How to Invest for the "Age of Abundance" 💰 - 10-Yr AI Plan

Here we are with Mr. Phil Bicell. Good morning. — I feel like I know you personally because we interact so much on X. Thank you for being here and thank you for all you do for the community as well. Do you want to give the audience a quick 2 seconds in the event that somebody hasn't heard of you? Oh, that's fine. your background. Very colorful background indeed. Yeah, I suppose. Uh, let's see. Uh, I live in Silicon Valley. I've lived the Silicon Valley lifestyle for quite some time. Uh, I've uh worked at Apple twice in my life. So, I like to say I had two tours of duty. It was my first job. Why I moved to California, actually. Did a bunch of startups between, went back to Apple. Um, was there for total 10 years. Various engineering roles and hearing management roles. Then I left in 2017 and joined this crazy company called Rivian. Um, and led the tech team here in Palo Alto. And uh basically architected the software-defined architecture of the Rivian vehicle. I worked uh built the cloud team, you know, the whole data side and uh the mobile team. And then I worked on um future innovations as well. And uh then I was there for 6 years and decided to move out. — Just stop doing stuff like that. So, um became very interested in Tesla. You know, ironically, I wasn't totally um swallowed by Tesla when I was at Rivian. I was so focused on Rivian. So, oddly, the architecture and the product we built really may it may resemble it in some ways, but it really wasn't intentional because we didn't spend a lot of time studying Tesla, but became very interested in Tesla after I left Rivian and um and invested in Tesla heavily. And uh you know, decided to start writing about all things in the Musconomy because they're interesting to me. Oh, yeah. Um in so many And it's — so many ways it's hard yeah. It's the most interesting company. And you did mention at the beginning before we were talking offline that you did see one of my early videos on Tesla, too. Yeah, that's sort of how I became introduced to you by proxy it's looking uh I don't remember who was may have been on Herbert's podcast or somebody, but I remember seeing one of your I think it was like 20 23 time frame seeing one of your um you know, business modeling presentations and I was like damn, there's a lot going on at this company, you know? So — even more going on now. So like every year the goal posts move and they get bigger, which is hard to believe. — 2023 was the year that I decided to really invest heavily in Tesla um and decided that you know, I had this basic thesis that it's like look, if I look forward um many years out, I don't while I worked in an automotive company and I understand automotive market texture, but I'm not I don't have an automotive background. So I have really a software you know, Silicon Valley kind of mindset background um I felt that um that Tesla was heading in a different direction and one that was based on more you know, it we see it now as AI and and all that, but I it was very clear to me that embodied uh autonomy um embodied AI would be an amazingly huge future uh, and uh, it's starting to obviously make its way into the world. Um, and so I was very, very interested in that. And uh, so I invested in the Tesla and I my basic thesis was that Tesla only rivaled by SpaceX would be the most valuable in the company in the world and you know, on a 10-year horizon. And I'm sticking with that and yeah, it's — It's so funny cuz uh, if you came across me five or six years ago when I first started doing YouTube, I used to say all roads lead to Bitcoin. Now it's a different kettle of fish. Like the whole narrative is AI. But let's jump in. I do have a whole bunch of questions about you and you know, hopefully they're very captivating topics. The end game is to help people uh, always understand where the puck is going to be and how to position themselves on things that are really life-changing. And the question is how real are they? So let's just jump in because we I've always been a huge believer in a couple of things, effectiveness, efficiency, and synergy. Those three. Now, when you look at all the stuff that Musk industry is putting together, it blows your mind. I mean, the synergy between the companies, you know, whether it be comms or bulletproof steel or self-driving vehicles or using the vision technology you know about from Rivian etc., to be the same platform that the Optimus runs off to the Terrafab. I mean, I calculate 18 I always I mentioned this in a video a few months ago. 18 different lines of business just within Tesla. Forget SpaceX. And I heard an analyst talk about that as well. He mentioned the exact 18 number as well last week. — That's interesting, yeah. But this is fascinating. So, tell us your view of wherever you want to go. You wrote some great stuff on SpaceX. You know terrafab. You know what it takes to pull

this off. I'd like to dig into how real it is. You know, I was always thinking like maybe space-based data centers 2030. I can wrap my head around how possible it is because they got over 10,000 satellites up there running off solar. Right. They all have chips in them. So, theoretically, it's possible. But when I heard last week that Google's knocking on the door or Monday it was, of SpaceX to get capacity to send stuff into space. It shows you how terabase data centers are so constrained. This is very real. But tell us about your view of how big Tesla is synergy with SpaceX, terrafab, any risks out there and when will things happen? — [clears throat] — So, it's incredible story and a story that is going to blossom over the next coming years in a way that is really going to fill out how Musk talks about it, I believe. Um it's almost inevitable and I'll tell you why. It comes down to how fast AI is growing and the compute demands are growing at 3. 4x per year on compute basis. Um this is something that will continue and probably increase. And what's extraordinary about this is that means that Musk's small little vision of 1 terawatt of compute, so like the world does about 20 to 30 gigawatts of AI compute today, okay? And we think about when he talked about terrafab only back in March, um that he said, "Okay, we're reaching for the gold, 1 terawatt of compute power. " And I think some people would look at that and go, "That's ridiculous. We're not getting there. " And I got news for you, we're getting there. And we're getting there we will run into a brick wall. Uh we will hit that demand irrespective of what whether terrafabric exist or not, whether Musk is here or not, we will demand a terawatt in compute in only 4. 