원문 보기
Wow, there's a lot to talk about regarding the SpaceX situation, particularly the blocking of shares that may not have expired yet. We'll talk about it. About the technical rebound we are seeing in SpaceX shares. We'll talk about it. We will also talk, of course, about Terafab. Actually, that's where we'll start, and we'll end by talking about Qute, Iron, and other Neo Cloud data center projects and their revenues.
We'll focus on the facts because, as always, there are a few fools on the internet who like to make a lot of noise. I do n't like Kevin's headlines. And then, you know, we forget that there's often a 30-minute video with the facts. And that's exactly what we'll do— focus on the facts, not on one sentence.
We will focus on what I think 99% of sane Tesla, SpaceX and Elon investors care about, including me , since I started sharing Tesla facts back in 2017 when I bought my first Tesla and I was a fan. Of course, I'm not so deep that I can't criticize. This led to several interesting results. Although I was at a Tesla shareholder meeting, convinced Elon to promote Tesla features, and then used AI to get Ryan Reynolds to promote Tesla, which went wildly viral, leading to a duel between Ryan Reynolds' Aviation Gin, who responded to my AI video of him promoting Tesla with his AI- generated Elon promoting Aviation Gin.
It was epic, until, of course, I received a very stern letter demanding that I delete my post, which I subsequently did. So, where did we stop? Let's begin. Terafab, fell energy, free electron lithography. If you're not familiar with this, by the way, "fell," the acronym that Elon used to confirm this " Easter egg" in the video, is...
something that Tolkien took from the Lord of the Rings series, yes, the author. This is also from World of Warcraft. It turns out that there is a bit of this filth energy behind me too. But free laser lithography is really cool. It's conceptual, it's academic, and I'll show you where it comes from. Overall, here you have this futuristic version of Terafab, obviously CGI.
We have Robo Vans, maybe Grok (Grok AI) put it all together. This is all very cool, this multi-level highway. You know, everyone always wanted this. We love trees, you know, it's cool. Eh, building upwards—we thought there would be more of that in the future . This is really selling us the future. We haven't seen much of this yet.
We're obviously looking at the giant Starship Terafab, but there's an Easter egg here. And this " Easter egg" is connected to the circle on the outside, which Elon has since confirmed. This is essentially a particle reactor. The light is amplified like a laser in a tunnel, an electrified, extremely bright laser.
And you use this to essentially etch wafers beyond ASML's current capabilities, at a level that has never been achieved before. If Elon can do this, it will be a decade-long leap forward in chip manufacturing. Of course, we've had similar resumes before, for example, regarding 4680. We're going to hire the best people.
We have the most ambitious battery technology ever . This will be a structural battery. This will make the car safer. The output will be higher. Costs will be lower. Boom. What happens in the end ? We are getting delays. We have nominal output figures , and reality is never as interesting as expectations.
Well, and that's why we have to be a little careful. SpaceX and Elon have no experience running a factory. They have a testing or research center at Gigafactory Texas. Eh, this is a research center where they're trying to start producing records, and that's cool. This is great. Well, and all this is very ambitious.
Tesla operates a research factory in Giga Austin, and this planned facility will come to Grimes County, Texas. 100 million square feet. Expected investment is somewhere between $55 billion and $119 billion. But, just so you know, that ring you see around it should be taken as a call option. In fact , to be honest, a significant part of this object should be perceived as a call option.
Well, the projected timeframe is that they will have test batches in 2027, sometime in the second half of 2027. Perfectly. I call this a call option for 2030, especially considering that data centers are not that easy to launch. For example, let me give you a quick forecast regarding the time frame.
Texas Governor Abbott recently ordered an audit of every data center requesting permits in Texas. No approvals until audit is complete due to concerns about energy and water supply. Now SpaceX's lead developer at this facility, Kyle Wiley Treadwell, has suggested that we're going to build a massive natural gas-fired power plant, or rather, several of them .
