CNET: Inside the Luxury Robotaxi Uber, Lucid and Nuro Are Testing

[exxxviii]

This is a naive take. It does not cost Lucid that much to build a vehicle. The 'loss of $X per vehicle' argument has been bandied about for years now. It's a dumb argument because Lucid is spending billions building out their infrastructure, so of course they're in the negative. I'm positive they're making some money per vehicle sold, but the capital expenditures far outweigh the sale profits. People really need to stop using that clickbait argument.

No, Lucid is nowhere close to making any money per vehicle sold. Gross loss is very different from net loss. You are probably thinking about net loss, which is much more negative because they are building out infrastructure. Gross loss is in a different place in the financial statement and reflects the actual costs just to build the car and does not include the overhead and captital expenditures you are thinking of. Lucid is losing $100K just to build a car, and they are desperately trying to solve for that.

Simultaneously, Lucid is also losing around $325,000 net loss per vehicle, which includes all of the operational costs, overhead, SG&A, R&D, etc.

Gross Loss Per Vehicle = (Total Revenue - COGS) / # Vehicles
Net Loss per Vehicle = (Total Revenue - Total Expenses) / # Vehicles

In Q1 this year, Lucid recorded $594.2 million in cost of goods sold. They delivered 3,093 units. Therefore, it cost them $192,111 to build each car in labor and materials.

 

[Adnillien]

No, Lucid is nowhere close to making any money per vehicle sold. Gross loss is very different from net loss. You are probably thinking about net loss, which is much more negative because they are building out infrastructure. Gross loss is in a different place in the financial statement and reflects the actual costs just to build the car and does not include the overhead and captital expenditures you are thinking of. Lucid is losing $100K just to build a car, and they are desperately trying to solve for that.

Simultaneously, Lucid is also losing around $325,000 net loss per vehicle, which includes all of the operational costs, overhead, SG&A, R&D, etc.

Gross Loss Per Vehicle = (Total Revenue - COGS) / # Vehicles
Net Loss per Vehicle = (Total Revenue - Total Expenses) / # Vehicles

You do know that COGS includes depreciation and inventory impairment which do not reflect the incremental cost (material plus labor). While your calculation can be useful, it does not accurately reflect the incremental cost of producing another vehicle.

 

[exxxviii]

You do know that COGS includes depreciation and inventory impairment which do not reflect the incremental cost (material plus labor). While your calculation can be useful, it does not accurately reflect the incremental cost of producing another vehicle.

Yep, but so does everybody else's COGS. Maybe Lucid made some horrible decisions and it is costing them $100,000 more, per car, than their competitors in depreciation and service. But probably not. Rivian is in a similar stage of life as Lucid, yet their COGS per vehicle produced is only $97K.

 

[Adnillien]

Yep, but so does everybody else's COGS. Maybe Lucid made some horrible decisions and it is costing them $100,000 more, per car, than their competitors in depreciation and service. But probably not. Rivian is in a similar stage of life as Lucid, yet their COGS per vehicle produced is only $97K.

Having a second factory before the first is fully loaded does increase depreciation. Lucid has two factories because Saudi Arabia wanted a factory there to develop manufacturing and automotive supply chain in Saudi Arabia.

Fun with math... Lucid is currently losing about $100K per car gross. That is not SG&A, just materials, labor, and the costs to just build a car. So, if they sold 50,000 cars, that would be a $5B loss for Lucid.

You statement above ignores the impact of depreciation.

 

[merveille07]

One thing that still amazes me on this forum and it pertains to this discussion is how people like to play armchair ceo or manager for Lucid. Constructives criticisms are great and welcome, but what is the point of babling through every thread of how bad Lucid decisions making are? It seems we forget that people that work at Lucid are just like us and have a working brain and probably better understanding of business. Besides if you have a great idea about how to run Lucid, why not email them or better yet submit your resume?
******End of venting******

 

[segbrk]

Lucid is losing $100K just to build a car, and they are desperately trying to solve for that.

Where are you even finding this figure / what are you deriving it from? As far as I can tell their cost of production per vehicle is not disclosed, even indirectly.

 

[exxxviii]

Where are you even finding this figure / what are you deriving it from? As far as I can tell their cost of production per vehicle is not disclosed, even indirectly.

It is from their 1st quarter earnings release. It is simple, standard math, illustrated a couple posts above. They reported 3,093 vehicles delivered in the Financial Highlights on page 3. They reported $285.5M revenue and $594.2M Cost of Revenue in their statement of operations on page 15. Cost of Revenue (COGS) / Vehicles Sold = cost per car. Loss per car is (Revenue - Cost of Revenue) / Vehicles Sold.

