---
title: 'No Driver, Chaotic Traffic: Riding China''s Tesla Robotaxi Rival in Shenzhen (Pony.ai)'
source: 'https://www.youtube.com/watch?v=ZPVcbsq0JUA'
video_id: 'ZPVcbsq0JUA'
date: 2026-11-15
duration_sec: 770
channel: 'Blueprint IoT'
---

# No Driver, Chaotic Traffic: Riding China's Tesla Robotaxi Rival in Shenzhen (Pony.ai)

> Source: [No Driver, Chaotic Traffic: Riding China's Tesla Robotaxi Rival in Shenzhen (Pony.ai)](https://www.youtube.com/watch?v=ZPVcbsq0JUA)

## Summary

This video documents a first-person ride in a Pony.ai autonomous robotaxi through the busy streets of Shenzhen, China. The creator highlights the vehicle's advanced perception, its handling of chaotic traffic, and the integration with the city's connected infrastructure, offering a firsthand look at the current state of autonomous driving technology.

### Key Points

- **Booking and Access** [00:07] — The robotaxi was booked through a mini app in WeChat, highlighting the integration of the service with China's dominant messaging platform.
- **Safety Features and Assistance** [01:22] — The car provides voice prompts for safety, such as seatbelt reminders, and offers an assistance button for contacting a human or AI support, though the steering wheel is covered, indicating full autonomy.
- **Visualizing the Car's Perception** [02:06] — The main screen displays what the car 'sees', including other vehicles, pedestrians, and the planned path (shown in blue), which reassures passengers by demonstrating the car's awareness.
- **Recognizing Brake Lights** [02:33] — The system can identify and display when other cars, like a BMW, are braking, providing additional situational awareness and safety confidence.
- **Operational Area and Route** [03:16] — The ride covered 7 kilometers in 12 minutes, showing that the robotaxi service operates over a substantial area of Shenzhen, not just a small test zone.
- **Handling Chaotic Traffic** [03:43] — The car navigated intense traffic, including scooters and unpredictable lane changes, without being overly cautious, effectively negotiating its way through the flow.
- **Advanced Decision-Making** [05:06] — Unlike standard adaptive cruise control, the robotaxi assessed complex situations, such as a car merging, and chose to maintain speed rather than simply slowing down, demonstrating a higher level of decision-making.
- **Lane Positioning and Maneuvering** [06:19] — The car actively maneuvers within its lane to give more space to nearby vehicles, showing a proactive approach to safety and comfort.
- **Connected City Infrastructure** [06:33] — Shenzhen's connected traffic lights provide data to the car, allowing it to anticipate red lights and plan accordingly, a key advantage for autonomous driving in the city.
- **Pony.ai App and Integration** [07:12] — The service requires the specific Pony.ai app, though DiDi (China's Uber) is reportedly working on integration, indicating the service's current distribution model.
- **Overtaking and Merging** [08:07] — The car successfully overtook a slower vehicle, though the merge was cautious and sometimes paused, reflecting the challenges of dense traffic and unyielding drivers.
- **Driving Style and Speed** [09:27] — The robotaxi's top speed seemed to be 60 km/h, and it accelerated confidently but braked relatively late and intensely, which could be a comfort consideration for passengers.
- **Smooth Arrival and Drop-off** [10:15] — The ride concluded smoothly, with the car handling the drop-off well, despite the presence of various road barriers and obstacles common in China.
- **Sensor Suite** [11:29] — The vehicle is equipped with a comprehensive sensor array, including multiple cameras (three on each side, four in front), and likely LiDAR and radar units, enabling its advanced perception.
- **Overall Verdict** [12:24] — The creator concludes that the ride was smooth and not scary, and believes that if the technology can handle Shenzhen's traffic, it can handle almost any traffic situation globally.

### Conclusion

The Pony.ai robotaxi in Shenzhen demonstrates that autonomous driving technology is already capable of handling complex, chaotic urban environments. The integration with connected city infrastructure and the car's advanced decision-making suggest that robotaxis are a viable and increasingly practical mode of transportation.

