even broader AI computing performance.
“Just had a great design review today with the Tesla AI5 chip design team! This is going to be an epic chip,” Musk said on X. “And AI6 to follow has a shot at being the best AI chip by far.”
Musk explained that Tesla has consolidated its silicon roadmap, moving from two chip architectures down to one. “Switching from doing 2 chip architectures to 1 means all our silicon talent is focused on making 1 incredible chip. No-brainer in retrospect,” he added.
Earlier this summer, Tesla officially scrapped its Dojo supercomputer platform in favor of focusing all of its silicon resources on AI5 and AI6. Musk has said these chips will serve as a “convergence architecture,” powering both inference in cars and data centers, as well as model training. According to the Tesla CEO, AI5 will be “the best inference chip of any kind” for models below ~250 billion parameters, delivering unmatched cost efficiency and performance per watt. AI6 will take that even further.
Currently, Tesla’s vehicles use the AI4 chip, a custom SoC manufactured by Samsung, as the backbone of its Full Self-Driving (Supervised) software. AI5, which just completed its design phase, will be manufactured by TSMC, first in Taiwan and later at its Arizona fab. Mass production is expected to begin by late 2026.
Tesla has already lined up its next step with a massive $16.5 billion deal with Samsung for AI6 chip production starting in 2026. The South Korean tech giant is preparing its Taylor, Texas, fab to meet Tesla’s needs, with Musk himself planning to personally visit the facility to “accelerate the pace of progress.”
With the AI5 design now locked in, Tesla’s in-house silicon team — primarily based in Austin, Texas — is pushing forward on what could become one of the most important chip platforms for AI-driven vehicles and beyond.
