Waymo’s Master Plan

Analyzing Waymo's current strategy and what's to come..

Waymo is now giving 100,000 rides per week but they’re primarily operating in San Francisco, Los Angeles and Phoenix. Rideshare passengers in Phoenix have been able to hail a Waymo via the Uber app since October 2023, but the latest iteration of the Uber and Waymo partnership made waves because Waymo rides will be exclusively offered through the Uber app in Austin and Atlanta at the end of 2024.

This is a change in strategy for Waymo but from my perspective (as someone who’s covered the rideshare industry for 10 years), it makes sense since they still need to figure out their business model before they scale. Obviously, Waymo has a lot of dry powder to do so, but I think their strategy is clear. Focus on the technology and offload all of the hard parts of the business to other people. Running a rideshare company is difficult. Owning and operating a fleet of expensive vehicles is even harder. Waymo doesn’t need to do either one. 

Waymo vs Uber (First Party Rideshare)

Locations: San Francisco and Los Angeles (and Austin for now)

Logistics: Riders need to download the Waymo app, and they can only call for a ride via the Waymo app.

With this business model, Waymo is competing directly with Uber. That means they need to run a rideshare company (a la Uber) and they are ALSO responsible for acquiring, owning, operating, maintaining, charging, storing, etc (the list goes on and on) their vehicle fleet.

I dislike this model because of how hard it is to own and operate a fleet. Want proof? 

When Uber made its foray (no paywall) into the physical world of vehicles with Xchange leasing, they estimated a loss of $500 per car, but ended up with a loss of $9,000 per car. Multiply that times a 40,000 vehicle fleet and you get a $360 million loss for Uber (in 2017 dollars no less). Ouch.

Waymo has 700 cars on the roads right now, but imagine what would happen once they get to 7,000 or even 70,000 cars. The problem with fleet economics is that as you scale, your risk increases linearly, and there are a lot of factors beyond your control. Just ask Hertz who has had major issues with their fleet of Teslas as the value has crumbled thanks to continual price cuts. And remember, Hertz is an experienced fleet operator who is supposed to ‘know what they’re doing’.

Waymo is still losing money on each ride, but I’m also not sure there’s a ton of upside with a first party model. During busy times, Waymo can charge ‘surge pricing’ a la Uber, but they have a fixed fleet so there’s only so much they can increase prices. Could you imagine the headlines if Waymo starts charging $500 for a ride home on a Saturday night vs UberX at $100 (on a normally priced $20-30 ride)? And of course there’s competition from companies like Uber and Lyft, so Waymo can’t charge too much more otherwise folks will opt for competitors. Remember, AVs are supposed to make rideshare cheaper :)

The other issue with Waymo right now is high ETAs. There’s a lot of demand and not enough supply, so the common refrain I’ve seen is that all Waymo needs to do is scale up their fleet and ETAs will come down. But since demand is elastic, this couldn’t be further from the truth. As soon as Waymo scales up their supply, this will induce more demand. And if you scale up supply enough to meet peak demand on a Saturday night, you’ll have too many cars on Tuesday afternoon. This is why there were only 13,000 taxi cabs in NYC for so long (there are now over 100,000 rideshare vehicles in NYC).

So with this model, Waymo has limited upside and massive downside if they screw things up.

Waymo and Uber (Blended Approach)

Locations: Phoenix

Logistics: Riders can request a Waymo ride from either the Waymo app or the Uber app.

From Uber’s announcement (apparently Dara himself wrote this blog post himself):

How to ride with an AV: Starting today in Metro Phoenix, riders who request an UberX, Uber Green, Uber Comfort, or Uber Comfort Electric may be matched with a Waymo vehicle if the route is part of Waymo’s newly expanded operating territory—which includes Sky Harbor International Airport’s 24th St and 44th St SkyTrain locations—and a dedicated Waymo vehicle is available. 

Update your Ride Preferences: Uber customers in Phoenix who are excited about riding in an autonomous vehicle (AV) can increase their chances of being matched with an AV by opting in via the Ride Preferences section of their Uber app under Settings. 

What’s interesting here is that there’s no ‘Uber AV’ option launching, but instead it’s an opt-in/preference. Request an UberX ride and if you’ve opted in; if there’s a Waymo available/with a comparable ETA (I’m guessing); and IF your route is in Waymo’s service area, you will get matched with a driverless rideshare vehicle. 

Those are a lot of restrictions but it also highlights why Waymo plugs in so perfectly into Uber’s marketplace. Uber is great at maintaining liquidity in their marketplace. They’ve been doing it for over a decade and their number one KPI has always been a 3-5 minute ETA when riders open the app. That’s why they have surge pricing, which seems customer unfriendly on the surface, but means you’ll always have a driver available. I wouldn’t normally drive for Uber, but if I can earn more money on a Saturday night, I’ll get off the couch and hit the road.

