$TSLA · Solving The Money Problem · 2026-10-27

The Monro Life team have posted a nearly 90 minutee long cyber cab deep dive inside the Tesla facilities. Today I put together some highlights, just maybe the top 10 or so percent of insights from an engineering and Tesla investors point of view, what makes the Cyber Cab so special, how it's being produced, some of the massive efficiencies in producing a Cyber Cab compared to Tesla's conventional vehicles.

As already mentioned on a recent earnings call they'll be pacing the production rate roughly in line with how quickly they're actually scaling their robo taxi fleet.

Speaking of scaling the robo taxi fleet in case you missed it 3 days ago the amount of foresight and planning that went into this. Tesla registered 69 additional cyber cabs for its robo taxi fleet which took the total to 420.

And as soy merit pointed out, yes, these are real numbers. But yes, we're not done today. A massive update with Tesla's robo taxi service area in Houston, Texas, now expanding enormously.

So although it's still early and these are still relatively small numbers in terms of total cybercaps, operational cities, total square miles being served, we can see the trend quite clearly, paving the way for rapid expansion and back to the cyber cab production.

The last thing you would want would be for the number of cyber cabs Tesla can produce to be falling behind, lagging how rapidly they're actually expanding. At some point, this will almost certainly flip, but it's much better for production to be capable of outpacing expansion than for Tesla to be held back in their ability to expand simply by the fact that they're not able to produce enough cyber caps.

Model Y is already a hyperefficient vehicle to produce. This is why Tesla can sell them at a shockingly low price point relative to the value offered and do so profitably while not a single other company producing high volumes of electric vehicle can make money selling them. Period.

Cyber Cab is a quantum leap forward. Reducing parts complexity in this case 5 to 10 times fewer stations means that these things can be produced much faster, much cheaper, much more consistently.

There's less points of failure and also out of a much smaller factory footprint.

We also heard Lars describing the fact that with the smaller parts modules, robots can get in and around. So just think what happens a couple of years from now as Tesla's Optimus robots are more capable, more dextrous robots which essentially can work three to four times as many hours per week as a human for near zero cost, just a little bit of electricity plus the hardware to actually build a thing in the first place.

The unboxed manufacturing system developed by Tesla for the Cybercap puts them light years ahead of any other quote unquote automotive manufacturer in terms of cost, speed, efficiency, factory footprint per unit of output.

They are so far ahead. They're essentially embarrassing the entire automotive industry.

Any company could have conceived of this system for production. It didn't need to be a cyber cap. It didn't need to be an electric vehicle. I mean admittedly electric vehicle less parts lends itself more to this system.

But any automotive manufacturer in theory over the last century could have come up with a similar production system and massively increased their efficiency and lowered their costs.

Yet, it took Tesla coming up with this to actually make it happen.

When it comes to autonomy, dedicated robo taxis, Tesla's so-called competition, example, Whimo, don't have expertise manufacturing vehicles. They're utilizing other companies to build and retrofit with a fuckload of ugly sensors their vehicles.

There is no conceivable reality in which a company trying to compete with Tesla autonomy will be able to come even close to matching Tesla's actual cost to one put a vehicle on roads and two to actually operate that vehicle on an ongoing basis.

This means that no company will have a reasonable chance of matching Tesla's cost per mile offered to consumers to use their service if the service is comparable without losing money on every single ride.

While Tesla at the same price point has enormous profit margins built in, this all comes down to the cost to build the hardware.

Tesla's a massive advantage here and some of the additional ongoing operational efficiencies due to an efficient powertrain, less parts, less mass, blah blah blah. It's GG.

I just want to make this as clear as possible. There's no conceivable reality in which a company that isn't Tesla is going to be able to match or approximately match Tesla's cost to consumers per mile to use their robo taxi service if it's a comparable service without losing money on every ride.

Imagine your Tesla and you're scaling autonomy wildly and rapidly. CyberCap will enable Tesla from the same factory footprint floor space to produce twice as many cyber cabs as they could say Model Y's.

Alternatively, they could produce as many cyber cabs as Model Wise from half the footprint.

So, if they're scaling, they need to build new facilities, new production capacity. Less floor space is less cost, less time constructing, less everything, less walking on the factory floor from point A to point B, less movement of the actual products.