4 years from now. Yeah, let's weave in another couple of threads here. The Anthropic deal with XAI SpaceXAI that also took me aback. I mean, we got all these hyperscalers with all this capacity and Anthropic couldn't get their hands on any. That supports what you just said. Did that blow you away, too? That now we have AWS and EWS Elon web services? Yes and no, but I really looked at what Elon has done with uh essentially giving Colossus one compute. You know, so he built Colossus one and Colossus two in Memphis. And Colossus two is the real big deal there. That's about 5x the compute capacity of Colossus one when it's all — So that's gigawatt versus 200 megawatts. Yeah, so that's a big deal and it's and Colossus two is based on the the latest chip architecture, so it has more you know, it's basically think of it as just 5x the compute capability over Colossus one. So by leasing, if you will, Colossus one to Anthropic, to me it's um it's what I call the prototype deal. What I mean by that is it's sort of like to me um it's not sort of a funny lumpy thing in the corner. It's really the business of what SpaceX will become. SpaceX will become a provider of compute to the world through the orbital data center model. And this is just, you know, okay, so this piece is terrestrial. Uh but as they move into space, that's exactly the kind of service they'll be providing. In other words, if you were to fast forward many years from now and you were to ask well, where did Anthropic get its compute? Where did Google anybody get its compute? It's probably SpaceX in space. Um whether people are launching their own constellation through the SpaceX railroad or frankly just using SpaceX compute. I love that SpaceX railroad. I'll remember that one. Yeah. But when we look at SpaceX just for a second

before you continue there is at least I know very little about space. I was obsessed with it as a child with the moon and everything else. But is there anybody remotely close to pulling off their scale or their technology or their ability to move so much tonnage into low urban orbit or whatever No. low, medium, high Is there anybody? No, simple answer is a fat no. I mean not even governments, you know, I mean China is constrained, has no capability to do this at any scale. Now, SpaceX sits alone and will for many years. I'm not suggesting that they've completely run away with it and other people will not show up to play. But the reality is all paths to space go through SpaceX. Yep. At least in the next uh foreseeable future, 5 10 years. That's you know, especially if you think about the scale they're building here, right? I mean Starship in terms of its tonnage to space is a completely different platform and it's an enabling platform. So, without that platform um you aren't industrializing space. I mean we are at what's really fascinating about this is we are at this moment in time where we are starting to look at space differently. In the past, the past you know, 50 plus years since we've gone to the moon blah blah, all we thought about is space is um, you know, science projects and things that were always sold to the taxpayer as, well, you know, if you support NASA's mission, you know, we'll get some great new Saran wrap or something, you know, it was never connecting. And now what we're seeing is uh a way to expand our industrialization beyond, you know, terra firma, beyond Earth-based type of industrialization to both orbit in terms of industrialization of intelligence and then ultimately, you know, moon, Mars for industrialization of of physical product. So, I think that's what, you know, that's what SpaceX is enabling through its platform and um, you know, I don't I don't see how it's catchable, honestly. — I spend uh quite a bit of time looking at Elon's tweets and he always talks about the engineering difficulty of what is Starship. You as an engineer, you want

to maybe highlight the top three things that you think are mind-blowing, whether it be the Raptor version 3 or the steel or the way they continually iterate so fast on this thing? Yeah, I mean, I think you know, I think the key the ultimate key to Starship beyond its size, I mean, size is enormous, but it the fact is that we have this platform that will be sized as it is and it will get bigger, but will be reusable, right? Reusability is the key issue here and how do you do that? And, you know, part of it is the reliability that they develop through this Raptor engine, right? They're they'll be debuting the Raptor 3 on Starship V3 coming up this Tuesday uh if if it sticks to schedule. And that's an extraordinary refinement. You know, I think one of the things about what I about from an engineering mindset, what I like about Musk's approach is that it's a fast fail approach. In other words, you dive in, — [clears throat] — you build a prototype, you in some ways publicly fail because, you know, there's you can't really hide your failures when you're launching something like Starship or um but it's all part and parcel it's how you do it. It's how you get there. It's um and what I think is the beauty is at the velocity at they're moving. Like, you know, a prudent person might say, "Why would you um why would you launch on this coming Tuesday when you probably have a month to go before the IPO, the biggest IPO, and risk, you know, potentially um uh pushing it off? " And the reality is I think Musk looks at it from a different lens and his lens is "I'm in a race of time. " Yes. Even a month matters. And while that risk has certainly entered his his calculus, it's not what's in front. And I think that mindset tells you why SpaceX is where it is today. It's this incredible velocity of achieving these goals and um it you know, it in physical rocketry, you know, there's a lot of uh stuff in engineering where, you know, you design you can design in simulation, various different ways, and sometimes those things only fail in small ways in front of your customer, frankly. You know, you put out a product and you like a vehicle might have a recall. Uh a part, you know, isn't designed and even through its rigorous testing uh doesn't survive the, you know, five-year customer journey of the vehicle. But rocketry is kind of different and it's where that fast iterative process of that fail fast mentality is what is yielding the dividends for SpaceX, I think. Uh and you know, besides I mean, you look at V3, one of the things that's amazing about this new Starship is just the number of changes that they've made since the V2 platform. I mean, it's absolutely extraordinary and when you when you step back and you think about it, you're like, "Would you really do this? " Yeah, you would do this if your goal is to really push the accelerator pedal down and get there as quickly as possible independent of all these other things. I'm sure Musk feels on a daily basis that he's he's way behind it way behind his own plan. Um and so that's probably the number one pressure in his brain is probably how do I get there faster? Yeah, I remember it was an interview he did. I can't remember David Faber, I think. And I haven't seen the movie. I don't watch movies, but it's thing called The Princess Bride or something like that. Quote, he said, "Offer me money, offer me power, I don't care. " And I think it was in regards to somebody at Disney or something. He just doesn't give a toss about money or wealth. He is You nailed it in that one sentence where you said it's all about speed. Speed, time, time. Just go as quickly as