저자: Kevin · 외국인
요약 보기
스페이스X IPO는 밈 콜옵션 성격이며 상장 직후 제한된 유통주식(float) 때문에 단기 급등 후 락업 해제 시점에 큰 폭 하락 가능성이 크다. 우주 데이터센터·달 기반 제조 비전은 토큰 가격 급락, 우주 냉각 난제, 2035년 이후 실현 가능성 등으로 설득력이 낮고, 로보택시·옵티머스 목표 미달과 Starlink ARPU 하락을 근거로 스토리주라고 주장. 본인은 $3000억 시기 투자분을 가능한 한 빨리 매도할 계획이라고 밝힘.
다듬은 글
Things are going downhill over here. The question is, are things going to go downhill for that SpaceX IPO? Let's go inside this cave and talk about it. Literally, we're going to get inside this cave and we're going to talk about it. All right. So, let's talk a little bit about whether or not we're going to get hosed on the SpaceX IPO or what the play is here.
So, here's the scoop. Welcome to my cave. Uh this is where I'll be hanging out. So, here's what we got. We got 100 billion dollars coming from retail, 4X over subscribed. The Saudis want in, BlackRock wants in, every suit under the sun wants IPO shares. And so does every retail dude. Famous story, Bloomberg's like, "Yo, WHAT ARE YOU DOING?" I'M TRYING to get a bank loan FOR MY FRIEND TO BUY SPACEX SHARES.
Everybody wants SpaceX shares. So, what do you want out of this stock? To make money. Duh. Are we actually going to make money on this stock? Well, let's think about it. So, let's go through the facts. First, we're going to start with tokens, then we're going to talk Edge AI, space data centers. We're going to talk about what Elon is pointing out, what the goal is for the company, what you're actually investing in.
It's not just Starlink, it's not just rockets, you're investing in much more. We're going to keep it concise, we're going to get into the facts, and then we're going to go play-by-play in terms of how we think this is actually going to end up well, playing out. All right, let's get started. So, first, data centers in space.
That has been the catalyst for actually getting SpaceX to IPO. And they've really only recently come into vogue. In fact, in a very recent SpaceX interview between Elon Musk and two other individuals from SpaceX, they say, "Yeah, we didn't really have a reason to go deep into space. We've had these low or low orbit satellites that we've used for telecommunications, basically the Starlink network, which has massive pricing power.
I mean, airlines are begging to get their hands on Starlink. When I had a jet that I bought, and I'm like, "Hey, I want Starlink." They're like, "No problem. It's approved." And I'm like, "Great. When am I going to get it? Well, pay a $150,000 deposit, and we'll let you know." I waited 2 and 1/2 years.
They ended up giving me my money back because it still wasn't ready. But, anyway, it's kind of remarkable because it's a It's a massive pricing power business on Starlink. Now, profitability, something else, we'll touch on that. But, I want to talk vision here. The vision is all about what is going to be done in space.
And right now, the vision is data centers. We want those tokenomics handled in space because why? We can basically get electricity for a fraction of the cost once we're established because we can have 24/7 sun. In the terminator zone, where it's never dark, and float around with our solar panels, and basically print as many GPU tokens as we want with satellites, which Elon tells us it's actually going to be easier to make GPU-based satellites for computing purposes, whether it's training or inference, than SpaceX Starlink satellites, which is pretty remarkable.
But, let's understand a little bit. If the point of SpaceX is to get tokens mass-produced in space, then we're going to need a lot of exponential growth in token usage. Jack, walk with me for a moment. We're going to go a little deeper into the cave over here as we're getting deeper into the weeds, well, the cave.
Here's the problem. Right now, the last set of data we have on tokens comes from the Google I/O event. Google I/O told us that between 2024 and 2025, token usage 50x. That then between 2025 and 2026, 6.7x. Now, why is that a potential red flag? It's a red flag because during the same time frame, we've actually seen token costs completely collapse.
In fact, we expect token costs to fall 5 to 10x per year. ChatGPT-4o has seen its price fall from $20 per million tokens all the way down to 40 cents per million tokens. That is a collapse in pricing for tokens. We're getting more efficient with chips, we're getting more efficient with our LLMs and our usage.
That actually creates a little bit of a problem for SpaceX because if the purpose is to create this infinite supply of tokens in space using SpaceX data centers, where not only we're going to spend a lot of money to get them there, we'll talk about that in a moment, we'll talk about slingshotting satellites to space from the moon.
Yeah, we'll talk about that. But, if we're going to do data centers in space, we really need to have costs for tokens that are not depreciating faster than growth. In other words, if the price of tokens is falling faster than growth is rising for tokens, then we are approaching a commoditized product.
