SPACEX IPO: EVERY AI GIANT WILL NEED ELON

追加: 2026-05-09

[00:00:00]So yesterday, Brian Wong helped us understand the massive implications of Anthropic's deal with SpaceX and XAI for the use of Colossus 1. Now SpaceX is planning to launch tens of gigawatts of orbital AI data centers over the next 3 to 5 years, which then means that every single AI company will have to partner with SpaceX just like Anthropic did. Brian is back with me today to look over the horizon and analyze what this means for the future of SpaceX AI as a company. Brian is a futurist thought leader and he runs the prominent science blog next big future.com. Welcome Brian. Good to have you back.

[00:00:40]>> Great to be here.

[00:00:42]>> So help us expand to the entire industry now. Um Anthropic steel for Colossus one was a a great first step that's terrestrial. um when you're looking at what SpaceX is planning to do with AI data centers in orbit um over the next 3 to 5 years, how how do how do we look uh at what's coming over the horizon?

[00:01:10]So the um anthropic deal was for 300 megawws of um of compute, right? Which um the Colossus one probably has 445 megawws of compute. So they didn't give the whole thing yet, although they may give the whole thing to Enthropic. And then they're making Colossus 2, which will probably be 2 gawatt once it's fully built up. It'll be like um a million um B200, B300 chips that will be in Colossus 2. Um and it could be more than two gigawatts. They actually need more power, meaning 3 gigawatts to fully power up a million chips. Um but at 300 megawws, that cost them $7 billion to make. And now they're renting it out, you know, possibly for even more than that amount of money. So we know that um SpaceX launches 160 times or so of Falcon 9, right? So that's um everyone else is down around five or 10 or or less um or or tiny rockets like Rocket Lab uh or China, the whole country, you know, able to launch many disposable rockets and then getting up to in the range of 50 to 60 launches. So we're so SpaceX is triple even China and with bigger rockets and more payload. That's why they're at 70% 80% of all payload into orbit. And that's >> as the numbers stand today >> as numbers stand today before you get to Starship.

[00:02:46]>> Right. When you get to Starship then you know say this month they get the the flight 12 to go and all they have to do they're not going to go to orbit.

[00:02:57]They'll they'll go near orbit with their version three, the new rocket because they're not launching version two anymore and then restart that engine.

[00:03:05]The rest of the stuff, you know, you know, they don't land anything, they don't do whatever. It doesn't matter.

[00:03:10]Once they restart the engine, they know they can go to orbit and they can then start an engine again and de-orbit because they don't want a big Starship hanging around in orbit. Um, potentially >> these are the vacuum raptors on Starship, >> right? Right. Right. the vacuum racks, they need to restart those rocket engines, right? Once they can do that, that means the next launch, then they can go finally go to orbit. Um because before they were still testing version two, they didn't want to go to orbit with version two. They tested, we're testing other things, right? Because that version two could only launch 35 tons of material, a payload, which is only a little less than double a Falcon 9. This new one will be launch over 100 tons reusably, which is five times the payload, right? So now with flight 13 they can instead of launching dummy satellites that are suborbital they can launch real satellites right so that's how close we are >> so the flight after this next one that we're expecting in May >> the flight after this they can start launching real satellites version three it doesn't matter again they can do more testing with landing and doing whatever other stuff they can start launching real satellites full payloads whatever they want to do with that so then even with the number of test flights they did last last year, that would still be like four or they can do like three or four test flights, but launching real satellites. So they can launch like a 50 satellites per payload, they can launch 200 version three satellites while still in a air quotes test phase, >> right?

[00:04:43]>> Per Well, no, 50 per per launch, four launches, 200 by the end of the year, >> right? But if they get their cadence up and say, "Okay, now we're kind of doing stuff. We're launching once a month.

[00:04:53]We're launching twice a month." They could get to 10 payload launches this year, 50 satellites per that would be 500 satellites. And those satellites would be 20 times the bandwidth of them. So right now we have 700 terabs per second of capacity from 10,000 11,000 V2 minis with 500 V3s, right? That would be 500 terabits per second. A terabit per second per satellite. So I nearly double just from this kind of launching a little bit. Right.