you can. Forget the IPO. You know, does a Tesla earnings call poops the bed. Yes. But he doesn't care. And that's what I really admire about him. It doesn't And this is same insane desire to be desire and need, you know, like to be as deeply vertically integrated in his businesses as possible. Yeah. I mean, here's a thing where you look at it and you say, "Why did Tera Why does Terafab exist? " Okay, why is this a thing? Not enough chips, right? He's done that compute and he you know, he's done the back of the napkin math on that and he realized that in his own businesses alone he'll run out of chips for Optimus and but you know, robotic vehicle FSD robo taxi by 2029 2030. Yeah. And there is no way he's going to get that production out of the leading um fabrication houses. And so on one hand Terafab is a necessity. On the other hand, by doing Terafab he's going to change the entire chip industry as we see it, right? I mean nobody's going to do it at the scale and people may predict success or failure or whatever you want, but I think the thing is again we're going to see you know, I've seen like within he announced Terafab in March he we talked about it in it was talked about in I think a September earnings call or at the shareholders meeting where he talked about the need. Uh but I think he's moving rather quickly. I mean not only is he building a a research center for but he's decided to use Intel technology 14 and a 14 a tech stack that they're building. Which is radiation proof, isn't it for space too? Um you know, I don't know many of the details about that honestly. I think the you know, as he's spoken about that I mean for first off it there needs to be some consideration for that but the sensitivity is not as high with it if it's within infinite compute anyway because you know, small numerical mistakes in a math model that is you know, trillions or billions of parameters um with funny little values is less important than um you know, uh piece of discrete code that controls the descent of Apollo to the moon, right? So Yeah, interesting. Yeah. So slightly different consideration, but um nevertheless, yes He must build chips that are more resilient to the space environment running at a higher heat threshold. Um Yeah. And he spent all of 2025 hinting at it, too, saying, "Samsung, give me everything you can make. TSMC, make. " So, it it's you know, a lot of people Sometimes he plays really coy. I try to get inside his head. And sometimes he pretends like, "Oh, I didn't really think of that. " But if you look at the way he always thinks ahead 8 to 10 years, I know he's thought about it. So, do you believe he sometimes plays dumb? Like he did this whole convergence thing. Like there's no way he didn't know about Starlink and connected vehicles and Optimus and the need for chips and FSD. You know, that's a great question that runs through my head a bit. I think I would advance the argument as simply that however I think about that convergence, I would say that he's simply a year

to 2 years ahead. In other words, I don't think 10 years ago he could see a future where SpaceX and Tesla are colliding in ways that make them near almost inseparable, as you know. But I think at least 2 years ago, before I thought this would be something, um you know, I wrote an article about SpaceX colliding with Tesla literally 2 years ago. And I read it recently, and I thought That was impressive. — good, but it wasn't great, you know? It It didn't really explain this compute convergence, which is what it's mostly about, you know? Um And so, yeah, I think he's I think he is way ahead, and I think And sometimes I think he's trying to build an articulation of those of those points for consumption, if you will. You know, in other words, I think he trial balloons it a bit on X and — Yes, yes. sees how people react to it and then I think it it's just I think his velocity and where he starts ahead of us and his velocity is faster than us. But, you know, I mean, he's not going to predict something I mean, you know, he's going to predict rational things that right in front of him, clearly. He also, I think, the other thing that should be said is that he has insight into his businesses that we don't have insight into, you know? We don't have those things at our disposal. He doesn't know He knows — when he When he realized there was a shortage of transformers. Nobody was talking about transformers 2 years ago. He said, "We need transformers. " Well, I mean, because he probably experienced a shortage of trans- uh transformers, right? I mean, it's some of these things are just um realities that come back through his teams. And they, you know, what's amazing about Musk is his ability to this day to, you know, essentially run most of the prime engineering meetings at both SpaceX and Tesla. And when those meetings are had, you know, he's very skip level in those meetings where he's basically engineers are front and center. Uh managers and senior managers are, you know, not necessarily locked out of the room, but not the important voice there. And so, what Musk is getting is he's getting direct, unfiltered information and he's a bit able to consume that information at, you know, his incredible bandwidth. And um and I think, you know, what he from a culture point of view, he he sort of he's trying to get at the truth and I think in because he doesn't penalize people for telling the truth, I think he gets the truth. And the truth in many cases is delivered to him in a set of limiting factors, things that are holding back progress in one place of the business. So, whether it's transformers, chips, uh chip design, you know, something about uh reusability on Starship, that stuff comes to him very, very early in the cycle because of how he how open he is from an engineering mindset. And I think that that's paying dividends. Okay. Now, we've This is so exciting to talk to you for the first time live. We've covered a lot of ground, but we got to get back to the agenda and tackle the big topics and I got to drill into your engineering

