저자: Kevin · 외국인
요약 보기
Meet Kevin 유료 멤버십(알파 리포트) 홍보 영상. QQQ 저점 반등 및 반도체(AMD·마이크론·브로드컴·엔비디아) 강세를 자찬하며 지난주 콜 성과를 나열. OpenAI 2030년 대규모 적자·지출 우려와 Nscale IPO 자금조달 불안을 언급하나 버블 붕괴는 '아직 아님'으로 봄. 모델 압축으로 구형 GPU 가치 상승 주장, Figure 휴머노이드 로봇·AI 디플레이션 이후 부동산/자산 선호 발언. 마지막에 멤버십 가격 인상 전 가입 유도.
다듬은 글
You won't want to go to sleep until you watch this video in its entirety. Actually, we have a lot to discuss in this video. We're going to talk about what just happened with OpenAI. Is this a “bullish” or “ bearish” signal, especially in the context of what happened earlier with Anthropic? Then we'll talk about model compression, which is extremely important, especially with regard to high-speed memory and bandwidth.
This will be very important for who will be the winner and who will be the loser in the field of artificial intelligence. Then we'll talk about Figure Robotics and what they just showed us this week, my initial reaction to it, and some hints about where I'm looking for opportunities in the supply chain when it comes to robotics.
Then we'll talk about Buffett. Then we'll talk about real estate. This is the plan for this video, so I have a lot to say. It's Friday night, and you might be asking yourself, "Why?" Why do we do this in the evening? Why should you stay awake for this? Well, in exchange for all the information in this video, I'll take this opportunity to say that we are literally raising the price of the Meet Kevin Alpha membership by $900 once this coupon expires.
So, you go to meetkevin.com or meetkevininvest.com, it doesn't matter. Join now and you will be generously rewarded. We have some really cool things for course participants. You get all nine rates and every trade notification. You already know the gist of the proposal. You can read about it on the website.
But I will do something better for you. I'll actually show you what we've been talking about in the Alpha Report every day this week. And after that we will move on to the most interesting part. First of all, remember how we started the week and said, "Guys, you should buy QQQ." QQQ, which is an index, closed at 721.
We literally sent an alert when the price was one step away from $700, right on the rebound. That's the lowest point you see right there. We simply tore the market apart into "Qs". But I will show you the wins and the losses this week. That was on Monday, okay? On Monday, we started the week by saying that we expected software growth today, but not chips.
And since the opening, some of these stocks have risen by 6%, and some by 14%, from the levels they were at at the opening. So, in other words, as we sent in the alpha report, the market opened up and these things took off. It was a fantastic forecast on Monday. We predicted soft support at 702.87, the low of the day was 702.74.
We said that Monday and Tuesday are the days of peak fear . The Alpha Report hit the mark on Monday. We did n't make any mistakes on Monday. There are days when we make mistakes. You will see it. We made a mistake on Tuesday. I made a prediction in the alpha report, and by the way, here's how it works , I'll show you the wins and losses this week.
That's what you get, okay? Every day at about 5:30 am I send an alpha report before the markets open: levels, tests, short-term, medium-term, long-term plans, deals, ideas, investments. Buy-Sell Alerts, this week we sent two buy signals of $100,000 each. We sent out , uh, we did a live stream for the course participants where we added details about what we thought about the day and added more context to the alpha report.
Short-term, long-term and medium-term analysis, you know, there you will see advice like "don't buy OpenDoor." You will be left with depreciated assets , and the price has dropped from the $6–10 range to $2.50 . All courses plus an app, okay? Here's what you get. Although you already know that. Tuesday, this is what happened on Tuesday.
We predicted no takeoff and no crash. We were uncertain about that day because it was the day before the Fed meeting. And what happened that day? Qs actually dropped by 0.7. We stated that there is a 95 % probability of a rate increase, and don't be afraid of that. Be optimistic. Note that the market rose 3% 2 days after the increase occurred.
Quite important. Last days to buy on the decline, right? And we said it would be today or tomorrow. We ended up buying on Wednesday. Wednesday was the minimum. The prediction was: hardware is slowly moving up. I thought the cyber sector would lose some of its enthusiasm on Tuesday. In fact, it even grew a little on Tuesday.
So, we were right about that . I was wrong about the cyber sector. I thought that after a 14% increase in one day, we would definitely see a pullback. I was wrong. Fed Day. So, the Fed day was interesting because we set a specific condition: we could be 80% optimistic if we understood how far we were from neutral.