[00:05:27]>> So what's the ratio of um equivalent terabytes uh between V2 and the V3? So how many V2s roughly would a would a single V3 um account for?

[00:05:41]>> I think it's about like 20.

[00:05:43]>> Wow.

[00:05:46]because there's also difference in hardware. There's also differences in what spectrum they can use. There's a bunch of uh and also being able to use full power and stuff like that. So, it can it can do.

[00:05:56]>> Now, just to be clear, this is we're still talking Starling um connectivity >> connectivity. We haven't we haven't gotten into we haven't become talking about >> AI data centers.

[00:06:07]>> That's right. That's right. So, but the the thing is you need to get through and launch they will make you know say hundred billion dollars a year maybe $200 million a year from launching their full communication satellite system right so that gets them to Nvidia levels of profit and money >> which would then make them >> you know worth5 to10 trillion dollars right so they need to do that first but that you know for them to launch 20,000 V3s or 30,000 V3s if they launch a 100 launches next year of Starship which they you know three three launch towers um you know fully FAA approved right they can theor in theory launch 145 I'm assuming that they only launch 100 but at 100 they can launch 5,000 satellites V3s which is the equivalent of a 100,000 V2 minis >> V2s wow >> so but thing Once I've done that, now I can start using more launches. I can go to 500 launches per year in 2028. Now I finished launching all that other stuff.

[00:07:16]I can now launch a 100 launches for my AI satellites. I could also do a couple test launches in in next year, but you know, the year after now I can start seriously launching my AI satellites because I want to make sure I've got everything sorted out before I launch my AI satellites.

[00:07:33]So just so the audience is clear, the the orbit for Starlink is lower than the orbit that the AI data centers, the AI satellites will have, right? Because the AI um satellites in the AI orbital data centers will be in a sun-synchronous orbit and it will be further out in space. So, it's going to be a much higher orbit for which again, as you pointed out, we need those space raptor engines to fire so that Starship can get to that elevation.

[00:08:04]>> Yeah. Yeah. They brought them the the communication satellites down to about 300 330 miles of altitude. And then they were up at 500. Now they're going back up to 500 600 miles of altitude >> with the with the bigger ones, the V3.

[00:08:17]You mean?

[00:08:18]>> No, no, no. With the um AI satellites.

[00:08:21]the ASL, >> you know, the the the B3s will still also be the lower altitude, >> right?

[00:08:26]>> For the communication >> for latency.

[00:08:28]>> For communications. Yeah. For for I want to have a faster ping. I want to be able to play video games, whatever. Right.

[00:08:33]>> Yeah.

[00:08:35]>> So So then Yeah.

[00:08:36]>> So you're saying that um probably um you know this year maybe about let's say 50 to 100 satellites if you're lucky.

[00:08:50]Um, no. Um, no. I think between 200 and and 600 satellites, E3s, 200, 600 E3s.

[00:08:57]Um, assuming that we get, um, flight 12 so that we can start launching real satellites in the next launch or so.

[00:09:05]>> And then, of course, next year it's going to be exponentially higher, maybe 100 launches. Right.

[00:09:10]>> Right. 100 launches between 5,000 10,000. They could also switch over to V4 rockets next year. Doubles the payload to 200 tons. So, I'm not expecting that until happen late in the year. I'm not expecting the switch over till the year after.

[00:09:25]>> Okay.

[00:09:26]So then we're really looking at 2028 for like once Starink V3s are done, once you have the new uh the new booster and the new uh Starship uh by the end of 2027, you'll have everything set up for a massive scaling up of launch capacity for the AI data centers, the AI satellites.

[00:09:51]>> Right.

[00:09:51]>> Right. Now that's 2028 and I would imagine as you as you've also said that they're probably going to do a few um test launches with the AI satellites sometime next year.

[00:10:04]>> Right.

[00:10:05]>> So then you have this explosion in in orbital AI data center capacity.

