100,000 robotaxi rides from the company you’ve never heard of

Sergei Kirillov:

Our estimate for this market is that it’s over a trillion-dollar market, right? And there is lots of space in it for multiple players.

John Koetsier:

When you think self-driving, you might think Waymo, you might think Tesla, but Avride has completed more than 100,000 autonomous rides through Uber in Dallas.

Today we’re chatting with Sergei Kirillov. He’s the chief commercial officer at Avride, about what that means, what they’ve learned from scaling, and where autonomous mobility is going.

Welcome, Sergei. How are you doing?

Sergei Kirillov:

Hey, I’m doing great. How are you?

John Koetsier:

Doing well. We chatted about this earlier. People think Waymo, they think Tesla. I hadn’t heard of you until maybe half a year ago or something like that, and then I got this pitch from your PR people, and you’re doing 100,000 rides.

This is not small. You’re doing it well. You’re doing it safely. Talk about who you are, where you come from, and how you got into these 100,000 rides. It’s probably more now, by the way.

Sergei Kirillov:

Yeah, totally. Well, first, thank you for this opportunity. Really great to be here.

And yes, I agree. Mostly people think about the key players in the market, but the actual landscape of the mobility market is much more diverse. And you’re absolutely right in saying that we are a new name on this landscape, but we are moving extremely fast.

We are more in the camp of the guys who are actually doing things rather than promising. So we started sharing about our progress and achievements when we actually had them.

Basically, the stage we’re currently in is a huge learning curve, right? So we are learning a lot. The deployment we currently have is not mostly for profits. It’s mostly for learning, for understanding customer preferences, for understanding riders’ behaviors and potential future challenges for the scaling stage.

Scaling is the next step for us. But the learnings we’ve made during these, how many months, like nine months of actual operations on the platform, are significant.

First of all, we actually see huge enthusiasm from the riders. Basically, when riders see that they’re matched to an autonomous vehicle, they almost never decline. They’re glad to have this experience, and I think we’re getting quite positive feedback.

Our positive feedback is around 92%, which is very good even compared to the platform average, the platform numbers for all the drivers present on the platform.

And you’re absolutely right. We’re already past the phase of 100,000 rides.

John Koetsier:

I rode in a Waymo in San Francisco, and I enjoyed the experience. I’ve also driven a Tesla, and I’ve tested Full Self-Driving. I didn’t trust it enough to buy it or rent it or use it frequently. But in the Waymo, I felt completely safe.

You’re doing Dallas right now. I believe it’s Dallas, correct?

Sergei Kirillov:

This is correct.

John Koetsier:

Yeah, you’re doing Dallas right now. Have the rides been entirely safe so far? Any accidents? Any challenges?

Sergei Kirillov:

If you’re in the mobility market, you actually have accidents. Since the cars are on the road, you are having accidents.

But I think I’m probably not ready to disclose the particular metrics. We think we’re approaching or even exceeding some human-driven safety metrics already.

We just need to collect some more data to make them fully statistically proven so we will be able to share them publicly. But we’re doing pretty well.

John Koetsier:

Because if there’s anything major, it hits the news for sure. I’ve seen that from Tesla. I’ve seen that from Waymo as well, when there’s been a kid that darts out on the road, that sort of thing.

Haven’t seen anything about you.

Sergei Kirillov:

Yeah, you can’t actually hide these things because now, in the social media era, everything is public, right? And since the cars are on public roads, absolutely.

So we didn’t have anything major so far.

John Koetsier:

So you said, “Hey, we’re not in a commercialization phase right now. We’re learning, we’re gathering data.”

You started, and now you’re at 100,000, 125,000, whatever the number of rides is. What has that taught you?

Sergei Kirillov:

Well, I actually wanted to add some more color on this deployment because there are some other figures that make this case even more interesting.

The 100,000 rides are our actual commercial rides that we have made. But apart from that, the majority of our fleet, which totals over 200, is mostly used for advancement of our technology.

And we are not only present in Dallas, but we’re also doing R&D rides in Austin, Texas. This is where we are headquartered. This is where we’re based, where I’m speaking from.

And what’s more important, we are past two million autonomous miles in total, which I think speaks even better about how quickly we are moving forward.

We still think that technology needs to be improved before massive scaling, and this is where we are mostly concentrated now.

