Tim O'Reilly explains how self-driving cars will bankrupt Uber

Originally published at: https://boingboing.net/2017/10/02/externalizing-psychopathic-cor.html

No, the cars will be very lightly utilized because they have to have enough cars for peak demand. It only works if those autonomous cars belong to someone else who will take on that risk.

He has a valid point. Then again, Uber could simply serve as a shared platform for owners of autonomous cars to auction off the underutilized time of their vehicles.

(I freaking hate Uber, on principle. Just pointing this out.)


Then the question is - in urban areas, are people even likely to own autonomous cars? Because I can see car ownership easily becoming something that’s only for more rural dwellers (where things like Uber don’t work well) and the wealthy who don’t want to share a car with anyone/collect cars. Most urbanites will find that it’s much cheaper to “hire” an autonomous car when needed than own one (especially if rising income inequality makes car ownership that much less affordable). Owning a car and “renting” it out through something like Uber probably wouldn’t make much sense, either - the owner presumably would have the responsibility of keeping the car clean, which means taking time throughout the day to do that, for what would be quite little pay (or paying someone else to do it, which makes renting out the car even less financially worthwhile). They’d be competing with a company that does have a fleet of cars and a cleaning/servicing team that can benefit from economies of scale and offer a more consistent service. From what GM has been saying, they’re likely to be one of those companies.


Yeah, I understand the dynamic. But somebody has to own the cars everybody will just be hailing through their apps. Businesses, even small ones, will presumably own some vehicles for their business use (a plumber’s ride by day, automated taxi by night). Some people like long trips/hiking/have a country house. There will be legacy stocks from the transition period. All of these cars - plus the large fleets of professional operators - could be plugged into the sharing system. And then, ideally, the market could help us decide how many cars do we really need. And this number will probably be noticeably lower than status quo, decreasing the need for parking spaces etc. So potentially good news all around.

Given that Uber is apparently a Ponzi scheme losing $2B a year and being kept afloat by investors, I don’t think there’s a good reason to think about what will happen to the company when driverless cars become practical. It will be a wonder if Uber is still around by then. Then again, it’s kind of a wonder that it’s around now.


It’s already happening with younger people and conventional cars. Of course, corporate America is spinning it as “Millenials don’t want to grow up,” but if you live in a city with good public transit and car-sharing services like Zipcar why put all that money into buying and maintaining a wasting asset that sits idle 90% of the time? Heck, I’m in my 40s and that’s the way I think, even though I’ve been driving since I was 16.


I recall reading an article a while back which basically said, “Uber is in a race to invent self driving cars before it goes bankrupt”.

How that solves the problem, I have no idea. They’ve shuffled off the costs of buying and maintaining a fleet to the drivers, whom are probably their major expense now.

So they remove their major expense and, what, then switch back to owning and operating a fleet? I don’t see how they’ll be around in a few years either way.

A novel theory on how they could still exist, though: if they can hold off until Tesla achieves full automation, and as well a much larger market share, Tesla could acquire them for a song and then use the platform to let their customers rent out their cars while they are not using them.


The model to look at is electric utilities. They have a daily/seasonal peak cycle and have to own/maintain infrastructure to meet peaks. Home owned solar that can sell back to the grid is home-owned cars that Uber could lease.

The time the home owned cars are most likely to be needed by their owner are during the peaks. So like a utility has a quick-starting peaker plant they bring online to meet needs, Uber could own a smaller fleet they use to boost vehicles during peak, but then park somewhere off peak. Even offline their is a cost but it’s less than operating cost.

And since the need for cars off-peak is lower they could pay car owners they rent from lower rates because of high supply of cars.

I use “Uber” as a generic for Uber/Lyft/even taxi’s in this case. For Uber specifically they’ll still need to raise rates to make money.

I was going to point this out. Uber courts investors with autonomous vehicles, because otherwise it wouldn’t exist. There are numerous famous individuals forming famous companies that exist basically to be invested in and speculated on before they close their doors suddenly and the people buying the hype uncritically are taken for a wash.


I agree. I think of it like this:

Why park your car, when you can just order it to become a taxi while you don’t need it? This means tons of taxis on the road and the cost of taxis goes way down.

Once the price goes down far enough it becomes cheaper to taxi everywhere and many stop buying cars. This continues until a balance is met between car ownership and taxi pricing.

Less cars on the road mean less traffic and shorter trips. Shorter trips mean lower fares. Lower fares again change how many people bother to buy cars.

Car manufacturers will start to design cars specifically for taxi use. They can cut out the middleman and operate fleets of cars, taking advantage of economies of scale to keep costs down.

I see a role for Uber for acting as the marketplace. I don’t want to have to check the GM app, the Tesla app, etc.

Right, but what I’m saying is, the few people who find it useful to own a car, will they be enough to be the auto suppliers for Uber, and would they find it worthwhile to do so? Does the plumber really want to empty out their van each night and clean it up in the morning (and give a cut to Uber) when the pay for that service is being set by a larger company (like General Motors who are seeing cars as a service rather than product in the future) that can do the same thing for a lot less thanks to economies of scale? Also cars designed from the ground up as a service are going to be different from a repurposed personal/business vehicle, so it likely wouldn’t just be a matter of plugging a plumber’s van into the system. (And would you really want to be driven around in a plumber’s van, when autos that were actually designed for that specific purpose would also be available during hours when there would likely be a glut of vehicles?) “Legacy stock” are likely to be not-fully-autonomous vehicles, as I can see this dynamic taking over really quickly once fully autonomous vehicles hit the market; companies are preparing for this future already. I just don’t see the economics working out for anyone with less than a fleet of devoted stock designed expressly for this purpose.