[00:10:12]What does SpaceX do with that?

[00:10:15]So the thing is um they were talking about making 100 kilowatts um per satellite initially, right? So that means if I can launch 200 tons, which I have a version four in 2028, then I can launch a 100 if they weigh two tons each, right? So that means each uh rocket launch, Starship launch, I'm launching 10 megawatts of satellite AI into space each each time.

[00:10:45]>> So then with 30 launches, I've launched 300 megawws that is equal to the amount of data center at Colossus one.

[00:10:54]>> Colossus one. Yeah, >> the anthropic deal.

[00:10:57]>> The the anthropic deal. Correct.

[00:10:58]>> Right. which is worth seven billion dollars plus, you know, maybe like 14 billion dollars um for that. And the fact that it took two years to do it, it's taking two years to make data centers right now on Earth, >> right? Yeah. So, if I can build them and then launch them in two years, that'd be great, right? And I'm not stopped at 30 launches. I'm thinking that we go to like, you know, >> in the range of 500 to a,000 launches in 2028. We just need to improve the launch cadence. There's a bunch of things there. And then with the money we're raising this year, they'll make more launch towers, more sites. Right. Right.

[00:11:33]>> So then with six sites, you know, 12 to 14 or or 20 launch towers, I'm launching and catching launching and catching multiple in parallel, I can get to this higher cadence rate. Um, and then get to, okay, I've done my 500 launches that it took to launch all 30,000 45,000 of my communication satellites. Now I'm focused on the AI satellites and like you know if I launch just a hundred that would be a gigawatt of of power right of AI data center power which is three times as much as what we just did with Enthropic and that is worth 20 to50 billion or more you know so it's yeah so that >> how much is the entropic deal worth again? Is it about 16 billion a year?

[00:12:22]>> We we don't know exactly. It depends if they're paying Amazon rates. So if a Amazon charges more when you have a dedicated facility to it and we're saying, "Hey, it's a dedicated facility versus I'm sharing with a bunch of other clients and other stuff like that."

[00:12:37]Right. If I'm sharing then I don't want to pay as much because I'm not I can't count on it. Right.

[00:12:42]>> Right.

[00:12:43]>> So So then and then also that the fact that um is it 300 megawatts? Is it 445?

[00:12:51]If it's 445 megawatts and it's $7 billion times 2.4 then it's about 16 billion. If it's 300 megawws then maybe it's $5 billion time 2* 2.4 for then it's between 10 to 12 billion right but it's still serious money that they're doing from that and if I triple that that's you know $30 billion right if I do one gigawatt in 2028 and and so >> yeah complete that thought >> and then the other thing is that um the cost of me launching it because I have fully reusable rockets means that you know it's like $2 million per launch instead of like you know the you know, like have to to rebuild rockets and do whatever. Um, so then that's um more like $100 per kilogram or whatever like that, right? So, so my cost has gone way down and the building the satellites might only be two to five million. Well, for the solar and for the not the chips, the chips cost the same, right? Although I may switch over to AI5, AI6 chips instead of >> So that's that was my next question is and what which of what is it AI6 or AI7?

[00:14:07]because I if I'm not mistaken um at his terapab presentation Elon indicated it will probably be AI7 that will onwards that will be for the orbital data centers >> and there's a good chance that those will be built in terapab >> right >> so then the costs will be a lot lower >> well I think they could also just be built by Samsung and TSMC or Intel >> okay >> so the production of those chips AI6 AI7 um probably not be terab in 2028. I wouldn't count that. I would think that they'll be making it from um the three other suppliers that they have. But the thing is I do not think they launch Nvidia chips. The reason being is Nvidia chips cannot run at 300 degrees. You want to run them hot so I can make the radiator small, >> right? So it's for AI intention in space. I want to switch over to these other chips that have different physics so that I can run them hot which current chips are not designed to do. Right? So whether Nvidia starts making chips that do that for you, designing them, having them all made by TSMC or whether Intel with uh Tesla, SpaceX do that with AI6, AI7, right? It's probably not be it's not AI5. It's going to be these new chips because they need to run hot. They need to do different things than what we're doing now.

