Tesla is currently at the forefront of developing cutting-edge chips tailored to advance its vision of autonomous driving. Musk announced that the engineering review process for the AI6 chip is proceeding seamlessly, with expectations to set an unprecedented benchmark for usable computing power per wafer. The AI6 chip’s application ecosystem is impressively broad, encompassing diverse sectors such as commercial autonomous driving taxi services, civilian passenger vehicle Full Self-Driving (FSD) software, Optimus humanoid robots, and space data centers.

Presently, the AI5 chip has successfully completed its tape-out phase and is slated for mass production in the latter half of 2027. This chip is projected to deliver computing power five times that of the combined capabilities of the current two AI4 chips. Building on the advancements of the AI5, the AI6 chip is anticipated to double its performance yet again, ushering in a comprehensive upgrade to the processor computing and memory management architecture. Production of the AI6 chip is scheduled to commence in the second half of 2028.

Beginning with the AI5, the new-generation hardware platform will feature enhanced memory capacity. Nearly half of the TRIP AI computing accelerators in both the AI6 and AI6.5 chips will be integrated with SRAM, while the main memory will leverage LPDDR6 technology. Tesla has forged a deep collaborative partnership with Samsung, with Samsung’s new facility in Texas undertaking the manufacturing of the AI6 chip as part of a deal valued at up to $16.5 billion.

Furthermore, Tesla is co-developing the TERAFAB project alongside Intel and SpaceX, with the ambitious goal of achieving autonomous integration across the entire semiconductor production supply chain. It is noteworthy that the new-generation AI chips will initially find application in Optimus robots and supercomputer clusters, rather than in vehicles. These advanced chips will subsequently be deployed in civilian passenger vehicles. Currently, the AI4 chip’s performance is deemed sufficient to support vehicles in attaining driving safety levels that surpass those achievable by human drivers.