brain. Let's talk about Terafab, 1 terrawatt capacity. Where is the bottleneck to pull this off? Now, he's hinted about eating cheeseburgers and smoking cigars in a fab. We all know the process of how semiconductors are built. It hasn't fundamentally changed with all the five or whatever steps it takes and all the different places over the last 30 plus years. Right. Can he unthink how semiconductors are made all together at once and do it much faster? That's the big question. Or is it pie in the sky? Cuz when you talk to semiconductor experts, they say, "Nope, not possible. It's our way or the highway. " Can they re-engineer exactly how chips are built? Well, the There's a very mature process here with a lot of technical complexity, like the lithography process itself is extraordinary and amazing. Um and so, I don't think some of those logjams or some of those ways of doing things are going to take many, many years to rethink, you know, to first principles think the problem. And I think Musk will I think because of the pressures required to get to production earlier than later, I think he will both I think he'll take a dual strategy. I think what he will do is he will use a tried and true processes and suppliers for a good portion of the first phase of the development and then he seeing that live, and again, this is we go back to iterate fast, the concept of how he works. As he starts to build the and then that's why one of the reasons he's building the um Terafab research center on the Austin property, right? Because what he wants to do is he wants to see it for himself. what are the limiting factors in this process and um and so I think as he starts to see that he will both he will reimagine on a first principles basis maybe how to do this. But I do think because of the deep um processes that are so thus developed that he will sort of take a hybrid approach. Or so early stuff I think will come out based on how Intel does business here. Yeah. You know, he'll and then he'll start to see that and unravel it. Because you know, the one thing that I will tell you is that the end point here isn't one Terafab. It's multiple Terafabs. In fact, it's not possible. The Terafab itself in production producing a terawatt of compute annually will by 2035, but 2034, 2032 actually, hit a brick wall itself. So he will need multiple Terafabs and I will tell you the next terafab won't look anything like the current terafab. Yeah, let's just try join the dots here. I always try and figure out what's in his mind, and he does that all the time. He's talking about thousands of Starship launches a year to put data centers into space. Yeah. Have you thought about taking that back to how many chips are required for that? And the chips required for the 10 million Optimus humanoid robots and all the millions of cybercabs, etc. Have you done the mental math of calculating how many terafabs they would need? Well, look, I think the the reality [clears throat] is this. I've looked at the demand side of compute and AI across the entire, you know, at its growth rate today, which is about 3. 5x annually, and realize that in 4. 4 years from now, the world will demand at its current growth rate on AI, uh will demand a terawatt of compute. And in a mere 10 years total, will require 100 terawatts of compute. So, I mean, that's just do the math on the growth. It's growing at an enormous rate. So, so if that if that's the case and you know this and we hear the pieces, how come firms like TSMC, etc. and Samsung are so reticent to deploy capex? When you compare the capex deployed for TSMC and Samsung and all these other players and you compare that to Meta and It's Amazon, it's not the same realm. It's a fair, fair question, and I think it comes down to their sort of scar tissue in the boom-bust cycle of semiconductor. You know, I think they've gone through these boom-bust cycles. They see, you where are we? Are we in a Are we on a different path? And I think that they're you know, while they think they're moving as Elon says as fast, but they're not certainly moving this fast. Um and uh actually not even close. And so I think that you know, the the ironic thing here, James, is that they're um their reticence to their uh by them not investing, they could realize their own reality of the bust cycle themselves because — Yeah. Somebody else will run ahead. Well, it's not just that. I think what happens is the industry will slam into a a brick wall of demand. Like or a supply, you know, in other words, there'll be no supply for the demand that's coming, and therefore you'll go into a an artificial bust cycle because there simply isn't um anywhere to expand. Yeah. Okay, let's take Let's put a bow on this whole topic regarding Terafab. If Terafab delivers, Tesla becomes the most leveraged way to own the compute layer, period.

Would you allocate assets based on just this thesis alone? I know I do. Yeah, I think it's extremely compelling. And I think that's one of the things that I've realized is that my realization about the compute demand side of this and that Anthropic eagerly wanting to do a deal with SpaceX and Google talking about, "Look, we cannot satisfy compute demand on planet Earth. We cannot build data centers fast enough independent of chip supply. So put a pin in the Terafab and say, 'Yes, it produces the chips. Now we can't power them. Okay? And the only place to go to have production and to then power these very chips is space. It's inevitably obvious. And so at this point, yeah, that's like I my thesis is that through Tesla and through SpaceX you know, Musk controls the space. So, you have Do you have SpaceX exposure? I know you have Tesla exposure. Not enough. — You know? Yeah, I went into EchoStar. That's kind of my role. I had options to buy. Uh I'm always very afraid of these firms that have "Oh, we'll give you access to our shares. " And then they got a 5-year lockup period. They charge you 5% per year. There's no liquidity. And I hate This is one thing I hate in life is locking money up. Yeah. — to be able to I need dump my whatever it is in an instant. And the thought of I like I don't even know if I'll be alive in 5 years. So, why would I lock anything up for that time frame, you know? Um so Yeah, you know, I think that for me I've set aside a good amount of cash to invest in SpaceX eventually. In the IPO with Ron Baron, a billion dollars. What's that? dollars day one. He has a limit order in. Yeah, oh, yeah. I don't know if I can compete with him. But um But you know, look, I I'm not worried about Look, if SpaceX goes out at 2 trillion 2. 2, easy. — Um yeah, sure. I mean, you know, and it may come in then in a bit after that. You know, things will go wrong. Things will change. You know, we know how this goes. Um you know, it may fall by 50%. I don't really know. I mean, there's no way to for me to know, but the reality is I will leg into it slowly over time. I'm not going to be I'll have a bit of patience associated with this because look, I think that the reality is let's go back to what um what that man said, you know, he said that SpaceX ultimately will be a 20 or 30 trillion-dollar company. Yep. And I can see based on this sort of this, you know, tight noose around uh Musk's ability to control compute that the reality is that SpaceX will will will on an accelerated basis get to those bigger numbers faster than say Tesla did. Yeah. I think Peter Diamandis, by the way, said a hundred trillion-dollar company. That Tesla will be the first hundred trillion-dollar company. And let's just assume it goes ultimately um you know, two to 20. Let's keep it simple. Um, you know, you've got 10x there and I think that 10x is realized in um in um a five-to-ten-year cycle, somewhere in that range. So from my point of view, it's like if you bought in at two trillion uh and you know, you just sort of put that money aside and just forget about it and you know, you can wake up uh in five years and it probably will be uh 5x that. Yeah. And in fact, I've done the math a hundred ways to Sunday, SpaceX versus Tesla. And I'll give you the two-second answer. First, I