Waymo has no viable business model
provscons.com
provscons.com
The other points that the author raises are obvious things known for a decade, and Waymo doesn't even have to solve them, unless it really wants to.
- The cost of Lidar has fallen a lot already, and is likely to continue to do so.
- Waymo doesn't have to manufacture cars at scale. Carmakers would love to license FSD tech from Waymo and sell premium cars. I would happily buy such a car for $100k.
- Uber/Lyft will fight over themselves to license the FSD tech, Waymo has no need to build its own taxi services.
That's way less intensive than having to build the whole thing from scratch or having it be fully automated.
Humans are good at exceptions. Let the humans do that while the machines do the rote task of "Yep, road still looks like the map. Nothing to see here."
[0] https://www.businessinsider.com/apple-has-7000-fewer-people-...
If amortized cost of hardware + backhaul & processing costs < professional driver wages, then the model works. This is especially true as tech improves and wages increase. There will be a crossover eventually.
I would also point out that there are 2M truck drivers earning $47k per year [1]. That's $94B annually in pure labor cost, and doesn't even account for rideshare & taxis.
[1] https://www.bls.gov/ooh/transportation-and-material-moving/h...
I’m gonna go out on a limb and guess they considered this a decade ago and have a reason to think these aren’t a long term problem.
the idea of tesla having unlimited funding is a joke. They are barely spending across r&d for the whole company. what portion of their r&d is dedicated to self-driving do you thing? even if 100% is self-driving it's still less than the real players in the space.
Who are the real players in the space? There might be bigger players focused more on general AI, but Tesla seem like the biggest player by far, focused on automotive mainly.
In my opinion, despite the cost of the self driving hardware, which is much more than just Lidars, it can still be profitable. Before uber, taxi medallions could fetch $1.5M in New York City. Certainly the costs for Waymo are far lower and scale better with volume.
3D mapping is another cost, but it's also potentially doable since it's a big upfront cost with small ongoing costs. Taking it one (or a few) city at a time could well work.
At least nothing I've read here convinced me otherwise.
This isn’t a cheap strategy, but it’s believable that waymo could own most IP for self driving 10 years from now if they just outlast everyone else. They’re already on year 14 what’s wrong with another 2 decades?
> Before uber, taxi medallions could fetch $1.5M in New York City
I'm not defending the medallion system and it was corrupt af, but there were ordinary decent people who were destroyed by the subversion of this (completely corrupt and anti-competitive) system.
As for the software, eh, the Google Streetview vs. Waymo requirements were one of the weaker points of this article. We don't actually know what data Google collects with their cars (we do know they collect WiFi data[1]), and it isn't inconceivable that they were collecting precise data. It is true that this data ages faster than Google can refresh it, though.
I really believe that the software can scale but it is the real world/physical issues that will prevent this from ever becoming a real business.
[1] https://www.wired.com/2012/05/google-wifi-fcc-investigation/
I don't dispute all their points but I sure disagree with that one.
First of all, you need to own and insure a fleet of rolling depreciating assets. Maybe not as much these days with the crazy used car prices, but still, you have put these cars on the books. You can't even really lease the vehicles because of all of the LIDAR and other self driving modifications that are required.
Next you have to keep these vehicles fueled - either gasoline or electricity, so you are going to need to pay people to handle that. And you have ongoing maintenance for brakes, fluids, etc. And space to park them overnight and in low demand times.
And you also have the interior maintenance. Someone has to clean up the puke and garbage and just general messiness of having humans in the backseat of this car. Throughout the day, every day.
And this has to be repeated in every single city while competing against Uber and Lyft and traditional taxi companies that already know that the driver is probably the least expensive part of this whole system.
Somehow they turn a profit on renting you a car for $80/day which I imagine is less than a tax takes in for fares per day.
$80/day is probably a decent estimate, but once you include the optional insurance upsells and other add-ons (full tank of gas?) I bet they make way more than $80/day per car.
They also have very limited liability for what the driver does in their car while Waymo basically has unlimited liability.
So $260-270/week? Not sure if they have any extra fees applied, however.
You are actually trusting Uber's calculation, especially when they recently had to spend $250 million[1] to try to recruit new drivers?
