Why Elon Just Unlocked MASSIVE New Cash Flow for SpaceX AI

Added: 2026-05-28

[00:00:02]Today we are going to take a little bit of a strategic strategic excursion into the world of AI to be more specific into the world of AGI because a lot of things are happening and I just realized probably 48 hours ago that something incredibly big is happening in AGI that impacts Tesla, that impacts SpaceX, that impacts Anthropic and OpenAI and Google and most importantly all of us on a massive scale. And what that is has something to do with a cursor deal. And the fact that Elon made one of the most ingenious decisions in quite a while, which means something in Elon Land because Elon from time to time makes pretty smart decisions, but this one is a banger. So, in order to start this, I want to just share a little presentation I'm working on. Not the whole presentation. I just give you a little excerpt of basically three slides to kick this off to give you a bigger picture of the age of AGI and why this might be a very important moment. So you see it right here. I you know this is the slide where I start this and I'm basically saying AGI is here. I said it I remember exactly the place where I recorded that video in beautiful Vancouver, Washington. But I recorded this probably in January 2025 was definitely in January 2025 and said AGI is actually here. The singularity has started recursive self-improvement not 100% recursive AGI only but recursive improvement has started and we on the track of AG to AGI and I predicted in 2025 that this has started and will be completed economic AGI the complete replacement of all white collar human work and with all I mean 95 99% because there's always some tail end but it has started and will be completed by 2027 in 2027 and as things stand. Back then, most people thought I'm crazy. Now, I would say I was probably overestimating the amount of time it needs. I think we will be there potentially in the next 6 months. So, in 2026, economic AGI means that you can replace basically all mid-level and even higher level white collar jobs. All doctors as long as they're not doing blue collar stuff or like physical stuff, which doctors of course do. the doctors are a little safe as long as they have some patient contact and that patient contact is actually necessary but all the intellectual work forget about it all lawyers all accounts and sorry to all the doctors I'm always starting with the doctors because I have like a lot of doctor relationships and you know backgrounds with a lot of doctors so uh I like to bash doctors a little bit even though I have great appreciation but and they are considered one of the highest level white collar jobs but lawyers same thing and then everyone below that which is basically everyone else gone Why?

[00:02:58]Because AGI has that capability and we are crossing the Rubicon as we speak.

[00:03:03]Okay, this is happening. This is very big. Now the job replacement actually will happen not when we cross that Rubicon which is now in the next 3 to 6 months. Not the 99% Rubicon but the 50 60 70 80% Rubicon 50 60 70 80% of jobs affected. It will happen once this Rubicon of capability of AGI is then also extended into deployment because there's a difference between having the capability and deploying that capability which is an infrastructure problem. It is a change management problem. It is the problem of being in all Fortune 500, all Fortune 1000 companies, all mid-level companies uh and actually going through the departments and actually replacing everyone by training the AI to do that. But in case you thought, well, Joe, maybe you're right.

[00:03:51]Maybe economic AGI is here end of this year, but then it takes 10 years to implement. No, because when you think about the massive investments being done, why are these trillions of dollars being invested in AI for one simple economic reason? Because AI is defined and is destined and it is intended to replace all human labor. That's the only reason why people invest now infinite amounts of money. And that means these are not dummies. In fact, these are the smartest people on the planet who have the most money and the best technology understanding and they are betting on it now. And they're not just betting, they are doing stuff. We'll get to this in a second. So the deployment, the mass deployment of that capability that will be completed in my opinion latest by December of this year, absolutely latest and in many many areas is already being completed by July is then being deployed at scale and that deployment is being prepared across all companies with more than 100 employees in the United States right now. Okay. If we look at some facts here, if we look at some important facts and we get to cursor now look at Anthropic, look at the massive massive scale. They are now at $45 billion revenue run rate and they are doubling. Let that sink in.

[00:05:08]Currently, they're doubling the annualized revenue run rate every six weeks. As an example, if they continue doing that, anthropic alone will be larger than the Swedish GDP by end of this year. I said 2027, but by end of this year, just picked Swedenberg because why not? So, they will overtake Sweden not in enterprise value. That's always a dumb comparison. You can't compress uh you you can't compare market cap with the GDP. But in terms of actually comparable things, GDP is revenue of a country kind of this is revenue of the company. They will be bigger than Sweden. One startup and Sweden is a pretty strong country with a pretty good GDP. That is completely nuts and that shows you this graph how exponential these things are now developing. And now we are getting to the point of this video. So we have to understand this is exponential growth.

