Taylor fab initiates test production for next-generation AI silicon ahead of full-scale commercial manufacturing in 2027.
Samsung’s advanced semiconductor manufacturing plant in Taylor, Texas, has commenced running prototypes of Tesla’s next-generation 2-nanometre (2nm) AI5 chip, according to reports from the Seoul Economic Daily. The specialized processor is designed to power Tesla’s Full Self-Driving (FSD) platforms, the Optimus humanoid robot, and the Cybercab robotaxi.
According to Business Korea, Samsung aims to complete final production verification for the AI5 architecture by the end of the year. The factory—which possesses an installed capacity of 50,000 wafers per month—is scheduled to begin initial shipments in early 2027, before ramping up to high-volume commercial production by mid-2027.
The AI5 processor delivers a major performance leap over its predecessor, the AI4. Industry specifications indicate the new chip will operate 40 times faster, providing eight times the raw computing performance, nine times the memory capacity, and a three-fold improvement in energy efficiency per watt.
Tesla remains the anchor client for the Texas facility following a $16.5 billion supply contract, which grants Chief Executive Elon Musk direct access to inspect the fab floor alongside a separate production co-operation agreement with Intel. Industry sources report that Samsung’s 2nm capacity at Taylor is already fully booked through orders from Tesla, Broadcom, and Arm. Looking further ahead, Samsung plans to begin construction on a second Taylor facility later this year, targeting mass production on a 1.4nm process by 2030.