Having a second factory before the first is fully loaded does increase depreciation. Lucid has two factories because Saudi Arabia wanted a factory there to develop manufacturing and automotive supply chain in Saudi Arabia.

Maybe that is it... maybe depreciation and costs from an unused factory accounts for an additional $100K cost per vehicle. On the other hand... if the factory is not online yet, would it count under cost of revenue? Probably not, because it would just be a capital expenditure at that point.

 

[exxxviii]

One thing that still amazes me on this forum and it pertains to this discussion is how people like to play armchair ceo or manager for Lucid. Constructives criticisms are great and welcome, but what is the point of babling through every thread of how bad Lucid decisions making are?

The relevancy for me is it informs whether I would consider buying one... I had the Fisker Ocean high on my list a few years ago and happily dodged that bullet altogether. Now I am a bit more interested in a company's long-term prospects before I would consider a straight purchase. I would buy a Rivian R2 straight out, if the deal was right and it gave me flexibility to sell early to get into something I might like more. But there is no universe in which I would buy a Lucid and risk a depreciation event like Fisker's (or Polestar is about to have in the US). Lease is the only option for me.

Plus, as a strategy and product consultant and automotive insider by profession, I am super curious how this all plays out. On one hand, the Saudi PIF seems like it will give Lucid a very long runway. On the other hand, they would need confidence that profitability was at least a possibility this decade, and PIF did pull out of LIV from what seemed like a long term plan.

 

[RandomDesign]

It is from their 1st quarter earnings release. It is simple, standard math, illustrated a couple posts above. They reported 3,093 vehicles delivered in the Financial Highlights on page 3. They reported $285.5M revenue and $594.2M Cost of Revenue in their statement of operations on page 15. Cost of Revenue (COGS) / Vehicles Sold = cost per car. Loss per car is (Revenue - Cost of Revenue) / Vehicles Sold.

Maybe that is it... maybe depreciation and costs from an unused factory accounts for an additional $100K cost per vehicle. On the other hand... if the factory is not online yet, would it count under cost of revenue? Probably not, because it would just be a capital expenditure at that point.

They have three factories now as they also bought the former Nikola factory in Coolidge, AZ (as well as their former HQ and product development center in Phoenix, AZ).

 

[segbrk]

Cost of Revenue (COGS) / Vehicles Sold = cost per car.

I think the thing you're glossing over there may actually be significant though? Cost of revenue != cost of vehicles sold. That cost is counting all kinds of things including write-downs. On the other side, I don't think we actually have a "vehicles sold" figure either - we have a deliveries figure. There may be a number of sold cars that have not been delivered. Are those cars that were purchased by agreement with Nuro/Uber or the PIF actually delivered, or just sold but accumulating? So sure, Lucid is losing money for reasons, but I still think it's misleading and unhelpful to assert that the cost per vehicle produced is X, as if Lucid desperately needs to cut the costs physically going into each vehicle. We don't know that.

 

[exxxviii]

I think the thing you're glossing over there may actually be significant though? Cost of revenue != cost of vehicles sold. That cost is counting all kinds of things including write-downs. On the other side, I don't think we actually have a "vehicles sold" figure either - we have a deliveries figure. There may be a number of sold cars that have not been delivered. Are those cars that were purchased by agreement with Nuro/Uber or the PIF actually delivered, or just sold but accumulating? So sure, Lucid is losing money for reasons, but I still think it's misleading and unhelpful to assert that the cost per vehicle produced is X, as if Lucid desperately needs to cut the costs physically going into each vehicle. We don't know that.

I will start by saying that I am not finance or accounting guy. But I have spent a lot of time consulting in that space and I have led a few P&Ls. So I will do my best to answer, but I may mess things up because it has been a minute.