[00:06:00]There's unlimited TAM because first you replace all human labor which is a 30 to40 trillion opportunity uh including blue collar but white collar is actually a bigger chunk of that opportunity and that's happening starting this year at absolute scale and what does anthropic and open AAI the leaders in that space clear leaders much better than Grock by the way they're both completely dwarfing Grock in capabilities and I know we have a lot of Tesla fans among you Guys, a lot of Elon fans. I like Elon. I think he's very smart. As you will see in this video, he's actually incredibly smart. But we also have to be realistic on this channel. This is Pioneer Lance. We are not a Tesla fanboy channel. We are not an Elon fanboy channel. We are a fanboy of the pioneers, which is all of us. It is Elon. It includes many people, everyone who wants to engineer the future and take steps to do that and continuously improves themselves on the iterative engineering uh cycle. We are all reality engineers here. Constantly learning, constantly understanding the future, constantly predicting the future, and constantly calibrating our mental models to be constantly more right than we were before. That is what makes a pioneer plus a little determination to get things done. But so we are not just fans here. We criticizing Elon too. And Elon clearly built a great LLM with Grock, but as an AI AGI, it totally sucks because they basically have no product team, right?

[00:07:23]There is no product team. They're bad at product. And they were bad in building.

[00:07:28]Now I get to a very important point there. They were bad in building what I call the scaffolding other people called it the harness around the LLM. Okay, when you look at anthropic secret, how did anthropic outperform OpenAI so much? And of course Google, but Google is not even a real player in the space. Google is constantly falling behind. They have problems with LLMs and they have problems with all kinds of stuff, productization and being just smart because they're too large. So, it's Entropic versus OpenAI and now versus SpaceX AI. What was Anthropic's true secret? And now listen very closely because this is very very very important. What was Anthropic secret last year when they crashed OpenAI suddenly out of nowhere? Don't forget a year ago cursor was the hottest kit on the block. Maybe 18 months ago, maybe a year, maybe 15 months ago, Cursor was the hottest thing. The second hottest was OpenAI. OpenAI was the best model.

[00:08:23]and cursor had the highest adoption rates, right? It reached two billion in sales faster than any business software before it. So, cursor was hot. And then suddenly both OpenAI and cursor got killed. They became incredibly unhot because Anthropic took the crown. How did Anthropic do that? Well, no one knew at that point. It just worked better.

[00:08:44]You use claude, you cla very recently, and the minds were blown.

[00:08:50]what is h why is this thing so advanced so smart and then they had some leaks and I had to think about this a lot like what is actually happening why are they so smart and the answer was found not in the LLM not in the level of intelligence but in the scaffolding it was found in the surrounding prompt system of anthropic where enthropic basically was able to understand that the LLMs only are intelligence and int Intelligence is overrated. Surprise. Now comes my big insight here. Intelligence is slightly overrated. If you have a model with 150 IQ and another model with 130 IQ, or even more important, if you have a model with 165 IQ and another model with 145 IQ, it says nothing, or at least not much, nearly nothing about how good that AI is for you to deploy. You can have a 145 IQ model that easily beats 165 IQ model on all kinds of metrics. You can even have a 120 IQ model, Kimmy. I will get to this in a second because that is what Kursa is using and beating everyone right now with Kimmy that has probably 125 IQ. I'm just approximating IQ's here. Some lower IQ than CHPT or Claude.

[00:10:09]So, how is this all possible? What is the secret? The secret is the scaffolding because here is where the analogy with a human brain breaks. If you have a human with 165 IQ and a human with 145 IQ, or let's make it more realistic, a human with 145 IQ and a human with 120 IQ, the 145 IQ human, assuming a similar, you know, emotional functional structure will kind of beat that 120 IQ guy in most cases across the board because he's smarter. With LLMs, it is very different because an LLM is not a human. And here's the key. An LLM is not one singular entity with one singular IQ. An LLM or let's say an AI structure that you use, your chatbot, your agent, all that stuff that you're using is basically a giant organization of agents, each one with that IQ because you can run many instances and that's exactly what everyone is doing now for like a year. OpenAI, Anthropic, Claude and of course Grock. They are running multiple of these many many many of these things in parallel. Okay. And that introduces another dimension. Hey, on this channel we do something very important. We deeply analyze, understand and predict capital, AGI and politics.