## Transcript

booked. So, it had to come from a little bit of further distance. And yeah, all bit of further distance. And yeah, all via a mini app in in WeChat. So, yeah, I have no clue how to open it actually. Yeah, open the door.
&gt;&gt; Yeah, if somebody speaks Chinese, let me know what you're saying. But there's a button in the app actually. &gt;&gt; Hello. We even sit in front. I have no &gt;&gt; Hello. We even sit in front. I have no clue.
put the last ditch of my phone number. Oh my goodness. So, now you guys know my phone number. The last footages. Um, fasten your seat belt. Okay, I didn't fasten my seat belt. So, it
noticed &gt;&gt; your vehicle is ready. Please depart &gt;&gt; Assistance if you need help. &gt;&gt; It's definitely a little bossy.
Everything is fastened. So, what do we do now? &gt;&gt; Start. Right. Okay. &gt;&gt; We are about to start. Please make sure your seat belt is securely fastened. You can contact us by tapping the assistance
&gt;&gt; All right. Letting us know that we can contact an assistant. &gt;&gt; Not sure if that would be human or AI, but at least it's good to know there's somebody. And we do have a steering wheel, but there's a fullfledged cover
And you can see the steering wheel moving, which I think is actually kind of kind of reassuring that you see what's going on and that the car is actually steering and reacting. So I think that's that's definitely something
that's like yeah from a passenger experience a good solution. now okay so there's some video stuff going on but now we can skip to the main
screen so we can see what a car can see. That's I think I saw this now with a couple of self-driving cars from other videos and I think that's the best you can do because you are reassured the car sees everything because usually the car
is aware of the situation and you can just see what's going on. So there are a bus pulled up here and actually if you look outside a bus pulled actually next look outside a bus pulled actually next to us. So, it definitely sees that. And
here next to us in the on the right. I think uh let me show you this car. You see that it has red back lights. And if you look here, it's this BMW that actually engage the brake lights. So, it can even recognize which car is braking
or not. So, that feels very reassuring. And we see in the blue where the car plans to go. So, that's I think also good. Even better than with a human depending on if you can see his navigation or not and you're just like
now a left turn and you need to yeah maybe secure yourself a little bit in the car and just prepare yourself for the for the turn. So that's fine. So
besides the front screen we also have a rear screen here. Bit smaller but definitely convenient to look at. 12 minute drive, 7 kilometers. So, it's actually quite a big area in Shenzhen, you can use those robo taxis. It's not
just like one super small, very organized area. And I mean, if you know anything about China, the traffic is definitely intense like with those little bikes and stuff trying to cross, but it handled very well. So, it's not
overcautious. It's definitely negotiating.
very fast. I don't know why, but yeah, you can see just the guy in front just merged into our lane without any indicator or stuff and quite quick. So, was definitely the right call. I'm not sure if it was by
that somebody's approaching fast from the back.
switch lane. Okay, now there is a Oh my goodness. Ah, no, it it recognized now. You can see here it recognized the construction site, but that came a little unconveniently late.
fast. It's not the German highway, not German autoban with 200 km/h. Okay, the the bus tries to merge, but okay, it wasn't showing any
indicators on the on the bus on the screen. Not sure it was just ignoring or negotiating. Well, but the car didn't stop. It kept going. And I think that's something that's pretty advanced for me. Like it's easy to drive very cautious,
but actually to be useful in a yeah busy traffic situation as we have it in Shenzhen all the time is actually I think something something that's more impressive. Now this guy merging into our lane, but
about like your normal like speed control, um, cruise control with a with a distance control as well, you if somebody merges in front of you, you usually slow down to get your distance back. But the car definitely assessed
the situation on a more complex level and kept going. straight for quite a while, we're going only 45 km an hour. Before we did 60 and
you can see the traffic next to us definitely faster. And since we went faster before, I'm not sure like it's not the top speed obviously. So, it did
already more. Okay, now it's getting a bit more busy, but handling quite well. It's also not just staying perfect in the middle of the lane. It's like always trying to give cars a little more way if
they come too close. So, it's like even maneuvering within the lane. And we have a red dot, whatever that means. That's maybe the anticipation of the traffic light. Yeah. One thing you need to know about China or or like not
