On September 9, Ashok Elluswamy, Tesla’s head of AI, revealed during last week’s Cybercab launch event in Austin, Texas, that the rollout of FSD v14 Lite for Model S and Model X vehicles equipped with Hardware 3 (HW3/AI3) had originally been scheduled to begin a few days prior. Automotive journalist Ryan McCaffrey shared this information on X after speaking with Elluswamy in Texas. In June 2026, Tesla launched FSD v14 Lite, initially limited to early testers on high-volume models. In July, Tesla released an updated version of FSD v14 Lite with additional features. Following this, Tesla began publicly rolling out FSD v14 Lite, restricting deployment to Model 3 and Model Y vehicles equipped with HW3 hardware. Subsequently, Tesla initiated international rollouts in South Korea and confirmed development of FSD v14 Lite for Australia and New Zealand. FSD v14 Lite is built on the mainline v14 architecture and uses compressed model parameters to enable operation on older vehicle computers. The memory bandwidth of Tesla’s HW3 computer is approximately 15% that of the current HW4 chip. Tesla has confirmed that HW3-equipped vehicles cannot support unsupervised autonomous driving and will not be compatible with the upcoming FSD v15 update.