The efficiencies compound and compound and compound.

Again, this is super nerdy stuff, but this is just one of the many ways why this unbox manufacturing system is a massive win. A huge leap forward will allow to a scale faster and at lower cost than any other so-called competition.

Well, in fact, it may actually help with some of the bonding if it's not coated, right? You don't have to worry about that. Now, is this something that you view as backwards compatible with a Model Y or a Cybert truck?

But does it actually make sense to retool and reconfigure the production systems for the Model Y when it's already a highly profitable high volume vehicle? The answer may not be yes, even though this could in theory good high enough volumes save a little bit of time and expense.

It's a fair amount of energy required from Tesla to actually reconfigure these. There'll be downtime.

does it actually make sense to retool and reconfigure the production systems for the Model Y when it's already a highly profitable high volume vehicle? The answer may not be yes, even though this could in theory good high enough volumes save a little bit of time and expense.

It's a fair amount of energy required from Tesla to actually reconfigure these. There'll be downtime. So, the question probably isn't could this be adapted to Model Y? It's more does it actually make sense to invest the time and energy or to just let Model Y continue pumping out being the world's bestselling vehicle by revenue and printing money for Tesla?

I think it had to come up front. Um but like certainly like as we move forward we're going to try and put that in because it's part of the mission is like a sustainable future and so like the less you know we can put into the product that is bad the better we'll be on the back end.

And this is all with doubling the useful life at a minimum of the car. Like the car is going to run half a million miles at least.

Yeah. I mean, you think about a robo taxi, it's going to go more utilization every day. So, totally up time. On to the next clip and a little bit more about deleting the paint shop.

We're going to talk plastics. We talked earlier about how we got rid of the paint shop. Um, this is our paint shop. Wow. So, so when first of all, it's nice to meet you. I'm very excited to talk to you.

So, just to help the non-extreme nerds out there, when Lars pointed out this is our paint shop. This is where they mold that gold plastic. The joke is there is no paint shop. This is a molding shop. talk us through like when that conversation started, why it's there, because there's other vehicles like you look at the BMW X5 injection molded front fender, Saturn way back in the day, right?

They had plastic panels. What was the rationale behind this? Obviously, paint booth is a big thing, but

Yes. So, the deletion of the paint was the first motivator, but then you've already discussed the unboxed architecture. Sure. So when we were firstly talking unboxed and how do we get the panels to GA, the biggest challenge if we had a paint shop was we had to box it back in together to ensure it goes through paint and comes out.

So then we said what else could we do to eliminate those processes and make the panel in one shot and then we started discussing multiple options of how can we get color on the panel and that's how yeah we actually this isn't the only idea we had.

We you know we talked a lot about a lot of the work we were doing. We had like uh in mold lamination which we actually use on semi now where we put like a film like vinyl film with PUR on it and then put it inside the mold and mold into the film like IML, IMGL.

Yeah. Yeah. Okay. And then we had uh we also had film over panel where we actually put the paint onto a stamped panel and stamped it. So like a film on like adhesive with you know just just pressure and then when we stamped it and then we would hem it over the edge.

And so we actually got pretty far on all of those one of those technologies. Yes. And I it's never a competition. No, I don't think and I think we I can we can proudly say we've done both the technologies now.

We have IML and semi and but then when it came to cyber cab and the need to do it fast but at the same time cost effective this was what was the best way to approach for all the multiple panels different geometries because um that because it is just we still had to injection mold and this is just adding another layer within the machine so we get no orange peel you know we don't get like the defects inside of it I mean we still have issues with dirt and dust But that means it's a cleaner panel to begin with.

And if you do scratch it, you can repair it just like you would normal paint shop. And then the plastic underneath is similar to any fascia you've ever used. It's like a tpo. Yeah.

Yeah. Okay. They're different by panel just from like the upper ones, the roof has slightly different requirements, but yeah.

You'll note if you're listening closely, the goal for Cyber Cap, why they use this particular technique and the semi is using a slightly different one. The goal with the Cyber Cap was speed and cost above all else. And by the way, these panels look fanfantastic.

Next, the battery pack. Is it? No, I'm kidding. I'm No, it's it's Yeah, it's the same potting, but we we're doing modules instead of packs. You probably want to know why. I do want to know why for sure.