1. Cost of Revenue is the total of all direct costs involved in producing the products that were sold. It is labor, materials, shipping, distribution, etc. to actually make the goods. It does not include other COGS stuff like marketing and support tied to the sales, but not part of the actual creation of the goods. So yes, Cost of Revenue equals cost of vehicles sold.
2. Vehicles Sold is Vehicles Delivered in this context. In the financial realm, companies generally cannot book the revenue until the product is delivered. So that is why they use the term "delivered" in the financial statement.
3. There is a direct relationship between Cost of Revenue and Vehicles Delivered. A vehicle that was Produced (Lucid reported 5,500 in Q1) is not necessarily associated with the revenue earned that period. They did incur cost for vehicles produced but not delivered too, but that is booked separately. Some of the Vehicles Delivered may have actually been produced during a different reporting period, and those costs are tracked and held for when the vehicle is delivered.
4. All these figures exist in financial disclosures so that stakeholders can understand the operational performance of the business, and that is why they are linked.
5. If Lucid delivered a car to Nuro/Uber, then it would be in both Vehicles Delivered and Cost of Revenue. If not, it would just be an asset accounted for outside of those two numbers. They cannot book a sale as revenue until it is delivered, even if the buyer paid in advance.

Lucid has disclosed the reasons they are losing money on their cars in their quarterly earnings calls. There is no speculation. They have been discussing their strategies for reducing cost of revenue for a long time.

 

[merveille07]

The relevancy for me is it informs whether I would consider buying one... I had the Fisker Ocean high on my list a few years ago and happily dodged that bullet altogether. Now I am a bit more interested in a company's long-term prospects before I would consider a straight purchase. I would buy a Rivian R2 straight out, if the deal was right and it gave me flexibility to sell early to get into something I might like more. But there is no universe in which I would buy a Lucid and risk a depreciation event like Fisker's (or Polestar is about to have in the US). Lease is the only option for me.

Plus, as a strategy and product consultant and automotive insider by profession, I am super curious how this all plays out. On one hand, the Saudi PIF seems like it will give Lucid a very long runway. On the other hand, they would need confidence that profitability was at least a possibility this decade, and PIF did pull out of LIV from what seemed like a long term plan.

So you're praying that Lucid fails and you are just putting your mouth where your head is on Lucid. I get it; you're not the first and certainly won't be the last.

 

[exxxviii]

So you're praying that Lucid fails and you are just putting your mouth where your head is on Lucid. I get it; you're not the first and certainly won't be the last.

Not even remotely. I'm not sure how you could possibly have read that. I have loved the Air since I first saw it, and the Gravity has been high on my list since I learned my current car is discontinued and I will need to buy something else this year.

However, I will not make a transaction blindly.

1. I will only consider a lease at this time, because there is material risk Lucid could radically transform during my ownership period. I think that risk is low to medium, but not zero, so purchase is not an option.
2. Lucid has replaced its CEO and most of the ELT over the past few months. I am super curious to follow how they turn things around.
3. I am a realist and pragmatist. I do not have a excessively optimistic and always cheerful view of Lucid's future, even though I think the cars look great and are amazingly engineered.
4. There is no universe where I would buy Lucid stock. But that has nothing to do with my interest in leasing one of its cars. (By the same token, I would not short their stock either.)

 

[segbrk]

I will start by saying that I am not finance or accounting guy.

Absolutely same here. I am not an expert by any means, I've just done more reading on this than is reasonable since this point keeps coming up.

Cost of Revenue is the total of all direct costs involved in producing the products that were sold. It is labor, materials, shipping, distribution, etc. to actually make the goods. It does not include other COGS stuff like marketing and support tied to the sales, but not part of the actual creation of the goods. So yes, Cost of Revenue equals cost of vehicles sold.

Except Lucid sells things that are not vehicles. This is even in the disclosure. It's not a large part of the revenue, but it is there. Total revenue in Q1 was $282.5M, vehicle sales revenue was $264.7M. So $17.8M of revenue is from "other stuff," which has its own costs. They don't split the costs by vehicle / non-vehicle as far as I can tell, so as far as we know that $17.8M could have cost them $100M. The cost of revenue also includes write-downs as I mentioned, meaning parts and materials that were purchased for products but are not going to make it into sold products for whatever reason. These could be used in the next quarter though as far as I know, they're just written down as costs for this quarter because they were expected to be used. It also includes provisional warranty costs. Not real warranty costs, but estimated future costs for sold products. It also includes fixed production overhead - not building the factories, but operating the factories. The whole factory has to be operational, lights on, safety staff on site, etc. despite running at a fraction of its capacity. That cost naturally factors out with production scale. Again, this is all still not a good thing for the company, but this is why we can't be sure of the cost per vehicle.

Vehicles Sold is Vehicles Delivered in this context. In the financial realm, companies generally cannot book the revenue until the product is delivered. So that is why they use the term "delivered" in the financial statement.

It might be, but that delivery number is not from the actual SEC filing. It's from a slide deck. It's not audited, it's not beholden to any particular legal definition other than not being a big fat dirty liar (legal term), it's just a number Lucid shares without definition. That's my understanding anyway.