[1] https://www.marketwatch.com/story/uber-spending-250-million-...
So their rental business is limited, at best.
My browser’s search function spotted only one instance of that string, and it was this question.
[1] https://www.marketwatch.com/story/bankrupt-hertz-shares-soar... [2] https://www.bloomberg.com/news/articles/2020-06-18/what-it-m...
The whole system - whether Uber/Lyft drivers who "relied on their unfamiliarity with capital depreciation calculations"[1] or traditional cabbies who, let's be honest is not a glamorous job, is tenuous at best. Cutting out the lowest cost factors and replacing them with technology and also incurring the fleet costs doesn't really seem like a great business.
[1] https://pluralistic.net/2021/08/10/unter/#bezzle-no-more
I don’t understand how someone who could solve the self driving problem could solve the “plug the car in via simple robotics after it parks itself” problem.
They are also able to park near the city where parking is cheaper, though I guess it could also turn out like it does in China (if anyone is out at 2AM and see all the taxis pushed over to the side of the road with sleeping drivers).
Anyone who lives in a city with a heavy taxi presence already isn’t going to be deterred by these challenges, just the Americans and some other drivin heavy countries will need to adapt.
There have been plenty of proposals, like Tesla's snake robot or inductive undercar charging plans, but no actual production-ready solutions. But even if you had the magical autocharging robot that will not help with gasoline powered vehicles (like all of Waymo's) and you still would need dedicated real estate and charging stations to recharge these vehicles.
The only reason no one finds automated plug-in charging necessary right now is simply because those cars don’t really need it yet.
As far as fueling/charging- Even if you had to pay a few people in the short-term to do the work, it isn't a high-skill job. Being generous, I can't fathom a situation where it could cost more than $2 per car in physical labor. Assuming you could drive more than 200 miles on a charge, that is less than a penny per mile in cost. Also, there are already gas stations in some states with people that will fill your gas tank for you. I am also sure an autonomous solution would be devised if that became a limiting factor. You could pretty easily design a purpose-built system for it.
I agree that regular maintenance needs to be performed, but I don't agree that the companies themselves need to solve the problem on their own. This seems like it would be something to easily contract out to a local company to handle. For standard maintenance and cleaning, there are plenty of companies that clean cars and plenty of companies that perform maintenance on cars already. It should just be a matter of getting contracts in place and making sure that they have software to do their jobs.
I didn't do the math to figure out the cost of a human driver vs the total taxi cost, but even if true the cost isn't trivial. Truckers, for example, can make upwards of $80k per year and that's not even at 50% utilization for the vehicle. If you assume a taxi driver would make $40k/year and they have 50% utilization of the car (I'm making up numbers for the sake of discussion), that is still a significant amount of cost. I would be surprised if maintenance and fuel would be $40k per year regardless of vehicle.
You're right that a Waymo would probably outsource the interior and (probably) mechanical maintenance to local providers, but they would need to deal with that in every city and with the additional technology involved in self-driving, you're probably not going to trust that the local shade tree mechanic will be able to handle that. So you have at least two problems now.
Over The Road drivers do make surprisingly good money, but that is a skilled job with a significant amount of training and regulation. In-town cabbies are often people with few other opportunities and are easily exploitable. The taxi cab business is notorious corrupt and is a big part of how Uber/Lyft could easily replace these services.
The biggest expense people are not seemingly considering is the depreciation of the vehicle. taxis don't last long because the drive so much and so much stop and go. There could easily be $10,000 in depreciation a year.
The diver does things a robo car will not be able to do like help people with their bags, clean the car, responds flexibly to situations as they arise.
Assuming super expensive fuel and terrible efficiency...
Gas car - 20 miles/gallon, $4.00/gallon ($4.00/20 = $0.20/mile. $10,000/$0.20 = 50,000 miles)
Electric car - 200 miles/74kWh [Tesla M3], $0.25/kWh (74 * $0.25 = $18.50. $18.50/200 miles = $0.0925/mile. $10,000/$0.0925 = 108,108 miles)
So assuming a gas car since the miles are the most expensive there, you would be looking at 50,000 miles for $10,000 in fuel. That checks out for me.