We received a lot of conflicting comments from the Fed. So, Mr. Powell, we knew the rate would go up. And we clearly named $700 as a reliable support level. And the price of QQQ literally bounced off the $700 mark , and we bought as soon as we saw it . Signal to buy 100 thousand. Signal to buy 100 thousand.
Here is the time we sent them. Words are backed up by deeds, baby. OK. Good. Almost everything. Thursday. And I'm serious, we're significantly raising the price because we 're changing our approach a little bit . And course participants who make it to this deadline will benefit significantly. Thursday.
I said that the peak of fear was behind me. The markets relaxed and we went up, as I said. Nasdaq rose 1.7% —its best day of the week. I called it equipment day. Hold 715, and that equipment will grow. We had 6% growth in AMD, Micron grew by 5.5%, Marvell by 5%, Nvidia by two and a half. We said to watch the 246 mark for a bounce on Marvell.
We actually reached 247, bounced back and went down, but we still got there. Avago still lost 355, although we were quite optimistic about Broadcom. It took a few more days for it to really work. Broadcom closed at 357 on Friday. So, we're back to that line. We got a nice boost on Broadcom that we were expecting.
However, at this point we haven't reached 227 on Nvidia yet. I thought we were testing this level. We eventually got up to $222. We didn't get to 227. I was hoping for this breakout test by Friday. This was wrong. We haven't achieved that. We're only at 222. Meta, my goal is to reach 700 by next week.
Okay, we closed on 666, the number of the devil. So, let's see. I am optimistic about this. Friday, today was a good alpha report. The test was to hold the 715 all day. We closed at 721. I actually said this morning that it's unlikely we'll get to 227. So I was trying to clarify and say, " Okay, you know, it looks like it's starting to weaken or something." "We still ended up at 222.
But we were right to lower the forecast." " Wait and see" is what we called for today. And the Nasdaq rallied, but mostly at the end of the day. If you were trying to trade today, it would be difficult because we only got a buy signal at the end of the day. So that's consistent with our optimism in the medium term.
But in the short term, we thought, "Oh, today is going to be a little tough." Well, yeah, the triple witch , you know, people are scared of the weekend, those are the things we're talking about. The fear of retail investors makes me optimistic. Okay, a lot of people are scared right now. We'll talk about some of that fear in a moment.
Marvell, which we talked about. Again , the rejected call. And we weren't wrong today, as far as my summary goes. And I looked to make sure, I checked all the Alpha reports. I don't like it. Am I missing something? Are we fair here? So , we're fair. Okay. That's it. Alpha Report is going to go up $900 if you don't use the War Room's Bulls coupon.
That's it . Now, let's talk about all the nonsense that really needs to be discussed. OpenAI. First, OpenAI. OpenAI just announced that they expect to spend about $280 billion by 2030. That's crazy. That's a lot of spending. And it's not because they don't expect revenue to grow. They think revenue will grow to $350 billion by 2030.
It's a little complicated the way they present it. So I'll draw it on the board. But at the core of it all, they expect to spend a lot more money than they make. That shouldn't come as a surprise. And I think that's part of the reason why Anthropic's IPO keeps getting delayed, because now is not the time for euphoria.
You may have seen my tweet. You may not be following me there. You can always follow me. in Real Me Kevin. There you'll find the good old "buy on the dip" face ringing the New York Stock Exchange bell. And here's what I'd like to point out. It's weird, right? But it's common practice these days. We talked about it in the video about the Anthropic delay.
I won't go into all the details, but when a company like Nscale goes public, hoping for a miracle, it seems strange that they literally say, we have no plans to finance the $44 billion, or we have no commitment as to how we're going to finance the $44.6 billion deal with Anthropic, but Anthropic is essentially saying, " We assume you'll get it." We're going to market as if we already have that computing power.
And Nscale is advertising the Anthropic deal as if it's already done, but they don't have the money yet, right? So the funding has to come from somewhere, and when you put it depending on the IPO, I'm starting to get nervous that Nscale is going to go public, the stock is going to crash, the funding is not coming in, and the bubble is going to burst.
I want us to understand one thing. I don't think the bubble is going to burst, but I also don't think these companies are going to be profitable anytime soon. It's unfortunate because the more they make and the closer they get to profitability, the longer this bubble lasts. We have to be very clear about that, and it's going to be obvious throughout the video.