[00:11:33]Not just to beat the markets and make a lot of money, even though that's important, but also to build a new society and make the future good for all of us, the pioneers. If you want to help me on this journey, you can do that by doing a very simple thing. Like and subscribe. It helps me a lot. Thank you so much for your support. And let's get back to the video. Here I explain it the first time very clearly what the difference is. If you have an IQ only person, that's one thing. But here you're dealing with a giant organization that is self-organizing all the time and the code the culture of organization is the scaffolding the harness the prompts that we cannot see when you type something in the chatbot what is open AI what is enthropic what is gro what is it actually telling itself after you type this thing in and some of that stuff leaked from enthropic but there's a giant complicated question what actually is this thing now doing why is claw design infinitely more capable than claude code in designing your stuff. Why? It's the same model. Claude code is infinitely inferior to claude design in designing stuff even though it's the same IQ and the same model because the scaffolding is totally different. And why is entropic so superior to OpenAI and to Grock? Because the scaffolding is more uh advanced. You can if you want an analogy or a metaphor more analogy. It is like instead of hiring a human, you are hiring an organization. And if you have a perfectly organized, highly functional, very good culture organization and everyone has IQ of 120, that organization will beat the other organization where everyone has an IQ of 145 easily. If that 145 IQ organization with, let's say, a,000 people is badly organized. So IQ becomes a little bit less important. Of course, if you have an 80 IQ organization, they can never compete with 120 IQ organization. But 120 versus 140 doesn't matter. The culture takes over. Much more important.

[00:13:31]So in a way that's the key. Now to build that what I call culture here, the scaffolding, the prompting, the secret sauce, what you instruct that thing to actually do, how you organize all these agents. That is where Enthropic wins.

[00:13:45]That's why they're winning. And that's why the LLM where I'm pretty sure Enthropic doesn't have the best LLM.

[00:13:49]Probably OpenI has a better one.

[00:13:51]Probably Grock has a better one, the raw LLM. But Enthropic shows how important that is. Okay, now let's get to cursor.

[00:13:57]What is the ingenious thing here that Elon just did? The ingenious thing is that cursor of all these players had by far the worst position on IQ. Cursor had the lowest IQ. Sorry, Cursor founding team, the founding team has a very high IQ, but cursor itself had no IQ. They don't have an LLM. They were just piping this whole thing, frameworking, scaffolding. But they showed recently that even using Kimmy, which is a low IQ model, they can nearly beat or approximate the other guys. Like look at this thing. Look at just this thing. Um, so here Composer 2.5, which by the way now is using rock. Now they're mixing it up because they just had this deal with Elon. I will get to this in a second, but they are basically beating OPUS and CHBT, which is crazy. Which is crazy. They're even approximating now the max and extra high chbpt and ous 4.7 which is absolutely nuts. So what is happening here? What is happening is that cursor is the one company in the world when you understand the secret sauce you need an LLM with higher IQ and then you need the framework around it and these two make together the power. cursor had by definition the best the world's best framework because if they can use a lower IQ model to actually get to better or the same results as the higher IQ models it by definition means their framework is much more advanced. Elon saw that and Elon also understood that with Grock they might have a very good LLM but a sucky little harness and you can't build that harness if you don't know how and Elon probably knows he's not great in hiring these teams from scratch. So what Elon saw is hm if cursor is the best but the worst on the LLM side and the worst on compute anthropic is like in the middle open is better than LLM and anthropic probably but worse than anthropic and scaffolding and then Grock is pretty decent or good at LLMs but totally sucks at frameworking if Grock and XAI Space XAI would partner up with the best framework they could leaprog everyone and that is how this crazy looking deal came together that Elon is offering or did pay $10 billion with SpaceX to cursor as an option. It's just a call option and a collaboration strategy and has a call option that he needs to use within 60 days after IPO which is probably by August. By August he has to pull the trigger and buy cursor for 60 billion which is peanuts if you have a $2 trillion stock with SpaceX which you probably will have by then.

[00:16:32]So, if they combine, and here's the point of this video that is so important. If they do this successfully, and of course, everything we see, these preliminary results you're seeing here, they show this equation is absolutely working.

[00:16:49]If this all works out the way I think, it could be that SpaceXi will leapfrog anthropic and land next to it, maybe a little behind or even a little before.

[00:17:02]Anthropic in no time and the deal they just struck with Enthropic SpaceX compute with Anthropic you have probably heard about that where Anthropic is paying 1.5 billion per month to SpaceX for compute for Colossus. That deal can be cancelled I think within 60 days or something. So it has a notice period that's pretty short and they could kick out Enthropic if they figure out that they can actually beat Enthropic. And now we get to the final piece of the equation. Okay, the final piece of the equation here is it's about deployment.

[00:17:35]How do you get then once you have the superior white collar digital human digital human? How do you get that thing into the companies? And here's another huge advantage that uh cursor has. Cursa has actually the widest far-reaching implementation status across Fortune 500, right? They are the company because they were the first uh one to do that that has the biggest adoption across Fortune 500 companies. So that is absolutely crazy.

