Tesla Exec Just Revealed Optimus Production Plans

[00:00:00]Let's talk Optimus and physical AI.

[00:00:02]>> Sure.

[00:00:03]>> So, well, I'm glad you're excited.

[00:00:05]This is what we're very excited about it. So, >> Cyber cabin physical AI, too, right?

[00:00:09]>> Yeah, true.

[00:00:10]>> Robot with wheels.

[00:00:11]>> Um So, Fremont factory, we're sitting here.

[00:00:15]>> Mhm.

[00:00:15]>> Near near Fremont factory. Uh the we said goodbye to the Model S and X.

[00:00:20]>> Yeah.

[00:00:20]>> Your babies. And now we're expecting that Optimus >> Like teenagers.

[00:00:24]>> lines. They're teenagers. All right, let them go.

[00:00:26]>> They're done.

[00:00:27]>> Uh we're expecting that there's a transition to Optimus production. Is there something you can kind of update us on that effort of how far away further away now with the full transition to Optimus >> Sure. I mean, obviously, we're done.

[00:00:43]>> [laughter] >> Got it.

[00:00:45]>> Got it.

[00:00:45]>> I believe you.

[00:00:47]>> I saw your face.

[00:00:48]>> I believe you.

[00:00:49]>> No, I mean, it's hard, right? Like the factory is pretty old.

[00:00:52]Um but the construction teams did a great great job clearing out the space.

[00:00:57]Um you know, we're saving what we can, robots and things we can reuse. Um and and we've already our first line has landed. So, we've already started to install that um from our groups our automation groups in the Midwest and in Germany.

[00:01:10]And uh you know, I think even on the the uh Sentra event we showed all the machines that were already already uh being built to build Optimus. And so, those are going through, you know, last stages of of uh field acceptance testing, um FAT, we call it. So, >> On the automation.

[00:01:27]>> On the automation in the in the groups and like, you know, getting the parts there to you know, test the machines before we get them over here.

[00:01:33]And uh you know, but the cool thing is we developed a way with this first line because we knew we'd have to change a lot.

[00:01:40]It's quite modular. And also like the bring-up It's kind of nice cuz Optimus is smaller than a car. So, like the bring-up uh can be quite quick. So, we we do the testing at the at the automation groups. And then we can, you know, take that line down there, ship it here, and then hopefully, like theoretically, you'd get it powered on and hooked up and back together like in 2-3 days, and then it only takes a week to get it going. And so, that's proving out in the first line now, and and and you know, we got 40 more to go.

[00:02:09]So.

[00:02:10]A lot of ways to go.

[00:02:11]>> [snorts] >> Okay.

[00:02:12]Wow. That was uh 40 more to go.

[00:02:16]>> I don't know. You think about all the lines. Like you know, >> You can't just do a number out there, but it's it's kind of a lot. It's a lot of lines.

[00:02:22]>> of lines. There's like a lot of actuators, right? There's a number of actuators. You know, and then there's like each limb and like you know, subcomponent of the limb. Torso and battery, you know, there's a lot of lines. Sublines.

[00:02:32]>> Is it Yeah, so let's say is it is it more modular cell-like or is it like more serial super long line?

[00:02:39]>> It's more Well, it's not really boxed cuz it but like it's like it's definitely like uh tributary to the small final.

[00:02:46]>> Yeah, I like that.

[00:02:47]>> Yeah.

[00:02:47]>> Yeah.

[00:02:49]Well, um speaking of like these decisions, how how does How does Tesla approach vertical integration?

[00:02:57]I know at a high level you guys There's engineering vertical integration, there's And how do you think about it? When you encounter a new part, you encounter a new subsystem, something's nobody's made, how do you think about it?

[00:03:10]>> I mean, I you know, I'll I'll say our supply chain partners like Roshan and Karan and their teams are like also super integrated with the engineering organ. I I think it's really like we we just kind of assess the problem at a high level.

[00:03:22]You know, does anyone know how to do this?