A token that doesn't eventually have pricing power. You become more of a utility like uh you know, a railroad or a utility company that supplies power than a company with big pricing power. And for this type of valuation at 97x, uh we're probably going to want to see a little bit more peppy growth. But, nobody's buying SpaceX for the valuation, they're buying it for the vision of what Elon is selling.
And that's where the interesting part evolves. See, how are we actually going to do data centers in space? See, usually when we think about space, we think, "Oh, it's cold in space, right?" The problem with space is you actually have really wild temperature swings because there's no atmosphere, so solar flares makes temperature skyrocket.
So, you have to deal with really, really hot and really, really cold temperatures. Problem with that is they're very challenging concepts built around how do we radiate heat away from heat that we are creating. There's literally no air. You're in a vacuum. So, it's not like you can just blow a fan and cool something down.
What do you have to do? You need ammonia-based liquids that can sustain those high and low temperatures. Those happen to be really deadly and noxious to humans. So, of course, we're going TO NEED THE OPTIMUS ROBOT TO HANDLE maintenance around these because spacewalks are going to get really expensive to service radiation or radiating cooling for all these data centers that really Google tells us probably won't be viable until 2035.
Which that so far introduces two problems. WE'VE GOT TOKEN COSTS FALLING FASTER than token usages is growing. That's a problem. And now we have really no path to viability for data centers in space until 2035, and we're already accelerating towards a commoditized product, which means we're kind of pulling in opposite directions here.
THE SPACE DATA center vision isn't dare I say the most promising right now. But, Elon sells the most gorgeous sex appeal for it because we talk railguns, launching satellites into space that we manufacture on the freaking moon. Talk about a stock going to the moon. SpaceX is literally going to go to the moon and harvest resources from the moon to make radiators and solar panels right there.
Maybe even just make all the satellites. I'm not quite sure who's going to work in those manufacturing plants. Obviously, Optimus robots, which we don't have yet because in January, Elon told us we've got too many problems with the hands on the Optimus robots and the batteries, which are a whole 'nother issue in the maintenance form, and we're still in the very early R&D stages, which is kind of where we sit with robo-taxis, both those over obviously at Tesla somewhat related to SpaceX.
Even Elon Musk trips up on the wording calling SpaceX Tesla, which he accidentally did at the JP Morgan conference, his presentation at that fancy new JP Morgan building in New York. But anyway, what's crazy about it is Elon's robo-taxis were projected by Bloomberg Intelligence to conservatively have 35,000 robo-taxis on the road by now.
Folks, we have 42. We have just 1% of what Waymo has on the road, and I'm not here to make a comparison to Waymo. I'm just saying, okay? Some of the goals on Optimus, multiple Optimus, right? And some of the goals, frankly, on, you know, getting our Optimus robots going have been a little bit of a miss.
So, the current vision to help sell that SpaceX is going to go to the moon is that obviously we're going to use the resources on the moon. We're going to colonize the moon. We're going to manufacture the satellites directly on the freaking moon. And then, because the moon has 1/6 the gravitational pull of Earth, we're not actually going to use rockets to launch satellites into space.
We're just going to use like an electric railgun and go and send the satellites into space from the moon. Which means the very rocket company that we are thinking about investing in, or already invested in, in my case, is actually saying they might not really need rockets at all to deliver satellites to space because they're just going to manufacture them on the moon and then send them from the moon.
So, we're not going to the moon, we are coming from the moon. I hope that's not like the waterfall where we start at the moon and we go down from there, okay? That's not That's not what we want. [sighs] So, doing this with a railgun, electromagnetic propulsion, you know, we we can get into that in a more scientific video.
Really a topic for a different video, but the point is when we actually look at some of the metrics of like Starlink and we try to get into the numbers. We've done this before, so I'm not going to rehash the old stuff. You can watch my other SpaceX videos on some of the numbers. But, something that I haven't mentioned before is ARPU.
Yeah, like, are you pooing? I'm about to go poo in a cave. No. Uh we Stay focused, KEVIN. STAY FOCUSED. ARPU, which is average revenue per user at SpaceX for the Starlink, was $99. Well, when they decided they wanted to IPO, they rushed to cut prices to boost user growth. So, now they're going into the IPO going, "We're going to use railguns to send satellites from the moon after we go to the moon and we're going to generate all the tokens in the world that we are going to need for all these robotaxis that aren't really being pushed hard right now and the Optimus robots don't really exist right now, which really call into question Edge AI.
We're going to talk about Edge AI in just a moment because it has It's like the opposite of data centers, which again, it's like this whatever. I'll talk about that in just a moment. We're going to try to coordinate that. But, Starlink saw their ARPU go from $99 down to $66. So, average revenue per user plummeted 1/3 when they were trying to advertise the pump that well, we have 50% user growth.
