# Elon Says $100T Is LOW: How SPCX Compute Could Break Every Model 🚀

Source: [YouTube](https://www.youtube.com/watch?v=kmL-jIAApCQ)

## Welcome and today's focus

**Speaker 1:** [00:06](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=6s) Hello. A little different today: we're not going to do Bitcoin/MicroStrategy because nothing basically has happened. Every week is the same. But we will talk about Bitcoin at the end of this video, and we have a live poll going.

I want to thank everybody for coming. I really do appreciate you all coming up here every day, and thank you for taking the time. I think about this one a lot: SpaceX compute moats. I have been burning my brains on this one since even before the IPO. I'm just going to look at the business, the rumors, the mergers, the odds, the valuations, and yet another price prediction model. We're up to like 12 or 13 now. Just on compute alone, that's a very interesting thing.

Big thank you to CyberSniper, Sean D, TND, K8 for being here. And, of course, Bookhorn, GQTrader, and B-Man. Love you guys. Let's get into it.

This sounds like copium, crazy bullishness, etc., but I've done the numbers, and I've also analyzed the competition. I tell you, the results are insane. Anyway, hope my mic is working. I think it is.

[01:27](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=87s) Let's go. Not financial advice, just a guy on the internet. Shout out to Donna Felix: “I love the information you share with us, and I enjoy being part of your Patreon family.” Thank you, Donna. We love to have women in the community. We're still dominated by men, probably 92, 93%. But one day, hopefully, we'll get to 10% female. Anyway, let's move on.

## Musk's bullish SpaceX statements

**Speaker 1:** [01:47](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=107s) First of all, the big story is just how much Elon has been tweeting and how extremely bullish he is. It is almost bonkers. It's almost like he's hallucinating or something. But when you look at all the numbers and back them all down, maybe he's onto something.

Basically, he has never, ever in history sounded this bullish on SpaceX. He now is talking about SpaceX being worth over $100 trillion in 10 years. That's a 50x from here, from where we are, and he backed it up with post after post after post.

The reason I'm making this video is: should you own some SpaceX, and how bullish should you be? I'm going to share all the receipts in detail. This video today is backed up by many real, live tweets as well, so you can go dig them up yourself. Sometimes the best value of a tweet is looking at the comments below it too. Little tip there.

## Compute roadmap and deployment speed

**Speaker 1:** [02:53](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=173s) This is the SpaceX compute roadmap, Q4 2026, and I want to take a second on this. Not only is this compute roadmap insane, but they're ramping way faster than anybody else. They're the only people that can build coherent compute. They're the only people that can build the energy, the cooling. They can build a facility in 20 weeks. They can set up 100,000 GPUs in like 90 days—or whatever days it was that Jensen said it was superhuman. Most other competitors take two to four years to do this. This is the big problem.

I'm not talking bottlenecks. I'm talking scarcity. We're moving into a world where I can prove a thousand ways to Sunday that the demand for compute is going completely parabolic. But what's not going parabolic is the ability to scale to deliver that compute. That's where this story gets very interesting.

There's one player on the planet that can do it fast and efficiently and in a coherent manner. Nobody else has been able to stream more than 30,000 GPUs together. Elon does 220,000 GPUs again and again and again and again. That's the point of this slide.

Before we do anything else, I just want to lay the groundwork with this. Not hopium, but real facts. We have the earnings call as well tomorrow, if I'm not mistaken, and then Tesla on Wednesday. It's going to be a big week, an exciting week.

[04:22](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=262s) Colossus 1: 150,000 H100s, 50K H200s, 30K GB200s. Colossus 2: 110,000 GB200s, 440,000 GB300s. That's monstrous. Then he's building another 220K GB300s, which will be fully operational next week. That was October 1st. That's up and running. Another 220K in November. We're heading into November in, what, two weeks? And if we get lucky, another 220K—it must be that sweet, sweet spot where he can ramp these very fast—by late December.

![SpaceX Q4 2026 compute-roadmap slide showing the Colossus GPU configurations and proposed additional GB300 deployments, at 04:36.](figures/frame_0276000.png)

Putting this all together is a monstrous amount of compute. As I say, he or she who has the most compute wins. That's the simple story, but there's more.

