Autonomous Car Progress

[EVNow]

Robotaxis don't even have to be EVs so I'm not sure why it's relevant.

Because, by the time Robotaxis are ready, lot of countries / cities would have banned ICE.

 

[diplomat33]

But I think K’s point (can I call him that???) is that there isnt just some dial any of these makers can turn and suddenly turn out vast numbers of BEVs .. for a start, there is a huge battery availability problem (why do you think Tesla have their own battery facilities?). Sure, anyone with deep enough pockets is capable of building EVs at scale .. eventually. But that takes years of infrastructure build-out, logistics, contracts and 3rd party OEM ramp-up.

That is true. And I get that switching to EV manufacturing is not easy. But I think there is an assumption here that a successful and profitable robotaxi company must have their own manufacturing and must be able to crank out millions of robotaxis. The fact is that we are still in the early days of robotaxis. We have not even solved the tech part yet. So I think it is a bit premature to be talking about mass producing millions of robotaxis. Right now, any company that could just deploy 1000 driverless robotaxi would be a winner. Sure, at some point when we have thousands of robotaxis everywhere, the company that can produce millions of robotaxis themselves is going to beat the company that has to outsource production and maybe only gets a few thousand. But we are not there yet.

 

[Supcom]

Making a big fleet of RTs that aren't EVs is economically nonsensical at this point.

Please help me here. Existing taxis/Uber/Lyft are economically feasible with a human driver and non-EV cars. So, if it not economically feasible to have robotaxis without EVs, then the AV technology must cost more per year than the human driver, thus requiring additional cost savings.

If that's the case, then a human driven EV taxi should be able to undercut the robotaxi and still make money.

 

[EVNow]

Please help me here. Existing taxis/Uber/Lyft are economically feasible with a human driver and non-EV cars. So, if it not economically feasible to have robotaxis without EVs, then the AV technology must cost more per year than the human driver, thus requiring additional cost savings.

Uber still doesn't make money. Waymo claims it eventually will turn a profit.

 

[Knightshade]

That is true. And I get that switching to EV manufacturing is not easy. But I think there is an assumption here that a successful and profitable robotaxi company must have their own manufacturing and must be able to crank out millions of robotaxis.

MUST? Nope. But there'd be obvious, significant, financial advantages if they could.

The fact is that we are still in the early days of robotaxis. We have not even solved the tech part yet. So I think it is a bit premature to be talking about mass producing millions of robotaxis.

Agree completely. But once it IS solved, assuming it's solved GENERALLY and not "Solved for one tiny area in one town, and maybe in another 5 years it'll be solved for another tiny area in one town" then you DO need millions of robotaxis.

Right now, any company that could just deploy 1000 driverless robotaxi would be a winner.

A winner of what?

Certainly not profitability.

Sure, at some point when we have thousands of robotaxis everywhere, the company that can produce millions of robotaxis themselves is going to beat the company that has to outsource production and maybe only gets a few thousand. But we are not there yet.

I agree we're not there.

But now it seems you've come around to agreeing we care about having an advanced factory when we ARE there at least

Please help me here. Existing taxis/Uber/Lyft are economically feasible with a human driver and non-EV cars.

No, they aren't.

Both companies have been losing billions of dollars for their entire existence. Uber DID have ONE profitable quarter last year under GAAP rules thanks to gains from an external investment- not from their actual business. Further, nearly half their revenues now are from food delivery, not taxi service.


Hope that helps you!

 

[Daniel in SD]

Because, by the time Robotaxis are ready, lot of countries / cities would have banned ICE.

That won't happen until there is enough production capacity though. Voters will not accept that mandate if it means fewer annual vehicle sales (and therefore much higher prices). California plans to go EV only for passenger cars in 2035 and for commercial vans in 2045. Of course I'm a bit skeptical we'll see widespread robotaxis before then too. haha.

 

[diplomat33]

A winner of what? Certainly not profitability.

Why not? If the cost of operating of 1000 robotaxis is less than the revenue generated by the 1000 robotaxis, they would be profitable. I think it should be possible to be profitable with 1000 robotaxis if costs are low enough and you are in a good service area that generates lots of rides.