China specifically Shenzhen, everything is connected. So all the traffic lights and even even though you can't see like a countdown on the traffic light here, you have this for for pedestrians. But for the cars, even though we don't have
for the cars, even though we don't have a countdown here, the cars and the apps and everybody knows like in Europe it would be or in the west I would say with Google Maps you would see on Google Maps a little countdown of the of the of the
traffic light how long it will stay red. You can see this on on a map. That's like Google Maps here. And you can see that as well if you call an Uber which would be um Diddy here in China for this car by by the way you need a specific
app um application. Diddi is apparently trying to get it integrated but you need a specific app. It's called Pony AI and that's like the yeah the autonomous robo that's like the yeah the autonomous robo taxi um yeah provider. And so since
everybody knows the traffic light situation how long it will stay red also sure why it's not showing us. Usually, you would get like a little counter next you would get like a little counter next to next to the traffic light, but that's
you can capitalize on when you're doing a robo taxi that you have such a connected city. So, we have the green light and very impressive, I think, is that it's recognizing all the the brake
lights. I mean not impressive as such but like a nice feature that is showing it to us which is just as more as more awareness that we know the car has of the situation the more safe I guess we will feel.
So now the guy in front of us doing some weird stuff and we decide now to take over this overtake this guy. So, we see the spot where we want to go, but we can't negotiate into that spot.
But now we're doing it. And we merged now. Now, we didn't merge. We tried, but then halfway we stopped. And the reason, I guess, because the car in front of us got got away. So, we're good in the middle lane.
We drove in the right lane before. Left lane, middle lane. So, it's not lane specific. It's just going whatever it thinks is best.
left hand lane. When it's trying to merge, it's like a little bit cautious, but also I would say people are not very Yeah. forgiving and not allowing some space and stuff like this. So yeah, and it's like yeah,
randomly breaking a little bit where you can see just a decision was maybe can see just a decision was maybe yeah on the verge of being option A, option B or something. But overall so far quite well. It seems like 60 km an
hour is the top speed. It's accelerating quite confidently and also it's breaking and like when it's going when it's rolling towards a a car at the traffic light or something, it's breaking relatively late and intense, but it's
And now you can see again that the car is recognizing that there's a traffic long before you can see it with any camera or sensor, whatever. That's purely based on the on the connected traffic lights that you have here. So,
we're trying to merge to the right, I guess, because it's not fast enough here. And that worked pretty well.
&gt;&gt; Belongings for a lighting. &gt;&gt; So, &gt;&gt; So, we arrived. That was a very smooth ride. we arrived. That was a very smooth ride. And also this was handled very well.
And one thing about China here is also you have a lot of like things on the road like barriers and stuff like this. But it's noticing all those things very But it's noticing all those things very well. And let's see how the drop off
So it's just parking. So there like about 10,000 dedicated drop off spots. somewhere. But &gt;&gt; yeah, I mean
&gt;&gt; the thing is it's like middle of the street &gt;&gt; and but nice. You can see the right rear camera. So we are like kind of Yeah, but it's just Yeah, you can't see a lot. But anyway, it's a nice idea. So let's try
to not annoy the robots. You never know. &gt;&gt; All right. &gt;&gt; Okay. Okay. Yeah, it's definitely definitely a bossy type of robo taxi. So, goodbye.
So, here you can see some of the sensors. So, we have a camera here. I guess that's a radar or late lighter. That could be ultrasonic sensor, but I'm That could be ultrasonic sensor, but I'm h maybe not. Not sure.
have a bunch of cameras going on. three to the left, three to the right, or two to the or four to the front in total. And then on the corners, and then I guess it's lighter and or radar.
radar. And another run on the front here. And another run on the front here. And it's a iron. Let's see. More cameras on the corners. And it's driving across here. I mean, it's just
was just like 75 cm away from me, but managed it quite well. again not overly confident uh sorry not overly cautious which I think is a cool thing. So yeah,
that was the robo taxi ride in Shenzhen, China. Not scary of was my first robo taxi ride. So definitely some moments. But um yeah, besides besides having
having my moments, it was definitely smooth ride and can definitely recommend. Technology is already there. I would say if it can handle a traffic situation like Shenzhen, I guess it can traffic handle any other traffic
traffic handle any other traffic situation as