## Brownsville facilities and revenue estimates

**Speaker 1:** [05:05](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=305s) This is a place called Brownsville, Texas. It's near Starbase, and they are building there and moving fast. I think this is where they're going to house not one of those little “micro harders,” these smaller buildings that have the same amount of compute, but a bunch of them at the same time.

According to CERN Basher, friend of the channel—shout out to CERN—he looked into this and some other data available. They have 444.45 acres. They have water, they have the land, they've got the buildings. They've got the 400 megawatts of wind turbines on site for one “mini heart,” and they'll do more.

They can make at least $22 billion a year on the 220,000 “mini hearts,” GB300s, and $30 billion on 220,000 Vera Rubins. When are they getting the Vera Rubins? Very soon. I will get to some rumors later that I heard from very good sources that we may be able to take to the bank.

They can spin all these up in 20 weeks. Six buildings could be $132 billion to $180 billion in revenue per year. That's monstrous. $180 billion is basically twice the revenue that Tesla makes after being in operation for nearly two decades.

Again, they've found a way of pivoting rocket scientists into building data centers better, quicker, faster than anybody else. That's the big story.

## $50 trillion, $100 trillion, and the addressable market

**Speaker 1:** [06:44](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=404s) Then Ron Barron came out, and he sees SpaceX being worth $50 trillion in 10 years. That's a 25x from here. Yes, 10 years is a long time. I know, I know. Stay healthy, my friends. That's what I say. Thank you for coming.

Ron Barron knows that he has huge bags in both SpaceX and Tesla. He's a big believer in Elon Musk and has been for a long time. Again, this is a combination of things. Starlink alone would be worth $14 trillion to $15 trillion of that $50 trillion. Consider one-third Starlink, but I think there's more value in a whole bunch of other things too.

This is 25x, but that's not the important tweet: $50 trillion, 25x. It's what Elon said. He said, “Unless there's a major world war, SpaceX probably gets to more than $100 trillion in less than 10 years.” More than $100 trillion, less than 10 years.

![Slide quoting Musk's forecast of a SpaceX valuation above $100 trillion in under 10 years, conditional on no major world war, at 07:35.](figures/frame_0455000.png)

This is my favorite Dr. Evil: “$100 trillion.” It's bonkers numbers. We'll get to the valuation model and the price predictions in a minute, and we'll get to why I've said for a long time how many SpaceX shares you need to have a different future. This is crazy, crazy.

He knows that tweets don't go anywhere. There are people that have bookmarked this a thousand times, maybe 10,000 times, maybe a million times. Again, greater than $100 trillion in valuation in less than a decade. Insane.

[08:10](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=490s) The other thing that's completely insane too is the SpaceX total addressable market. They had it in the S-1. I read the S-1 twice: $28.5 trillion.

James Katz said, “SpaceX claims its TAM is $28.5 trillion. No offense, Elon, but I will need to verify that independently from the comfort of my own bedroom.” Elon Musk said, “We were wrong. It's vastly higher than that.”

The TAM of $28.5 trillion, which blew people's socks off—they said it's wrong. The key thing is I build a lot of my price prediction models off of total addressable markets. Then I calculate market share. Then I calculate revenue and profitability and all that type of stuff. If the TAM is wrong, that means every model I have built so far is wrong. It means it's completely understated. Very important to understand how I model things out. Insane. Hey, Mike, good to see you.

## Orbital inference and Terafab

**Speaker 1:** [09:11](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=551s) Morgan Stanley says everybody has ignored the real engine. Think about inference from space, data centers in space. But also terrestrial data centers are going to be huge because of the point I mentioned at the very beginning. Nobody has enough, demand is off the hook, and you can't build these things. You can't hook up the power. You can't get the professional trades in, nothing.

Morgan Stanley was in a room full of people, and he asked, “Does anybody own SpaceX?” Not one hand went up. Not one.

He also said the whole SpaceX inference, data centers in space, orbit, etc., is completely ignored. He said it's very close to breaking out, gradually then suddenly, from Sawyer Vera 2. Musk responded to this too and said, “This is a very big deal.” Musk is basically—I shouldn't say pumping his SpaceX, but he's literally pumping his SpaceX pretty hard.

