---
title: 'The Start of the AI Boom (Or Depression) [How Cooked are We?!?!]'
source: 'https://youtube.com/watch?v=SLG4LOeBjek'
video_id: 'SLG4LOeBjek'
date: 2026-11-03
duration_sec: 2083
channel: 'Meet Kevin'
---

# The Start of the AI Boom (Or Depression) [How Cooked are We?!?!]

> Source: [The Start of the AI Boom (Or Depression) (How Cooked are We?!?!)](https://youtube.com/watch?v=SLG4LOeBjek)

## Summary

Kevin Paffrath analyzes the current state of the AI economy, contrasting strong labor market data with massive off-balance-sheet AI commitments from big tech. He argues that while token prices have collapsed, hardware costs are rising, signaling a shift toward hardware and platform (SaaS 2.0) winners. He provides a bullish outlook on AI hardware stocks, despite market volatility and concerns about circular financing.

### Key Points

- **Labor Market Strength** [00:03] — JOLTS data showed 7.523 million job openings, beating the 7.1 million estimate. Prior month's data was revised up by 1.5%. Quits rate increased from 1.9% to 2.1%, a positive sign.
- **Two Economic Legs** [02:25] — The economy is supported by two legs: the labor market and AI. If either collapses, the economy is in trouble. AI recession could trigger a labor market recession.
- **AI Commitments Soar** [04:04] — Big tech has $4.11 trillion in off-balance-sheet AI commitments, up from $3.1 trillion in August and $1.8 trillion in June. This is more than double in five months.
- **Token Prices Collapse, Hardware Costs Rise** [05:02] — Token prices are down roughly 50% year-over-year, making AI cheaper to use. However, H100 rental prices are up 6.2% since May 31, and newer chips like B200 and B300 are seeing even steeper price increases.
- **Palantir's Earnings Call Criticism** [08:15] — Kevin criticizes Palantir's earnings call analysts as 'shills' and 'memers', noting only one analyst asked a substantive question about competition with Anthropic. He sees this as a sign of a transition.
- **Anthropic's Chip Ambitions** [10:13] — Anthropic is working on designing its own chips using Claude. The company could be valued at $1.5 trillion with $80 billion ARR, but is not income positive, unlike Palantir.
- **AMD's Strong Earnings** [12:48] — AMD's revenue was $13.37 billion, with gross margin expanding to 54% from 51.7% last year. They beat expectations and guided for 27% sequential growth next quarter.
- **Winners: Hardware and SaaS 2.0** [13:29] — The winning businesses are selling hardware/chips or being a 'SaaS 2.0' platform (like Palantir) that locks in users. Token selling is seen as a losing business model.
- **Valuation Bargains** [17:08] — Nvidia (PEG 0.67), Broadcom (0.63), and Micron (0.33) are considered bargains. AMD (PEG 1.2) and Apple (PEG ~2.1) are more expensive but still attractive.
- **Hyperscaler Capex Surge** [22:26] — The five hyperscalers burned $31.4 billion in free cash flow in Q3, with capex up 118% to $230 billion. This shows they are reinvesting all profits into AI.
- **Capex Projections** [30:13] — Bank of America projects total capex to rise from $240 billion in 2024 to $1.2 trillion in 2027, turning free cash flow from positive $240 billion to negative $180 billion.
- **Credit Default Swaps Stable** [31:06] — CoreWeave's credit default swaps have risen but not to new highs, indicating market stress is manageable. Oracle's 40-year bond yield is 8.43%, while Microsoft's is 6.08%.

### Conclusion

Kevin concludes that despite market volatility and concerns about circular financing, the AI boom is just getting started. He recommends diversified exposure to AI hardware and platform stocks, viewing current valuations as post-bubble opportunities.

