Tesla's official Robotaxi account recently announced that Robotaxi service in Austin now runs until 11 PM, extending its previous operating hours by one hour.
As we previously reported, Tesla Robotaxi could begin operating around the clock as early as October. Tesla's head of AI, Ashok Elluswamy, previously indicated that 24/7 Robotaxi operations could become possible once the next technical milestone on the FSD v15 roadmap is successfully integrated.
However, Tesla still needs to overcome several technical challenges related to nighttime driving before achieving that goal.
Elon Musk recently highlighted one particularly difficult scenario: detecting small animals in low-light conditions, such as a gray kitten lying on gray pavement at night.
Musk's Nighttime Driving Challenge: Step on no pets
Musk explained that one of Tesla's current priorities is preventing Robotaxis from hitting pets that are difficult to detect at night. He specifically mentioned a gray kitten blending into gray pavement, highlighting the challenges autonomous driving systems face when identifying objects in low-light conditions.
In darkness, cameras capture fewer visual details, and the contrast between objects and their surroundings can become extremely low. For Tesla FSD, the challenge goes beyond simply capturing visual information. The system must identify objects from limited image data, predict their movements, and make safe driving decisions in time.
Different sensors can approach this problem in different ways. LiDAR uses laser pulses to measure the distance and spatial structure of surrounding objects, making it less dependent on differences in color and brightness. Thermal cameras, meanwhile, can detect temperature differences between animals and their surroundings, potentially making them easier to spot in darkness. However, both technologies have limitations, and neither can guarantee the detection of small animals in every situation.
Tesla continues to pursue its vision-based approach, focusing on improving software and AI perception rather than adding more types of sensors. Musk has long argued that autonomous driving systems should work much like human drivers, relying primarily on visual information to understand their surroundings. Neural networks then handle object recognition, behavior prediction, and driving decisions.
FSD v15 Could Feature a 10x Larger Neural Network
One of the most anticipated improvements in FSD v15 is a significant expansion of its neural network.
According to reports, the model's parameter count could increase from approximately 1 billion to 10 billion, making it 10 times larger than the current version.
A larger model could improve the system's ability to handle complex driving scenarios, including challenging nighttime conditions.
Some Tesla Robotaxi vehicles are already reportedly testing early versions of the software on public roads. Tesla plans to deploy FSD v15 to its Robotaxi fleet first, followed by a gradual rollout to consumer vehicles.
Austin's Robotaxi service recently extended its operating hours by one hour, to 11 PM, bringing Tesla another step closer to its goal of 24/7 commercial operations.
Musk's focus on protecting pets at night also adds a human dimension to this technical challenge. Accurately detecting low-contrast objects in dim lighting is not just about expanding Robotaxi's operating capabilities. It's also about protecting every life on the road, no matter how small.
Frequently Asked Questions
Q1：What time does Tesla Robotaxi operate in Austin?
As of the October 2026 announcement, Tesla Robotaxi operates in Austin from 6 AM to 11 PM, seven days a week.
Q2：Why is nighttime driving difficult for Tesla FSD?
Low-light environments can make small, low-contrast objects harder for cameras to detect. Autonomous driving systems must also recognize potential hazards and respond with enough time to avoid collisions.
Q3：Will FSD v15 improve nighttime driving?
FSD v15 is expected to introduce significant neural network and perception improvements.