Autonomous cars have been the buzz of the energy and tech sectors for years now, but how close are we as consumers to actually having fully autonomous cars in our driveways?
In this week's episode of Industry Focus: Industrials, analyst Sean O'Reilly talks with Motley Fool senior auto specialist John Rosevear about the state of autonomous driving tech today, and when consumers will start to see the fruits of the industry's labors.
Tune in to find out what the different levels of automation are and why they matter; where companies like Tesla (NASDAQ:TSLA), Ford (NYSE:F), and Fiat Chrysler fit into the story; where Alphabet's (NASDAQ:GOOG) (NASDAQ:GOOGL) Waymo and Apple's (NASDAQ:AAPL) alleged car project seem to be heading; and more.
A full transcript follows the video.
This podcast was recorded on March 2, 2017.
Sean O'Reilly: Welcome to Industry Focus, the podcast that dives into a different sector of the stock market every day. Today is Thursday, March 2, 2017, so we're talking about energy, materials, and industrials. I am your host, Sean O'Reilly, and joining me today via phone is Motley Fool senior auto specialist, Mr. John Rosevear. Good morning, John!
John Rosevear: Good morning, Sean! How are you, today?
O'Reilly: Extremely well. It's good to hear your voice. It's been awhile, a month or two, since we last had a podcast, and I missed you, buddy.
Rosevear: Yeah, it's been a few weeks.
O'Reilly: I had to get you in here to answer an all-important question.
Rosevear: What's the question?
O'Reilly: You're the only guy I know who can answer this for me. I hate driving, I live in D.C., it's very tiring, I want to know when I'm going to get my self-driving car.
Rosevear: Well, it depends on what you mean by get, and really depends on what you mean by self-driving, [laughs] unfortunately.
O'Reilly: I'm not touching the wheel in my driveway.
Rosevear: Let's back up and define some terms here. This system of classification for things that assist drivers has become pretty universal. It was developed by the Society of Automotive Engineers in the U.S. The U.S. government has adopted it, other governments have adopted it, it's the way all automakers talk now. They categorize it in six levels that are helpfully numbered from zero to five. A level zero system, there's no automation, the driver does everything. This might even include warning systems like lane departure warning. But, the driver actually has to do everything. Level one is when you get some basic level of driver assistance, assistance that parallel parks for you, that kind of stuff. That's a level one system.
O'Reilly: Does that include anti-lock brakes? I'm surprised that's on here.
O'Reilly: Eye of the beholder?
Rosevear: The driver is still executing that. They count that as an intervention system. It's when the car decides to break for you that it starts to become assisted. These emergency stopping systems that they have now where the radar is on the front of the car, the levels one through five require some level of sensing what's going on around a car. So, yeah, these parallel park systems where the car spins the wheel for you and backs into the spot, that kind of thing, that's a level one. Level two is where it starts to get interesting. This is the system where you're on the highway and the traffic jam assist systems, maybe the earliest versions of Tesla's autopilot, where you can kind of take your hand off the wheel under certain circumstances, and the car will maintain and adjust its speed relative to the cars around to you, and it will keep itself in its lane. It's doing a little steering and a little accelerating and decelerating on its own. But there still has to be a human driver, the human driver has to be ready to grab the wheel. In fact, a lot of these systems require you to actually touch the wheel every minute or two, or something like that. Level three is when it kind of gets into self driving. This is a system that can drive under certain conditions, we might say highway driving, this is where Tesla's autopilot is starting to go, this is where some of the systems, GM's (NYSE:GM) Super Cruise that will come out on Cadillac soon, this is where it ends up, level three.
O'Reilly: Super Cruise? [laughs]
Rosevear: That's what they call it. That's what they've been calling it for a few years, I'm not completely sure that's what they're going to call it when it comes to market. Anyway, the gist here is, it's self-driving until suddenly it isn't, the driver has to be ready to grab the wheel. That makes it a little tricky. We'll talk about that in a minute.
O'Reilly: The Motley Fool's Dan Sparks did that article about a year ago, he lives out in Colorado and he has a Tesla, how far did he get, 90 miles? I can't remember, but, I don't think he touched the wheel on the freeway with that thing.