For Uber, this partnership with Waymo adds additional, high quality, consistent and reliable supply (frankly areas where the Uber product has deteriorated over the past 5 years). Which is valuable for Uber since it fills in baseline demand, and then their variable supply of human drivers can handle peak demand times like it always has.

Even though Waymo has shown with their first party operations that it’s been relatively easy to get tens of thousands of people to download their app, I like this blended approach since Waymo is plugging into an existing well-oiled machine. 

I’m not sure there’s a lot of benefit to Waymo here by having their cars available on both Waymo and Uber though, since there’s not a demand problem on the Waymo app. Traditionally for new rideshare players, it’s been easy to recruit supply (a lot of drivers have issues with Uber/Lyft) and tough to get demand (why switch to a new service when Uber/Lyft work great?!). Waymo has tons of demand, but if they scale their fleet to meet peak demand (about 50-100x on New Year’s Eve per an ex-Uber operations source), they’ll have way too many cars on the road at less busy times. So I wouldn’t be surprised if Waymo ditches this model soon.

Waymo Provides Vehicles to Uber (Third Party)

Locations: Atlanta and Austin 

Logistics: Riders can only request a Waymo ride from the Uber app and Uber is responsible for cleaning and maintenance of the vehicles.

This is the most recent evolution of Waymo’s master plan and I think it’s the best business model so far since it allows Waymo to shed pesky operations, maintenance and cleaning, and instead, focus on improving their product AND reducing the cost.

I’ve always liked to ask AV experts the question, “Who’s going to clean up the puke?!” Well now Waymo won’t have to! Not that Uber knows a ton about cleaning and maintaining vehicles either, but I’m sure they can figure it out (hint: Uber drivers won’t want to clean these vehicles). 

For Waymo, they don’t have to be a rideshare company anymore and only have to be a partial fleet owner. Currently, Waymo vehicles are empty 55% of the time, compared to Uber/Lyft, which are as low as 20% in SF. By plugging into Uber’s network, Waymo vehicles can increase their utilization big time for a fewer reasons:

1. Lower ETAs - Uber operates at a high density with tens of thousands of trips per day in tier 1 markets, so that means any request for a Waymo will be within a 3 to 5 minute ETA (we call this Period 2, drive time to the passenger pick-up after a request)

2. No downtime - On Uber’s network, Waymos will never have any downtime. Period 1 (online, waiting for a request) is costly since you’re literally just sitting there waiting for a ride. Uber knows their baseline demand in all scenarios and can effectively forecast for Waymo to fill all of those rides. This also leverages the asset well since there’s no human driver to get tired and you really only need the Waymo Driver to take a break for cleaning and charging.

3. Expanded domain - Waymo is geo-fenced in every city, can’t go on freeways, struggles with certain pick-up/drop off locations, and has no airport access (except for PHX). This is a bad experience for passengers if they’re looking for a reliable rideshare app. But with Uber, you can set a Waymo preference and any time a trip falls out of Waymo’s operational domain, you’d get a human driver without having to check another app.

If I had to pick a winner between Uber and Waymo with this model, I would give the nudge to Uber since they get exclusive AV access in two cities and have a great answer to investors who are banging down the door asking, “What happens if Waymo disrupts you?” Looks like Waymo might need Uber way-mo than Uber needs Waymo. 

This deal also bolsters the ‘Uber as a mobility platform’ case where AV operators can just plug their vehicles into Uber’s network whenever they need high utilization. It’s one thing to tell people this will work and another when you can show it. 

Waymo Works With OEMs

Locations: ‘Coming Soon’

Logistics: OEMs purchase the ‘Waymo Driver’ stack up front and can then plug their vehicles into Uber, Lyft or build a mobility platform of their own (terrible idea but I’m sure a few of them will try).

What if I told you there was a way that Waymo could cut all of their capital expenses, operating expenses and just focus on selling a software product with infinite margins?

Well Waymo hasn’t publicly announced anything like this yet but they’ve hinted at it, and have been inking new OEM partnerships left and right. From the recent Waymo/Hyundai partnership announcement:

The companies declined to disclose financial terms of the partnership, but confirmed Waymo will purchase and own the vehicles.

Right now, Waymo’s tech is expensive so it makes sense that they are still taking most of the financial risk in these deals with OEMs. But I wouldn’t be surprised if each successive deal has more and more favorable financial terms for Waymo. I think we’ll know that the ‘Waymo Driver’ has a potential path to profitability on a unit economics basis when an OEM buys the hardware stack from Waymo, puts it on their vehicle, pays a monthly SAAS subscription to Waymo and starts giving driverless rides. 

This type of partnership seems like the best outcome for Waymo since they don’t have any control over the cost of the vehicles and I’m not sure they want to get into the car business with a vertical integration. OEMs on the other hand have shown that they love investing billions of dollars in the next big mobility thing so why not Waymo?