There is a direct relationship between Cost of Revenue and Vehicles Delivered. A vehicle that was Produced (Lucid reported 5,500 in Q1) is not associated with the revenue, though they did incur a cost for those too. But that is booked separately. Some of the Vehicles Delivered may have actually been produced during a different reporting period, but those costs are tracked and held for when the vehicle is delivered.

"Cost of revenue — LCNRV" is a carve-out of that Cost of Revenue, and reflects part of what I was saying above. These are material costs that explicitly did not go into actual vehicles sold. The logistics department messed up and bought a million of the wrong bolt? That ends up here. Engineering decided to replace some component at the last minute because it caused an issue in crash tests? The cost of the whole inventory of that part ends up here. And yeah, again, this is not a good thing. But it doesn't mean the car will cost that much to make tomorrow, when they don't buy the wrong thing, or change their mind, etc. It's not a design flaw of the car or a read on its manufacturability.

 

[exxxviii]

Absolutely same here. I am not an expert by any means, I've just done more reading on this than is reasonable since this point keeps coming up.

Except Lucid sells things that are not vehicles. This is even in the disclosure. It's not a large part of the revenue, but it is there. Total revenue in Q1 was $282.5M, vehicle sales revenue was $264.7M. So $17.8M of revenue is from "other stuff," which has its own costs. They don't split the costs by vehicle / non-vehicle as far as I can tell, so as far as we know that $17.8M could have cost them $100M. The cost of revenue also includes write-downs as I mentioned, meaning parts and materials that were purchased for products but are not going to make it into sold products for whatever reason. These could be used in the next quarter though as far as I know, they're just written down as costs for this quarter because they were expected to be used. It also includes provisional warranty costs. Not real warranty costs, but estimated future costs for sold products. It also includes fixed production overhead - not building the factories, but operating the factories. The whole factory has to be operational, lights on, safety staff on site, etc. despite running at a fraction of its capacity. That cost naturally factors out with production scale. Again, this is all still not a good thing for the company, but this is why we can't be sure of the cost per vehicle.

It might be, but that delivery number is not from the actual SEC filing. It's from a slide deck. It's not audited, it's not beholden to any particular legal definition other than not being a big fat dirty liar (legal term), it's just a number Lucid shares without definition. That's my understanding anyway.

"Cost of revenue — LCNRV" is a carve-out of that Cost of Revenue, and reflects part of what I was saying above. These are material costs that explicitly did not go into actual vehicles sold. The logistics department messed up and bought a million of the wrong bolt? That ends up here. Engineering decided to replace some component at the last minute because it caused an issue in crash tests? The cost of the whole inventory of that part ends up here. And yeah, again, this is not a good thing. But it doesn't mean the car will cost that much to make tomorrow, when they don't buy the wrong thing, or change their mind, etc. It's not a design flaw of the car or a read on its manufacturability.

Generally yes to all the above. But it does not materially move the needle. Their cost per vehicle is somewhere close to $192,111, but not precisely $192,111. Lucid is a lot different from Rivian, which has material revenue and costs separate from vehicle sales, so they break those out with more definition. Analysts and investors recognize the differences. And analysts can determine Lucid's cost per vehicle and loss per vehicle just like they can with Rivian. And at the end of the day, if an analyst has any doubt, they will ask the CFO or CEO directly and get a sound answer.

Your callout about write-downs is probably related to their excessive direct costs. They have a huge materials problem. (And labor problem.) If they are purchasing materials that they cannot use and must discard them or write them off, then that would be totally non-productive cost of revenue that is harming their performance.

I do not believe any kind of warranty cost is included in cost of revenue. Cost of revenue should only be direct costs associated with the actual creation of goods.

Lucid's cost of revenue per car has been very high for a very long time. So if they have been repeatedly buying a million of the wrong bolt or changing components last-minute every quarter for the last twelve quarters, then they have a really, really bad problem. Looking at notes from their earnings calls, they have plans to bring materials costs down, and they are already working to reduce labor costs, but they have a while to go. It did not sound like they could slash cost of revenue by 50% any time soon.

 

[segbrk]

Generally yes to all the above. But it does not materially move the needle. Their cost per vehicle is somewhere close to $192,111, but not precisely $192,111. Lucid is a lot different from Rivian, which has material revenue and costs separate from vehicle sales, so they break those out with more definition. Analysts and investors recognize the differences. And analysts can determine Lucid's cost per vehicle and loss per vehicle just like they can with Rivian.