Rosevear: Tesla's official instructions with the original system where you're supposed to touch the wheel regularly, they considered it an advanced level two system. I think now they're into level three territory. The border here is a little bit fuzzy. The border to level four is not fuzzy. Level four is, the car is driving itself, period. It does not need you, but it's limited. Usually what it means is it's limited to an area that's been mapped, we say it's been geofenced. These systems tend to depend on very highly detailed 3D maps that show stuff that normal maps don't show, like exactly where the curb is within a couple inches and so forth, so the car can navigate precisely, especially on a city street. So, it's full self-driving, you don't ever have to touch a steering wheel, but it only works where the car has a map. Level five -- and we'll get back to this in a minute -- is the full-blown thing. It'll drive itself wherever you want under any conditions. If it's snowing, it'll go up a mountain, whatever, anything.
So, what we're talking about in the near future is level four, a car that will self-drive itself. This is coming soon. But it's limited by its maps. The idea is, gradually the maps will expand over time and you'll be able to drive --
O'Reilly: Are there cities that are being mapped right now? I'm actually surprised to hear this.
Rosevear: Yeah, Pittsburgh is being mapped. This is where Uber is doing their testing, a lot of it. This is actually where Ford's new company they created as a start-up with a veteran of Uber and a veteran of Google's self-driving car project running it. This is basically Ford's software team for self-driving, and they're in Pittsburgh. There's been a lot of mapping done, interestingly, in Pittsburgh, because the artificial intelligence lab at Carnegie Mellon is the source of a lot of the research that went into this, so, a lot of the people are there. There's some mapping going on in San Francisco. I know GM's cruise automation subsidiary is doing testing of level four assistance in San Francisco, so they have mapping going on. Then, there are companies mapping all over the place. Tesla, I think, is automating some of this mapping with the sensors on their existing cars. A company called Mobileye (NYSE:MBLY) is going to start doing mapping with sensors on their customers' cars, these are sensors for level two systems but they can help do the mapping for level four. This is going to start very soon, probably next year for Mobileye. There are private companies trying to do this, too, hoping to be able to sell the maps to the automakers or whoever. And, of course, the big wild card in this is, of course, Waymo. Google's been mapping for years and years. [laughs]
O'Reilly: You've seen the cars? They look ridiculous, they have the pole and they have the ball of cameras on the top, and they just drive around all these cities. That's how I can see what my dentist's office looks like from across the street. They're mapping cities with that, too?
Rosevear: Yeah. I think everybody is. And it's not just mapping for Google Maps, it's mapping at a much more fine level of detail, where is the curb, where is the speed bump, and so forth, so the car knows what to expect as it's driving. The idea is, a map is a fallback from its sensors. If it's dark out, if it's foggy, if the lighter system is being buggy for whatever reason, you can kind of fall back on the maps. Because these are safety systems, there has to be redundancy built in. That's the thinking.
O'Reilly: So, when is this stuff coming out?
Rosevear: Good question. Tesla constantly does these upgrades, and they'll break into level four at some point. If I had to guess, I'd say 2019. But I don't know for sure. They may not even know absolutely for sure when they'll be ready to ship it. There's also an interesting question there that with their owners of their current models, they've been shipping stuff that maybe has some bugs still in it for the owners to bang on. And because Tesla owners tend to be early adopter types, they're mostly cool with that. As Tesla moves into the mass market with its upcoming Model 3, something else we've talked about, I don't know if they're going to tighten it up a little bit. They might. We'll see. But that might push that data off. But, anyway, this partnership of giant auto supplier Delphi, Mobileye, which is a machine vision specialist that has done a lot of work around--
O'Reilly: And but they are publicly traded, right? For our listeners.
Rosevear: They are. The ticker there is MBLY. Look them up, because if you want a pure-play on self-driving, that's as close as it gets. The third company in this is Intel, which is designing the processors. They're working together on a level four system. We've talked about this before. They will have this available to any automaker who wants it by the end of 2019, they say. Now, the cars are going to ship in 2019. These things are incorporated as the cars are developed. So, probably late 2020, early 2021, people will be shipping cars with this system in, but those cars may not go to retail. [laughs]
O'Reilly: What?! Will they hang out and not do anything? [laughs]
Rosevear: Here's what, for instance, Ford is saying. Ford says, "We're going to mass produce a level four vehicle starting in 2021." Great. Then, they say, "Yeah, but we're not going to sell it at retail. This is going to be for car sharing use in cities and ride hailing services, Lyft and Uber and so on." It's because of the maps. They only want to sell it where they have the maps. CEO Mark Fields at Ford has said, "We're not even going to give it a steering wheel in this," so obviously it has to have the maps. Again, as the maps expand, there'll be more. Ford is saying right now that they think they'll have level four functionality in their retail vehicles 2025-ish. But they always say that, Bill Ford, executive chairman of Ford, always says, "We throw out these dates, but it's always earlier than I expect." So, we'll see. Right now they're saying 2025. So, yeah, he says, "This stuff moves much faster than we expect." There's also some other stuff going on. The thing to emphasize here, I should back up a little bit, not all the automakers, not all the tech companies talk about what they're doing. There will be surprises here. This is just, what has been announced, what we have seen, what we generally know is coming.
