$25K Tesla and unsupervised robotaxis are starting this quarter—and BYD, Xiaomi, and legacy auto won't keep up. With BYD's 4% margins a sham (thanks to 3% dodged interest and other non-GAAP accounting), Tesla's cheaper EV and 5,000+ robotaxi fleet by 2025 and one million Tesla robotaxi network in 2026 in will dominate. We break down the numbers, the supply chain, the tech, and the tariffs hitting China's carmakers hard.

Only 1% of LIDAR are robotaxi grade lidar. 5 are needed for every LIDAR using robotaxi. 1.5 million total LIDAR in 2024 are mostly inferior ADAS grade LIDAR.

3-5 Years to scale to 100,000 LIDAR base robotaxi if the LIDAR based robotaxi solution was solved first. Billions in investment. 2-3 billion per year R&D by dozens of companies. Still tele-operated mainly. Scaling lidar factories and scaling retrofit kit processes.

Technical Differences- ADAS LIdar vs Robotaxi Lidar
1. Resolution and Range

ADAS-Grade Lidar (Huawei, BYD): Designed for tasks like adaptive cruise control, lane-keeping, and emergency braking, ADAS lidar typically features lower resolution (e.g., 16 or 32 laser channels) and shorter ranges (100-150 meters). This is sufficient for detecting nearby vehicles or obstacles in controlled scenarios where a human driver remains in the loop.
Robotaxi Lidar (Waymo, Apollo Go): Built for Level 4 or 5 autonomy, robotaxi lidar offers higher resolution (e.g., 128 or 256 channels) and longer ranges (200-300 meters or more). This enables detection of smaller objects—like pedestrians or debris—at greater distances, critical for safe navigation without human intervention in complex environments.