# Is Waymo About to Dethrone Uber? Maxinomics · 2024-11-01 Uber stands to capture the $25 billion that US riders pay drivers each year by partnering with Google's Waymo, which has scaled driverless rideshare from under 300 rides per day in Phoenix last January to 15,000 rides per day across four cities, on track for 40,000 within a year. Phil Andrews argues Uber's exclusive Waymo deal in Austin and Atlanta lets the historically asset-light company pocket $0.70 per mile, since riders pay $3.20 per mile while a Waymo costs at most $1.50 per mile to operate, and to redesign the door-to-door experience it currently can't control with human drivers. Using Chicago's full 136-million-ride dataset, he estimates covering peak demand would take roughly 10,000-20,000 cars per city, or a minimum of 300,000 Waymos across the 30 largest US cities. At a $0.15 per-mile profit and 50% market share, Waymo could earn $4 billion a year, though Tesla FSD, Zoox, and Nuro compete, and the next city choice—snowy or Sunbelt—will signal the software's confidence. ## Questions this episode answers ### How much do riders pay Uber drivers each year, and why does that money matter for Waymo? Phil Andrews says riders paid drivers $25 billion in 2023 — the portion of Uber and Lyft fares passed on to drivers. Driverless rideshare like Waymo is scaling fast, so that $25 billion is set to move to companies operating or building driverless vehicles. [0:00](https://maxinomics.podhood.com/a8cca3fb-5319-4169-b815-e62c7527b421?t=0) ### How fast is Waymo scaling its driverless rides? Phil Andrews notes Waymo gave fewer than 300 driverless rides per day last January, all in Phoenix. Now it's 15,000 rides per day across 4 cities, with Atlanta coming online early in 2025, and another 1,000 cars being built should push it above 40,000 rides per day within a year. [0:19](https://maxinomics.podhood.com/a8cca3fb-5319-4169-b815-e62c7527b421?t=19000) ### Why did Uber sign an exclusive agreement with Waymo, and how much money can it make per mile? Phil Andrews explains Uber customers pay on average $3.20 per mile, with $2.20 going to drivers and Uber keeping about $1. Since operating a Waymo costs at most $1.50 per mile all-in, Uber could pocket a $0.70 per mile difference, which is why it agreed to exclusively host Waymo in Austin and Atlanta and manage fleet cleaning, charging, and maintenance. [3:06](https://maxinomics.podhood.com/a8cca3fb-5319-4169-b815-e62c7527b421?t=186000) ### How many Waymo cars would it take to cover a major city, and what would that cost? Phil Andrews calculates that covering Chicago's peak New Year's Eve hour of 28,359 rides would take about 20,000 cars, while a typical Saturday night peak needs roughly 10,000. At about $120,000 per car, covering one major city costs billions, and covering the 30 largest U.S. cities would take a minimum of 300,000 Waymos and tens of billions. [6:33](https://maxinomics.podhood.com/a8cca3fb-5319-4169-b815-e62c7527b421?t=393000) ## Key moments - **[0:00] Intro** - [0:00] Uber drivers received $25 billion of riders' fares in 2023 — the money driverless rideshare is now chasing - **[0:19] Scaling fast** - [0:19] Waymo scaled from under 300 driverless rides a day in Phoenix to 15,000 across 4 cities, heading for 40,000 - **[1:00] Fifteen years of R&D** - **[1:55] Why riders switch** - [1:56] Riders choose Waymo because there's no driver to cancel, judge you, or play bad music — even in Barbie pajamas - **[3:23] Chicago dataset** - [3:23] Chicago's full rideshare dataset: 136 million trips, 26% of riders tip, average tip $3 - [3:44] Uber's exclusive Waymo deal: riders in Austin and Atlanta can only order Waymo through the Uber app - **[4:07] Uber's exclusive deal** - [4:18] Uber riders pay $3.20 per mile, drivers get $2.20, Uber keeps $1 — driverless could cost just $1.50 per mile - [4:38] Uber could pocket $0.70 per mile running Waymo fleets — and will clean, charge, and store the entire fleet - [5:11] Why flying is miserable: Phil Andrews' analogy of Delta building its own airports explains Uber's door-to-door opportunity - **[5:38] Unit economics** - [6:21] Uber could sell Kind bars, Diet Coke, and ad-supported cheaper rides inside driverless cars, says Phil Andrews - [6:42] Only 700 Waymo vehicles operate today, and wait times are too long for full city coverage - [7:09] Chicago's busiest Uber hour: 28,359 rides on New Year's Eve would need roughly 20,000 driverless cars to cover - **[7:11] Airport analogy** - [7:49] Covering the 30 largest US cities needs 300,000 Waymos at $120,000 each — tens of billions upfront - [8:37] Waymo partnering with Uber instead of building its own app: routing algorithms and 75% market share fill the gaps - [8:39] If Waymo earns $0.15 per mile on 50% of big-city rideshare, it adds $4 billion a year to Google's profit - [8:57] Tesla FSD, Zoox, and Nuro are real competition, but Waymo leads with paying customers, says Phil Andrews - [9:08] Waymo's next city will reveal robotaxi readiness: snow cities like Boston or New York signal software confidence ## Speakers - **Phil Andrews** (host) ## Mentioned Google (company), Magna (company), Nuro (company), Uber (company), Waymo (company), Zoox (company), Jaguar I-PACE (product), Tesla FSD (product) ## Transcript ### Intro **Phil Andrews** [0:00] What happens when Uber doesn't have to rely on coaxing drivers onto the road to drive its passengers? This is a $25 billion a year question, because that is how much we, the