This bubble is going to burst at some point. And if it happens under Kevin Warsh at the Fed, it's going to be really bad. I'm not trying to be scary. I'm trying to say, just keep in mind that we can work together, we can make money, but somewhere in the subconscious, you know, honey, maybe this month we're going to start paying off the car loan.
Maybe this month we're going to make an extra payment on the mortgage instead of buying in the downturn. It doesn't matter, right? But now you're reducing your debt risk a little bit. And that gives you the freedom to choose. You know, people always say, "Oh, but you're only saving 2% ." I could make more money in the market.
That's not guaranteed. And these numbers from OpenAI suggest that we're in for a real challenge. This Nscale IPO is probably going to happen before Anthropic even thinks about going public. And what is OpenAI telling us? OpenAI tells us that they expect to raise $350 billion . Uh, and that's... Oh, sorry.
The numbers are expected to be negative. What are they saying here? Expenses are going to far exceed revenues, which are projected to grow tenfold. From $36 billion this year to $350 billion in 2030. Oh, right. Okay. And then they expect to have $840 billion in revenue by the end of 2030. But , let's say the period from now until the end of the decade.
That's $280 billion. So by 2030 they expect to be in the red by $278 billion, but they're looking at a profit, right? $840 billion, although it's worth noting that the numbers are missing. If they expect to be in the red by $278 billion for OpenAI by 2030 and they're looking at a profit of $840 billion, that's crazy—to make that kind of far-fetched prediction.
It's just crazy. It's going to lead to a lot of lawsuits. Maybe Karp was right. That means you're actually looking at $840 + 278 . You're actually looking at $1.118 trillion in spending . That's how much you expect to spend. That's how these calculations work. The Financial Times didn't tell you that.
But that's how you can flesh out that idea, essentially. Now, where are they going? this money? Well, the vast majority of that money, probably about $400 billion, is probably going to Nvidia. Nvidia is the undisputed leader and the driver of what's happening right now. In the entire AI ecosystem. And I'm going to show you something really amazing about the Nvidia situation.
It's called model compression. This is going to sound a little weird, and I'll probably go into more detail in a video over the weekend because it's going to be technical , and I can't do it very well tonight . So we're just going to touch on it briefly today. There it is. I have it on my iPad. I opened it up for us.
I like to cook this up for you. So here's the gist of it. Oh, look at this. As soon as I said that, everything went black. Okay. So what is this? In the top right corner, you see a table that shows the GPU platform distribution on the left. The GPU platform is important. The GPU platform is the history of what we've seen here.
From the A100 to consumer chip RTX Blackwell 5090. You have a professional chip like the RTX 6000. These are great chips. H100, which Elon has a lot of on his Colossus facilities. We've simplified things a little bit here to get an effective output speed to compare these chips. And we're going to compare them at around 200 tokens per second.
The reason we compare them at 200 tokens per second is because you're going to see a high percentage of what's called "idle memory" at that speed. You see these chips have 72 to 91 percent of their memory being idle . As you increase the number of tokens per second, your ability to utilize the memory decreases.
The reason is that the more tokens per second you can process from a throughput perspective , the less memory you actually need. So you're left with a bunch of memory sitting around . Some chips do it better than others. And that's where, to simplify, we get to what we usually call the practical maximum tokens per second for a chip.
So a practical number for a chip model would be something like 36 tokens per second for an RTX 5090 or a Blackwell 6000. For Vera Rubin, where Vera is actually a reference to an astronomer. Vera Rubin was an astronomer. Vera is the CPU , Rubin is the GPU. So let's focus on the Rubin. The Rubin processes about 10 times as many tokens per second as the RTX 6000.
That means that despite having a lot more RAM, your ability to process tokens is much greater than the difference between a gigabyte of memory here and there. Think about it. The Rubin has 288 GB. To make things simple, 288 divided by 96 is exactly 3 times as much. So there you have it, I'll press here, has three times as many tokens as the RTX 6000.
And that's in terms of the installed video memory. That's almost three times, or 3.3 times more than the H100, to be exact. But its performance in practice, because of its much higher throughput , is actually 10 times better than the A100. In terms of tokens per second, it's about 8 times better than the H100.