We are going to build large battery arrays to store this energy off-grid. This is great. If you can do it off- grid, you'll bypass some of those Greg Abbott audits, but you still need permits for those gas turbine power plants. Those gas turbine plants look like a real natural gas facility, unless you use something like Bloom Energy , which can be done very quickly.
You know, Bloom Energy, these things can be installed in about 3-4 months. If you're really going to build a true peaking natural gas power plant, you're probably looking at 3-4 years to get the capacity you need. This is obviously a downside risk, and that's why I see 2030 as a realistic option for this Terafab.
As attractive, desirable, and exciting as it is, there is a real risk of delay. Keep in mind that this is very different from micro and macro hard. Micro and macro hard are not for orbital computing, while Terafab is expected to be 80% used for data centers in space. Orbital chips that will be used in satellites.
20% potentially for full autonomous driving and Optimus robots. This is where the partnership with Tesla comes in. But these, you know, these would be AI 6 chips, whereas the orbital chips are supposedly called D3 chips. Cool ideas. I hope we can get them. But what is much closer in time and more important to our conversation, I think, is what is happening with the lockup shares and Michael and Macrohard.
So, Michael and Macrohard are objects on Earth. This is a calculation for Earth. These will most likely be mostly GB300, Rubin and subsequent generations. They are expected to appear between 2027 and 2028 . This is great. This is a land purchase of more Nvidia chips, which Elon confirmed. They are going to buy more Nvidia chips.
This is actually great for Nvidia stock. Nvidia shares responded by rising, while AMD shares responded by falling. Nvidia is currently close to its all-time highs. What's driving stocks today? We obviously also have SpaceX moving, which we'll talk about. But you know, Nvidia still, if you can assume that their growth rate will remain relatively high, then they look relatively cheap given that growth.
I still argue that they can reach $300. It's not my favorite investment because I think we're in for maybe another hardware cycle, another good boom in hardware stocks . And then we're essentially approaching the end of the hardware cycle. I'm still pessimistic about memory stocks themselves, mostly because we expect high cyclicality and supply is increasing, but that's all for another video.
More interesting here is not only Nvidia's potential success, but also the situation surrounding SpaceX's stock, lock-up periods, and our thoughts on the competition, especially with respect to macro, microeconomics, equipment, and rental rates. So, we need to talk about this . So let's talk about this lockdown.
First of all, we are approaching the point where we can draw correction lines. We see that the first test of the correction will be at $133.34, if we take it from peak to bottom. We expect SpaceX's free float to increase from 4.9% to 11.8% on Thursday. A lot of people have asked me, " Kevin, by the way, yesterday morning on the live broadcast we were talking about this.
A lot of people have asked, " Hey Kevin, you know, the end of the lockdown is coming." Are we expecting a big sell-off in Tesla shares? Yesterday morning during the stream I said, "No. The news of the lockout is probably already priced in. In fact, the price is probably going to go up, and selling is not as easy as it sounds.
Boom. What do we have? The stock bottomed yesterday because people were afraid of the huge selling pressure on Thursday. That didn't happen. The stock went up 6% because of buying on the dips. People thought the bottom had already been passed. If this is the first unlock and there is no damage, then the bottom has been passed, and the stock is up another 11.1% today.
That is, to be fair, what we talked about on the member stream. By the way, you can always join us. You get lifetime access here. Join us before the big price increase in August; we have two big price increases coming up. Remember, you can probably write off the cost of this membership on your taxes and participate in the member streams.
You get all the trading alerts, courses, analytical reports that we do every morning. Most tax professionals say you can write it off as an expense . You get access to what we do here as outlined on mekevin.com. So come on, join in and take a look. But let's talk about this. So why is this actually happening?
Why didn't the sales restrictions work? Well, a lot of it has to do with the serious multi-step operational hurdles to selling. For example, I got an email regarding our stake in SpaceX, and we have a sizable stake that we acquired back when the company was worth about $300 billion. The email that I got indicated that they actually need to remove the restriction stamp from the stock certificates at the capital registry and at the transfer agent.
