[georgespencer]

> Uber is demonstrating that they can sell $2 bills for a buck and that there's demand for that.

I think the consensus is that rides are subsidised by around twenty cents on the dollar.

Ubers bet is autonomous cars. They pay a 75% fee to the driver. When autonomous fleets launch in the next few years they'll normalise prices at around 50%, meaning your $10 ride costs you $5, and each ride has wafer thin profitability (the additional 5% margin will be chewed up a little by maintenance costs).

And let's be real: Lyft did $600m of sales in 2016. Uber did $5.5bn. This isn't just a race to the bottom on price, it's about building a brand.

[prostoalex]

> Ubers bet is autonomous cars.

Out of the following contenders:

1) GM (+Cruise Automation), Mercedes or VW with their enormous supply networks, vendor lock-ins, economies of scale and ability to massively ramp up a car production when needed.

2) Tesla with their (allegedly unique) battery storage tech, working manufacturing facility and massive amounts of data harvested from thousands of Tesla vehicles on the roads today.

3) Google/Waymo with their unlimited budget, ability to partner up or buy whoever they choose, and nearly limitless amount of software engineering talent they can suddenly relocate to this project when desired.

4) Avis/Budget, Hertz or other fleet management company that already has a relationship with car manufacturer and can own/maintain/wash/clean/park/dispatch a large number of vehicles on the cheap.

5) Uber.

Why is #5 the strongest candidate? Does the dispatch app that displays available cars on the map create such a strong moat that #1, #2, #3 or #4 will never be able to replicate? Does Uber have a bunch of aces up their sleeve that will allow them to build/buy automobile manufacturing facilities on the cheap? Do they have a know-how of maintaining a large fleet and take care of simple things like changing a tire or cleaning the car after the previous passenger puked in it? And do that at scale?

[code4tee]

Quite simply at the rate Uber is going it will crash and burn long before self-driving cars are a thing at scale

[georgespencer]

I'm not arguing that it's a sensible bet. I'm arguing that people shitting themselves laughing because Uber has a model whereby it loses 15% of the cost of each trip do not appreciate what Uber is trying to do.

[dragonwriter]

> Ubers bet is autonomous cars.

Autonomous cars being generally available lowers the barrier to entry for competitors to Uber, and makes their position worse, not better. Uber's bet is Uber controlling the autonomous car market, which is a different and much harder thing.

[georgespencer]

> Autonomous cars being generally available lowers the barrier to entry for competitors to Uber, and makes their position worse, not better.

Why do you think that's true? Uber is investing heavily in autonomous cars because they want to be very early to the market (hence "Uber's bet is autonomous cars"), and because when you have autonomous cars high utilisation is important, they are in food delivery, and interested in shipping and courier services.

I think you're also overestimating the complexity of recruiting a load of former private hire drivers and underestimating the complexity of developing scaled fleet management solutions for entirely autonomous vehicles.

> Uber's bet is Uber controlling the autonomous car market

I think most reasonable people would be able to impute some sort of first mover or early mover advantage into what I was saying.

[jonathankoren]

I think you're underestimating the difficulty of getting a system in tens of thousands of vehicles cheaply and easily.

Do you really think with this many players, that anyone is going to get a huge market? They're not. Uber, even if they manage to survive their lawsuit and build a working system, has plenty of fast followers in the self-driving systems market. So the competition will be at most 24 months behind. Not nearly enough time to acquire and retrofit a fleet of vehicles, and then deploy the fleet in numbers big enough to make a difference to their current cost structure. (Keep in mind the fleet acquisition is a capital expenditure.)

Second, they won't have a large market to distribute this technology, because they don't make cars, and the people that do make cars are already players in this space. I find it hard to believe that Volkswagon, Toyota, or General Motors would simply ditch their internal projects and license with a newcomer that has no manufacturing experience to be a supplier, especially when they themselves would be close to making their own solution which would be able able to be manufactured at scale.

Look, everyone knows that auto manufacturer navigation and entertainment systems suck, yet they still roll their own instead of wide partnering with Apple, Microsoft, Google. My Acura's infotainment system is awkward, slow, and outright user antagonistic, yet there it is.

At most, Uber's self driving kits could be what Alpine is to car stereos. Sure people buy them, but most don't, and if you already have one that works, you probably won't swap it out.

[Nacraile]

>> Uber's bet is Uber controlling the autonomous car market

> I think most reasonable people would be able to impute some sort of first mover or early mover advantage into what I was saying.

I think most reasonable people are deeply sceptical that Uber is going to win first-mover advantage against the likes of Google (who's been working on the autonomy tech for what, a decade, now?), traditional automakers (who have deep experience actually building vehicles), and Tesla (who've been working on the autonomy for at least as long, while also developing manufacturing experience, and have many more vehicles on the road gathering data). Let's not even bother quibbling about that nasty lawsuit.

[georgespencer]

> I think most reasonable people are deeply sceptical that Uber is going to win first-mover advantage against the likes of Google

Sure. Not anything to do with my point, though.

[true_religion]

Autonomous cars are not going arrive for 10 to 15 years. If Uber can't reduce costs by 20% or increase prices by 20% or some combination of the two, then they are making the wrong kind of bet's.

Personally, I believe their bet is on not needing a large advertising budget once enough people know about their service.

[georgespencer]

> Autonomous cars are not going arrive for 10 to 15 years.

Tesla believes its cars are already equipped with the necessary hardware and that the software, including Tesla Network (their announced autonomous Uber competitor), will be going live at the end of 2017.

At this point five years seems like a long time horizon for autonomous cars being on the road.

[rgbrenner]

Saying they're going to release details at the end of this year is not the same as launching at the end of the year.

[georgespencer]

This might be the perfect Hacker News comment. It's a shifting of the premise to something which is axiomatically indisputable, but clearly has no relevance to the argument.

OP said that self driving cars are 10-15 years away.

I observe that the hardware has already been developed and that Tesla anticipates its software being ready within the next 12 months, suggesting that a ten year horizon feels conservative.

HN pedant wades in to make an irrelevant point.