So what this means is that these chips are becoming less dependent on RAM. I'm not saying that high-speed memory is a bad thing. I'm just saying that as these GPUs get more bandwidth and more efficient, they technically need less headroom. They need less RAM because they can process so many tasks so quickly.
For quick calculations, because it's useful for understanding, I think this makes things a little clearer. I want you to think for a moment about what gives you 1 million source tokens. Got it? 1 a million output tokens, just for fun, on a Fable or Astra model will cost you about $50. That's pretty expensive.
If you use, say, a Qwen model or some other Chinese open model, you're paying about 40 cents for the same 1 million output tokens. 1 million output tokens is about 750,000 words. So that's a fair comparison when you think about it: $50 for 750,000 words. You pay 40 cents for 750,000 words when models are compressed.
Now, why do I bring up model compression in the context of price compression? Because smaller models that aren't cutting-edge actually allow the price of older equipment to go up . That means that data centers don't necessarily have to be at a disadvantage, because traditionally--and this is just mind-blowing, seriously--it's so much information.
I know it's mind-blowing , but that's the way it is. Traditionally, we think: a new chip comes out, so the old one has to be thrown out. Why would you use the old one? But what's really happening now is that the new chips are just helping with the advanced models. Then the distilled Chinese models are just sitting on the old chips, even though they didn't exist before .
So all of a sudden the advanced technology is going from the old chip to the new one . But then a new product comes out , and the distilled advanced models are starting to work on the old chips. And so, in the most perverse sense, from an industrial revolution perspective, we're literally seeing the value of both the old and the new chips go up.
It's actually creating -- and it's a revolution. As soon as I started to understand this, I thought, because we have our own chips, right? We have our own RTX 6000s, 5090s. I mean, we have, I don't know, 50 machines or more that are processing data for our Reinvest members, you know, our alpha members, and the software that we're building, and everything that we We do.
We have a lot of equipment, and it's expensive. Of course, we don't have the equipment that GPT has, we're a small startup. And I realized that these chips have doubled in price , and I look at it and I'm like, why are they more expensive? Are we getting twice the value out of them? And then I talk to our developers, and we're like, we're not getting twice the value out of them.
We're getting more than twice the value out of them. And I'm like, how is that possible? So I take what we see in practice and I apply it to the teaching and the research that I do. It's actually... these are all little sheets that I've added to the SemiAnalysis article. I usually have questions about some of the things that I see on SemiAnalysis, so I look at them and try to figure out, " What do I agree with?" What do I disagree with?
"And the problem is , there's a lot of information out there. It's a huge amount of data to process. So I'm trying to summarize for people. And I've noticed that yes, there are indeed cases where these data centers, especially the ones that are least at risk of going bankrupt, right? So the ones that don't die and don't go into debt, in other words, the ones that have other businesses to support, are more likely to survive, and not just survive, but probably recoup their capital expenditures.
Even if Elon is wrong, and I think he is wrong. I don't think he's going to get $50 billion per gigawatt. I think he could get $18 billion per gigawatt with long-term contracts. But even if he gets that, what? The chips pay for themselves in about two to two and a half years. It's a big deal. Although the big players that can support this with other revenue or trust and financing—Elon through the stock market and his friends at Morgan Stanley, and the income from Starlink and Starship, and of course optionality, when you can add Tesla to it.
Optionality for survival. Who else? Well, Meta is just another example, right? I want to be clear that I have investments in some of the companies that we're talking about. I'm not trying to promote or promote anything. It's Friday night. Got it? I don't think my words are going to change the market in any way.
I'm just trying to share information. I think Meta is an interesting case because their advertising business pays for all their chips . So even if you don't get ROI, the chips are already paid for. You get ROI as a bonus. So Meta is also a very interesting option. I'm starting to get nervous. We've talked about this before: watch the Anthropic video where I break down the metrics of N Scale, Coreweave, and Nebius.
Nebius is the best of them. Watch that video, about the Anthropic IPO delay and the Nebius review. I think it's a very useful video. It'll give you context so I don't have to rehash it all here . But it's important. Model compression is what keeps this bubble going. So yes, these big players like OpenAI are spending money, and they're losing it.
But that doesn't necessarily mean that the chip owners are going to go bankrupt. The chip owners are probably going to get their money back, at least the ones who own the chips now. They're probably going to make their money back before these companies go bankrupt. I'm not sure about the viability of the OpenAI and Anthropic business models , especially in the context of open- source models, and we'll see what people do with software.