Yeah, I’m seeing autonomous cars as essentially amplifying existing trends. There are multiple financial, social, etc. trends why people don’t have cars, and those trends will likely also continue as well, independent of autonomous vehicles. Plus, these autonomous vehicles won’t just be autonomous versions of the automobiles people own now, but autonomous public buses, shuttles, single-passenger mini-cars and vehicles carved up into separate compartments (so you can share a cheap, late-night ride without worrying the other passengers are going to abuse you, for example).

Yeah, I mean - General Motors has come out and said they see their future as providing cars as a service, not a product. So even if Uber (etc.) did turn their model upside down and had their own car fleet, they’d likely be competing with whoever they bought that fleet from…

But that presumes that people privately own cars in urban areas - and given the existence of autonomous vehicles, why would they? I don’t think just how big a game-changer autonomous cars will be has sunk in. It feels like a discussion happening at the introduction of automobiles where people are saying, “When the car runs out of gasoline, you hook your horses up to it and have them pull it…” You either have private car ownership or autonomous cars, not both (at least, not in urban areas, for the vast majority of people).

I think the thing is, there will almost instantly be a strong incentive not to have a car (if you’re urban, etc.) the minute fully autonomous vehicles are introduced. Now, in American car culture, it may take a while for people to figure that out, but…


Well, that’s exactly what Uber is calculating with and what that O’Reilly guy is disputing. If no one besides specialized mass operators (which Uber wants to become) can keep this up, there will be no competition since the oligopolies will be able to undercut everyone’s costs and Uber will be doing just fine - they just have to switch from the independent contractor model to fleet ownership. Which is not such a big deal since their main advantage is the app service. (Labor exploitation, fooling people and skirting regulations are just temporary cherries on top of their model.)

I think the core overlooked principle is that as long as there is demand (i.e. as long as humans are around), somebody will provide the services. The peak fleet will be exactly of the size the customers are willing to pay for (this is one of the problems capitalism actually solves pretty well).

Uber has negative capital…

1 Like

No it doesn’t. I don’t even think “negative capital” is an accounting possibility. What it has is a negative revenue stream net income because it’s deliberately using dumping prices to undercut competition and build up a customer base, partially masking this dumping tactic through use of independent contractors and its murky status as a “mere app service.” As far as capital is concerned, they are still swimming in investors’ money.

Their model probably wouldn’t be currently competitive without all the dirty tactics and it’s questionable if they ever have a chance of achieving the desired ultimate stability and market position through a combination of new tech, economy of scale and networking advantage, as planned. But even if it’s a long shot, it’s not fundamentally impossible.

1 Like

Why will people in urban areas want to own when autonomous cars are available?

  1. They want to. Breaking people from thinking they need to own a vehicle is going to take way longer than people think. And that’s after the way longer than people think time it’ll take to perfect autonomous vehicles to the level of service we’re talking about. “But we’re 80% of the way there!” Like all tech projects the last 20% takes 90% of the time.
  2. I think the biggest reason is that currently Uber and Lyft pricing is far too low to be sustainable in an effort to keep out competitors and build loyalty to the brand. In other words, I think for a number of people owning will be cheaper than taking Uber and Lyft’s everywhere.

People still buy vinyl and CDs and digital music instead of streaming. People still buy desktop computers when laptops are available. People still buy laptops when tablets are available. There is still a market for buggy whips. Markets may shrink and become more niche but rarely disappear.


They don’t have negative revenue, that is an impossibility. They hold negative margin and owe money to their stakeholders. They are less than no profitable, their brand does not hold value to anyone looking to buy. They have no assets to sell to fund buying a fleet without selling their brand.

Uber in no way whatsoever has the ability to buy a fleet of vehicles.


My mistake - they have negative net income, not revenue.

That said, their current market valuation is still somewhere around $ 50 billion. Which is basically the value of their intellectual property, customer base and goodwill. I don’t think they will have a problem financing a gradual purchase/lease of their own fleet.

Except that Uber will be competing with General Motors - so how do they do better than the company that supplies them with their cars? Uber becomes an unnecessary middleman. At best, they get bought up (at a tiny, tiny fraction of current value) by a car company to provide the ride-hailing software, but I don’t even see that, as it’s the easy bit if you have local fleets of cars everywhere. Sure, someone will provide the service, but it won’t be based on other people’s vehicles, and there’s really no path that allows Uber to be the one that does so when they’re up against the actual car manufacturers.

Sure, because the two things are very different. I repeatedly specified where autonomous vehicles are going to be the game-changer - car ownership and non-autonomous vehicle ownership will continue, but increasingly in the minority and, importantly, where Uber doesn’t operate now anyways. (The dynamics of a niche business are radically different, too.)
As I said, US car culture will prevent some people from immediately realizing that they don’t need to own a car anymore, despite what are likely to be some pretty obvious advantages. How long it takes for autonomous cars to show up in the first place is irrelevant, as the premise of the conversation here is that it happens.
Autonomous car rides are going to be cheaper than Uber/Lyft - there’s no driver cost, maintenance will be lower (because they’ll be electric), and various new forms of shared rides (e.g. autonomous buses that change route based on where people are and their destinations, making them far faster and more direct) will drive down costs. There’s certainly no way that owning a car that sits doing nothing for 90% of the time will be cheaper than paying for just the hours you need. If the ride-organizing company is making that much profit, you’ll see car co-ops start up.

Problem is it will have to invent roads they’re safe to drive on and other drivers tolerant of their shenanigans, before then, too.