[00:15:45]>> Okay. So, how does two questions there's this there's this whole debate about the shortage of memory, right? How does this come into play? And this deal that Intel has with um Tesla and SpaceX, how do these two things come into play?

[00:16:05]Because um in the green room before we started recording, you were telling me about the potential for Intel. Now in the the Trump administration has a stake in Intel, right? So Intel has a special kind of a >> he's in a special place of its own, right? Is that >> right? So um the fact that Intel can build both chips and memory, >> right? That makes it uniquely placed uh in this entire ecosystem that's developing. How do you see that playing out?

[00:16:40]So the overall demand like if I can make more gigawatts right um and they're in space versus gigawatts on the ground I'm increasing the demand for memory because the ones in space also need memory right all the AI all the agentic stuff needs more memory so I will be able to deploy and power up CPU GPUs and memory right so I'm increasing the overall TAM market right so then we'll need more memory and we'll have the ability to bring online more memory, right? And more chips, right? So, if I can get towards by 2030 10,000 launches, then I can launch 100 gigawatts, that will be about five times more than the 20 gigawatts per year we're doing now, right? So, even if we somehow find ways to increase the 30 gigawatts with distributed AI or something like that, that's still three times more. So, I've gone from a little to a lot. Once I've crossed over into tens of gigawatts, I'm massively deflecting the market where it's like instead of 20 gawatt, we're at 30 because I added 10. If I add 20, I go from 20 to 40 because the amount of power added on Earth flatlines because we can't build more grid and more more power, right?

[00:17:58]>> I can make more solar and just toss it into a rocket and launch it. I'm not limited anymore. I can make because each one of those solar solar on earth is unreliable, you know, day, night, clouds, whatever, right? Pollution in space is like a nuclear reactor. It's running like 99% of the time.

[00:18:19]>> Okay. So, let's now look at all of this in the context of the SpaceX IPO.

[00:18:25]>> Mhm. So uh in our last chat we talked about the concerns the cash burn concerns um for the XAI part of SpaceX AI and how the deal with anthropic alone um tackles a lot of those concerns right >> now when you look at um when you look over the horizon at this entire ecosystem coming into play and given all the constraints on on on the ground for expansion of data centers, construction of new ones and you have this entire blank canvas so to speak in space that Tesla um SpaceX XAI will have almost a monopoly on because of their launch capabilities, right? Mhm.

[00:19:17]>> How should people view approaching the SpaceX IPO given all of these factors and keeping the next the next 3 to 5 years in mind let's say till 2030.

[00:19:31]So one um just with continued expansion of communications they will make more money and be super profitable, right? So already they have like 12 million customers um with um you know highspeed internet, right? All they have to do is keep making more dishes and they're increasing from 25,000 per day to 50,000 per day from the factories. they could stop launching satellites and they would be able to go to 30 million to 50 million customers no problem. Right? So they can go to 50 million customers five time four to five times more than they have now without launching another satellite. Suddenly they can't launch anymore. Something happened. They they can't launch anymore. All they have to do is keep making dishes and then they will have more customers and they will, you know, I'm presuming that they'll charge less in order to get more customers, you know, like do deals to undercut um, you know, fiber or other things. What they're already trying to do, they they have a $50 per month tier, $80 per month tier, $120 per month tier, right? So they can go to, you know, 40, 60, 100 and get more customers to to do that. So that's less than the $100 per month they were charging before, right? Um, so they'll have more customer, let's say at the $60 level, you know, average, more, some higher, some lower, $60 per month, $700 per year times 50 million customers, right? And then that's $35 billion plus another $10 billion or more from launch, you have $45 billion. It is and a lot of that's profit. Like say $30 billion per month, per year is profit. So you're already at onethird the profitability of Nvidia just from that nothing nothing launched but I'm going to launch more you know I'm going to double my capacity even without V3 right so then with V3 I can go to even more I can get to you know 200 million customers or something like that which is what um ST Micro electronics a supplier says that they can do um so at 200 million customers you know you're looking at 150 million dollars per year money your Nvidia money which means um but you probably grew faster five years to grow faster um and you have these other growth curves where I can do AI where I can do like the forward will be continued growth right so so that's how I see them like it's kind of like worst case you're at Nvidia or double not investment advice Yeah, I was just going to say that. Not investment advice. Do your own research.