The next year is going to be transformative. Like, what else is transformative? Honestly, four hours of this robot doing something is pretty cool. So what Figure did, and I invested in a robotics startup called Apptronik through our venture group, and we seem to have made a pretty good profit on that.
We did two venture capital. One is in SpaceX at a valuation of about $350 billion . The other is in Anthropic at a valuation of about 1.8. In fact, before I had a venture fund, I also invested in Anthropic. No , not Anthropic. I also invested in Apptronik. Apptronik. Uh, that's different. Oh, I must have mixed that up a few times.
So let me clarify. Apptronik. It's a developer of robots. Humanoid robots. But anyway , it's Figure, Figure AI, a humanoid robot company. Apparently they rent 30 houses in the San Francisco Bay Area and put these robots in unusual situations where the robots walk around and do things . I think some of it looks a little weird.
For example, I'm not sure who puts the nightstand on the left in the room and turns the drawers so that they're against the extension cord, which is there . But never mind . I'm not going to go into too much detail about it . What I find fascinating is that they're apparently training 90,000 people who are wearing special glasses or something, collecting data from people.
And they rightly point out that the thing about humanoid robotics is that humanoid robots act like people. So they're working on human data. How useful is that for training? Very, damn it , useful . So very useful. Okay, well, good for them. So how do you invest in the robotics supply chain? What I did for the course participants, I'll show you some of it, not all of it.
For the course participants, I added a lot of information in the research tab about stocks that, if they all had a PEG ratio of 2 and a P/E of 2, would have the greatest upside potential. You can see it in the "suits" section of the app, and then in the research tab. And we have a table that includes, let's see, 1 2 3 4 5 6 7 8 9 10 11.
Yeah, probably about 20 names. 20 names that have potential in robotics, all calculated with a PEG ratio of 2. And there are forecasts of declines as low as minus 63% for some of these companies, and some have potential for growth as high as 265% . And then it all varies within that range. I'll give you a few examples now.
One of them is definitely going to take over the robotics market. I have no doubt about it, because they're already everywhere in this industry. And people don't believe it. People don't even think about it, or I don't think they notice it. It sounds corny, but it's Nvidia. Because Nvidia is literally the brain and the system that runs most of these robots.
And it's weird to even think about it. Of course, there are other companies that we have on this list. Teradyne, Applied Materials, Lam Research, Analog Devices, great. Lattice Semiconductor is also here . These are examples of companies that have had varying degrees of impact on the robotics industry, and I'm really excited about them.
I think they have a great future. That's cool. Some of them have a lot of growth potential. But let's understand for a moment what Nvidia is cooking up . Nvidia wins because these robots need to be trained. Where do they train? On Rubin chips. We just saw Enscale, the same company that I did a separate video about , sign a $3.5 billion deal to deploy 100,000 Vera Rubin GPUs.
I think they're going to need some more money, right? This is an incredible chip, and Nvidia is the beneficiary of that because Enscale wants to license these chips to companies like Figure so they can train or lease them to companies like Figure so they can train their robots. So Nvidia wins on training.
But then you have to go to the device. But Nvidia has that too, because now people are training on CUDA and run robots on ROS 2 software. They train them in synthetic worlds on Omniverse. And the chips that actually run these robots are Thor and Orin, which are the computers and the brains of the robots.
So you have a chip for edge computing. They literally have a CPU stack and a GPU stack for training. They have ConnectX and BlueField DPUs for communication between the robots and external systems . They have the software for the robots. They have Omniverse for synthetic training of the robots. They have CUDA on which you train and run edge computing on Orin and Thor.
Incredible. Tesla is a devilish competitor with vertical integration. Although, to be fair, they closed Dojo. So they expect to get AI 5 in 2027 from Taiwan Semiconductors. Samsung is reportedly the partner for AI 6. But they want to build their own simulators, their own training tools, and their own capacity.
That's why they have a research factory in Austin, Texas, and their own circuit boards. They want to do everything themselves. Now they depend on Nvidia for training. That's it. It all comes back to Nvidia. So it's crazy to me that Nvidia is so cheap. But it really is lifting everyone up. AMD, Broadcom, Marvell, ARM, Qualcomm, MediaTek—they're all benefiting from robotics workloads.