I agree we're not there.

But now it seems you've come around to agreeing we care about having an advanced factory when we ARE there at least

Oh I agree that when we are "there", having an advanced factory will be key. But it could be 5-10 years before we are "there". Are you saying that in 5-10 years, Tesla will still be the only company with an advanced factory able to mass produce robotaxis? That would surprise me.

And maybe that was not your intention in your original post, but my "who cares" was just pushing back against the notion that Tesla will have a ride-hailing network of millions of driverless robotaxis by 2023 just because they have advanced factories that can mass produce lots of EVs. Like I said, there is a big difference between mass producing EVs and deploying a driverless robotaxis in a ride-hailing network.

 

[Supcom]

No, they aren't.

Both companies have been losing billions of dollars for their entire existence. Uber DID have ONE profitable quarter last year under GAAP rules thanks to gains from an external investment- not from their actual business. Further, nearly half their revenues now are from food delivery, not taxi service.

I guess Uber and Lyft have shown that you don't have to turn a profit with your services to be economically feasible. You just need investors! I expect Waymo will be the same.

 

[Knightshade]

Why not? If the cost of operating of 1000 robotaxis is less than the revenue generated by the 1000 robotaxis, they would be profitable. I think it should be possible to be profitable with 1000 robotaxis if costs are low enough and you are in a good service area that generates lots of rides.

Because making only 1000 of a custom car is insanely expensive. You don't just need to make revenue > operating costs, you also need to pay for the car

Further- you need to do all that while keeping your prices low enough that you get enough business.

Oh I agree that when we are "there", having an advanced factory will be key. But it could be 5-10 years before we are "there". Are you saying that in 5-10 years, Tesla will still be the only company with an advanced factory able to mass produce robotaxis? That would surprise me.

In 2013 VW said they'd be the market share leader in EVs by 2018.

I'm sure some people were surprised when they weren't- but it still happened.

In October 2017, GM announced that it would release 20 new EVs by 2023- so far they've released....one.... (the EV hummer) and are making roughly 1 a day. Technically I guess you can consider bumping that to 2 with the Caddy Lyriq they haven't delivered yet but are supposedly starting to build.

In Q4 2021 the President of the United States publicly claimed GM was leading in electrification of cars. A quarter in which GM sold 26 total EVs.


Folks have been "surprised" by how badly legacy has been doing EV manufacturing for a lot of years now.

And there's been lots of press releases about how IT IS DIFFERNET NOW, but relatively little on the ground evidence that it is.


VW has probably done the LEAST awful job in actually making some moves finally, but besides being MANY years behind their stated goals, they've got massive headwinds on the labor and cost fronts among others--- to the point they've said they're massively scaling back models in the coming years to focus on just their highest margin stuff.


Now, EVENTUALLY, you'll have other companies that can genuinely do EVs at scale. I'm unsure that'll be true in 5 years, but should be by 10. Of course that's only half the issue.

If one company has a 30% gross margin on their vehicles, and runs their own RT service-- while another has a 5% gross margin and is a middle man for an RT company-- that first company is going to be able to either make massively more per RT, or heavily undercut pricing on RT services, compared to that second company.

And maybe that was not your intention in your original post, but my "who cares" was just pushing back against the notion that Tesla will have a ride-hailing network of millions of driverless robotaxis by 2023 just because they have advanced factories that can mass produce lots of EVs. Like I said, there is a big difference between mass producing EVs and deploying a driverless robotaxis in a ride-hailing network.

I think best case Tesla has wide rollout L2 on city streets by 2023 personally.

But hey, if ANYBODY has working-at-scale RT tech by then, and Tesla doesn't, they can always offer to mass produce RTs for THOSE guys.... faster and cheaper than your GMs of the world could

 

[diplomat33]

Because making only 1000 of a custom car is insanely expensive. You don't just need to make revenue > operating costs, you also need to pay for the car

Further- you need to do all that while keeping your prices low enough that you get enough business.

Paying for the cars and paying for the retrofits is a one-time initial cost. If you can generate consistent revenue above operating costs, you will recoup that initial cost and make a profit going forward.