Oh, and we're now running Super Bowl ads. Which Elon hates advertising. So, he hates advertising, RUNS A SUPER BOWL AD, CUTS prices for Starlink, so he can pump the numbers and say, "Hey, we're up 50% on our user base." And then what? Then secretly after they provide their financials and say their user growth is up 50%.
They raise the prices 10 bucks a month to make it seem like they're going into the IPO with pricing power, but we have no idea what the churn is going to be. Now, the good news is the churn for like a United Airlines is basically going to be zero, but that user growth probably ain't going to be 50% anymore.
We will see. But I think the numbers going forward are going to be problematic. Now, the bigger issue that I face is what we're actually going to use all these data centers in space for because here's the reality, if we're going to use them for Optimus robots, look at what Figure AI tells us about their robots, the Figure robots.
We don't use data centers for our inference. We train robots. Yes, we will train robots. Let's go over here, Jack. We will train robots with data centers. Of course, we're going to train the AI. But, what's actually going to happen is when these robots have to infer or respond to what they see, they're not going to send that information to a data center and wait for it to come back.
There's too much latency for that. So, what do we need? We need GPUs inside the robots so they can perform on device inference. That's what Apple's trying to do. That's what the robot makers are trying to do. So, my point is we're going to go down here. It'll probably be a little echoey. It's kind of cool.
All right. The point is as we transition over here, the point is is really difficult to say. Can you come a little closer just so we don't have too much echo? That Meet Jack, by the way. Shout out to Jack. He's got a YouTube channel. He's probably going to post a cave tour. Posted a video on our Copenhagen stay.
It was a great Copenhagen stay. But anyway, the problem with these edge computing tasks is they actually killed the very need for inference at data centers. Yes, you're going to have to train a fleet of robots, but come on, man. We don't even have the freaking robots yet. So, we're a little early on this.
So, what does that mean for today's stock and should we buy it? Let's get back out of here and we'll talk about it, Jack. You guys need help getting out of here. You need to get out of here while I'm talking. Oh, he's getting the toilet. He's getting the toilet. All right. Good job, Jack. It's great.
So, here's what this means. This [snorts] IPO is the biggest meme call option since GameStop. You remember Vlad with five sigma event? Yeah, this is that all over again. Is We are You've already heard this from a million other videos. People like to go really basic level on this kind of stuff. So, I'm just going to keep this fast.
We already know that the Nasdaq bent over loosening index inclusion rules for Elon Musk to basically get convinced to list on Nasdaq. He was probably going to list on Nasdaq anyway, but you might not know that there are multiple stock exchanges. I mean, obviously you've got the New York Stock Exchange where I got I had the pleasure of ringing the bell.
And then you've got the Nasdaq, which is in Times Square. And they just opened one in Texas. You have choice when it comes to listing. That's America. Now, it's kind of a duopoly because Nasdaq and NYSE are usually where companies go. Now, Dow companies usually NYSE and tech companies usually Nasdaq.
Both of them were vying for Elon to list with them. Elon, in my opinion, basically took the NYSE bids or offers and shopped [snorts] them to Nasdaq. And Nasdaq's like, "Yo, we'll just change the rules. Oh, you guys are only going to have like under a 5% float for your stock? Hmm. And you want to be in our indices, which usually require at least a 10% float.
No problem. We'll change the rules, and instead of waiting 90 trading days, we'll try to get you in within the first 15 trading days, which happens to perfectly align with Elon Musk trying to get a 30% retail allocation. That is an allocation of 30% of the stock to retail, who are going to be locked up for 15 days.
So, in other words, the people who get an actual allocation, the people who are lucky enough to actually get an allocation at 135, who are locked up for 15 days, get to sell the same day that indices start picking up the stock. It's kind of manufactured not to be able to go down. They're low floating this.
You can look at a low float stock like Weble, SKYROCKETED WHEN it first hit, because it was like a 1 or 2% float. It was so low. And then, it bled out and tanked. The thing is, the tank or the bleed out is going to take some time for SpaceX, because the lockups from IPO tomorrow, where less than 5% of the company is available.
It's like 4.2% or something like that. To next year, you good? Still recording? All right. To next year at this time, 53% of the company is expected to be public, which literally means there are going to be 13 times as many shares of SpaceX available, assuming they don't actually issue even more shares, which they might.
There will be 13 times as many shares available a year from now as there will be tomorrow. Which means, if you are Here's the playbook, okay? If you get an allocation at 135, the sucker's probably going to pop up, okay? It goes to the moon, okay? Most people won't get an allocation at 135. They're going to be like, "Yo, I want 100 shares." They tell Robinhood, they tell Fidelity, they tell Charles Schwab, "Yo, I want 100 shares." Then they're going going an email, it's going to be like, you got 10.