All right. So, it's it's good to be self-critical. It's one of our one of they would say our cultural attributes here. It's okay to look at yourself and I didn't do a good job.

I'm going to give our Model Y structural battery a grade D. Okay. To the customer, it was an awesome product, but internally just sort of the like how hard it was to build and how much it costs.

Yeah. Cybertruck structural pack. We We made some big changes. I'd call it a maybe a B minus. I was going to give you a C++, but okay. It like it was definitely passing. It was for sure passing a passing grade.

You've seen the pack. You've seen the line. Like I'm proud of that. What we started to realize is like in the limit the grade A structural pack. Yeah. And the grade A, we'll call this non-structural semistructural pack.

Everything we could tell is like pretty equivalent cost for you're going to get maybe 10% more volume for cells in a structural pack when you delete the structure, but you're going to save a little cost as a result, but the the issue is the service cost, right?

And like you have to get just infinitely low cell reliability to win on structural packs because we get to salvage three of these modules if you know 1,500 bucks a piece if if one goes down on a Model Y.

You're replacing the entire thing. So like as a thought experiment, a mega pack at 3 and a half to 5 megawatt, you're never going to pot that whole thing. If one cell goes down, you don't want to lose that asset.

An Optimus battery, it's it's one module. It's 2 and half kilowatt hours. Three, you know, it's that's kind of okay. It turns out in that in that product, it's actually not the largest cost at all.

And that too. Yeah, exactly. So we found Well, that's a juicy little side nugget. I mean, we could probably have guessed, but can confirm it's on the record now. The battery pack inside Optimus is not the most expensive thing in the bot.

Now, the reason for this, the four different modules as opposed to one gigantic pack for anyone's a little bit lost. If something goes wrong with one of these, you can just swap it out instead of having to do the entire pack.

We talk about serviceability and longevity just once again illustrates how serious Tesla is about Cybercap operating for years, if not a decade plus, with very minimal servicing and maintenance.

In the event that something does go wrong with one of the modules, quick, easy, and much more affordable swap out. So they're really thinking for vehicles that got to be operating hundreds and hundreds and hundreds of thousands if not potentially half a million maybe toward a million miles in that situation what's the best design that it was going to be too much engineering drag to maintain two architectures in parallel let's just like go all in on making this the best it can be yeah so whether it's on the engineering design and development side or andor I would say the economics of service and getting to it and is are we are we scrapping a whole pack or just a module so that's really What drove us here?

Exactly. So, what So, what grade is this one? If this I got to know the grade of this one. I'm feeling pretty good about this one. I would I actually think this is uh Robbie and the whole battery scene like most elegantly best designed pack, most integrated in the vehicle.

I can't ever say you like perfect, but I'd give it an A minus. Okay. Always a little room to improve. I'll take it. Yeah. I mean, that's that's a that's a good grade from where I sit.

So, it's also worth noting when you think of TCO as we've talked about, replacing an entire pack before half a million miles is a killer on cost. So, we have to think of remand, we have to think of like pack serviceability, you have to think of all those things that actually get us to our cost target for the product.

Even maybe the insurance side of it, I would imagine, right? If you're if you're going to insure a product like this and Tesla selfinsures and does all sorts of things, that's got to be a benefit.

Anything else you want to see on this? No, this is great. Thank you for taking us through. Can I show you just a couple more things? Yes, you can. Anything you want. That's a really funny moment.

Proud engineer wanting to flex. I was going to jump ahead to the next thing, but we got to watch this now. Bro, so excited to show this off. Let's see what he's got. Okay, there's some stuff in here I'm really excited about that I think you'll like.

So, one thing you'll notice is there's no harness. Yeah. From the B&Bs. Yep. It's all BMBB is battery management board to the to the penthouse that what we call the uh inline ancillaries.

This is we weren't the first to play around with wireless, but we implemented wireless communications in here. Okay. And it's just it's been huge. It allowed us delete harnesses which are super messy to automate and cost money intraac wireless communication.

Yeah. So, normally this is a dumb circuit board that sends data to the HBC communication wires through a harness. BMS battery management system. It's a package. Pay for it, it fails.

So, now we have an RF chip on here. It drives the cost up a little of this circuit board. But now we get, you know, to delete that harness and and now we have edge compute here.