Once that is removed, and the process could have started yesterday, then the shares can be transferred to the fund managers. And then from the fund managers they can be delivered to the actual brokerage accounts of the investors. So, for example, my venture capital fund is an investor. It's not a fund in the sense that it's a manager that manages individual shares of individuals.
It's a venture capital fund, right? So it has its own, essentially, brokerage account. So what happens is that when the transfer agent stamps those shares, which legally could have started yesterday, they're sent to the fund manager, who can then distribute them to the end investors, who can then sell them.
We have no idea when that's going to happen. And that's just part of it, right? Because obviously the sales restrictions are ending. That's just going from 4.9 to 11. We expect the free float to be about 50%. So that's effectively 10 times what it was before the IPO. Also, some of the other shares, especially from the XAI merger, actually have to wait until February 2nd because of the 12-month holding period under Rule 144.
So in practice and in summary, I think many people think that The lock-in period went smoothly. Wow, all of that was absorbed by the market. But the pressure from sellers who were locked in has n't really even started yet. And that's okay, because it takes time. To be fair , the markets have already priced in a lot of that .
So the question now is, can the buying back of the drawdowns that we're seeing on SpaceX offset the real pressure from sellers that has n't really started yet? That remains to be seen. So expect more volatility, in short. You can use technical analysis to help you, mainly because I think people are going to use TA to pick a good time to get out, you know, when they hit different levels of divergence.
And who knows, we're actually going to go higher. Honestly, there's one more thing you can do : you can set up, you can draw an extension down, uh, drawing up from the low, instead of drawing an extension up. And so if we do that let's do it, let's see. Let's draw our high of 64 at 1.382. There are slightly different lines here, which is quite expected.
But, uh, yeah, I don't think we're going to go down to 30. So I don't know if that's a useful line. We're going to get rid of this set of Fibonacci levels right here. And that's it. Let's stay with it. Let's stay with the extension. Mainly because if this asset goes down to $80, $60- $80, that's a better place to buy, in my opinion.
I'm not saying you shouldn't buy at a price a little above 100, but I think it's better to buy there. I think a lot of people will have bought by then. Uh, anyway, what do we have with the cloud situation? Okay. That's where things get really interesting with Act and we have to talk about that, because the strong demand for cloud services that we're seeing for SpaceX Colossus 1 and 2 from XAI, is quite reasonable .
We're seeing a massive acceleration in demand through Codex and Fable. The first quarter really led to a tipping point in annual recurring revenue for these companies. And Groq's ARR today is up 50x, you know, 50 billion ARR. And GPU rental prices have also jumped dramatically. That's actually good for SpaceX.
So if you look at spot prices and GPU rental costs, you see that there was a certain plateau in 2025. We really saw a rapid increase in the first quarter. The hope is that that will continue, especially with the GB300 chips, which are obviously the newest, so they have the highest premium value. You can see the A100 here, they're still making money, which is great, because these are 8- year-old chips.
Which is good. But data centers, you know, take a long time to get ready to go. So it's only natural that they'll hold price advantage for a while until we actually build this computing power that everyone is expecting, including Terafab. So the big debate that's going on right now, or should I say the big debate that's going on right now, is whether Elon's projected rental rate is reasonable for SpaceX?
And Elon's projected $30 billion to $40 billion per gigawatt or $30 billion to $40 million per megawatt. You know, it does n't really matter. It's just that 30 to 40 is something to think about per year. And some people say that's three to four times the industry average . And that's true, it is. Other people, like one person at X who I really respect, who doesn't get into emotional comments , because some people get very emotional about these things.
This person, Southern Value, says, "Oh, you know , if AI labs pay Elon's rates of $30 billion to $40 billion dollars, they're not going to have any money left over." They're not going to be able to pay those rates. They're going to have no money left over because that's effectively all their money. So that's not good.