O'Reilly: Yeah, once you go public with something you get held to it in some capacity, and nobody wants to get egg on their face.
Rosevear: Right. And GM originally said they'd launch Super Cruise in 2016 --
O'Reilly: They need to get a different name for that. [laughs]
Rosevear: Yeah, OK. They're a level three system. They came out a while back and said, "Wait a minute, we're not happy with it, we're going to do more work and make it more advanced. We're not going to set a firm date, but we think late 2017-18, probably around there." But again, they are not holding themselves to a firm date on that. In fact, they're not holding themselves to a firm date on a lot of this stuff. Mary Barra's thing is, "We'll ship it when it's ready." But what we do know is that GM is planning to actually start building a whole lot of level four vehicles, reportedly in the thousands. These will be the electric Chevy Volts. They will be fitted with prototype level four systems, and they will go into this massive test with Lyft in several different cities. So, when will you get your level four car?
O'Reilly: Which is why they invest in Lyft.
Rosevear: Right. They own 9% of Lyft, it's something like that. It's a major investment, certainly, for GM. But, when do you get yourself a driving car? If you're in D.C., you might get one in several months from now or a year from now, when you summon Lyft with your smartphone and a Chevy Volt comes. Uber is already doing this on a tiny scale in Pittsburgh. They have 8 or 12 cars or something that are prototype self-driving vehicles in their fleet. With these things, there are two engineers on board. One is ready to grab the wheel and the other one is taking notes on everything that's happening with the system. And it may be the same deal with these GM Lyft self-driving vehicles. So it might not seem all that exciting because there's somebody in the front seat, but the car is driving itself, is the idea. This is a way to rack up a whole lot of test miles quickly, if you have a whole bunch of these things out there, you know?
O'Reilly: Yeah. I can't wait to see all these Lyft driverless cars handle all the roundabouts in D.C., that'll be a treat.
Rosevear: Yeah, they'll teach them. We call them rotaries up here in New England, we have lots of them too. It'll be interesting to see how they handle Boston, which is colonial era streets in some areas of town. [laughs]
O'Reilly: Yeah, they're made for horses, not driverless cars.
Rosevear: Yeah. Well, they were made to walk in, or cow paths, definitely not driverless cars. But, there's other stuff in here, too. The wild card is Waymo, which is the Alphabet subsidiary that used to be the Google Self-Driving Car Project and has now been set up as a company to go into business and commercialize this in some way. Who will partner with them? They had this little deal with Fiat Chrysler where Chrysler built them some Chrysler Pacifica minivans as test vehicles that took on their latest prototype system. They've been talking to Honda, reportedly. But of course, there's a certain view in the auto industry that a deal with Waymo is a deal with the devil because Google will want all the data. And then you're an Android phone maker, and Google's making all the money. That is a great fear. I think every auto CEO in the world is mindful of the example of the phone industry. [laughs] They don't want to give up the data. I think that is actually a lot of the selling proposition of the Delphi-Mobileye system, this is a system developed within the established auto industry supplier structure. We'll share the maps among our fellow automakers and so forth, but we're not selling the data to Google. [laughs]
O'Reilly: What do you think has been going on there? Correct me if I'm wrong, but Google has those tiny little white go-kart looking driverless cars for years. The laymen, we haven't gotten a hint of them trying to make a buck off of them at all. It's weird.
Rosevear: They hired, a while back, and executive named John Krafcik. He had been the president of TrueCar, and before that, he had been in the business for years. He was with several automakers, in management capacities, executive capacities. He's a guy who knows the business, a good guy, smart guy. His job is to lead Waymo and find a way to bring this stuff to market somehow. He's been out there talking to automakers. He actually did a really big presentation at the North American International Auto Show in Detroit in January -- I was going to say last month, but it's March now. He was talking about where their system is and so forth. They have a solid prototype level four system. I think he was there, in part, fishing for automakers who might want to have conversation. It's possible those conversations have happened and nobody is talking about them. The ones that have made it into the media are Honda and, of course, Fiat Chrysler, which is a much more overt thing, because the latest test vehicles are all Chryslers.