believe all the TAM the biggest TAMs are for Tesla over the next five years. Post 2030, the biggest TAMs are for space. But if you want a horse to ride, that is Tesla now for the next five years and then pivot to SpaceX, but still have a position or layer into SpaceX Uh up until that time frame. I totally agree with you. I think uh you know, one of the realities that we're under today that people don't seem to recognize is that the SpaceX is going into an IPO. There's a hype machine going here. It is not that um Elon is you know, ignoring his other child here. If the reality is that, you know, you focus a lot of attention, you do deals, you do things as you're leading up to this IPO. That the success of this IPO raises a massive amount of capital. That capital gets deployed into these businesses that benefit Tesla. And I think the reality is that we are going to see Tesla's businesses, all the businesses start to explode. And so, if you just take them and put them in a pretty little box by themselves, the future of robotaxi, the future of Optimus, the future of semi, the future of energy, uh and of course, just its contribution into Terrafab alone, I think, gives you a reason to want to be in Tesla stock. And I also believe that that Tesla's just essentially like having a bit of a call option on SpaceX anyway. Yep, exactly. So, we agree. And all the people out there thinking, "Oh, SpaceX is going to apply to acquire Tesla, etc., or they're going to merge before the end of the year. " I think that's complete BS. That's 2 years out minimum. But it could only be accelerated if there's a regime in place that would force that level of convergence. Okay, let's switch gears. We're going to switch We've kicked all the big picture stuff to death and positioning. Let's talk about Optimus robots. Um you, again, engineer, you have worked at Rivian, Apple, seen a lot of stuff happening. We'll get to Rivian and Apple later down the line, too, but a lot of people are concerned uh about Optimus robots. Uh we've seen what's happened with white-collar jobs. Some people think it's kind of cosmic. a nothingburger. I'm beginning to edge on the side of the former. But, the big question, do you think, uh, first of all, Optimus 3? I know of two people that have seen Optimus 3, and they one person said it was creepy. And they couldn't say any details because they were sworn to secrecy. And the other one person said it's sublime. Uh, what is your gut feeling about how good this Optimus 3 is? And that people are saying, "Oh, look at whatever. " That one folding laundry right now live. I think Figure AI Yeah. is folding laundry. Like, who cares? Sorting packages. It's got to be much more. Look, I think Optimus is a is I think Musk's target for Optimus is a very refined robotics product here, you know, at a level that right now the industry is copying the thought of what it is, not necessarily what it is. And I think I am delighted that Tesla has I have seen a shift recently in Tesla being more quiet about how it talks about Optimus publicly. And as an investor in Tesla, I am actually delighted by that. I do not think it is necessary for me to have information about Tesla. I prefer the Apple model. Deliver it. Um, surprise and delight. You know, when it's ready for order. I I, you know, I think the I think Musk got early on that, and, um, you know, basically created an industry here. Uh, the reason Figure exists is because of of, you know, Optimus version one sitting in Tesla stores, you know, nicely displayed. Um, the reality is I think it's a very refined product, and I think it requires something that nobody else has, which is the ability to build the product in the factory along with the product of Optimus itself. So, to me, I think both building a supply chain for this product and building a uh uh a mechanism to manufacture at scale is probably as much of the secret sauce and the advantage that Tesla has than anything else. Okay, supply chain. Let's talk rare earths for a second. What type of limiting factor could that be? And I'm sure, you know, the US is making moves to secure rare earths in different parts of the world. Uh they're in China now, probably talking about that stuff exactly as well. Magnets. Um uh does that concern you for the future of Optimus? And will they get enough rare earths supply, magnets, actuators to be able to pull off 10

million a year? Well, that's a great question. I mean, I am no expert on rare earths, and I you know, I defer to others that are. I Those problems seem to solve themselves though. I don't see you know, that's the one thing I would say Yeah. Yeah, I would say that the reality is, I mean, when I mean, there's a real big difference between say a rare earth uh supply problem and an actuator supply problem, you know, it regardless of of inclusion of materials. I mean, the actuator stuff it is is extremely hard and requires uh Tesla to go pretty deep in its own vertical integration of building, you know, to In other words, I think where you see convergence of these two thoughts is the reality is that that Tesla will in fact do the because no supply chain exists, there's no, you know, it's not like a cell phone market, it's not like the automotive market. Um there are simply no suppliers here that at scale. So, um as necessary, Musk will have to go as vertically integrated as possible. And I think Yeah. Even with the automotive market, they started to build everything themselves, including seats. So, that would not be strange for them. But, I'm glad about that. And then the fun question next regarding Optimus, again, a lot of people are worried about the white-collar jobs, but do you see Optimus robot impacting blue-collar jobs, like plumbers, welders, people that do flooring, et cetera? Or are they safe forever? No, I mean, I think Optimus in terms of repetitive tasks that humans do, Optimus is clearly the replacement um vessel for that. I mean, you know, anything that's highly like the dumber the robot and the dumber the task, uh what I mean by that is there is a notion that Optimus will show up as a highly refined — [snorts] — C-3PO style product that will immediately enter the home and interact with us in a way that you know, is very futuristic, et cetera. And that may be somewhat possible. But, the first version of Optimus is most likely uh to be it is sort of prototypically uh as you see what I mean by prototypically is what you start to see the competitors show off is a type of uh labor device and and, you know, starting in manufacturing and moving up through, I think, the service industry. Um so, yeah, I mean, repetitive tasks in factories, uh stocking shelves. I mean, this is this is what this product will do. And it will do it at scale. — was probably the last time I went into a Starbucks. I boycotted the store for many reasons. Uh not to get political, but — it's kind of ridiculous. I walked into a store uh and they were very thinly staffed. There was two people there. Uh I asked the first person one question, "Can you do this? " He said, "No, I can't. I'm new here. " And