Waymo

[DanCar]

- Technology cost is "company secret" but cost of Waymo Driver is significantly less than expectations. Roughly in the range of cost of car. Not extraordinary expensive technology when integrated with electric vehicle. Goal is to have 100% EV fleet.

I think he said cost of "an electric car". About 22 minutes into the video. They are doing Jaguar electric which is north of $70K MSRP.
I think he is saying the hardware cost for the Waymo driver is about $70K. If you take into account Waymo has lost billions of dollars and continues to lose billions, then the total cost is much higher.

 

[diplomat33]

- Technology cost is "company secret" but cost of Waymo Driver is significantly less than expectations. Roughly in the range of cost of car. Not extraordinary expensive technology when integrated with electric vehicle. Goal is to have 100% EV fleet.

I think he said cost of "an electric car". About 22 minutes into the video. They are doing Jaguar electric which is north of $70K MSRP.
I think he is saying the hardware cost for the Waymo driver is about $70K. If you take into account Waymo has lost billions of dollars and continues to lose billions, then the total cost is much higher.

No, I think you misheard him. He says the cost is "in the range of the cars we are driving". He does not say "electric cars". He does say that it is not "extraordinarily expensive when integrated into an electric car". So he mentions electric cars but in the context of the cost not being high when integrated into an EV.

Here is the direct transcript:

"The technology costs is something we keep tight within Waymo for competitive reasons. But the cost of the Waymo Driver is significantly lower than I think the expectation is. Just to give you a general sense, it's in the range of the cost of the cars we're driving. So it's not an extraordinarily expensive piece of technology when integrated into an electric car, which is our priority right now."

Waymo is using 2 different cars right now, the I-Pace and the Pacifica. He just says "cars we are driving" without specifying which one. So it is unclear which one he is referring to.

The Pacifica only cost $35k. The I-Pace is $70k. Since he says the cost is in the range of the cars they are driving, the cost could be around $35k - $70k.

He also mentions that they expect to get 300k, 500k or even as much as 1M miles out of the cars (presumably talking about EVs since he mentions going 100% EV). So I think he is taking into account the cost spread out over the entire life of the car.

 

[EVNow]

Each city has different driving challenges. Waymo Driver does best when we have time to learn and master the environment. It takes time to make sure we are driving well in those locations.

Exactly. That’s why Waymo can’t easily scale. Their technology is simply not scalable by Silicon Valley standards.

Waymo is using 2 different cars right now, the I-Pace and the Pacifica. He just says "cars we are driving" without specifying which one. So it is unclear which one he is referring to.

The Pacifica only cost $35k. The I-Pace is $70k. Since he says the cost is in the range of the cars they are driving, the cost could be around $35k - $70k.

So, the driver costs the same as cars I.E. total cost in 2x the cars ?

 

[diplomat33]

Exactly. That’s why Waymo can’t easily scale. Their technology is simply not scalable by Silicon Valley standards.

Making sure that FSD works well in different driving environments is a challenge for all AV companies, not just Waymo. Tesla can say their solution is general but we know AP/FSD works better in some areas than others. So Tesla has the same challenge of getting FSD to work consistently well in different cities. It takes time for Tesla to get FSD right in different driving environments too. Tesla's FSD also needs time to master a different driving environment.

So, the driver costs the same as cars I.E. total cost in 2x the cars ?

Right now, it would seem so, yes. But costs will surely come down in the future. Also, Krafcik might have been talking about the 4th generation hardware since that is what Waymo One in Phoenix is using now. Waymo has said that the 5th generation is much cheaper and is designed to be manufactured at scale. I suspect the 5th generation is much cheaper.

Also, if the car lasts up to 1M miles, costing the same as the car would not be too bad.

 

[EVNow]

Tesla's FSD also needs time to master a different driving environment.

Absolutely.

In general I think with Waymo and Tesla, we can see where they are at. There is no secret version that is better. Waymo can’t yet handle rain and Tesla can’t handle turns well.

 

[EVNow]

Waymo can scale this up. They are starting with a high end system and reducing costs as they go, like Tesla did by starting with expensive cars and trying to make them cheaper.

Scaling is not an issue from HW perspective. Atleast that’s not what i’m Talking about.

How fast can they scale geographically is the question.

The answer is not quickly - considering they left out important retailers just a hundred ft from their boundary.

 

[diplomat33]

The answer is not quickly - considering they left out important retailers just a hundred ft from their boundary.

I think that is a nitpick. They just happened to stop the geofenced area short of a particular retailer store. I don't think it necessarily means anything.

After all, there are stores all over Phoenix. If Waymo had extended the geofenced area a few 100 ft to include that Costco you want, there would be another store somewhere nearby that would be left out. The fact is that any geofenced area is going to leave someone out no matter what.

 

[ThomasD]

What does Waymo charge to ride in their Autonomous Vans with no backup driver

 

[diplomat33]

What does Waymo charge to ride in their Autonomous Vans with no backup driver

It varies but I calculated the average is $1.95/mile.

