Tesla has extended its Austin Robotaxi service by an hour, but the modest change underscores a larger technical hurdle. Elon Musk says detecting small, low-contrast animals after dark remains a safety concern, putting renewed focus on Tesla’s decision to rely primarily on cameras rather than lidar and radar.
Tesla’s effort to keep its Robotaxi service running later into the night is being held back by a deceptively difficult problem: seeing small animals that blend into the road.
Elon Musk, Tesla’s chief executive, said the company was working to ensure its autonomous vehicles could reliably identify pets in poor lighting. Responding on X to an announcement about the service’s latest schedule change, Musk described the challenge as avoiding “grey kittens on grey tarmac in the dark”.
The Austin pilot has now been extended to 11 PM, one hour later than its previous 10 PM cut-off. That remains a narrower operating window than the one offered when the Robotaxi service began in June 2025, when vehicles were permitted to operate from 6am until midnight.
The difficulty Musk identified goes beyond cats or other domestic animals. It illustrates the broader challenge faced by systems that depend on cameras to interpret their surroundings when illumination and visual contrast deteriorate.
Why darkness matters to Tesla’s approach
Cameras can provide rich information about road markings, colours, signs and objects, but they need enough light to distinguish one feature from another. An animal whose colour closely resembles the road can therefore become difficult to separate from the background, particularly at night.
More from TechTesla has made cameras the centre of its autonomous-driving strategy, rather than combining them with lidar and radar as many rival autonomous-vehicle developers do. Lidar actively measures the environment with laser pulses, producing detailed depth information, while radar uses radio waves to determine the position and movement of objects. Both technologies can provide information that does not depend on visible light in the same way as a conventional camera.
That difference has become particularly relevant as Tesla seeks to expand Robotaxi operations beyond limited pilot conditions.
Musk has long rejected lidar as unnecessary for Tesla’s approach. In 2019, he dismissed the technology as a “fool’s errand”, while Tesla subsequently removed radar from its production vehicles in 2021 and ultrasonic sensors in 2022. The company has continued to argue that a vision-based system can ultimately achieve the level of performance required for autonomous driving.
The approach contrasts with that of Waymo, Alphabet’s autonomous-driving subsidiary, whose driverless vehicles use cameras alongside lidar and radar. Waymo operates its service around the clock, giving the comparison particular relevance as Tesla attempts to lengthen its own operating hours.
Waymo and the wider safety question
Waymo co-chief executive Dmitri Dolgov highlighted the value of multiple sensors during a presentation at Y Combinator’s Startup School in August. Footage shown during the session included children and dogs moving across an unlit road. According to Dolgov’s demonstration, the camera view became difficult to interpret while the lidar data continued to register the subjects.
The comparison does not, by itself, establish that lidar is essential for safe autonomous driving. Tesla’s system and Waymo’s technology have been developed around different engineering philosophies, and operating conditions also vary. But the episode illustrates why sensor choice remains a central point of disagreement within the industry.
Tesla’s approach is also attracting regulatory attention. The US National Highway Traffic Safety Administration is investigating about 3.2 million Tesla vehicles following collisions involving Full Self-Driving in low-visibility conditions. Investigators have sought internal company material as part of the inquiry, including a document titled “Radar Saves Us”.
The document has drawn attention because Musk reportedly acknowledged in 2021 that combining vision with high-resolution radar could outperform vision alone. The investigation remains separate from Musk’s latest comments about Robotaxi operating hours, and its findings have not established that Tesla’s camera-based strategy caused the reported incidents.