Congratulations. Or you got 25. The rest of the money they have, if they go put market orders in on the stock, it's going to pump to the freaking moon. That pump over the next 30 days, I'm not calling the top in the next 30 days, but over the next 30 days, I think will be like that waterfall. It will mark the top and we will bleed out between now and next year when all the shares are available.
Not because of the fundamentals of the stock or the story, because the story is sexy. Come on. Rail guns on the moon? With the [clears throat] robots and the radiators and using the resources on the moon? It sounds straight out of a movie because it's basically just as much science fiction. But, with that said, people love [snorts] it and they got to have it.
And so, there are going to be a lot of people buying market orders and putting a lot of money into the stock. As the lockups release, the sucker's probably going to bleed. I would say, uh if not within a week after IPO, you know, once we get that index index exclusion, let's call it maybe 30 days after IPO to the end of the year, to next year, this sucker could be on sale 30 to 50% off.
So, if you're like a long-term holder and you're like, "Look, man, I don't care. I just want in." Keep some powder dry. You'll probably have opportunities to buy it cheaper. Now, that is strategic. Let's finish this up though with a little bit more information. So, first, we talked about that retail 30% fixed as soon as they get to sell.
When? 15 days after IPO, that's exactly when the index inclusion occurs. Not a coincidence. Major lockups step up throughout the end of the year. Again, a year from now will be about 13X. Now, as far as, let me just make sure I covered some other notes here. Ah, yes. A problem that I wanted to address that I forgot is that Elon tells us that making GPU satellites is going to be cheaper and easier.
But, the problem is, and you have to know this as a long-term investor, SpaceX satellites are manufactured with a 5-year life. If you take Blackwell chips, and let's just say they actually even work in space, and you shove these chips into space, and the satellites are going to stop working after 5 years, which is where their expertise is, and they deprecate them, and they leave them floating around like space junk, and let's say we're actually still booming and need more tokens and more GPUs, I'm going to create a space junk collection service, and in 6 years, I will fly myself to space, and I'm going to start collecting all of those deprecated SpaceX satellites, because those GPUs are still going to have a lot of freaking value.
Now, that part is probably a joke. But, if they're shoving GPUs into 5-year lifespan sats, that's a red flag. That's problematic as well, because now we have Now, we're taking technology that was for telecos, 5-year life, strapping on GPUs we should be operating for a lot longer. So, don't love that.
I want to see I like if I had a question if I had the opportunity to ask Elon a a question about that, that's what I would ask. Elon downplays how easy it is to cool in space, or how hard it is to cool in space. He downplays the difficulty. He says it's actually easy to cool in space. Most scientists disagree with that.
He does gloss over the details on how he's actually going to pull off cooling in space, because most people disagree. Cooling in space is surprisingly hard because you're in a vacuum. Again, airflow doesn't work. So, we talked about my expectation for the share price. If you get an allocation, great.
You're probably good with the allocation. If you've had SpaceX stock, like we bought it when it was back at $300 billion. If you ever want to see what we're buying, what we're doing fundamentally, you can always go to meetkevin.com, learn more about that. But, most importantly, we uh we got to look at what Elon has done recently.
Robotics and robo-taxi as our leading tell for what to expect with SpaceX. Not to dish not to dump on the the rocket launch business. Rocket launch business is great and sound, but even Elon himself says we might not even need the rocket launch business anymore for the satellite side of the business, which is exactly the side that people are excited about.
Now, don't get me wrong. There'll be other purposes for rockets. We got to get people or Optimus robots or whatever to the moon. Fine. Okay, great. There are a lot of ideas here though that I would say make this not something that we can really look at and say, "Hey, fundamentally, here's what the actual price of this company is worth.
This is a story stock. That's it." Unfortunately, the last time I heard somebody call something a story stock, it was JP Morgan saying that Nikola stock was a good buy because it has a great story. We're going to move up as a civilization. We actually look and and think about what we've been just sold, it should make a lot of us nervous.
And I've been transparent about this since the S-1 filing came out and frankly even before that. As soon as our lockups are available, oh, I'm going to sell. And I'm sorry. I just want to be transparent about that. Uh it is a venture capital bet that we made when the company was $300 billion. If we can get out at a billion bucks, great.
That would still represent probably a 50% decline though from the top. That would be my guess. So, that's my take. We'll see what happens. Who knows? Maybe we'll make even uh more money because maybe the meme lasts longer. But to me, the story is quite far-fetched. I almost think it's kind of on the side of lunacy.
But hey, then again, in order to do and make crazy things happen, sometimes you have to be a little crazy. So, maybe I'll be wrong. But I'll keep sharing the insights with you and you can make your own decision. . Go to meetkevin.com. Check out the coupon we've got for the program.
Comments 0