 

[EVNow]

I expect Waymo will be the same.

Only if they spin off. I don't think Google wants to endlessly support a cash sink .... even monopolies have their limits.

 

[EVNow]

That won't happen until there is enough production capacity though.

True. Good thing Waymo/Cruise aren't going to be ready with robotaxis anytime soon

 

[Daniel in SD]

Making a break for it.

https://www.reddit.com/r/SelfDrivingCars/comments/tzk3f4

 

[Bitdepth]

Funny but looks more like remote assist telling the car to cross the traffic and pull over in the next section of the road so as not to impede left turn lane and traffic. Still funny though as the cops don't know how to react.

 

[Daniel in SD]

Funny but looks more like remote assist telling the car to cross the traffic and pull over in the next section of the road so as not to impede left turn lane and traffic. Still funny though as the cops don't know how to react.

Yeah, could be. It's funny how fast it takes off. My interpretation was that the cars behavior was entirely autonomous until remote assistance hit the big red button causing it to pull over and turn on the hazards. Really not sure though.

 

[Daniel in SD]

Instagram post. Pulled over for driving without headlights on. Hidden danger of LIDAR. SMH.

B-RadAbroad on Instagram: "Confused SFPD pulling over an #autonomousvehicle in the Richmond District!! Then it tries to take off!! 😆 🚔 @londonbreed @cruise #sanfrancisco #bayarea #highspeedchase #copwatch @waymo @kron4tv @generalmotors @sfgate @only

3,844 likes, 286 comments - b.rad916April 1, 2022 on : "Confused SFPD pulling over an #autonomousvehicle in the Richmond District!! Then it tries to take off!! 😆 🚔 @londonbreed @cruise #...".

www.instagram.com

Cruise's guide for first responders: Driverless Deployment Program Guidance for First Responders (Updated 10.28.21).pdf

 

[Daniel in SD]

https://twitter.com/x/status/1513181598140796936

 

[Jeff N]

First responder training….

First Responders

How Cruise works with first responders and partners in public safety to keep our communities safe | Cruise

www.getcruise.com

 

[Doggydogworld]

Because making only 1000 of a custom car is insanely expensive.

Agreed. That's why Waymo's first 1000 weren't custom. And IMHO their next 10,000 shouldn't be, either. GM and AmaZoox jumped the gun with podcars. Better to run a service a while to learn customer needs and encounter unexpected operational issues.

Folks have been "surprised" by how badly legacy has been doing EV manufacturing for a lot of years now.

Now, EVENTUALLY, you'll have other companies that can genuinely do EVs at scale.

Legacy has zero problems building EVs at scale. They just can't sell them.

But hey, if ANYBODY has working-at-scale RT tech by then, and Tesla doesn't, they can always offer to mass produce RTs for THOSE guys.... faster and cheaper than your GMs of the world could

Tesla, a low cost manufacturer? Hilarious. Their cheapest car is $50k. They have to hard-wire manufacturing lines for a single variant and run one color per day. Manufacturing is not their strong suite, to put it kindly. Their overwhelming advantage is marketing. Put a Ford, Buick or VW badge on a Model 3, cut the price 10k and you still couldn't sell 100k of them.

Doesn't matter, though. Your messiah says Robotaxis will gross 100k per year for 10 years. A 3-5k upfront cost differential is a rounding error.

 

[Knightshade]

Legacy has zero problems building EVs at scale. They just can't sell them.

Complete nonsense.

GM, we keep being told, is the "leader" but has never made more than ~25k EVs in a year. Tesla was making that many in a week by end of last year, and just opened 2 more factories.


Ford isn't doing any better, and keeps shutting down MachE production for parts shortages.

nobody in legacy, apart from arguably VW (who is YEARS behind its own stated schedule) can build EVs at scale today.

Tesla, a low cost manufacturer? Hilarious. Their cheapest car is $50k.

You are confusing cost and price.

They're making >30% margins on those cars, while GM has been losing money on every Bolt they build.

. Manufacturing is not their strong suite, to put it kindly

Or more accurately, to put it hilariously wrongly.

. Their overwhelming advantage is marketing.

On which they spend $0 a year?

Weird take.