Speaking about the learnings, basically one of the things that makes the commercial deployments so special and so different is the actual interaction with the customer, especially at the pickup and drop-off points, which is probably the trickiest one.

Even more famous competitors are having challenges managing the picking up and dropping off of passengers. So basically, that’s probably the biggest component of the learning during actual operations.

John Koetsier:

I know that’s challenging sometimes with just a human driver as well. What side of the road are you on? Where exactly are you? Because it’s not necessarily exactly where the map says you are. All those things can be challenging.

So as you scale this, you’re talking about scaling to Austin, adding more there. You’ve gone two million miles. Where are you going next? What other cities are you going to go to next, and what’s your vision for this?

Sergei Kirillov:

Well, the current plan is first to continue developing, continue growing in Dallas with our cars, and continue working in Austin for advancing the tech.

We have a short list of 10 next markets, but we haven’t defined which particular one is going to be next. We want to pass a couple more scaling milestones before we get to massive scaling on the car side.

Not everyone knows, but we are not only present in the car segment. We’re also heavily present in the autonomous robot segment, and this is where we’ve proven our ability to scale by launching over 20 new locations just during this summer.

So our robots are present in more than 30 markets in the United States and in Japan, and we more than doubled our fleet during this summer.

John Koetsier:

Super.

You think about fleet, you think, “Hey, there’s real opportunity here.”

So you have the self-driving cars, that’s great, autonomous cars. You’ve got the little mobility delivery robots which trundle along on a sidewalk or something like that and carry some groceries or other packages or other things like that.

That’s wonderful. You think, “Hey, what’s the other leg of that tripod? Is it drones? Is it flying through the air?”

Does that ever cross your mind, or do you not want to bite off more than you can chew?

Sergei Kirillov:

Yeah, we mostly want to stick to driving on the surface rather than in the air.

But basically, we acknowledge there are many more applications possible for that type of technology, right?

Basically, as a human learns to drive, they can ride a bicycle or drive a passenger car or drive a truck or a delivery van, whatever. The same applies to autonomous driving tech.

I think autonomous delivery has huge potential. We are currently already very well represented. I think we have all the chances to become the largest autonomous sidewalk robot operator in the United States by the end of this year.

We’re already very well represented here. Adjusting this technology to do autonomous delivery by cars could be a very good next step in our development.

John Koetsier:

I might want my Wendy’s hamburger via drone because it comes quickly and it’s hot and it’s fresh. But if I’m getting groceries and there’s four or five bags, there’s no delivery drone right now that can go in a city that can carry that amount.

So that makes sense, and there’s room for it all.

I saw something the other day that was super interesting because you’re growing, you’re expanding, and I’m talking about autonomous cars right now, your fleet for that.

We’re seeing Waymo grow and expand. Tesla obviously has Cybercab, and that’s growing and expanding as well.

At some point, these fleets have to almost start talking to each other, don’t they?

Sergei Kirillov:

Well, are the current fleets of taxi drivers talking to each other? Is that so necessary?

I mean, self-driving technology is a new product, a new technology. But basically, the business isn’t new, right?

It has existed since people were riding horses, right? Seriously.

John Koetsier:

Pony.

Sergei Kirillov:

Correct. We’re just changing the shape. We’re just changing the approach.

But the mobility market is huge. Our estimate for this market is that it’s over a trillion-dollar market, right? And there is lots of space in it for multiple players.

And the beauty of the tech that we’re developing is that it actually does not depend on any connected infrastructure.

We don’t need that. Well, maybe somewhere in the future that will be helpful, but as of today, the tech works on existing roads with existing infrastructure, with traffic lights and regular signs, and coexists with other drivers and other cars on the roads.

John Koetsier:

Do you use safety drivers in all your cars?

Sergei Kirillov:

Yes, we do use safety drivers.

First of all, all the R&D operations are strictly with a safety driver because basically this is where you learn to drive. You’re actually mining for the corner cases or some challenging situations to advance the tech.

And in our commercial operations, we also currently operate cars with safety operators.

So basically, removing the safety operator and making our operation fully driverless is actually an important next milestone, which we are planning to hit next year.

John Koetsier:

You’ve got two million miles right now driven, and maybe you’ll be at three or four or something like that, then you can remove that driver.

How does that change the economics? How does that change the cost of a ride and the profit of a ride?

Sergei Kirillov:

Actually, the economics of the ride only work when you achieve fully driverless operations, right?