[00:22:17]But um so then you have you've got Starink in the near future that's going to propel propel um the earnings um for at least the next two years. And this is a a massive TAM. Um >> director cell >> I think I Yeah, director cell again. I think I think the TAM people don't really understand the TAM like take for example just India alone >> um in the rural areas is 250 million 2G uh users.

[00:22:52]>> Mhm.

[00:22:53]>> Yeah.

[00:22:53]>> That can easily switch over to uh direct south.

[00:22:57]>> Mhm. And the advantage being because in rural areas and in the mountainous zones it's very uh it's very capex intensive to lay cabling um for fiber or to set up towers. Right. Because of the terrain, right?

[00:23:12]>> Right.

[00:23:12]>> Very easy lowhanging fruit for stalling to solve that problem.

[00:23:15]>> Yeah. And then you've got like another five to 600 million users um in towns, cities and the urban centers, the metros, right? So just India alone potentially >> has a market of easily 500 million, >> right?

[00:23:36]>> And then you look at Africa and you look at all these countries and then we discussed in in one of our previous chats and I'll link that at the end. um the cost advantages that SpaceX can bring to bear and undercut the market >> right and they can also do some volume deals you know with the cell phone providers where say I will take out your need for some you know 50 to 80% of your cell towers because I'm going to cover all that out there. So instead of you spending $10 billion a year or more collectively on all these cell towers, now you can just ignore all those rural cell towers, all those suburban cell towers, and you'll just use uh SpaceX.

[00:24:16]So that is just, you know, cash money B2B enterprise to enterprise. I'm going to take out a bunch of capex that you don't need to spend. So that will um you know so launch 10 20 billion um another 10 billion from from from that from back hall moving data around for the so again these enterprisey things >> and then the um potential for high speed internet going to 100 billion $200 billion per year the cell phone 100 billion and then on top of that AI with this unlimited thing of trillions of dollars >> so it's it's a layered evolution and a layered explosion right >> of revenue that we're expecting from SpaceX. And I so as in my intro I I I I promised viewers that we would help them get a sense of how to approach thinking about Space X AI.

[00:25:08]>> I should have added Starink into the mix, >> right? But for the AI part of it, yeah.

[00:25:12]Yeah. AI part of it comes later, >> right? But the 2028 thing where that really takes off in a serious way, right? is that they bring uncompetit, you know, blue blue ocean, no one can compete on it because you can't really do it at scale without Starship.

[00:25:30]>> So once Starship goes, then they get to thousands of launches, which they can do by building more um launch sites, right?

[00:25:38]Even if I can't, you know, get okay, one launch every few hours, which maybe they can get to, right? If I but they need to get away from the air traffic. So >> yeah, >> they can get to thousands of launches just by saying I make five new launch sites, you know, a billion dollars a piece, $2 billion a piece, and I launch away from air travel. Then I can launch 500 times, like twice a day from the locations. That's can be five, you know, 2,000 3,000 launches just from that, right? So plus my 500 from my regular launch sites. so that I can get to this 3000 level, which would be 30 gigawatts of of capability, right? So 30 gigawatts is as much per year as everyone else. So now if I have 30 gawatt of extra compute that equal to everyone else then if everyone else is still short of AI compute then just like entropic every company and maybe just in big may be big enough by itself but other companies will say I can't get my compute I can't do the stuff I need because I don't have the energy in the compute.

[00:26:48]>> Yeah.