[10:10](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=610s) Then we have Terafab, which nobody believes they'll ever be able to pull off. But last week, we just heard that TSMC is in working with them. NVIDIA is in working with them. Intel is in working with them. Everybody wants to go to this party because it's going to be huge.

What they're going to do—again, of course, chips will be a bottleneck, power will be a bottleneck, etc.—but the value of the output of this factory will be in the top 20 of country GDPs in the world. I think there are 180 countries in the world. Think of the top 20: you've got Italy, Germany, France, U.S., Canada, China, Japan. That is his vision. Huge vision. Monstrous.

![Terafab slide describing the claimed GDP-scale output, first-phase investment and jobs, vertical integration, and partners, at 10:50.](figures/frame_0650000.png)

They've already talked about how much they're going to invest in the first phase: $17 billion, create 3,000 jobs, and it'll be completely vertically integrated. They're going to unthink how chips have been made in the past. They're going to do their own memory and everything else, but they've got help from some very powerful friends, which will help them pull it off.

Then I'm going to turn off my camera to the bottom right for this.

## Starlink growth and Starship launch plans

**Speaker 1:** [11:16](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=676s) Boo. Starlink, global comms, direct-to-cell, all that stuff as well. There's a guy called Jon Erlichman who's very good at predicting things like profit growth, revenue growth. Follow him on Twitter. He said Starlink revenue could grow 925% in five years.

![Starlink slide showing an analyst estimate shared by Jon Erlichman of 925% five-year revenue growth and Musk's “Of course” reply, at 11:40.](figures/frame_0700000.png)

That's a 10x in five years, my friends. Revenue, Starlink, global forums. Of course, Elon comes and gets on the bottom, and they said, “Of course.” Easy peasy, no problem.

Starlink, etc.—Ron Barron in the beginning said it's worth $15 trillion, and the growth rate is through the roof. Global comms: nobody will be able to touch them. Yes, you have countries like Europe and China and Russia that will all want to build their own satellite networks. They don't have a chance.

Now that Starship is running and moving huge payloads of 100 to 150 tons up in space with massive satellites—they put, I think, 60 of them in one of these Starships. When they open up, they're wider than a 787 or 747, whatever big plane it is. The world is so far behind. It's actually unfair, very unfair. Anyhow, that's Starlink too, and Starship.

[12:39](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=759s) Speaking of Starship as well, the FAA is reviewing a Starbase expansion and approving up to 456 launches per year. Get that: 912 landings per year. That means the Starship goes up with the booster and then the spaceship on top, and they both come down separately. That's why there's a two-to-one ratio between landings and launches.

I'm sure for the kids at home, we should multiply 456 by two. What do you get? Hopefully successful landings.

Anyway, that is more than one a day, my friends. There may be restrictions on weekends and evenings, I don't know, weather. But either way, it's nutty. That is in for approval, not approved yet. Last week as well, we had the big dip. Got time to buy some things because markets get rattled very easily.

## CoreWeave and the difficulty of building data centers

**Speaker 1:** [13:27](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=807s) We had what I call the CoreWeave kerfuffle. I hope I spelled that right. Their yields spiked, etc. But the point of this was, yet again, Elon responded to Zero Hedge, who talked about all the danger for CoreWeave and maybe bankruptcy risk probable in one or two years, all that stuff.

The thing that Elon tried to point out here, for everybody to make it abundantly clear what's going on, is CoreWeave kind of sells stuff, then it gets in money, then it buys stuff, and then it has to build the data centers and procure the GPUs, etc. It takes two and a half years, three years in some cases, to do it. It's getting harder and harder and harder to build these AI data centers. That's the point.

Elon's point is getting the money in to build is the easy part. What is incredibly difficult is to build the complex hardware. I know some people in the audience here work with data centers. I see you, and you know exactly how hard it is. But getting the money in to build them—the massive power, the cooling, the racks, the wiring, the connectivity, etc.—is extremely difficult.

Elon Musk has a crack team of SpaceX rocket scientists who are building the data centers. He said as well, “We even find it extremely difficult to do, but nobody does it faster.” If you buy these chips, the GPUs that cost billions of dollars, you want to be able to turn them on and monetize them as quickly as possible so you can get the money in the door.