Because basically, if you have to have the safety driver inside the car, that ruins the economics, right? Because you have to have a more expensive car with a driver. You still have to pay a driver, and then you have all the additional infrastructure around it.

So that basically kills the economics.

That’s why we are not currently rushing to put thousands of cars on the roads, because it just doesn’t make sense from an economic standpoint.

That’s why, as I said, we are learning. We’re basically studying the challenges that we will have to overcome at the scaling and growth stage.

That’s why the midterm plan is basically to become fully driverless and then scale from that point.

John Koetsier:

What are the hurdles to jump over before you’re fully safety-driverless?

Sergei Kirillov:

It is different market to market.

The markets that we are choosing for our short list are the markets where we do not expect any regulatory challenges or restrictions.

But basically, we are working very closely with all the authorities and regulators to ensure that our operations are legitimate.

John Koetsier:

Let’s talk about what car platform you use.

I was looking at some of your videos, and I saw a logo, an insignia, and I couldn’t quite make it out. What is the platform?

Sergei Kirillov:

We’re currently using the Hyundai IONIQ 5 as a platform. That’s our main platform.

But basically, we’re not exclusive to that platform, and there’s a very high chance that you may see other platforms with our AD kit on top on the roads pretty soon.

I can’t disclose which ones for now.

John Koetsier:

As you get closer to scaling, then the car will look less like this is obviously a different car because it’s got equipment hanging off it in different places, and it’ll just look like an ordinary car?

Sergei Kirillov:

Well, it’s a very good question, and there are many discussions ongoing in the industry right now.

The biggest challenge for now, I think, is that with several players basically having proven that this case is scalable, the automotive industry is still pretty conservative.

And it’s hard to make the first move.

We see some good attempts from car manufacturers entering this space, and we praise Hyundai Motor for their vision. They look to be one of the most active in the market for now.

But our vision is that in, let’s say, several years, most car manufacturers will have to have what we call AV-ready or AV-specific platforms in their lineup.

And as time goes on, the variety of platforms should only increase because, as I said, it’s not only one use case for that technology.

I think we should expect to see AV-ready delivery vans somewhere in the future, AV trucks, AV buses, and all varieties of body styles that are actually available in the market for human drivers now.

John Koetsier:

That brings up the steering wheel, right? And we see the Cybercab removed it.

I don’t know how they have a safety driver when there’s no steering wheel for them to operate. But will you do that as well?

Sergei Kirillov:

Well, actually, the steering wheel is not a deal breaker.

For the majority of the rides, you don’t need the full capacity of the car. Cybercab still has two seats, right? One or two people per ride is the most common case for taxi operations.

Using a car with a regular steering wheel is totally fine for now.

At least it helps to optimize the investment into the platform and doesn’t stretch the car manufacturers and AD developers too much.

John Koetsier:

Let’s look into the future a little bit.

We see Waymo, we see Tesla, we see Avride, and we see some other players in China, in Europe.

Let’s go out 10 years, and all these platforms are super widespread. What’s that change?

Is car ownership down? Is the cost of transportation down? Are there more cars on the road? Are there fewer cars on the road? What does it look like? What’s the future hold?

Sergei Kirillov:

Well, nobody knows how exactly the future will look. I wish I did.

But the way we see that is basically automation usually starts from the most routine tasks, right?

So it’s mostly shorter rides in dense areas with steady demand and some lack of parking places, where it makes it hard to use a personal car.

Automation definitely, theoretically, increases utilization of the car and should be taking the cost of the ride down, which will make rides affordable to a wider range of passengers.

This is theoretical. In practice, there are many challenges that you can’t see on the surface.

For example, the residual value of the cars. The residual value of an autonomous vehicle is totally different compared to a normal car, right?

You can’t just sell it after you use it for several years as a taxi. It’s different.

So that’s why the final future is unclear.

But I think in 10 years, for sure, I even think that in a five-year perspective, autonomous cars will take a pretty large stake in the ride-hailing business.

Especially in areas with steady demand outside of peak hours, for the most routine, shorter rides in dense areas.

John Koetsier:

Interesting. Cool.

Well, I look forward to seeing more as you enter new markets and do different things.

Thanks so much for taking this.

Sergei Kirillov:

Yeah, absolutely. Thank you.

John Koetsier:

Cool.