Using that assumption.
Uber charges $0.78/mile and $0.27/minute.
Assume only 75% of driving is charged to the customer (going to a fare, going to a charger, etc), that gives you 37,500 miles.
Assuming 45 miles per hour (probably too high) at $0.27/minute = (37,500/45)60 = 50,000 minutes of driving $0.27 = $13,500
37,500 miles * $0.78 = $29,250
So $42,750 in rev. - $30,000 in fuel, insurance, maintenance, and depreciation, you're still $12,750 in profit. Again, is pretty much giving every disadvantage possible. In an electric car scenario you could easily drive 200,000 miles on $10,000 in electricity. Which would multiply all numbers by 4.
I'm not trying to take your numbers too specifically, just illustrating for the sake of discussion.
Good color TV cameras cost thousands of dollars 25 years ago. The one in your cell phone costs about $10 and is much better.
The business model problem is real. A big disappointment has been the lack of self-driving shuttle buses. The cost problem isn't so bad, and at moderate speeds the sensing problem isn't that bad. Shuttle buses are expensive from a labor perspective - they don't carry that many people, and you want a lot of little ones running frequently all day and night. servicing airport parking lots and such. Not a big bus every half hour. There are a few companies selling them, and they work OK on fixed routes at slow speeds. But they're not selling well. There are a few in various places around the world, but not many.
A big name not well known seems to be set to change this. New Flyer is the largest big-bus builder in North America, and they have come out with an automated model.[1] This should be interesting.
>Waymo relies on 3D digital maps.
The cost of HD mapping is vastly overstated in popular press and blogs. It’s not at all a limiting factor. You can update it using the same cars that are part of your service. I've seen hard numbers that back this up.
Google’s unable to keep street view up to date because they have a fleet of hundreds of mapping vehicles spread throughout the entire world, vs hundreds of cars in a small service area that Waymo has.
>Waymo is LIDAR depdendent.
This has been discussed to death - the weather point applies to both cameras and LIDAR, and you can train ML models to deal with noise from weather the same way you can with models running on cameras. Cost has also been addressed a million times before- it’s coming down drastically.
>Waymo is GPS dependent
This one is particularly inane. HD map based AVs localize themselves to the HD maps with their sensors - they’re not actually GPS dependent. The 2-4m of accuracy GPS has is not enough for AVs.
>Internet connectivity issues >Waymo depends on 3D HD maps. Therefore, the vehicle needs to have onboard downloaded maps at all times. Waymo must download these maps in realtime, or you have to download these 3D maps on your Waymo vehicle before you visit a new Area.
I don't know why the author thinks you'll be able to drive robotaxis from city to city. Regardles, AV companies design their cars so that they don't need a continuous internet connection to function. They know cell connections can be spotty or blocked while going through a tunnel. See https://readwrite.com/2017/01/11/google-waymo-security-tl4/
>No Viable Business Strategy >Waymo’s primary business strategy is to sell its technology to the vehicle manufacturer.
This is just wrong and I have no clue where the author got that from. The model is robotaxis, and maybe partnering with trucking fleet operators to outfit their cars.
>Simulated driving is not real world driving.
The data stuff has been discussed before in detail but suffice it to say AVs generally aren’t a data problem - they’re a robotics problem.
>To be successful Waymo has to work very well not only in the USA but also in Canada, Mexico, etc
The revenue from ridesharing services is hyperconcentrated in a few dense urban areas in the US. If you got a robotaxi service working well in just a few of those tiny areas, you could capture a huge amount of revenue to fund tackling other areas.
Unruly passengers without supervision who destroy properties might just be the reason it won't be a viable business.
Who wrote this? Who funds them? How can I know that they aren’t introducing bias to their assessment in their own self interests?
You can easily get the cost savings elsewhere in the vehicle or just target the premium market.
I’m sure George Hotz also has some pithy takes on this.
Anyway, this all completely misses the point. Waymo was born of a DARPA challenge and continued to exist to search for even more DoD money. But we don’t need self driving tanks when we can pilot them remotely.
You could have a central office in a cheap location of the USA for professional drivers to take control as required. We already do it military for drone strikes...