In my last post, I shared the importance of LiDAR in autonomous driving and Tesla’s vision-only strategy. You may read it first if you haven’t:
A clip came across my recommendation on YouTube, which I would like to add to my point of view in that post. While expressing the opinion that LiDAR’s cost had come down a lot over the years (and Tesla should re-think about adding those sensors back to its vehicles), I missed the following consideration:
Each additional sensor adds cost and challenges to supply chain management, calibration, and integration into the overall system. This complexity can bog down development, diffuse focus, and ultimately lead to diminishing returns in performance.
Andrej Karpathy, who was the director of autopilot vision at Tesla, brought this up in Lex Fridman’s podcast. For those interested in the clip, you may watch it here:
The news that BYD is offering its drivers the FREE “God’s Eye” system sparked my interest in exploring the development of autonomous driving in China. I will also share my findings and learnings in two areas: Robotaxi service and personal self-driving system. The latter will be discussed in the upcoming post to keep things in bite size.
Robotaxi in China
You might not expect this fact: China is currently leading in the global robotaxi market, with over 16 cities operating autonomous vehicles. According to a report by S&P Global in Oct 2024, at least 19 cities are currently testing robotaxis and robot-buses.
I feel that China’s autonomous vehicle journey is different. There is a palpable sense of reinvention in the air - China isn’t merely playing catch-up but aggressively carving its path in the self-driving world. Major players like Baidu, Pony.ai, and WeRide are not just testing their autonomous vehicles but transforming urban landscapes through extensive road trials in multiple cities. While markets in the US and Europe tread carefully within well-established regulatory frameworks, China seems to be embracing an ambitious and innovative future (of course, with pros and cons).
Here are a few major players in the robotaxi scene in China:
Apollo Go by Baidu
Since its initial launch in 2019, Baidu’s Apollo Go service has been a trailblazer. Today, it operates in over 10 cities, including hotspots like Wuhan (more than 400 vehicles), where its fleet of autonomous vehicles navigates complex urban environments. Baidu’s approach leverages a robust fusion of computer vision, deep learning, and sensor integration to deliver a reliable autonomous ride-hailing experience.
Its ride-hailing service has reportedly operated over 7 million rides (as of July 2024) and more than 100 million km in testing mileage.
While concrete revenue figures for Apollo Go remain under wraps, the project aims to be the first autonomous ride-hailing project to profit by 2025. Check out the video below for an experience ordering a ride-hailing service from Apollo Go.
WeRide
WeRide has been active in domestic trials in China and partnered with Uber to launch a commercial robotaxi service in Abu Dhabi, UAE. This marks the first time autonomous vehicles are available on the Uber platform outside the US. As a safety precaution, all WeRide robotaxis currently have a human driver.
In this video clip of WeRide robotaxi testing, a human driver is in the seat to take over the steering wheels in an emergency.
Although WeRide is still far from breaking even in terms of cost, Nvidia revealed its investment in the company yesterday (14 Feb 2025) by purchasing 1.74 million shares.
Pony.ai
Similar to WeRide, Pony.ai is also a publicly listed company on Nasdaq. While it currently has around 250 fleets on the road in tier-1 cities in China, this number is expected to at least quadruple to more than 1000, eventually leading to cost break-even this year.
Well, this is just a tiny target. Given the reduction of operational costs and the support of authorities, it aims to expand to more than 10,000 fleets in the next three to five years. To launch its seventh-generation robotaxi fleet, the company has forged joint ventures with Toyota and strategic alliances with local Chinese automakers like BAIC and GAC Aion.
Expanding to the horizon outside of China, Pony.ai will soon begin the road tests in South Korea.
Still a long way to go
While everything seems exciting here, autonomous robotaxi services only account for roughly 1% of the overall ride-hailing market in cities in China. For sampling purposes, it was estimated that Apollo Go generates around 290,000 rides per month in Wuhan, a city with millions of daily ride-hailing trips.
While the autonomous tech behind China’s robotaxis is advancing at an impressive pace, we are only seeing the tip of the iceberg now. I genuinely believe that this early, modest penetration is just the beginning—a massive surge in growth is waiting to unfold in the coming years.
Rather than leaving companies to figure it out independently, the authorities have been rolling out clear guidelines since 2018. They set standards for testing intelligent connected vehicles and even created special zones in cities like Beijing, Shanghai, and Wuhan for autonomous driving trials. In Beijing, for example, new regulations starting in April 2025 won’t just standardize safety measures like real-time data monitoring and having a safety supervisor on board—they are also designed to fast-track the whole innovation process. The government is paving a smooth highway for companies like Baidu’s Apollo Go, Pony.ai, and WeRide, giving them the confidence and a safe playground to push autonomous tech forward. This firm, unified support builds trust among consumers and investors, setting the stage for what could be a transformative shift in urban transportation.
In my next post, I will discuss self-driving systems for personal vehicles in China, notably BYD's “God’s Eye,” which caught my attention in the past few days. I can’t wait to share the information with you.
If this interests you, subscribe to “A Curious Semi-Expert” newsletter to receive my weekly learnings directly in your inbox. I’ll keep you company on this journey as we adapt to these exciting technologies together.