So we can do local things like runaway detection. We get redundancy um within the like cuz each each board is, you know, not now singularly powered. But also like when you when you think about cost as I said as the whole vehicle, even if your board cost went up 20 or 30 cents for the communication chip, we deleted a harness that might have been a dollar.

You deleted the assembly of the harness might have been 50 cents. And on the system level, it's cheaper. It's huge. My my overarching takeaway when I look at this pack, you know, when I actually stare at it for a second, my question when we were walking up to it that I didn't ask, I'm like, are all the pieces in here or not, right?

And that's a good grade, right? That's why we get a good grade cuz it looks it looks like it's missing something. We're missing the potting over here. Yeah. So, you can see our our 4680s.

And then let's take a look at the ancillary back here. We got another one. So, I was trying to get you some more room, Jeff. It breaks when you open up a three Cybert truck ancillary is pretty good.

When you look at a 3 Y ancillary, it's kind of a rat's nest. I'm being self-critical. I worked on that. This is super clean. So, we took the approach of a single casting where you can get all the mounting features and build that idea.

Hey, I know. Yeah. A casting over the guys next door. Yeah. So, first things I I want to point out that are pretty cool here. The um the shrouds for the connectors just integrate them into the casting here.

Right here. Like this is our front drive unit. which allows the connectors to go down in costs I would imagine. Exactly. So the the part basically what we do here are we call it um instead of doing panel mount and jumper we're doing direct pass through on all of these things.

Right. So these three components plus this coolant connection. Those are actually like this part this part this part and then that that's the inside of the cold plate. So rather than have panel mount and a jumper we just pass through right here.

All the intermediates go away and you're putting a fork through a grid. Exactly. Yeah. Okay. Yeah. Exactly. Um and then we did some cool things you can't really see here, but yeah.

So, we deleted a lot of harnesses by directly mounting the contactors to the board. You know, you got sense, signal, and power. All those connectors are gone. The the BMS just mounts directly to those devices.

And the interrupting contacttors, we had to check this is out in the world. We filed the patent application for it. You can say it. the um we were able to delete the pyro the active fuse really because these interrupting contacttors now have the speed to to latency with reduction of all the connection longer relays fast enough interesting and not only that they can open under load a few times and keep working so often times you'll get more redundancy huge fraction of pyro fires are from like low speed crashes where the HP system is actually healthy and you don't need to you don't want to have to service the pack sure and so we can But with a with a one drive pyro, you have to Exactly. Yeah. Game over.

Onto the drive unit. From a from a actual traction motor and gearbox perspective with a cyber cab application, you don't need to do 140 mph. You're going to limit the vehicle speed as a result.

Just like quick visual. I'm trying to crunch the the gear math. Yeah. Yeah. How did you count? I can check you. How how much of a change on the gear ratio was there? And how did you dial?

We went to 11. It went to 11. That's actually Yeah. So, um, most of our cars are a gear ratio of nine. Nine. Um, we didn't need that top speed. We went to 11 for the gear ratio.

The motor would actually be a little happier on efficiency, a little bit further, but as gear ratio is not free on efficiency. So, we picked the best optimum. If we could make a 14 to1 be as efficient as 11:1, we would probably want to go a little bit higher.

And it's something that we'll be working on. So, so you're saying like theoretically if you went a slightly higher gear ratio without adding losses without adding losses you may actually be able to get more mileage.

Yes. Okay. A little bit. Yeah. So, so it is possible in the future that the already hyperefficient in terms of ongoing cost which all comes down to this CyberCap could become even more efficient meaning the actual cost to operate per mile could become even lower.

Meaning Tesla could have even greater a competitive advantage when it comes to pricing their robo taxi service. We're getting more mileage a multitude of ways. We design our own lubricants.

We're going to add we're dropping gear mesh losses by 7% later this year. The thought of like, okay, for autonomous vehicles, let's make the best thing we can. And then as we move more and more towards autonomy, that'll probably evolve into other vehicles.

Got it. The electrical improvements at the inverter level, as you can see, we kept it common. Another reason is that we can this this PCDA and power stage will ship in 3 and Y this year.