[00:03:23]>> No.

[00:03:24]>> Do we want to keep the IP in-house?

[00:03:26]>> Yes.

[00:03:28]>> Uh it does it require a significant amount of investment?

[00:03:31]>> Yes.

[00:03:32]>> Mhm, okay. Maybe we should do it. The classic make versus buy matrix.

[00:03:35]>> We do a lot [clears throat] of make versus buy, but there's definitely like it's not like a pure financial table.

[00:03:40]It's Cuz like if you do a financial table of make versus buy, you always make. Yeah. Right? Because you eliminate the profit, but like then there's the CapEx investment, and sometimes finance won't want that. But like the cost is always cheaper in-house. Like >> Right. Yeah.

[00:03:51]>> Like it save maybe if you're shipping it from some low low labor market, but um well, I think that you know, we we kind of have this sort of technology risk assessment process I don't even want to call it a process mentality.

[00:04:05]>> Yeah.

[00:04:05]>> Um where where we just like look, it's high IP, we're going to do it cuz we want to control.

[00:04:09]>> Yeah.

[00:04:09]>> There's high risk of change.

[00:04:11]>> Yeah.

[00:04:11]>> We're going to do it cuz we want control.

[00:04:13]>> Yeah.

[00:04:13]>> If it's like actually we're adopting something from another you know, part of >> [snorts] >> industry where they do a pretty good job.

[00:04:23]>> Okay.

[00:04:23]>> Yeah.

[00:04:23]>> Let's not do it. Like the composite bed for for Cybertruck.

[00:04:26]>> Yeah.

[00:04:27]>> It's like yeah, that's like not very common.

[00:04:29]>> Yeah.

[00:04:29]>> Um but you know, we had a found a good partner with Hanwha and and like they built us a factory near Austin and and okay, that works.

[00:04:36]>> Yeah.

[00:04:36]>> Like we could have built that in-house, but the benefit to us wasn't so high.

[00:04:40]>> Yeah.

[00:04:41]>> And and uh >> Also longevity, like where are you going to use this?

[00:04:44]>> Yeah.

[00:04:44]>> Yeah, and for how long, yeah.

[00:04:46]>> And so so when you have a good partner, you work with a partner.

[00:04:48]>> Yeah.

[00:04:49]>> But I think ultimately it comes down to controlling our own destiny and making sure that we have that ability to do it.

[00:04:55]>> Yeah.

[00:04:55]>> I mean, if we need it, we build it.

[00:04:57]>> Yeah.

[00:04:57]>> Seems like a lot of things have landed in a vertically integrated manner and it really I think it really helped you guys through COVID.

[00:05:04]>> Yeah, it definitely did.

[00:05:05]>> Yeah.

[00:05:06]>> The fact that we design all our own PCBs helped us through COVID.

[00:05:08]>> Yeah.

[00:05:09]>> Like the chip crisis.

[00:05:10]>> Yeah.

[00:05:11]>> I don't think anyone really had a level of understanding of the boards in their product except for a company like us.

[00:05:17]>> Yeah.

[00:05:17]>> You know, um a lot of that's contract manufactured, a lot of it's purchased.

[00:05:22]But it's like I mean, I can't remember being on phone calls with suppliers and be like, "No, no, no, no, no.

[00:05:26]We'll just change that from the 12, you know, output ASIC to the eight cuz we only use eight anyway."

[00:05:32]>> [laughter] >> And they're like, "What?"

[00:05:33]>> Yeah.

[00:05:34]Just over the phone we're doing this?

[00:05:36]>> You have the we have tons of supply of the eight you know, the eight IO ASIC.

[00:05:40]Okay, we'll use that.

[00:05:41]>> Yeah.

[00:05:41]>> And then we'll just change it to the map.

[00:05:42]>> Change them up, yeah.