riders, paid our drivers in 2023. Not how much we paid Uber and Lyft, but how much of the fares we paid Uber and Lyft were then paid onto the drivers. This is interesting today because driverless rideshare, specifically using Google's Waymo, is here. It's happening, and it's really starting to scale. Check this out: last January, Waymo cars were giving less than 300 driverless rides per day, all confined to Phoenix. ### Scaling fast **Phil Andrews** [0:33] But now that number has ramped to 15,000 rides per day across 4 cities, with Atlanta coming online early in 2025. With new cities coming online and another 1,000 of Waymo's cars being built as we speak, that number is going to climb above 40,000 rides per day by this time next year. Hello, Waymo. Hi, Philip. **Phil Andrews** [1:00] These 1,000 cars being built to satisfy all this new demand are the result of 15 years of research. Founded back in 2009 as the Google Self-Driving Car Project, they went through multiple iterations of hardware on a few different models before finally setting on the electric Jaguar I-PACE. ### Fifteen years of R&D **Phil Andrews** [1:15] Outfitted with radar, lidar, 29 cameras, and a variety of other sensors to do things like hear emergency vehicle sirens. People love these things. This growth isn't gimmicky; this is riders switching to a better product. When you order a Waymo, you don't have to wonder what kind of car is going to come get you, whether or not the driver is going to cancel, whether or not the music is going to be something that you like, the temperature is going to be too hot or too cold, whether there's going to be room for your legs or room for your bags in the trunk. It's also safe. There are so many ifs removed. But one of the things that makes this better: it's private. As soon as you get in one of these rides, you realize that the ride is entirely yours. You can say whatever you want to whoever you want, however loud you want. You can even look however you want. ### Why riders switch Listen, y'all can be scared of driverless vehicles, but I'm currently coming home from my boyfriend's house in my Barbie pajamas, and no driver to judge me for doing it. **Phil Andrews** [2:03] This technology is clearly working, and it's only going to get better. That $25 billion paid to drivers is going to move to a set of companies operating or building driverless vehicles. And we know it's about $25 billion because we have the complete last 12 months of rideshare data from the city of Chicago. Every single rideshare trip that happened within this area—what is the third largest city in the U.S.? It's 136 million rides. Pickup and drop-off locations, total fare, tip amount, mileage, duration, everything. It's about 40 gigs of data. It took 15 hours to put it into a database. It's wild looking through data this complete of this size. I know 26% of people tip their Uber drivers, and the average tip, it's $3. Chicago is such a good representation of a major U.S. city that we can expand this data across the U.S. and find that drivers in the 30 largest U.S. cities account for about $25 billion in the Uber fares we pay per year. Standingright in the path of this money shift is Uber. In a surprise move, they signed an exclusive agreement with Waymo last month, ensuring that if you're in Austin or Atlanta, you can only order a Waymo on the Uber app. Uber has historically been an asset-light company, operating an app and doing their best to own no physical assets. No cars, no maintenance shops, just an office building full of developers. The idea was it'd be more profitable to be the company that paired rider with driver, take a cut of the fare, letting drivers fuss over the car. ### Chicago dataset **Phil Andrews** [3:24] That has officially changed, and for good reason. As Uber customers, we pay on average $3.20 per mile when we book an Uber. $2.20 of that goes to the Uber drivers and their vehicle. Uber keeps about $1. We're paying Uber that price for ridesright now. There's no reason Uber needs to lower it, even if their costs go down. So that $2.20, however much Uber can shave off in costs from it, they could keep, which is why driverless is such an attractive business. The high-end estimate of what it will cost to operate a Waymo vehicle is just $1.50 per mile, everything included: insurance, depreciation, licensing, maintenance, energy, all of it. Immediately, Uber could pocket that $0.70 per mile difference. This is so enticing that part of the exclusive agreement they made with Waymo is to manage physical things, entire fleets, thousands of driverless cars, cleaning and maintaining and storing Waymo's entire fleet of cars. ### Uber's exclusive deal **Phil Andrews** [4:16] They'll need parking lots where they can charge, maintain, and clean these things. But that $0.70 per mile difference is really just the start, because now Uber, in partnership with the big company and what will become companies like Google, can design the entire door-to-door experience. Let me explain this door-to-door experience with an analogy using air travel and why air travel can be so unpleasant. Air travel is largely unpleasant because the airport experience kind of sucks. Half the experience is being in the plane. The other half is all of this: standing in line to check luggage, worried if you'll make it to your gate on time, the bad food, the uncomfortable chairs. There's very little anybody likes about the airport, minus it being