then I asked somebody else, "Could I get one of these? " They said, "I'm training and I'm still learning how to use a cash register. " They had two people behind the counter, neither was equipped to do the job. And this is a pretty busy Starbucks that I was so upset that I spent 6 minutes of my life queuing up to get no fulfillment. And I'm thinking, "Wow, how hard would it be for Optimus to pull off what they're doing? " The answer is not hard. And that, I think, is the big, big story. We have it's as if ever since C-19 hit, the world has taken a step back. People have become more and more incompetent. They're worse drivers, etc. Have you noticed a difference in the world since 2020? — for you. That could go on for a while. — That conversation can Let me redirect your question back to the one — we're positioning Optimus here and it's like things like AI and humanoid robots are coming at the appropriate time, whether you're talking about, you know, falling birth rates, etc. And I know these things are political, but they need to be discussed. And I think — Timing is everything. I want to respond in one particular way about your Starbucks experience and Optimus, right? Let's assume, let's stipulate that Optimus could be trained to do some functional task in a Starbucks, like make the perfect espresso. Now, you can't do this at Starbucks for lots of reasons, but we won't get into those. Um the thing about Optimus, unlike the human counterpart, is once that skill is trained, it can be replicated across the fleet instantaneously. Yes. You can't do that with humans. No. So, any skill at whatever level, whether it is the most primitive skill in a factory or you know, it reaches a higher level of in the service industry, wherever that skill gets realized by the product, uh the replication of that skill is near instantaneous. And I think that's the thing that people really need to factor in when they start thinking about Optimus is once the you know, the capabilities may be slow to realize, but any single capability given is [clears throat] one that is easily replicated across a fleet. Yeah, and I think about things like classical pianists and highly skilled electricians. I had to replace some electrical panels on a place and the cost and the time to do that. Shut up. I think you know, it takes a human I mean, think about it. It takes it takes 15 to 20 years to train a human and I mean from birth through a cycle to get them to even be a capable person to stand in a Starbucks. Exactly. — An Optimus bot once it's manufactured, having a prototype over here that already does that task, merely it's merely replicated across its brain instantaneously, right? So, uh that's the impact driver here that I think people are not really, you know, dialed into. And let's talk allocations again. So, we cover the kind of terrafab impact on allocations and what to buy. When you talk about humanoid robots, um and I think that is

the biggest TAM. It's easily 30 trillion. I think in terms of TAMs and I back into them to discover which assets can control most of the TAM, etc. The Optimus TAM is the biggest on the planet right now. Easy for Tesla, $30 trillion. There is no bigger TAM on Earth. Now, when like volume production, I'm pretty confident 2027 will be quite a special year if they're going to build 10,000, 50,000. I look at the Cybertrucks being spat out right now out of Austin every day. They went from 8 to 12 to 20 to 30 to 70. Like they're ramping fast, and that's what they do brilliantly. But in terms of Optimus and allocations, is there another humanoid player on the planet you would consider? Or how much if you imagine you've got a pie to allocate to what I call baby AGI, which is self-driving cars, humanoid robots, Terrab data centers in space, whatever synergy you want. How much would you allocate to this pie that is humanoid robot for Tesla? Well, I mean, I think, you know, I mean, it's if you consider all shots on goal, that's the biggest one that Tesla has coming up. I mean, you put that puck in the goal and you score 10 points. Um and I think Tesla will be very successful with Optimus, and I don't I am looking at just how they are building, and I realize that people had expectations. Forget your expectations, you know, um you can complain about oh, it should have been could have been, but the reality is I agree with you, 2027 will be a special year. You can see what they're doing. You can see that they're building what I'll call the prototype factory here in Fremont, converting the S and X line. That is under development as we speak. Uh you can see that they're developing a uh you know an annualized production factory in Austin of about 10 million units for Optimus as we speak and you can see that what I think is one of the things that's holding back Optimus is required for you know what I'll say customer level production is AI 5. So I think you know that chip will start to come to Tesla in volume you know late 26 and then it then explode in volume in 27. So you know all those things I think will line up pretty well and I'm sure part of the Optimus problem too is the specific you know there's a lot we don't know and I think a lot of the stuff the Optimus problem is physical development of a manufacturable scalable development of the of its actuators is something that probably is we hear little about we hear more about the AI side the training side etc. We think of those as being some problem set to me that's a compute problem and it's really you know requires AI 5 but I don't see you know reality is I don't see most of the competitors able to you know if you consider the Tam what it is reality is that you know Tesla will largely own the humanoid robot market and you can put China in a box because the reason I say that is Chinese robots won't be on US soil for national security reasons they will never be you know with with a potential agency that they have and the fact that they are literally moving spies in a sense, they're not coming here. Whether or not, you know, Tesla has opportunity within China with the Optimus, I don't know, you know. But, I think Tesla will build premier, you know, what I'll call the iPhone kind of style product here, all right? In other words, there'll be lots of copycats with lesser products out there, and some of them will be early entrants, and most if their products are good, they will be compressed hard by their ability to manufacture them at at scale and the demand that's that the market will Yeah, and there's three things that matter that only Tesla can do, the brain, the hands, and manufacturing at scale. Yeah, this is a big deal. This is it's a really big deal. I don't think people you know, it's like one of the things that if you look at what's happening with Cybercab, we are just starting to see the reality of what Cybercab is about to do here, and one of the pieces of that reality is the fact that by redesigning the manufacturing process, they're not only dropped the cost considerably for a product like that, um but they've their ability to deliver it in volume is will be unprecedented. And if you combine the two of those, and then you basically think of that as the prototype for Optimus, uh you can see what they're going to do from a manufacturing from manufacturing technologies point of view with and philosophy when it comes to Optimus. And just look at what look what the next year will yield in terms of Cybercab production, and then just extrapolate that to what they're doing from a manufacturing point of view for Optimus, yeah. This is a shocking conversation, Phil. Literally, I thought we were 15 minutes in. I'm looking at the clock here. It's nearly 52 minutes. 52 minutes and we haven't gotten a quarter of the way through my agenda yet. So, I am That's your fault. Yeah, I I'm going as fast as I can here, but it's just too much Every time we answer a question, we open up another Pandora's Box. So, let's wrap up. We're going to have to do a part two if you're open to that. — That's fine. In a week or two because the real juice is coming. Mhm. But since you dovetailed into Cybercab