[00:05:43]>> So, um that that kind of stuff I think definitely helped us and again, you know, the supply chain team through COVID was amazing. You know, you know, if if if Giga Hong Kong's our superpower is is manufacturing, it it also is supply chain because the flexibility and agility they showed during that period is unmatched.

[00:06:03]>> I don't I look when I look at it just back from my perspective, I look at how inventory is managed, how these recoveries from these disasters occur, the quick product changeovers in in the line. I I actually never seen anything like it. And and it shows in a lot of the statistics. It's a the supply chain the upstream supply chain team is it's exceptional.

[00:06:24]>> Yeah, I mean, they're great.

[00:06:25]>> Yeah.

[00:06:25]>> I mean, even if you just look at, you know, like the the American-made index, right? Which one and two Tesla.

[00:06:32]But, you know, I think that's a testament to the vertical integration cuz it goes to the tier three, tier four level like that we're understanding.

[00:06:38]And you know, localizing and and and not necessarily like in-house, but just keeping it close to production.

[00:06:45]>> Yeah.

[00:06:46]Yeah. I think I think it was instrumental in COVID, tariffs. You think of all these things.

[00:06:53]>> Right?

[00:06:54]>> Who cares?

[00:06:54]>> Yeah.

[00:06:55]>> Cuz you're localized. Mostly localized.

[00:06:57]>> Yeah, I think tariffs is funny. I think we Canada was worse. Like I think we flopped which factory sent cars to Canada three times this year.

[00:07:05]>> Yeah, yeah.

[00:07:05]>> Yeah, who cares? It was like >> Yeah.

[00:07:07]>> So, it's it's pretty well uh it's known, common knowledge, I think, that Tesla is a leader in car manufacturing.

[00:07:14]To what extent can you speak to how much of that transfers to manufacturing Optimus and other robots?

[00:07:23]>> I mean, I I everything.

[00:07:24]>> Everything?

[00:07:25]>> Sure.

[00:07:26]>> Like >> You know, like all the things you just said, Jeff. Supply chain robustness, you know, like the the the vendor management, the the material management, the inflex it's everything. And those are the hard things. And I think when you think about Optimus or even our energy products, you can see the same thing. Powerwalls and and Megapacks and and our charging products that we sell, those have the same benefit and we're able to scale those where other people can't. And I think we're able to do that because of the background we have in in high-volume manufacturing from automotive. So, when I think about, you know, those other products, it's it's everything. It's like what a luxury a robot manufacturer, like a humanoid robot manufacturer, would have to fall back on.

[00:08:04]Like this this foundation of understanding in terms of how do you get parts in the building? How do you move them through the building? How do you get them out? How do you How do you manage like multiple like levels of supply chain? How do you manage multi-sourcing? How do you manage crises when a you know, there's a fire at a a steel mill.

[00:08:21]Like that's the hard part of scaling.

[00:08:23]Like it it's it's like the force majeures are crazy. Like I can't even like the things I've dealt with in my career like termites under the you know, in a shipping boat or or or, you know, fires at tire factories or, you know, like [clears throat] earthquakes in, you know, the middle of your PCB core area. It's like you have to be robust to that. And it sounds like you know, crazy, but it all happens.

[00:08:48]>> Yeah.

[00:08:48]>> And I think that the you know, the whatever product we build, whether it's a car or a power wall or a charger or, you know, a robot, like it benefits from that foundation.

[00:09:00]And uh then you go into the line design and like the people designing the Optimus lines are the same ones that design our car lines. Like literally the same people. Like they are It's not like we hired a new crew.

[00:09:12]Um and so those people are have the battle scars, you know, that we talked about earlier. They get it. And so in the automation groups that build the equipment are the same.

[00:09:22]So, they understand, you know, the importance of precision and repeatability and, you know, OEE and uptime and and, you know, ironically, like the bomb cost of a car is more than a robot.

[00:09:35]So, like it's actually more critical to have those things be you know, repeatable and reliable and have high up time on a on a higher volume cost product.

[00:09:43]>> Right.

