Tesla's AI5 with 2nm-class node tapes out at Samsung Foundry — production starts soon, months after TSMC tape out

Tesla's AI5 chip is set to enter mass production using Samsung Foundry's 2nm-class process technology, according to a principal engineer at the company. The chip has recently completed tape-out and wi... Read original →

Tesla’s AI5 tape-out on Samsung’s 2nm-class node signals automakers’ direct entry into advanced-node competition. This move pressures SK hynix and others to accelerate GDDR7 adoption in automotive systems, pulling high-bandwidth memory roadmaps forward. Geopolitically, dual-sourcing mitigates overreliance on TSMC (Taiwan, China), yet Samsung’s Taylor fab remains in yield ramp-up, risking cost inflation. Competitors like NVIDIA may respond by opening more customizable ADAS chip interfaces, while Mobileye could fast-track 4nm automotive certification. Within 18 months, in-house AI chips will shift from performance showcases to supply-chain leverage tools, with sub-2nm nodes emerging as a new barrier—forcing Tier 1s to restructure foundry partnerships.