Another thing that's not good is using a wet wipe when you 're trying to clean a dry erase board. They're actually called dry erase markers. I don't know if that's why they're called that, because you have to write with them when they're dry, or because they're dry erase. That's probably more likely.
But anyway, the point is that this $ 30 billion to $40 billion prediction by Elon is really exciting, and it's leading a lot of people to assume that they're going to be able to actually create a profitable business in the AI vertical. The big question or the problem that we're facing is this. The SpaceX contracts that we're finding are for some reason significantly more expensive than the rest of the industry.
And that's the question. Why is there such a gap? So one possible explanation is affordability. The premium, so that Google and Anthropic want that availability. So they're willing to pay higher hourly rates for GPUs. And maybe everything in a similar cluster. And maybe that's also true for MicroMacro hard.
The big problem that we see is not that Elon has managed to get a premium now. That's not the problem. A few things to sort out, and then we can really see the problem, right? So first we have the revenue premium. Okay? Elon is at 30-40. And we have the costs. We'll talk about costs in a moment. So first on this revenue premium of 30-40, let's look at it when the market is now 10 to 15 million per megawatt.
Elon is easy on his projections, and his recent rates for Anthropic and Google are three to five times what the market can justify or prove. The only benchmark is Elon's deal with Google and Anthropic. And both have Reasons to be bullish on SpaceX stock. Mostly, what we talked about in our last video, Google's $ 94 billion exposure to SpaceX stock could lead to some preferential treatment, and Google's investment in Anthropic.
It's a bit of a vicious circle -- ensuring that some of these data centers are up and running. There's a lot of support here for Anthropic, which you know is a SpaceX customer, which helps Google. It's a big circle. We already know that. But the bigger problem is that companies like Enphase , Iron Core, Weave are leasing computing power for $9 million to $15 million per megawatt.
Let's use millions per megawatt. Again, you could count in billions per gigawatt. It doesn't matter. The ratio is the same. And that's where the concern comes in . Why is there such a big gap? People will say, " Oh, well, Elon gets a raise because he has spare capacity." But they forget to mention about the 90-day cancellation contract that we talked about earlier.
We already know from my last video that these leases have huge cancellation rights, which is not good. Now, about the $2.7 million per megawatt argument: that was in Tesla's IPO proposal, which they later removed, but Needham picked up on it and started pushing it. Remember , this is related to Bloomfield: no GPU, no service.
We already know that. This is old news. We've already talked about, how to say, the consequences of this being unrealistic. We're not worried about the cost of construction being unrealistic. We know that GPUs are expensive. We already know that. The big debate right now in the financial community is: Is this premium that Elon is getting sustainable?
And many will say no. Mainly because trying to prove that SpaceX's performance somehow sets the market benchmark. If it does, then SpaceX's benchmark will become the new market rate. If it does the new market rate, then Anthropic and OpenAI are likely to be far from profitable, which means they can no longer afford to pay those amounts.
I mean, we know they are far from truly profitable right now, but I'm just saying that the hope of that is fading. That's the risk factor—that people can't even actually pay those rates on a sustainable basis. They can only pay a fraction of those rates, not the whole amount. That's the big question.
But even if, say, they could sustain those rates, who is actually better off? If that's the new market rate for compute, and that's already built into SpaceX's pricing, wouldn't it be true that Iren, Ennemies, and Coreweave would have to raise their prices, which would make Neo Clouds potentially a better investment than SpaceX, because SpaceX already takes that pricing into account if they're the benchmark?
And these guys could catch up with the market when they renew their contracts. That's the big risk for SpaceX's compute— contract upgrades . Will they be upgraded? And the second part of the question is, will the new contracts actually be able to meet this benchmark? So if we can't upgrade the Anthropic and Google contracts at these levels, then the predictions are wrong by several orders of magnitude.