All to varying degrees and at varying valuations. Teradyne, for example, owns Universal Robots. It's incredible. So I'm really excited about robotics and this whole technology stack. Before I saw this Figure video, I thought robots were a 2035 prospect . In other words, it's still 9 years away, maybe even 9 to 15 years away.
But I have to be honest. Even though these robots that we're seeing are still bad at folding clothes, it's a significant step up from what we've seen before. Nobody's going to pay 30 grand to put away five toys for 5 minutes, right? But that's a huge, huge improvement . Assuming that this isn't a remote control, and we're actually autonomously giving these little guys the task of putting things away.
But when he tries to fold clothes, look at this. He takes forever to pick up one thing, like, pick it up already. Pick it up, man. It's not ... It's not that hard. Come on, man. But he knows to go back and try again. And so, frankly , even though he's 10 times slower than a human, he's working 24/7.
That's pretty impressive. Okay. And he's not going to make the bed perfectly. Is he going to pull the sheets like in a hotel? No. He's going to take it from looking like this to looking like it still looks terrible. But this is 2026. And that's what happens when people... get a little bit better. When people see that, they put more money into robotics.
And that's how you end up funding those the Nvidia chips themselves that make all of this possible. That's not bad. That's really impressive. So I'm excited about that. Figure is making a lot more progress than I thought possible . So that gives me some optimism. Do we have potential problems in the economy?
Of course. Remember, we made a video earlier, and that, that's true. Of course, there's always going to be those losers who just read the headline and say, "Wait a minute, Kevin." I thought you were an optimist. Why do you say there are two paths to recession? "Because they are. You can be optimistic.
I'm optimistic right now. I think it's a good time to buy. But at the same time, you have to understand that there are risks. The big risk is that AI spending will collapse, that people will sell bear skins before they've caught the damn bear. And if they go into the woods and there are no bears, no funding, they won't be able to sell the skins.
They won't be able to deliver their product. They won't be able to train their robots. Nvidia won't get the money. The computing power will run out. OpenAI and Anthropic won't get the resources they need. And you'll have problems. So the cyclical funding has to continue. There's no sign of slowing down yet, but if the Nscale or Anthropic IPOs fail, that would be the first red flag.
There's a second risk. I mean, that's not bad, right? That's not bad. There's a second risk. Towels aren't perfect, but they're kind of good enough. Enough is enough. good. Honestly, it's pretty good. It's... it's pretty good. I mean, they would n't hire them to work at Hollister because they have to make these perfect folds.
You know, it's kind of crazy. But it's a dish towel. Who cares? You wouldn't fold jeans or shirts like that, because they'd wrinkle. I love that they spent four hours on it . I mean, that's so brave. Well done . Honestly , brilliant marketing. I just left it running in the background. It's really funny.
Oh, by the way. So the second fear -- you just watched the recession video -- but the second fear of recession, and you'll see the data that you should be looking at, is the labor market. If the labor market falters, that's a real risk. So the bottom line is that I'm optimistic, but I'm watching two red flags .
And I think I've made that very clear. I'm very optimistic about what's happening . Model compression is a win for hardware stocks. It's a smaller win for memory stocks. I'm not trying to criticize memory stocks, but it's just a smaller win for them. Yes, the extra headroom allows you to potentially do more parallel computing, where you take, say, multiple task packages, but I think the high bandwidth on the GPUs themselves ironically reduces your need for what's called high bandwidth memory.
So, overall, I'm bullish on hardware. I still think software in Q3 and Q4 is where we hit the mark. And I think that's going to continue to be the case . And so I suspect that's where things are going to go. Of course, people are nervous. The election is coming . Will Trump win? Whatever it is, it's going to happen.
All this nonsense -- we'll figure it out. Now let's talk a little bit about Warren Buffett. So , Warren Buffett has stepped down. It's kind of sad. Now he's chairman emeritus or whatever it's called, and his 70-71-year-old son has taken over the company. One of the last deals that Warren Buffett was supposedly involved in with Greg Abel was Greg Abel's purchase of Taylor Morrison for cash.
Taylor Morrison is a real estate developer. And Taylor Morrison is making a pretty interesting transition into real estate. Listen, if you don't want to hear about real estate, just remember: tonight the price of the Alpha subscription is going up by $900. We're making a huge change. You'll be well taken care of if you join the course by the end of today.