acceptable, maybe even encouraged to have a beer at 9:00 a.m. If Delta were able to build airports for the exclusive use of its passengers, those airports would be built much differently. Drop-off would be elegant. The security line wouldn't be like checking into prison, if there was a security line at all. It would be very easy to want to become a Delta loyalist if they had their own custom-designed, comfortable airports with great chairs, lighting, Chick-fil-A, Sweetgreen, Carrabas, and a few pizza shops. The design of the airport would change because they'd be able to win over so many customers tired of using the public airports, and they'd be able to sell additional things they know travelers would want. This is Uberright now, only controlling half the experience. It is impossible to work with millions of drivers and vehicles to unify the experience. Once you get in the car, Uber has very little control over what happens. In that lies a huge opportunity. ### Unit economics **Phil Andrews** [5:44] Working with just one company like a Waymo, they can change how the car is laid out, how it feels, what you see, and what you can buy when you're inside it. Why keep all the seats facing forward, especially when you could put in some type of food and beverage offering? How many people would love to be able to grab a Kind bar, a Diet Coke, or maybe even a hot cup of coffee from the vehicle while riding to the airport? Or Uber could offer a subscription service that costs less as long as you're willing to have ads turned on inside the vehicle. Google is, after all, an advertising company, and a captive audience in a vehicle would be a very attractive offer to local businesses that are around a rider's pickup and drop-off points. But first, they'll need the initial set of cars. There are only 700 Waymo vehicles operatingright now, and as riders will tell you, the wait time to be picked up, it's too long because there just aren't enough Waymos to offer full coverage over the cities that they're operating in. So we start with a question: what's the busiest day of the year for Uber in Chicago? The answer is New Year's Eve, and if you're curious, the slowest day is actually Christmas. But on New Year's Eve, the peak number of rides given in Chicago in one hour was 28,359, with an average ride length of 23 minutes. Back of napkin math says 20,000 cars could probably cover 29,000 rides confined to the Chicago area over the course of 60 minutes. But to cover the average peak hour during a regular week during the year, you would need roughly half that number, about 10,000 cars to cover 17,000 rides on a Saturday night. Either way, it would be a huge upfront expense to build and deploy all at once. At about $120,000 per car, after being outfitted with all the custom hardware, it would be billions to cover an entire major city and tens of billions to cover the 30 largest U.S. ### Airport analogy **Phil Andrews** [7:22] cities. It would take a minimum of 300,000 Waymos to do that. This, and what will be a slow regulatory approval process going city by city, is why Uber's position to remain at the center of rideshare. 1,000 Waymos are being put together at the Magna Factory in Canadaright now to be added to the 700 on the road, deployed in different amounts across the five cities where Waymo operates. Those cars will eat up a portion of the current rideshare demand, but the rest of the demand will have to be filled with cars driven by humans. This is why Waymo is partnering with Uber instead of rolling its own rideshare network. Users want to go to one app to find a ride. They don't want to hunt between multiple, and because Uber accounts for almost three-quarters of the rideshare market, they can offer a multitude of options at different price points to fill in the gaps while the driverless fleet is built up. It also gets them linked into Uber's routing algorithms, which are second to none in predicting where riders will want to be picked up and dropped off, and Uber's dynamic pricing. Waymo needs its vehicles routed efficiently and priced competitively against all the rides being requested on Uber's platform. Being priced correctly and routed efficiently becomes increasingly important the more rides Waymo gives, because Uber is likely paying a per-mile fee to Waymo for each ride it gives. And when you realize how many rideshare miles are given in a year, that per-mile fee adds up very quickly. If Waymo is able to make a profit of $0.15 per mile and they covered 50% of all rideshare rides in the 30 largest U.S. cities, their profit, after vehicle and hardware expenses, would be $4 billion per year, an increase of 20% of what Google makes in profit today. Real quick, a couple footnotes I wasn't able to squeeze into the video. One, there is competition. Tesla FSD is very good despite what the internet says; it's come a long way. There's also Zoox and Nuro. However, Waymo is in the lead. They have paying customers. They're integrated into Uber. People are paying them. They're giving people rides. Also, the five cities Waymo is inright now, they're all Sunbelt cities. The weather is very nice. So one thing to watch to see how fast robotaxis will unfold is where the next Waymo city is. Is it in a city that has snow andice like Boston, New York, or Philadelphia? That will say a lot about the confidence Waymo has in their software. --- This library is powered by PodHood (https://podhood.com), the podcast website platform.