Your callout about write-downs is probably related to their excessive direct costs. They have a huge materials problem. (And labor problem.) If they are purchasing materials that they cannot use and must discard them or write them off, then that would be totally non-productive cost of revenue that is harming their performance.

I do not believe any kind of warranty cost is included in cost of revenue. Cost of revenue should only be direct costs associated with the actual creation of goods.

Lucid's cost of revenue per car has been very high for a very long time. So if they have been repeatedly buying a million of the wrong bolt or changing components last-minute every quarter for the last twelve quarters, then they have a really, really bad problem. Looking at notes from their earnings calls, they have plans to bring materials costs down, and they are already working to reduce labor costs, but they have a while to go. It did not sound like they could slash cost of revenue by 50% any time soon.

I think we've beaten this one into the ground and I'm fine with that. We have different reads, but overall it doesn't really matter as yes, there are clearly problems somewhere, we're just arguing about where they are. Since it was just released and is relevant I just wanted to point out: https://ir.lucidmotors.com/news-rel...ces-operational-reset-and-second-quarter-2026

 

[exxxviii]

I think we've beaten this one into the ground and I'm fine with that. We have different reads, but overall it doesn't really matter as yes, there are clearly problems somewhere, we're just arguing about where they are. Since it was just released and is relevant I just wanted to point out: https://ir.lucidmotors.com/news-rel...ces-operational-reset-and-second-quarter-2026

Agree, and thanks for the link. Reading it now. I like their three-pronged reset: Cash and Cost; Customer and Quality; and Culture and Team. Not to drive it further, but I was surprised to see their Cost of Revenue to increase relative to deliveries in Q2 to $210,498. I figured it would have improved.

 

[TahoeMichael]

I'm not an accountant. But I do know that the accountants at various companies I've been at have played a lot of games around what goes into COGS, and it is definitely not just a simple measure of material costs and manufacturing labor. I'm an engineer, and at some companies my labor has been attributed to COGS, and at some it has not, and from my perspective I was doing the exact same kind of R&D work at all those jobs. Maybe a forensic accountant could tear apart their public filings and figure out what's going on, but I don't get the sense that anyone here has done that exploration of their books. So maybe they're in dire straights, and maybe they're not, I really don't think we know. What I am quite sure of is that Lucid needs to grow to succeed, and I expect they wouldn't be agreeing to sell a lot of Gravities for robotaxis if they didn't think it made sense for them. If Gravity was fundamentally a dramatic money loser for them, they would have pushed for a Cosmos deal instead.

As to whether it makes sense to use an expensive vehicle for a robotaxi: my sense is that the importance of vehicle costs is overestimated by people outside the industry. Someday when we have truly autonomous vehicles, then vehicle cost will matter a lot. But all autonomous vehicles today are only semi-autonomous, and they rely to varying degrees on a huge amount of central operations and infrastructure, as well as ongoing R&D. All indications are the unsupervised FSD from Tesla currently requires a lot more central infrastructure that Waymo Driver does, and that may well be a big reason that Tesla hasn't expanded their nascent robotaxi offering. (I'm speculating here, I work in AI, and hear gossip, but I don't have direct knowledge.)

 

[Aka-none]

This means they will be the only company with a luxury robotaxi offering. There are absolutely people willing to pay a premium for this.

Town car or rando taxi
Town car every time I’ve had way too many stories about taxis
So yes I’d pay more

 

[illopp00]

No, Lucid is nowhere close to making any money per vehicle sold. Gross loss is very different from net loss. You are probably thinking about net loss, which is much more negative because they are building out infrastructure. Gross loss is in a different place in the financial statement and reflects the actual costs just to build the car and does not include the overhead and captital expenditures you are thinking of. Lucid is losing $100K just to build a car, and they are desperately trying to solve for that.

Simultaneously, Lucid is also losing around $325,000 net loss per vehicle, which includes all of the operational costs, overhead, SG&A, R&D, etc.

Gross Loss Per Vehicle = (Total Revenue - COGS) / # Vehicles
Net Loss per Vehicle = (Total Revenue - Total Expenses) / # Vehicles

In Q1 this year, Lucid recorded $594.2 million in cost of goods sold. They delivered 3,093 units. Therefore, it cost them $192,111 to build each car in labor and materials.

This also includes operating costs of a factory running at 25% capacity, or they are building like hand built Aston Martins instead of the Chinese Dark factory methods...