Musk explains why Tesla Robotaxi isn't running at night, and Lidar is the answer

Elon Musk says Tesla’s Robotaxi curfew - recently extended from 10pm to 11pm in Austin - is driven by a specific perception problem: the cars struggle to detect low-contrast animals at night, like “grey kittens on grey tarmac.” Cameras are passive and rely on ambient light and contrast, so small dark objects can vanish in nighttime imagery, whereas lidar actively pulses lasers to generate reliable 3D returns and radar detects motion even in fog or glare. Other autonomous operators already combine all three sensors: Waymo’s current fleet uses dozens of cameras plus multiple lidars and radars and runs 24/7; its lidar imagery has shown clear detection of animals in pitch-black scenes. Tesla, by contrast, removed radar and ultrasonic sensors years ago and has never adopted lidar, even as Robotaxi deployment has barely scaled since its June 2025 Austin launch and consumer FSD continues to run at all hours.

That stance clashes with Musk’s long-standing public rejection of lidar as unnecessary, calling it a “fool’s errand,” yet regulators are investigating millions of Teslas for low-visibility crashes and internal documents reportedly argued radar “saves us.” Critics note the hardware limitation is well-known, lidar costs have fallen to a few hundred dollars per car, and Tesla’s own vehicles already operate in darkness before the curfew - suggesting the company may soon confront the trade-off it once dismissed and might need additional sensors or other fixes to restore full overnight operation.