Elon Musk has unveiled the latest advancements in Tesla’s in-house AI chip development. The AI5 chip is projected to undergo sample trial production and achieve small-scale deployment by 2026, with full-scale mass production slated to begin in 2027. This chip will maintain Tesla’s dual-foundry approach, with both TSMC and Samsung handling manufacturing duties. It is expected to deliver computational performance ranging from 2000 to 2500 TOPS (trillions of operations per second), roughly five times that of Tesla’s current HW4 chip.

Looking further ahead, the AI6 chip is planned for large-scale production in mid-2028, aiming to double the performance of the AI5. Manufacturing responsibilities for the AI6 will remain with TSMC and Samsung. As for the more distant AI7 chip, given the extensive R&D involved, other foundries are anticipated to join the manufacturing process at that time.