Uber-Waymo partnership: Why robotaxis may not be so profitable

Uber (UBER) has expanded its partnership with Waymo to offer driverless rides in Atlanta and Austin in early 2025. Waymo is a subsidiary of Alphabet (GOOG, GOOGL)

Carnegie Mellon University associate professor Phil Koopman joins Asking for a Trend to break down the news and the overall viability of the robotaxi trend.

Koopman notes the significance of Waymo advancing its fleet: "The question is what kind of growing pains they'll have as they start operating in new cities." He expects "more surprises" for the company, pointing to incidents where its robotaxis have interfered with emergency responders, like blocking firehouse driveways.

"I expect Waymo to pay attention, and I expect them to try and fix things, but we won't know what the particular problems are until we see them," Koopman adds.

He explains that while robotaxi companies like Waymo say that safety is top-of-mind, "the reality is nobody knows how this will turn out." Driverless vehicles are a new development, and there is simply not enough data to come to a conclusion about whether they offer more safety than human drivers.

"There's enough data now to know that small crashes, fender benders, and minor injuries, there's enough data that says Waymo seems to be on track for that... They need another 10 times more miles, another 20, 50 times more miles before we know how it turns out for fatalities," Koopman tells Yahoo Finance.

While many investors are bullish on robotaxis, Koopman casts doubt on its sustainability as a business model: "I have trouble seeing robotaxis as a business model because it's a cheaper ride hail and the technology is really sophisticated, expensive. It requires a lot of support." He argues that this could change down the line, but as the technology stands, it may be difficult to scale up.

For more expert insight and the latest market action, click here to watch this full episode of Asking for a Trend.

This post was written by Melanie Riehl

Transcripción del vídeo

Ways Jaguar ev Robo taxes will soon be part of Uber's fleet in Atlanta, Georgia and Austin, Texas.

The two companies making the announcement, Friday, expanding a partnership that has already been the alphabet owned Waymo autonomous vehicles, fulfilling over 100,000 weekly paid rides across Phoenix, Los Angeles and San Francisco.

It is currently unclear exactly how many Waymo Robo taxis are going to be deployed, but Uber says there will be eventually hundreds of the suvs in Atlanta and Austin and joining us now is Phil Copeman Carnegie Mellon University, Associate Professor Phil.

It is good to see you.

So let's start right there, Phil because this uh certainly got a lot of attention today.

Uber expanding this partnership with Waymo uh to Austin to Atlanta made a lot of headlines, just high level, Phil, what did you make of that news?

What you make when you saw that headline?

Uh I don't think it changes anything fundamental about the technology.

It's a business arrangement and that's fine.

But the the real trend here is that Waymo is growing its fleet and the question is uh what kind of growing pains they have as they start operating in new cities.

And what kind of growing pains would you possibly expect?

Phil?

Well, we saw a lot of growing pains in San Francisco.

Uh and some of the things were, were things that safety folks like I hadn't really anticipated.

Uh So I expect more surprises, but to give some examples, there were a lot of incidents with interfering with emergency response equipment and firefighters and police scenes and things like this.

Uh There were cars blocking firehouse driveways uh and that was not only Waymo, that was also crews having those problems.

So the question is, what about these cities that will be different, that will cause problems?

Now, I expect wemo to pay attention and I expect them to try and fix things, but we won't know what the particular problems are until we see them.

You know, Phil, you've been studying this technology for a long time, right?

I mean, for the rest of us, you know, many of us are relatively new to see these vehicles on city streets, in your opinion, Phil, um commercializing this technology here.

And now does it make sense?

Is it ready for that?

I have been doing self driving car safety for more than 25 years.

So this stuff has been around a really long time uh back then.

It was, it was a college experiment kind of thing that uh some of my colleagues did.

Uh I I think that the question always has been safety Uh, that's certainly what people ask you can have nuisances where the cars get confused and stop in the middle of traffic and those can cause problems for cities.

But the thing everyone's really asking is, well, is it really gonna be safer than human drivers?

And what the companies have been selling on is it will be safer in terms of injuries, severe injuries and fatalities.

They keep saying it will have fewer fatalities.

And right, well, right now, if you look at the Waymo safety page, it's pretty rarely interpreted to say we're already saving lives.

But the reality is nobody knows how this will turn out.

There just aren't enough miles to know whether they're saving lives or not.

And it's gonna take a few years to find out.

So that's the big question.

Is it really safer?

That's interesting because that is I I that's a key messaging that you hear, right.

You know, we need this technology.

This is a lifesaver.

Your point is, you know, not to mischaracterize you, Phil.

So I wanna make sure I, I hear you right.

We don't have enough data right now to make that claim.

Right.

There's, there's enough data now to know that uh, small crashes, fender benders and, and minor injuries, there's enough data to know that that says Waymo seems to be on track for that.

Uh There's all sorts of discussions you can have about the data methodology but, but right now there's no reason to worry about the small crashes from what we know now, uh, maybe there's a change tomorrow that comes, that changes that.

But for, for when now it looks ok.

But we're another, they need another 10 times more miles, another 2050 times more miles before we know how it turns out for, for, uh, fatalities.

And, and it's a very simple numbers thing.

We o has something like 22 million miles.

Last time they, they had a number, 22 million.

Sounds like a lot.

But in the USA good round number is 100 million miles between fatalities.

And you need much more than that to get statistical significance.

So they need 2030 50 times more miles before we have any way to know how fatalities will turn out and, and I'm not saying anything good or bad.

I'm just saying that it's too early to claim you're running a marathon and you're two miles in, you're three miles in your splits look good.

It's too early to give yourself the gold medal.

And so anytime someone says we've already proven we're safe, we've already proven we're saving lives.

They're giving themselves that on gold medal and they're, they're still back, you know, they're only a couple of miles in simple as that Phil final question.

We got a lot of investors watching this show.

Um As you mentioned, you, you've been studying this area for a long time.

Do, do you think Robotaxis are ever gonna prove to be a, a profitable business model film.

I, I have trouble seeing Robotaxis as a business model because it's a cheaper ride hail and the technology is really sophisticated, expensive.

It requires a lot of support and maybe someday, you know, don't ever say, don't never say never.

Uh, but I, I find other business models such as middle model trucking a lot more compelling.

Uh, but, but, you know, we're gonna see, I'm, I'm the safety guy.

I'm not the, the investment guy.

So we're gonna see the, the one thing that has been true for a decade and will be true for another decade is that scaling up will take longer than everyone wants to believe.

Uh The key, the key limit to this technology is um machine learning.

The technology that's used is really bad at stuff it hasn't seen before and the world is full of crazy edge cases, rare events and that is why it's taken so many years and why it's gonna take more years.

The big question is how fast can they scale?

And even at 1000 robotaxis compared to the number of cars on the road, that's not many.

So we don't know we'll see what Waymo does, but keep your eye on how fast they can scale up.

We will.

And Phil, it was great to have you on the show.

How about us walk through a complicated topic?

Appreciate it.