O'Reilly: We don't have this as a note, so feel free to tell me you don't know a ton about this, but I'm sure all of our listeners are on the tip of their tongues with this, they want to hear, what is Apple up to? They have that big plan, and I've got a push notification from The Wall Street Journal a year ago that Apple was going to build a car within the decade, blah blah, then they pulled back on it, they're doing testing again. What is Tim Cook scheming right now?
Rosevear: Good question. Yeah, I remember when that Wall Street Journal thing came out.
O'Reilly: You got the notification too, I'm sure.
Rosevear: I did, and I actually poked a few sources and ended up writing a big piece for The Motley Fool first thing on Saturday morning after that broke, sort of, what does this mean, can they actually do it, that kind of thing. And the answer was, Apple has so much money that, of course, they could start a car company, the question is, do they want to?
O'Reilly: Yeah, I remember you talking about that.
Rosevear: Right. They can throw money at it. The full boat, it's about $10 billion to get to where Tesla is where they have a factory and they're banging out cars and it's really high level of quality and so forth. Full boat. That's Tim Cook's lunch money these days. [laughs] Their cash load is unbelievable. They can do anything. The question is, would they want to? The margins in the car business, by Apple's standards, are pretty lousy. A luxury car business might make an operating margin of 10-12%. If you're Porsche, you might make 16%. That's kind of the upper limit.
O'Reilly: So, is it possible that Google and Apple wind up dropping the brains in these cars and calling it a day? Google already, they must have tons of data on these cars operating.
Rosevear: Google has made it clear, John Krafcik came out and said, "We're not building cars, we're not going to build a car factory, we are not building cars. We're looking for partnerships, we are not building cars." Apple hasn't said anything, of course. Officially, this program doesn't even exist. But the word was, a lot of the people who might have been on the actual build-a-car side of the business have been let go. People working on electric self-driving cars are in very hot demand these days. So, none of these people are wanting for employment. But it was thought that a lot of those folks were let go, and they really ratcheted down the ambitions of the program. They may be aiming to do something where they come out, like Waymo, and offer a system that Apple-izes a self driving car built by somebody else. It's possible that they will find an automaker to partner with, and maybe offer that as a deluxe ride hailing option. My thought when they were originally talking about this was that they were going to build the deluxe Apple version of robot Uber, where it's a little bit more money but it's a nicer car, and it recognizes all your Apple devices and it's this seamless experience, it sees your iPhone and it automatically sets your seat and your climate and your music and everything else and all this stuff, it takes exactly where you want to go, and it's just a subscription model, $500 a month for the Apple Car service. That's where I thought they were going to go. Others disagreed. Some of my fellow Fools who are also very close watchers of this kind of technology and particularly close watchers of Apple thought they might actually build a car for retail Tesla-style.
O'Reilly: Cool. We're about out of time, John. I want to give you the last word. When am I magically going to get my driverless car in my driveway?
Rosevear: If you want a Tesla, you might get it by the end of the decade. If you want a car from one of the bigger automakers, it's probably a few years beyond that. Somebody may surprise us and have a level four car out by 2020 or 2021 that has a fairly big map. But I'm thinking it might be over the next couple years after that that these start to be available for retail. I suspect the first ones that come out, it will be an expensive option on a luxury car.
O'Reilly: Awesome. John, I cannot thank you enough, once again, for your time.
Rosevear: It's always a pleasure.
O'Reilly: I have to imagine that as this continues to ramp up, this will not be our last show about this topic.
Rosevear: Oh, heavens no. We'll be talking about this for several years to come. It's a big deal.
O'Reilly: Awesome. Well, have a good day, John! Thank you, again!
Rosevear: All right. Take care, Sean!
O'Reilly: Bye. That's it for us, folks. Be sure to tune in tomorrow for the Technology show. If you're a loyal listener and have questions or comments, we would love to hear from you, just email us at firstname.lastname@example.org. As always, people on this program may have interests in the stocks that they talk about, and The Motley Fool may have formal recommendations for or against those stocks so don't buy or sell anything based solely on what you hear on this program. For John Rosevear, I am Sean O'Reilly, thanks for listening and Fool on!