You can see the data yourself here: Waymo Rides

 

[ThomasD]

That is not that cheap enough if they want people to give up their cars, Even taking Waymo to work and home say 5 miles each way would run in the thousands of dollars every year. Also getting a Waymo to take you everywhere else you needed to go during the year would cost a fortune

 

[diplomat33]

That is not that cheap enough if they want people to give up their cars, Even taking Waymo to work and home say 5 miles each way would run in the thousands of dollars every year. Also getting a Waymo to take you everywhere else you needed to go during the year would cost a fortune

I don't think the ride-hailing in Phoenix is designed to do that yet. After all, they just really launched to the general public a couple months ago. It is still very early. It is not the "finished product" yet. I am sure, the price will come down as Waymo improves the tech, and scales up. Plus, the service area will expand too over time.

 

[ThomasD]

Even if Waymo scales up and has cars in all of the 50 major urban areas in the country they would have to keep track of millions of cars. That would be an expensive feat. I wish them luck. I would enjoy riding in one every once in a while.

 

[JJRicks]

Wow - much smaller than I thought. They are also carefully avoiding high traffic areas. There is a Costco right on 87, they have carefully left out. One would think if you want to run a robotaxi service, you want to include places like Costco. Looks like a gerrymandered congressional district.

Looks like the whole area they are covering is the size of a medium sized university campus.

Google Maps

Diplomat over here fighting the good fight, keep it up!
By the way, the service area used go way past 87 and all the way out to McQueen, the old map is on their website: Waymo One App Experience – Waymo
It's temporarily smaller right now until they let the safety drivers run with passengers again.

 

[JJRicks]

In case the full size of the service area doesn't come across easily on the map, this next video is a traversal of the entire thing. It's 40 minutes long.

 

[diplomat33]

Waymo CEO: Building safe driverless cars is harder than rocket science

Last year was the most significant yet in Waymo’s 11-year effort to develop a driverless car.

The Google sister company raised $3.2 billion, signed deals with several partners, and launched the world’s first truly driverless taxi service in Phoenix, Arizona.

Even so, the widespread rollout of fully autonomous vehicles remains slow, staggered, and costly.

“It’s an extraordinary grind,” said John Krafcik, Waymo chief executive, in an interview with the Financial Times. “I would say it’s a bigger challenge than launching a rocket and putting it in orbit around the Earth... because it has to be done safely over and over and over again.”

Why did you start another Waymo thread? We are already discussing this very article in the current Waymo thread.

But Krafcik is right. It's easy to put out "beta" software that requires driver supervision. Waymo is doing true driverless FSD where the car has to be able to drive truly autonomously with no safety driver and rarely get into any accident. That is a lot harder.

 

[diplomat33]

That is not that cheap enough if they want people to give up their cars

I think this is a false assumption. Why does Waymo have to convince people to give up their cars? Regular taxis have not convinced people to give up their cars, have they? People still own cars and also use taxis in some cases. Waymo is a robotaxi service. Just like any taxi, their business is simply offering rides to people who need one for a fee. As long as Waymo is able to meet the demand and make a profit on each ride, I would say their business model would be a success. Waymo is competing with other ride-hailing services. They are not competing with personal car ownership yet.

Now, Waymo has talked about offering robotaxi leases at some point. When they do, then we can see if the robotaxis leases convince people to give up their cars. But we are not there yet.

 

[JJRicks]

Apparently posting a link banishes my comment to the shadow realm, so I'll let Diplomat take care of that. Here's the video timestamps though

00:12
01:40 Unexpected slowdown for unknown object
04:15 Two traffic lights in close proximity, nice edge case
07:33 Beginning of main city streets
08:17 Demonstration of object permanance issues
11:05 Map glitches
14:42 Police cars and construction
15:41 Ambient noise demonstration (the sensors are silent, here's proof!)
17:16 Where the shaking noise is actually coming from
26:01 Cool demo of a car on a trailer
27:25 Heavier-ish traffic starts
29:16 Still seeing a car that was under a tarp
33:03 Unprotected left
33:10 Misses the destination, does a go-around
36:07 Pedestrian misclassification

 

[diplomat33]

Here is the video that @JJRicks tried to share. It is a 39 minute unedited driverless ride.

00:12
01:40 Unexpected slowdown for unknown object
04:15 Two traffic lights in close proximity, nice edge case
07:33 Beginning of main city streets
08:17 Demonstration of object permanance issues
11:05 Map glitches
14:42 Police cars and construction
15:41 Ambient noise demonstration (the sensors are silent, here's proof!)
17:16 Where the shaking noise is actually coming from
26:01 Cool demo of a car on a trailer
27:25 Heavier-ish traffic starts
29:16 Still seeing a car that was under a tarp
33:03 Unprotected left
33:10 Misses the destination, does a go-around
36:07 Pedestrian misclassification

 

[diplomat33]

Even if Waymo scales up and has cars in all of the 50 major urban areas in the country they would have to keep track of millions of cars. That would be an expensive feat. I wish them luck. I would enjoy riding in one every once in a while.

Once the FSD is even better, Waymo won't need remote guidance or chase cars anymore. So eventually, they won't need to track all the cars. They can just let the robotaxis do their thing unsupervised. That's kind of the whole point of a robotaxi.

 

[ItsNotAboutTheMoney]

Once the FSD is even better, Waymo won't need remote guidance or chase cars anymore. So eventually, they won't need to track all the cars. They can just let the robotaxis do their thing unsupervised. That's kind of the whole point of a robotaxi.

They'll obviously have to keep track of them, but that's just scaled software.