[15:05](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=905s) Elon Musk, CFO for SpaceX, said they can get full return on investment on a data center in nine to 12 months, which is absolutely unheard of. Most people, it takes them that long—two and a half times that long—to even get it stood up, never mind turning on the revenue. Again, just like Starship and everything else, they have a bonkers advantage.

## Power constraints and compute scarcity

**UNKNOWN:** [15:29](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=929s) Now,

**Speaker 1:** [15:29](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=929s) the other big thing, which goes to the very first point I made at the beginning: is there enough power on the planet to turn all this stuff in the first place? Because we know there's plenty of chips, and the chipmakers are building them as quickly as they can.

This is North America. Think of the production of compute: 33, 35 gigawatts is planned for next year. But there is a problem. They may not be able to turn them all on because the power and the transformers may not be there. The shortfall could be as high as 13 to 15 gigawatts of that 33, 35 gigawatts. That's the big problem.

China: 40 to 45 gigawatts. They could also have a shortfall of two. That is the big problem. You can have all the chips in the world, but if you don't have the facilities, the power to actually turn them on, which could be a big problem.

![“Can 2027 Compute be Turned On?” slide with Musk's infrastructure warning and a regional production, deliverability, and shortfall table, at 15:55.](figures/frame_0955000.png)

This is where this is important. If the demand continues to go up—and I'll repeat this again—and nobody else can deliver compute, what happens to the price of compute? The answer is it goes to the moon, and SpaceX make even more money per gigawatt. They'll be able to pull in maybe $50 billion per gigawatt or $60 billion per gigawatt, which is all built and up and running.

[16:44](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1004s) The simple graphic here I'm trying to share is: as compute demand goes through the roof and nobody has the ability to deliver in time, or at least not for two and a half plus three years, the demand will explode. There's only one company in the catbird seat to take advantage of this, and that is SpaceX.

That's why they've been building hand over fist and have huge crack teams of SpaceX engineers building out these data centers, because it is a gateway to massive revenue and profits. Massive. That's why it'll be very interesting to be on the earnings call this week to see exactly what they've achieved.

## Tesla merger: facts, rumors, and prediction markets

**UNKNOWN:** [17:24](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1044s) Anyhow,

**Speaker 1:** [17:24](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1044s) back to the stuff you're interested in as well. Merger update with Tesla: facts versus rumors. So much stuff flying around—Polymarkets and Kalshis and experts and analysts—and everything is kind of at this stage like, “Just merge already.” You're tired of hearing about it.

First of all, Barclays came out last week as well. They modeled SpaceX acquiring Tesla in an all-stock deal as the likelier structure. Again, it was speculation. No, it's not a deal. Tesla has their annual meeting on December 3rd. There's no mention of a merger in the ballot, etc., so that's another thing. It's not going to happen that soon, I don't believe.

Tesla has, of course, a big multibillion-dollar stake in xAI, which became SpaceX shares. That's good. There's already kind of co-ownership there. But again, there's no 8-Ks, there's no press releases, no nothing. Elon is also hinting at it in some very breadcrumb-like “SpaceX x Tesla” tweets and things.

Of course, we had the car show last week. The best demo I've ever seen in my life. Cost a lot of eds, but it still was fun. Anyhow, these two companies are going together. The question is when.

[18:39](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1119s) When we look at the prediction markets, the odds of a SpaceX-Tesla merger spiked over the last few days to nearly 60%, circa spring 2027. That is, what, five, six months away? Not far. A 60% chance is pretty high.

![Kalshi merger-odds chart with contracts for dates in early 2027 and a slide headline highlighting a rise to 59%, at 18:57.](figures/frame_1137000.png)

If you go to Kalshi, you'll see the same thing. This could be Kalshi. This is Kalshi. If you go to Polymarket, it's pretty much the same thing too. The odds are spiking, so something is cooking for sure.

## How bullish? Moats, chip allocation, and financing rumors

**Speaker 1:** [19:08](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1148s) Getting back, before we get into the numbers and some new price predictions too, this is where it gets interesting. How bullish should you be? They have so many moats. They have so much ability to do stuff so fast. They've got Jensen Huang in the pocket.

Rumors I heard from a visual source: they've locked up 30 to 40% of the Vera Rubin allocation from NVIDIA. Some of those are going to go into space next year as well. They're building a special version of that that is going to be—what's it called?—radiation capable. For ICU, Cypher Sniper. He's a space guru, understands the importance of radiation-proofing your chips up there.