uh you said something very profound. The world does not know what's coming and how transportation will be revised and how this They spent years optimizing the manufacturing of scale of this thing. I mean, and the most exciting part for me is not just the transport revolution, but it's also the inference in the car. Uh there's some expert online and he analyzed the cooling system in the front of the car. He said, "That's not for air conditioning. That is That's for inference. " And remember, 99% of all AI will be inference in the future. What happens if there's 20 million of these things running around with their own cooling, power, and internet connectivity? Yeah. I know. It's um extraordinary what's about to come. Yeah. Cybercab, I mean, it's Look, here here's some facts that I would say is clear to me that FSD is solved Yeah. driving problem, okay? Like, you know, I mean, not only do I understand the technology deeply, but I drive it or it drives me every day. And it is not a science project for me. It is how I want Do you find it a little bit intimidating that you've become redundant? It drives better than you now. Because I I feel that in my car. You know. No, not at all. I feel a sense of relief that I don't need to drive my car. In other words, it's become it's as much a sense of relief that I don't it's like I I'm not intimidated by the fact that I don't have to wash my dishes anymore. You know, I have a dishwasher. I love that. No, I I absolutely I [snorts] Look, I kind of got a Tesla vehicle just to kind of test out the thesis of these things from an investment point of view. Um You know, I own a Tesla. I own a Rivian. I obviously my own a Rivian for sure. Um — [snorts] — but uh as I started to drive this thing or it started to let it drive me over these versions, especially as 14 emerged, version 14 of FSD, I realized this is a problem that's solved. It is not uh there's no safety consideration really. Uh what what we're left with is a bunch of operational considerations and that's what's holding back the scale out right now of uh robotaxi is really just operational issues that they need to work through. — And before I forget, by the way, Phil's link is below in the description of this video. Please follow him if you want to get smarter in this age of AGI. One thing I track all of the breadcrumbs that I see across cyber cab deployment. I subscribe to Joe Tegtmeyer, watch his videos every day. Uh it's reminds me when I used to watch the Cybertruck production. I want one of those outside here as well. Um and a Y and an X. So, the whole fleet except for the three. I thought it was too small. Um Guess what's next? The Roadster. We'll see if that ever happens. I think it might be too fast for my old back. But looking at the last earnings call and looking at the actual scale, my thesis is Elon Musk is sandbagging because they are deploying or making plans to deploy in 33 cities right now. Like literally they don't put butts in seats. They do not waste a penny until something is imminent. I think it's imminent. Yeah, I think here's the thing that I think you This is the math that I've been thinking about here, right? What Elon is attempting to do is to have an absolute scaled solution to um to this entire ride share type of market and if you look at Waymo and you look at its operational issues, not its safety issues. It's a pretty good safety record if you think about it from a pure aspect of who's been injured, this is, you know, it's not awesome, but it's pretty darn good. And and but the reality is that the service can't scale. And it can't scale because it has a lot of operational issues. And what it what I mean by that is what Elon needs to do here is he needs to say, "If I put a thousand of these in a market, I have to compute how many people I need in a call or operations center. " Like the backside support of this, right? And if if Waymo's total deployment is, let's call it 10,000. We'll make that number up, you know, be aggressive. Um if Waymo taxi were to be deployed 10,000 units today, he's got too much operational burden still. So he needs to get to a place where that operational burden is dissipated down to another factor of 10 scale so that as he scales, he's not building massive, you know, call centers of of people that have to respond to customer issues or customer complaints about a poor drop-off or, you know, this pickup or so they're working through those set of problems. And in parallel, and I think that's what people need to understand is that this is a you know, walk and chew gum at the same time enterprise here. In parallel, what they're doing is they are developing their prototype into all these different cities. They're starting to get their foothold in all these different cities. And so I do think that this is going to be a slow but then snap your fingers in all at once type of thing. And I think it's it's coming. And I'm not too concerned. Uh you know, I made a a prediction and a very public prediction that um that um you know, Austin would be 100% um you know, 100% unsupervised like the natural way, the way we expect the product to be uh by the end of May. And I was wrong. Got news for you, I was wrong. But the reality is directionally, I think I'm very accurate here. I see what's happening. I see that they're starting to move that needle. And I think that they are just working through uh a very very long tail of operational stuff. And I would say Yeah. I would say it took a bit them a bit by surprise as well, to be fair. You know, it's sort of like the rocketry problem, you know, until you're in the business doing it, you don't understand it. And so what what's happened is since last June when they put their toe in this business, they're starting to understand this business. And believe me, they now understand this business all the way to close to the point where they can start to scale it. Yeah, exactly. Final point before we wrap part one. Don't forget this is may we got so much more ground to cover. It's I'm very excited. This is really good conversation.