[00:26:48]>> Then SpaceX has unique position of come do a deal with me. you want to become able to deliver this money, get make this make this profit, you have to do a deal with me. So they become >> there's not a single AI company that that would want to scale at the level that the competition is, whether it's Open AI or or even Google for that matter. I mean, once once you run out of space on Earth, you'll have to do some sort of a deal with uh with SpaceX. in the 2030 page report that Google did which u um it was um uh what was it um star something or other. So they um so they they made a report where they said okay we have a team of you know 10 50 guys and then we tested our stuff the AI in space we did a design right they made that report in the report they say and then in 10 years we'll be paying SpaceX to do the launches they say of course we're going to have to pay them to do the launches we hope they only will charge us you know 50 or 80% margin right but we'll pay it right so they just said our plan is you pay SpaceX to do this.

[00:28:01]>> So, that's it's pretty crazy when you actually begin to wrap your head around how all of this is kind of like layered explosions um in revenue. To to your point about um you know, being able to launch at a cadence away from air traffic, there was this recent news about um a potential SpaceX XAI campus in Louisiana.

[00:28:25]>> Mhm. Um, I think that was really interesting and uh maybe we can get together for >> I don't I don't think that's anything to do with launch. I think that's some factory or something like that because it's not in the right location for to do a launch. You need to launch over water where the if the thing fell in the sky doesn't drop in anyone.

[00:28:43]>> So Louisiana's the wrong spot.

[00:28:47]>> Okay. So then now let's look at the IPO.

[00:28:53]Most of the concerns have now been addressed, right, about revenue versus cash burn. And then you have these explosive layers of of technologies and markets and dam and growth.

[00:29:07]>> Um, Starlink would be the first rocket that powers it and then you'd have and and other launch and everything else as you said.

[00:29:15]>> Um, you'd have data, direct cell, all of it. And then that's going to push it till 2028 at least. Not that it's going to stop, but that's the first kind of, you know, the launch rockets, >> right?

[00:29:27]>> Um the seale uh Raptors, so to speak, >> and then the vacuum raptors, which is the AI uh orbital AI uh data centers that those kick in 2028 onwards, right?

[00:29:41]So then when someone's looking at SpaceX as an IPO, why >> would there be any doubts about the choice? It's pretty clear to me. I I would rather deploy my money into SpaceX for this for the rest of this year post IPO than Tesla.

[00:29:59]>> Um Tesla. Yeah, I think well the rest of this year is an issue because there's a bunch of um uh lockup periods. There's a bunch of financial things that can um mess things up because um people will be allowed to sell. So there's this supply demand thing that's kind of chopped off.

[00:30:18]>> People won't be allowed to sell. being >> well they they won't be allowed to sell and then they will be allowed to sell.

[00:30:23]So they're unlocking rolling unlock which is you know I think Larry has has described Goldber has described. So you know 15 days includes an index and you unlock something. So they'll be matching that up but there'll be a constant for the next six months of more and more people can sell right and then more there'll be matched buying with it. But during that choppy period the stock can be kind of like waving around based on you know a bunch of um chaos in terms of just financial chaos happening. But in terms of the underlying company, they'll go from like, you know, $20 billion run rate to $30 billion run rate. They'll because it's like that execution is um all I'm doing is is like printing out uh dishes which are the equivalent of small TV screens, right? So there's not there's no like magic to that. I'm just like the thing I'm already making, I'm making twice as many of them and I'm selling them. It's like it's like um brain dead level of growth. there's no risk to it, right?

[00:31:22]>> Yeah.

[00:31:22]>> Um so, >> so that will be happening and they'll continue growth through next year.

[00:31:27]>> Can you can you spend a few a few minutes on explaining this ro roll rolling unlock and >> what sort of a structured rolling unlock we we should expect. The first would be I think within 15 days of uh >> 15 days yeah >> of the IPO.

[00:31:43]>> Of the IPO 15 days of the IPO. Yeah. So um Elon owns 45% bunch of other insiders own the 10%. So you have 45% of 40% of the company owned by either venture capital companies other companies that are early into the stock as well as there's these secondary market things where um 10% or 15% were sold via things called special purpose vehicles where the shares were were in that. So people have been trading this thing for like you know six seven eight years where they've been accumulating this in each of the rounds right and then there's employees employees also got a bunch another 10% something like that so the employees and other people um the normal re high net worth retail that bought into this are selling but then they they say okay you can't sell until 6 months you can't sell until 3 months right so they're determining when this 20 30% of the company can sell as well as the 10% that are or 15% that are the um the funds. You know, some of those are has I have a fund I bought back in 2008, but then I I need to sell within 10 years because my fund has a 10-year lifespan.