If the new contracts don't get done at this level, it's still very misleading. It's a fiction. And we won't know that until we see what happens to compute over the next year or two. But these are real risks, and I think the markets are getting a little ahead of themselves in assuming that this new benchmark is the absolute standard.
Because don't forget, there's more capacity expansion coming. And it's not just from Neo Clouds, it's even Meta. I mean, Meta is planning to sell 1 to 2 gigawatts of compute capacity between 2027 and 2028. By the way , they're leasing a significant portion of that from Google TPUs and using their own internal TPUs.
So, we'll see. The bottom line is because we're still in the decision-making stage. It's too early to say with certainty that there's no risk, especially with the 90-day cancellation terms, that there's no risk to SpaceX pricing when you know how much they're paying per hour of GPU for essentially the same chip.
It doesn't make sense. And I think that's the point to pay attention to. Let's make it a little bit more obvious. Let's put a huge arrow here, okay? You see this GB300 in the deal between Iren and Microsoft or Nvidia and Share AI. These GPUs are being leased for less than $3 an hour. And these GPU lease rates have gone up.
Look, in May, Iren and Nvidia's lease rates were $2. But for some reason, SpaceX's lease rate for GB300 is $11, and the blended rate of the H100 and H200, which are being leased for significantly less than some of the 200s , is for some reason $5. So somehow there 's this huge gap. Nobody really has There's no good explanation other than it's an "Elon premium" or an " availability premium." Well, these companies like I-rent clearly had availability and their capacity is growing , but they haven't been able to command that premium.
So why? It seems strange. We already expect CoreWeave to grow revenue from $1.3 billion to $23.3 billion. They're expanding, and their annual revenue (ARR) is expected to grow from $551 million to $7.9 billion by the end of 2026. I-rent themselves think they'll see ARR of $4 billion by the end of 2026.
And of course you have Microsoft, Amazon, and Google, who are also pushing their own data centers, along with Meta. We know that. Meta is mostly because of their off-balance sheet leases, which also have termination clauses. The bottom line is that it's all obviously a vicious cycle. But with For some reason, and this was the topic of the last video, Elon is getting huge bonuses.
And the movement of SpaceX stock in the short term, or, you know, someone tries to be rude on social media and says, "But your headline, Kevin, your headline is just this and that ." This does not fix the problem. We don't know what the renewal rates will be. We do not know what the prices will be for new contracts, which have no cyclical connection to SpaceX and do not have a personal interest in SpaceX shares.
And the technical rebound after the end of the lockdown period, which hasn't actually happened yet, right? Legally it happened, but practically it didn't. That's cool, but it doesn't mean the bottom has been reached, computing prices are guaranteed, and I'm not pessimistic about SpaceX, I hope it grows.
The more I...I just want to make this clear. I think people forget about that. The more SpaceX grows, the more money I make. Okay? It's simple. SpaceX is growing— dad is making more money. Luigi is very happy. Oh, I'm copyrighting myself, okay? Let's be honest about bias. I would really like her to grow up.
But then, Kevin, why are you making this video and highlighting these risk factors? Because that's what I do. I cover risk factors and events in the financial community. If there are risks that you should be aware of, it's better to be aware of them , because maybe someone listening to this is thinking, " Okay, maybe you shouldn't put everything on margin in SpaceX, because the risks are still there." Isn't that right ?
That is my opinion. I'm not here to give you personalized advice on what exactly to do with your money, because I don't know your situation. But I think you should know about these things so that when people blindly say, "Oh my god, Iron, Enphase, and Corning are definitely going to hit SpaceX's targets." "Uh, metrics set by companies interested in growing SpaceX shares." Not the best thesis, you know?
Just saying. Like when you have an extra understanding: "Damn it, Kevin ." That's why, even though I hate you, I still have to watch you. But thanks for watching, friends. See you in the next issue. Goodbye, good luck, and join us at Meet Kevin Al Connie. Communicate with me directly during live broadcasts. Get Alpha Board every day. Goodbye and see you soon. .
Comments 0