conversation. If you had to clarify whether 2026 is the proof it works or scale it fast year for Cybercab, where would you lean? Cuz you're hinting Yeah. I I think more prove it works. Scale it hard is an operational headache. Yeah, I think you know, 2026 I think will be it's [snorts] it's the building year and it is the year, you know, I think what we will see is you know, it to me it's sort of like this. I think there's a there's a bit of patience that they developed here in understanding that they're going to own this market. All right. And based on that kind of realization and patience, I think they're trying to do it right. And I think it's going to take time and it's probably taken more time than they expected to be fair. Um and I'm sure they're a bit disappointed in that reality, but I think the thing is using Elon's kind of style of you know, his sort of engineering mind. Ethos is sort of like you just got to get in and do it. And I think 2026 will be largely in terms of deploying robot taxi a somewhat disappointing year for most. Right? — Well, I I see where they're going to make it up in other areas. I believe there's a huge backlog of bookings for Megapack, high margin 32% margin business. FSD uptake is through the roof right now. [snorts] Cars are selling like hotcakes. This is all pure gravy. So, — Yeah. you know, they I think the CFO and Elon said early, was it late 2024 early 2025 that 2026 would be incredible and then 2027 would be completely off the hook. So, we're seeing already all of the building blocks. And when it comes to goes back to Cybertruck, to put a bow on that, too, they've done all the hard stuff. — [snorts] — FSD, bloody hard 10 years. Cybertruck design and manufacturing at scale, bloody hard. And they've got all the charging equipment around the country. They got the service centers around the country. Now, it's just all operations. I don't understand anything about running a taxi business, but I think it would be pretty cool if they could create a clever way to turn, you know, normal people into the operation shop. That would be powerful. Instead of Tesla having to do all that low-margin type activity themselves. By the way, I totally agree with your analysis of yeah, the hard stuff is behind them. Yeah. You know, but uh the Let's put it this way. The hard stuff is behind them. The tedious stuff is not. Yeah. Okay? They're in this tedious mode right now of getting through this operational what I'll call mess of just learning how to run the service at scale. And I think, you know, that stuff is hard in its own particular way, but it's not as hard as developing the core technology of FSD. unboxed manufacturing process of Cybertruck. These things are definitely you can chalk them up as investments that will, you know, that are in the done box, you know, um Yeah. So, I think '27 will be an absolutely explosive year and we will look at the end of '27 and we will think to ourselves, "Who are these other people out there playing? " Um, I really do believe that. You know, I think I I don't think for example, uh, this so-called science project of Waymo, I think will be something that we to use an automotive kind of metaphor here, will be largely in the rearview mirror by late '27, '28. Well, what I've always said is I think if firms like Uber and Waymo, they need to pivot to being super high-end luxury limo type black car service. Forget the the playground that Tesla's going after with cyber cab, like sub dollar a mile. They need to think about going after their own niche. I think that'll be critical. But then they have the RoboVan, which I just always imagine like not probably not too far for you near white wine country here. Imagine RoboVans touring the wine country. And those party buses, I mean I mean, a limo for the day 20 years ago would cost 1,200 bucks. I imagine it's more north of 2,000 for 6 hours nowadays. So, like demand for the RoboVan use case would be also off the hook. But that's how I think the Waymos and the Ubers survive to carve out their own niches to go forward. Well, to me, you know, I mean, my analysis of it is I don't care what happens to them. I mean, it's I care as much of as an Apple investor in the day, I could care less what happened to BlackBerry and happy to see them not exist anymore, frankly. So, you know, I don't think Waymo survives and I don't think Uber or Lyft survive and I literally could care less what happens to them because the reality is that cyber cab will deliver incredibly almost bespoke kind of experience with these products to the consumer at the lowest cost literally possible. I mean, I predict in a sense that eventually you'll be you'll potentially just be a subscriber to a cyber cab style business and that, you know, you'll get as many miles as, you know, within some limit. Um sort of like, you know, the way uh you know, we used to pay for the call or per byte on cell phones and now we subscribe with a monthly fee and I think that's what will happen here as well. Um you'll pay, you know, some dollars per month to basically all you can eat transportation in your in your suburban environment. Yeah, exactly. Well, we have so much more to talk about. I'll tell you what's on the docket for the next video. You already know, but for the audience as well, we're going to talk about the Optus RAM strategy, competition, the abundance economy, UBI, etc. How we offset that. We're going to talk about how people position themselves for the future, uh agents, you know, macro hard a little bit, too. And when can all this trillions and you know, CAPEX be recouped by the other players, spend a little time on Apple and Rivian, your portfolio and how you're positioning yourself, and then more on human capital. But one of the ideas I did get from this video is to create a continuum and put along the lines of the continuum the time in which true scale happens, technological proficiency happens across all the different lines of business, and then identify when the big pops could happen for the stocks and when the synergy could happen with SpaceX, and when Terrafab will really start ramping, and all that type of stuff. I think that'll be a fantastic exhibit. I'll kick you a copy of that before I'm finished it and get your input, too, and then we'll showcase that hopefully whenever we can get you back on the docket and next week or the week after, whenever you're free. Phil, don't forget link below for Mr. Phil Bysell. Thank you so much for your time and thank you to the mods in the chat as well for being here, keeping us all safe from scammers, Clint Alexander, Zeno and VR Trev, Razors and K8 as well. Thank you, Phil. Any closing thoughts? How about that continuum idea? Uh no, well, thank you very much for having me on your show. I appreciate it. Oh, no. Anytime. I'm the one who is honored. Appreciate you, buddy, and we'll talk real soon. Thank you. Bye.