So if you see it available, buy it. If it's not available, it's not available. As for real estate, it's very interesting because real estate wins in two ways. It doesn't look like it's going to win right now, but if we get a massive deflation due to artificial intelligence. So, Let's say it's AI-induced deflation.
This will happen after the build phase is over. So this is the post-build phase, when most of the infrastructure is already in place. Once the AI deflation hits, interest rates will go down. Kevin Warsh is not a money printing press. When you're not printing money, cutting rates to zero is not enough.
You actually get negative bank rates. So the Fed cuts rates to zero , you're probably going to have to go negative bank rates, and 30-year mortgages will go below 2%. You're going to have 30-year mortgages at 1.8% just to stimulate spending when the AI deflation hits. The fact that they're not printing money means even less inflation, so rates will be lower than ever before.
I don't know when that's going to happen, but I firmly believe that real estate will dominate this mode. If the AI crash happens, well, then rates will also go to zero and you'll see a huge slowdown in construction, which means you'll probably get a 30-year mortgage for less than 2%. Real estate wins again.
Where real estate is struggling is during the endless war with Iran and the ever-higher oil prices, and also during the build-out of AI computing power. But it's not going to last forever . The bubble is not going to last forever. Real estate will come around . Just not now. That's why I say , if you have real estate now, I think there's an opportunity to gradually just pay down a little bit more of your debt.
And if the AI bubble or something lasts another 5 or 10 years, you'll have that same amount of time to build equity that you can use in the future. And that's exactly what we're also talking about when we talk about my company Reinvest, which is a small startup. And people ask me, " Kevin, you know, what about an IPO?" When is the IPO?
"So when we started the company in 2022, we were looking at, hey, maybe... and anything I'm going to say now is not a forecast, you know, we're not raising money right now. Uh, you know, I could be wrong about what I'm going to say. This is just a rough estimate for the sake of openness and transparency.
We started the company in 2022, and I was assuming that we could go IPO between 2028 and 2030. That's a phrase that I've been repeating over and over. There were times when we thought about going IPO earlier. Which would have been great. But I didn't think it was the best decision for the company at that point.
So I look at what we had 12 months ago. By the way, you know, this Warren Buffett concept was kind of a transition. I'm bullish on Berkshire Hathaway and Warren Buffett. I think he sees the same thing about real estate. That's why he's invested in construction companies. This morning we looked at a logging company.
And we're like, there's a catastrophic shortage of new homes being built. The shortage of new homes is only compounded by high interest rates. You're just going to inflate that housing market pressure. It's just going to explode. It's going to push prices up even more when rates come down. It's going to make housing even less affordable.
The wealth gap that I've been talking about for years is going to only get bigger. AI is making companies richer, and because of the structural deflation from AI—whether it's a boom and bust or just deflation—you end up making housing more expensive. So in 10 years, we could be in a situation where AI creates so much disinflation that you lose your bargaining power over your job, or it's minimal unless you're in a certain industry.
Or houses are going to be so expensive, or both, frankly. Houses are so expensive that nobody can afford them. And in this at some point you're going to have to implement socialism, which will cause a bunch of other problems. But this K-shaped economy is ultimately going to get worse. I believe that.
That's why I've been pushing this for the last few years. I say, "People, buy assets . Buy companies." Companies are going to benefit from AI. Real estate is going to benefit from AI. AI has caused deflation or a crash. Because it's not a housing bubble. There's not enough debt. Who's buying houses now?
Who's going to go all-in on real estate now ? Not many people. The stress in commercial real estate is almost over. So when they say that rates on shopping malls are going up. Shopping malls. Ultimately. Coming back to this conversation. A year ago we had virtually no software. I was actually looking at the flight plan.
I put this in our daily wealth review yesterday. I was looking at the flight plan, and it was my return flight, but I was meeting with our development team , so I had a screenshot of this plan because I brought us there myself . Now you know, I don't fly anymore. I don't think anyone should do piloting as a hobby.
I think if you're going to be a pilot, you should fly three times a week, although maybe that's not the recipe for everyone. Maybe it's more of a recipe for me. I think that skill is lost if you don't fly three times a week. And I thought I was a great pilot when I was flying. I mean, I was flying like five times a week.
It was crazy how much I was flying. But then I was flying less, and being a pilot became impractical. Although I'm still a pilot. I'm going to be a pilot, once I'm a pilot, it's forever. But I thought it was really interesting. I... I... I hope it's showing here, because my HDMI is connected. That's it.