There's also a rumor of $40 billion NVIDIA financing, which is in talks with Apollo. Remember, if Jensen and Elon want $40 billion, they just wave their hand and say, “Give it to me.” They have no problem raising cash, no problem at all.

ARK believe Tesla-SpaceX will get to 66 to 67 gigawatts by 2031. That goes into my valuation model, by the way. Musk said he would be very disappointed if that's all they get to.

Again, they've got the moats, they've got the advantages, the ability to build. Nobody else is able to keep up with them or monetize things quicker. This means you should be very bullish, and I'm only talking compute. Nothing more.

## Compute-only 2031 valuation and the “retire-on bag”

**Speaker 1:** [20:27](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1227s) This is my latest, number 12 or 13—I do like the number 13, by the way—of SpaceX Elon Web Services compute by 2031. Price target only.

Again, the more compute they have, the more scale they have, the more profitability they have, so bear that in mind when you look at the EBITDA. We're going to go into way more depth regarding these price prediction models.

Just my base case of 66.5 gigawatts of online compute by 2031—Elon says it's going to be a lot higher than that—would take the price of just the compute business alone of SpaceX. Nothing else. No value attributed to launch or Starshield or any of the other goods, Starlink. Nothing. Just compute.

![SpaceX EWS compute-only 2031 price-target table with base and bull scenarios, including gigawatts online, revenue, EBITDA, and share-price estimates, at 21:00.](figures/frame_1260000.png)

That would take the share price, at a very conservative multiple, to $825 from here. That's a 5.65x. If we get to the bull case of doubling the gigawatts online, that would take the share price to nearly $4,000, to 21.7x from here. That is just compute. Please don't confuse this with anything else.

[21:43](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1303s) Remember, as I said before many times, the retire-on bag for SpaceX needs to be about 655 shares. If we go back here, 655 shares times that price target, $825, gets you about $540,000 approximately. Of course, nearly $4,000, 655 shares is about $2.4 million just for the compute piece.

The idea of that target, by the way, is what gets you to $1 million of value in the year 2032. Two, the point I'm trying to make here is the compute business alone gets you there very fast.

My job was to build a SpaceX retire-on bag. I got that done. Tesla doesn't go away either. They're going to merge. There will be a premium. We don't know how much yet.

## Bitcoin versus SpaceX polls

**Speaker 1:** [22:40](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1360s) Let's finish with—this was inspired by So in the community. He put together: you have to pick one. Which will outperform the next four years? By the way, 0.18 of a Bitcoin is $15,000. One hundred shares of SpaceX is $15,000.

The poll results were one vote that Bitcoin would outperform SpaceX. One hundred shares of SpaceX: 78 of the vote. Ignore the last part, brothers waste time. But we did a different poll as well. Shout out to the team behind me, who should have sent it to me. Here we are.

Which will perform better? The poll said 0.18 of a Bitcoin. Twenty-six percent of you believe Bitcoin will outperform SpaceX and outperform 100 shares of SpaceX. 0.18 Bitcoin. Seventy-four percent of you said 100 shares of SpaceX.

It's good to have both, but I do very much believe that AI is the faster horse. Bitcoin is there in case the world goes to poop, which it is. That direction is pretty clear.

## Closing and earnings outlook

**Speaker 1:** [23:50](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1430s) Anyway, that was the story today, my friends. I hope you enjoyed it. I hope you all learned something new. Big thank you to everybody for coming. Of course, huge shout out to Boat Corn again, GQ Trader, B-Man, Booth McDuck, Nick M, I am Colt, Holly Akayle, long-term holder.

Never seen a company like this in my life, ever. I used to say I've never seen a company like Tesla either. But wow, can you imagine the two of them together? That's going to be wild.

Thank you, Sean D, CyberSniper, T&D, Tesla. Everybody take care. Back tomorrow for Okta.

[24:21](https://www.youtube.com/watch?v=kmL-jIAApCQ&t=1461s) Get—start building a bag of Teslax, SpaceX, if you haven't already. I know that the prices are different. There's unlocks coming and everything else. But this will be interesting from an earnings perspective. Let's see if we get some nice earnings surprises too. Thanks all. Have a good night. Bye.