[00:32:59]I need to then, you know, liquidate and distribute. So, there's reasons why people need to sell, right? And then >> and we'll we'll have a a clearer picture of of these timelines in the IPO documents, I imagine.

[00:33:11]>> Right. Yeah, there'll be some of the IPO documents. There'll be some of the road show where they explain how does that work because because the big finance people want to know, you know, you know, when this happens because it'll affect the price of the stock.

[00:33:24]They need to know how that plays out.

[00:33:26]>> Yeah. Because I mean, every every unlock is going to be a potential selling wave and then the stock falls and it's a buying opportunity for for pretty much everyone, especially if you're long-term into it.

[00:33:38]>> Right. Right. But as as we're seeing, you know, Goldman, um, Morgan Stanley, you know, JP Morgan, Bank of America, all these big banks and then like 20 other international banks are all involved in managing this where it's like they're trying to match up supply and demand unlock and then I get a inclusion, unlock, get inclusion. Six months, a big unlock and I get S&P 500 inclusion. So they're engineering and mapping out this thing so that they can keep the balance in check and ensure that the stock price goes up, right? So there's less available to chance around this thing because of how they're working on it.

[00:34:21]>> Yeah. But I would imagine there'd be a lot of high netw worth uh retail uh a lot of employees of SpaceX that would just want to say, "Hey, you know, I want my millions, >> right?" And I would and you know because they they've been living life on the edge like so many cases of burnout they just want to retire. So probably you're looking I've heard seen estimates of like 160 people with over $und00 million after the IPO and then probably like I don't know 40 are billionaires right so a bunch of >> from employees from SpaceX employees >> employees big investors coming in I would think half employees half other people because if you were an executive right you know not Elon like Gwen Shotwell probably has like you know five billion dollars of But another executive in the tiers below, say 10, 20, 50 of those guys, if they had gotten like >> a million dollar worth of pay or or or stock back in 2010, right after the IPO, then they have $100 million. A million dollars of stock back then became $100 million. So if you have an employee who's been there for like six, eight years at executive level, they probably have $100 million, right? If they were lowered and only got $100,000, right, of stock back then, they're probably at the $10 million for this level, right? And then that's assuming the stock doesn't keep going up, say to $5 trillion, then more people get to that level, right?

[00:35:56]And then similar, someone invested $250,000 back back then, they will have $25 million on IPO day, $50 million when the stock doubles, right? So, there's quite a number of very wealthy people coming off this thing.

[00:36:10]>> Yeah. And I'm sure I mean, pretty much everybody who has the opportunity will be thinking about taking some cream off the top while leaving, >> you know, while there's still a lot running. Still believing. Yeah, that's right.

[00:36:20]>> Yeah, >> that's right.

[00:36:21]>> Yeah.

[00:36:23]Well, it's going to be a fascinating u second half of the year. Uh I think we're still looking at June uh for the IPO.

[00:36:30]>> I think late June uh early July is the early >> July.

[00:36:34]>> Um we should be seeing the IPO documents come out in the next couple weeks. Next week is week after. Yeah, >> it's going to be a big exciting week and uh I'm looking forward to >> getting the analysis uh with you. So Brian, we'll leave it at that for now.

[00:36:48]Thank you so much. I hope we've given uh uh you've given our audience um a kind of a holistic 360 view of how to look at SpaceX as you move uh to 2028 and then 2020 and then beyond with the layered explosions of of revenue generation from different technologies whether Starlink or AI data centers or even launch and then of course the all important thing everybody wants to know how do we look at the SpaceX IPO and the next six months. So thank you this has been really helpful appreciate it.

[00:37:18]>> Thank you. Thank you.