[00:18:09]Curser is in all these companies already. They have tremendous adoption across America's largest and most important companies. I think they're in all of them at this point. I read somewhere.

[00:18:23]And that is absolutely crazy because that means this is not just a sales force for the digital human that XAI or space XAI will then have the white collar mega replacement tool, but they now have a deployment infrastructure where they not only can sell to all Fortune 500, but they have an existing standing infrastructure and are inside of all of these already with account managers and huge cash flow into Curs.

[00:18:53]That makes this whole thing absolutely crazy. And you know, of course, when you look at the numbers because people discuss about SpaceX 2 trillion dollar valuation wherever we get to 1.5 trillion then going up to 2 trillion probably quickly before then dropping.

[00:19:07]That's my prediction. But some somewhere around two trillion at some point.

[00:19:12]Uh how does the cash flow evolve?

[00:19:14]Everyone is looking at Starink. But now look again look again at the anthropic annualized revenue. This is much bigger than Starink and it's growing dramatically faster than Starink. And I'm always bashing Stalink, don't get me wrong. Stalink will make a trillion dollars in revenue relatively soon in my opinion cuz they will get to a billion users in the next 10 years. So it is a giant cash cow, but nothing compared to Space AI and probably nothing compared to Space Xai, the actual digital human deployment system. Because if Anthropic grows to hundreds of billions of dollars of revenue in the next 12 months, pretty assured that they will and cursor can overtake them, we will could look at a situation where SpaceX AI is on track of making hundreds of billions in revenue just with their new anthropic competitor starting next year or in 2028. So tens of billions next year and hundreds of billions in 2028. That is a completely unrecognized revenue stream with enormous margins. And I do think when you put together these pieces, this is more likely than not to happen. What exactly? That SpaceXi, the combination of Grock as a very strong LLM in conjunction with cursor, the world's best scaffolding system, which in my opinion is even more important than the LLM and will be more important in the future. the way how you organize, spin off, manage and improve the agents and in conjunction with cursor's massive market access that might beat anthropic as early as next year. Now I have to think this through more and model it out more. And this is not something easy to model because there are lots of unknowns here. But the strategic first principle understanding that I just developed over the last 48 hours is that these are the two components and that these components are represented by Grock and by cursor.

[00:21:10]Enthropic is only so strong because it has the cursor component inhouse and the Grock component in-house. But anthropic might be inferior on the cursor side to cursor and on the Grock side to Grock and only wins right now because no one combined these two things successfully.

[00:21:26]So if this is all true, Elon might be sitting on the holy trinity perfect scaffolding in the world, best harness in the world, cursor, which is proven to be the best, probably one of the best models, maybe the best model Grock as a pure LLM. And third, of course, last but not least, this thing, the crazy space AI situation.

[00:21:52]Uh, this is a little like maybe forwardlooking that we have a moon base, but you don't need a moon base to have massive amounts of orbital AI satellites. Elon promised to present the AI satellite very soon. I always said since the onset of this discussion in November 2025, November last year, I had the first video out first on Pioneer Lands in the world where I outlined the whole orbital AI business and why it's so hot and why it's much faster coming than people think and why it will replace terrestrial AI build out very very quickly before 2030 at scale. uh this is all happening and that's the holy trinity of course scaffolding LLM superiority and then the compute buildout and the scaffolding is probably the most important piece and the fact that you can buy the world's best scaffolding company cursor for 60 billion is an absolute bargain of course no one else would do that bargain only Elon can do it and it just proves again Elon is very very smart he doesn't he understands what he's bad at he was bad at product he was bad at scaffolding, but he understands he needs it. What does he do? He goes up to the guys who are the best in the world, who right now don't look very good. Now they look good again, but before the deal, they look just, oh, everyone's forgetting about them. Elon saw it, took them on, boom, has a brilliant team.

[00:23:09]So very cool. And then here, of course, uh that is the translation of this whole thing into revenue for SpaceX. So we will look at this more. SpaceX Starlink revenue is interesting. Then of course the the bare compute lease out that by the way within a week surprise week doubled SpaceX's outlook because it's now as big as Starling just the anthropic deal but now we're looking at the really big deal the new digital human business. They call it digital optimist I call it digital human. This is a direct shot at the anthropic core business which is basically infinitely large. So, I hope that was interesting.

[00:23:52]We have all these things being discussed on Pioneer Lands. Go there and check it out. And I hope to see you very soon.