[00:09:44]>> And so I think that all of that benefits optimus and I think you know, Elon said this before, I'll reiterate it. Like we have three things that a lot of other you know, robot humanoid robot manufacturers don't or design where companies don't have. We have the the scale manufacturing we just talked about. We have the design of actuators and high volume high volume high power powered dense electric motors that no one else has. And then we have the real world AI implementation.

[00:10:10]So, I I think those are the three key legs that we're going to stand on and you know, you only need three for a chair, so it's probably good enough.

[00:10:17]>> Yeah.

[00:10:17]>> The argument I've heard is they said, well, humanoid robot is more akin to a consumer electronic small mobile phone than it is a car. So, that's the argument that oh, then then >> That's interesting. I haven't heard that.

[00:10:31]>> Okay.

[00:10:32]>> [laughter] >> I mean like like, you know, take functional safety.

[00:10:36]You know, like our cars are are you know, from a electromechanical standpoint, we like go thoroughly through the functional safety design of the of the car to make sure it doesn't kill you.

[00:10:44]>> Yeah.

[00:10:45]>> Right? In whatever way, right? Your cell phone can't do that. There's no functional safety. I mean, there might be a little bit on the battery in terms of like overheating and charging.

[00:10:52]>> Yeah.

[00:10:53]>> Right? Um but not really.

[00:10:55]>> Mhm.

[00:10:56]>> That's not like in my mind, it's way more akin to a vehicle from a functional safety design manufacturing standpoint.

[00:11:04]>> Yeah.

[00:11:04]>> In terms of scale, volume, moving things. Like this is also a 2D product.

[00:11:08]Let's be honest. Like we're looking at your phone. I mean, it's 3D sure, but it's built in 2D.

[00:11:12]>> Mhm.

[00:11:12]>> Folded over a couple times, but like that's easy. Pick and place high speed.

[00:11:17]The robot has multi-axis dimensions, you know, that's like the suspension.

[00:11:21]You know, so we we I I I With respect to the investors and the people that are saying that, >> Yeah.

[00:11:27]>> if they walked the line, I think they would disagree.

[00:11:30]>> Okay.

[00:11:30]>> Yeah.

[00:11:31]>> That's a good to Can I ask you some more more questions?

[00:11:33]>> All right, yeah. We got to I don't know what time is it. What How long are we supposed to be going here?

[00:11:37]>> I just want to drill in a little bit on Optimus if you don't mind. So, Optimus has been You guys have always referred to the cars as robots with wheels. And every time Elon comes up on stage, often at the Tesla Owners Silicon Valley conferences, he'll say, "Everybody wants a C-3PO and they want an R2-D2." Now, when you first R2-D2, that means it's not a humanoid, it's another form factor.

[00:11:59]Should we expect that there's is like all sorts of form factors?

[00:12:03]>> I mean, well, you have the car, you have the robot the humanoid right now. I think right right now we're focused on like acceptance. Like we don't have robots in our world like Star Wars, right? They had C-3POs and R2-D2s and you know, the rolling ball thing, I forget his name.

[00:12:17]>> Yeah.

[00:12:17]>> Um >> BB-8.

[00:12:19]>> BB-8, yeah. And uh and I think that our world hasn't accepted that yet.

[00:12:26]Like, you know, even the automatic beer pourer they have in Japan in the lounge.

[00:12:30]Like people are oh, I don't know. So, I think that the most important thing that you know, Elon has been pushing with the first iteration of Optimus with the team is is limbic resonance. It's got to feel like it belongs in our space.

[00:12:44]And then people will get used to it, right?

[00:12:47]The car was easy. Cars already existed, you know.

[00:12:51]It's easy for people to accept that. But um once once people get used to the idea of having robots in the workplace or the home place, then I think you'll you'll see some expansion of of that into the R2-D2 world, but I think it's important we start with C-3PO.

[00:13:08]>> Okay.

[00:13:08]>> You know.

[00:13:09]>> That makes sense.