Yeah, that was... Oh, it's only showing part of it. That was my flight plan. Never mind. Oh, there it is . This is our flight plan. And I thought that was a really cool memory to look at , but it was my flight plan home. Uh, you can see the flight number there. N694 Papa Papa, cleared to land. But...
You know, it was last July. What's the date on it? And I remember that for a reason. It was last July 5th . We didn't have the software that we have now. There was no Homes AI product. The beta version of "In visual" didn't come out until the end of December 2025. We had nothing . I mean, we had our Wedge software, which we, you know, bought for a few hundred thousand dollars, which, looking back, was a bargain.
We made a lot more money. Um, we had nothing. And we built it from there , and on June 30th, we released a beta for evaluation, which was amazing and it's only getting better. We just released another backend update, and now you should see four times as many properties in the Homes AI app. It's incredible what's happening at Reinvest.
It's amazing. And in July we released our exchange software, which is V0 of our terminal software that's coming out in the future. And we're going to do something really cool for current course participants, probably next week. That's why I say , as I said before, as soon as there 's a $700 or $900 jump, you should probably jump in as soon as possible .
I'm just being completely transparent. I know it's like last minute on a Friday night, and that's fine. Now I just think about what we've released in the last year, and I can only imagine where we'll be in a year. But the thing is, people keep asking me, "Kevin, when's the IPO?" ", and the answer is—I don't know.
I think I'll have a lot more clarity by January 2028. And then maybe we'll still be investing in the 2028 to 2030 target period. But it all comes down to how we grow our annual recurring revenue and how our software grows. We think that in the time between now and then, we'll probably do another fundraiser.
Maybe we'll do a fundraiser at the end of this year. Maybe we'll do a fundraiser, I don't know, next week. Probably not next week, though. Those kinds of things take a little bit more time to prepare, but we're not raising funds right now. And we expect that the vast majority of those funds, I would guess 85-95%, we'll just invest in real estate, building in supply-constrained areas where it's hard to build, and we'll buy good deals.
Mostly because our programmatic direction and our security allocation —all of that is self-sustaining, which is just great, whether it's through our memberships, rental income or whatever. It's all self-sustaining. The company is operating at a profit. From a cash flow perspective, it's great. So it's positive operating cash flow .
And you have to be careful how you phrase these things because, you know, you're subtracting the interest on the convertible bonds that we expect will probably convert next year, and then we don't have that interest anymore. You have depreciation on the property, and the houses have only increased in value, but we're still taking depreciation.
Yeah, I'm pretty sure it's over a million dollars a year. Maybe even more. So whatever it is, let me make a quick guess. Honestly, I should just open the data. That's probably better. Oh, it's probably closer to $2 million a year. It's probably closer to $2 million a year, which is just crazy. But it's great, because that means that when we get revenue from the software, we're actually protecting it by depreciating the real estate.
Meanwhile, real estate is going up in value. That's just ridiculous. But we think we're going to use about 85 to 95 percent of our next fundraiser just to buy real estate, because we think that creates a very strong foundation. For example, a property that 's been bought for cash, is fully leased out.
That's a strong foundation with its own property management. So we're really excited about that. And we already have over 10,000 people using the Meet Kevin app every month, and that number is growing. We're continuing to reach new heights, new heights, new heights. But this is just the beginning . So we're working on tools to make things even easier and better for people.
And we think that real estate is the lifeline for all of this right now. Can we be optimistic about AI? Can we be optimistic about on chips? Can we be optimistic about software? Yes. But we're being cautiously optimistic . And I think that's the most transparent way to look at all of this. So if you found this helpful, like this video and subscribe.
Go to meetkevin.com , meetreinvest.com . It's the same site. And become an Alpha member before it's too late. Remember what you get. You get an Alpha report every morning. We went through our Monday report. We went through our Tuesday report. We went through our... What do we have here? I mean, we went through every day.
We went through the Fed meeting day, Thursday, Friday. We went through all of this. And of course, remember what you get. An Alpha report every day that the market is open. Buy and sell alerts. Live for course participants. In short-, medium- and long-term analysis. All courses. Lots of interesting stuff.
And seriously, a big price increase is coming. So , as I said, I advise you to take the opportunity and join in on these big changes. The course participants will definitely love what they have to offer . So exciting. See you soon . Bye and good luck. Why not promote what you ha.
Comments 0