Yeah. So you're saying the overall footprint the geometry even if it's different population you can still maintain that housing and it's the power stage the PCBA and everything.

You could you could just swap this into a a model Y drive unit. Oh the entire thing. Yep. Yeah. We kept we kept the entire interface. I mean it works. It's robust. It performed really well in the fields.

So are we going to I plead the fifth. Okay. We always we're always opportunistic with how we how we take an advancement from one project and try to expand it to others or build upon it for the next project.

Like for instance, this this is the same unit that goes in the semi, believe it or not. So our smallest vehicle and our largest vehicle are using not only the same technology, but in a lot of cases exactly the same assemblies.

But then we started talking about the door and I asked, well, why did we go with the G wings? Cuz they can be expensive to service. It's got it, you know, drives this big casting for the pivot point.

And you guys provided what I thought was the best answer, which was there was no technical reason. It was just because it's cool. Yeah. Which if we can't do that in our jobs, then our jobs get a little dull.

So, um, that's a great answer. I I appreciate that. And I and I will say as well, the like the door opening is ergonomically better. Like you can get in somebody who's in a wheelchair and they can get into the car because the seat is the same height as a standard wheelchair. the even somebody getting out.

The fact that the door pulls the underside out further first means you have a lot more room to sweep your feet out of the car if you're in a tight space. The the step in ingress egress lower and there's a little bit of rain

and there's a little bit of rain protection. So there's like some benefits and maybe we talked ourselves into those but the fundamental reason was cuz it's cold. So Jordan, in a normal assembly line, um the wheels are put on somewhere in the middle of the process and then you roll it out on the final line.

Um this is our final station here. Um we call that robot Hulk Hogan. Uh it got installed when the you know the legend passed away. So we wrapped it in gold and red, gave it a little Hulkamania look.

But we installed the wheels on the second to last station in the vehicle. So the first time a cyber cab is ever put down on its wheels is when it rolls off the line.

we have this infrastructure like we we have to utilize it. It's hard to change over like to a new um automated way, but like we have a charge port because we know we're going to need to use the the infrastructure we have.

Hopefully, Optimus will help us with that shortly.

Um but then, you know, we are looking at more automated ways with our expansion of infrastructure that the car the cab is like pre-built and ready for. Um and I think that that that will you know be the step change to not requiring uh you know a human or or or Optimus to to plug it in.

Yeah. You know might service the car might add in the delivery space. You know, like we we I mean They're obviously thinking about Optimus and Cyber Cab and deliveries, meaning Tesla Cyber Cab's not just taking people from point A to point B, but also food deliveries, packages, parcels, and obviously you're going to need a humanoid to actually leave the vehicle and do the final few feet of the delivery.

The whole physical AI world is going to come upon us very very quick. And you know, you think of Cyber Cab, Optimus, and Roboan like that trifecta will be incredible.

What is Tesla becoming? Maybe we're a transportation company. Maybe, you know, it's it's a and redefining how that everyone thinks of that. I think longterm the best description for Tesla is probably going to be an AI powered transportation, labor, and super intelligence company, which is quite a mouthful.

Obviously, one of the big ones. TZ are actually already thinking about utilizing their humanoid robots for the final feat of deliveries. So, parcels, packages, urgent items requiring a courier, food deliveries, you name it.

Here's the bottom line. Tesla Cyber Cap will be faster, cheaper, better than anything else on the roads. The implications are profound because there's no magic catch up to Tesla button that would allow any of the so-called robo taxi competition to meet their scale, their experience, their costs, their efficiency. It's G [ __ ] G. This only plays out one way.

Let's just say that Tesla's cost, not what they're charging consumers, but cost per mile is $1. That's how much it's actually costing Tesla to operate a Cyber Cap at scale.

Let's say that Tesla decides to charge customers. Got cost over here. The price that they're charging customers per mile on this service. Let's just say that they're charging $2 per mile.

The reason they're doing that is cuz they got a lot of cyber cabs. Any lower than that and their weight times will be too absurd. Any higher than that, it's just totally unnecessary.

In this situation, Tesla would be making a profit of $1 per mile.

while Tesla charging the same fair with a $1 cost per mile would be making a profit of $1 per mile.

What if instead of Tesla charging $2 per mile, Tesla instead charges a much more reasonable $1.50 per mile?