## Transcript

and of course, companies like Super Micro and others just had earnings. And a lot of people are wondering, Kevin, what is going on with the AI play and how is it going to affect the economy? Because just this morning, we got Jolt's
data that was pretty good. Now, that's just job openings data. You know, we've come we're coming off of last month when we had good S&amp;P and PMI numbers as well. Just little reads on the economy. We've got big numbers coming up later this
week, but they're the second half of our economic analysis. They're really two halves. There's the AI canary, which is really at this point more than just open canary, and then there's the labor canary. This video is going to be
canary. This video is going to be focused on AI. So what we had with the canary on jobs, just a quick side note this morning was we got Jolt's data that indicated 7.5 23 job openings. The prior estimate was
that uh or the survey rather for this section was that we were going to come in at 7.1 million. So we beat that by a substantial amount which is great. In other words, more job openings than expected to the tune of 5% over
course people say, "Oh, but Kevin, those are just going to get revised down." Well, they actually revised up the last Jolts numbers from the prior month by 1.5%. So September beat because we just got
weeks. Uh, and then the prior month came in better than expected. So for sort of a quick little side tangent on the labor market, this is good. The quitz rate market, this is good. The quitz rate also increased from 1.9% to 2.1%. That
is a good sign. People don't quit into a really bad labor market. Now, yeah. Is the labor market the fiery economic hot potato that it used to be? No. This is nothing like the labor market we saw after CO where everybody could get
massive pay increases and make a whole lot of more money. Now, less than half of people who quit are making more money than they were at their prior job. It used to be more than half were making more money if they quit and went to a
new job. After COVID, everybody was making more money when they left a job. making more money when they left a job. It was crazy. That's the labor market. Revisions are actually going positive. But what about that AI canary? Because
that's the second half of this economy. And both of these are the legs that hold up this economy. If one of these legs collapses, we're cooked. Labor market goes into recession, we're cooked. AI goes into a recession, we're cooked.
recession, which one could cause the other to fall over. We're mega cooked better hope that somebody like Kevin Worsh can come out with the money course, given that, you know, he kind of came in under project 2025, which is
sheet and don't run the money printer, I think your expectations and hopes that who's going to run the money printer are not good expectations to have at all.
Now, small extra note before we get into artificial intelligence and some of the data. I do think it's interesting to see how many hikes the market is still pricing in. Still pricing in those two hikes. In fact, this morning we were
sitting in per the uh Kali uh prediction market here, we were pricing in about a 27 to 28% chance that we would just get one rate hike. That's now moved down to 24% suggesting we're going to get that second hike. Probably because those
jolts numbers came in saying, "Hey, no, so far both of the legs of this economy are cranking, which is a good thing. We don't really want the economy to roll over, cause a lot of pain and suffering. So, what about AI?" Quick note, uh,
Koshi is a sponsor of my startup, Reinvest. When you sign up for Koshi, you get $25 at kshi.com/r/meet and you make your first trade. You could Here's the scoop with artificial
intelligence. Big tech now has $4.11 trillion of AI commitments that sit poof off the balance sheet. We just got AMD earnings that freaking crushed. They were fantastic earnings. and they increased
and beat expectations by about four and a half percent for their guide for next quarter with sequential growth that'll be around 27%. Really good. be around 27%. Really good. They have less capacity though to go
conduct circular financing like Nvidia does. They've got somewhere around $500 even hold a candle to the free cash flow that Nvidia has. And that's one of the reasons Nvidia is one of the big contributors to making uh these big old
commitments that don't end up showing up on their balance sheet. But $4.1 trillion of AI circular financing commitments, it's all got to go right for this not to go poopy dupy. The good news is so far things are going right.
And so when we look at the actual data for what's going on, there's a trend of what we see. First of all, token prices are down roughly 50% year-over-year, which is great because it makes it less expensive to do productive things. But
what's actually happening is over that same time frame, so from last November, we are seeing in fairness, uh what hold on uh forget the inferest comment there. We are seeing H100 prices moving up at
this slope right here, this red line that I've drawn. That's the slope roughly of your H100 pricing. There we go. uh increasing. So why are H100 go. uh increasing. So why are H100 rental hours going up uh from May 31st
up 6.2% but from over here you're at under a buck to now well over a buck about a buck 30. It's pretty impressive. You know that's actually good for underwriting some of that core debt or some of that nebious debt which doesn't
Cororeweave does. If I now go to the B200, you could actually see that I only have to move the right side of this. And what to move the right side of this. And what happens is the slope actually increases.
So we get more aggressive at the price increases for that newer hardware. But even 4 and 1/2year-old hardware is getting more expensive. And then if I getting more expensive. And then if I come over here to uh draw a slope for
come over here to uh draw a slope for the B300, the slope is even yet again more aggressive. So, not only are these chips getting more expensive on a per hour basis over the last year, but the rate of acceleration is increasing. So,
they're getting more expensive at a faster pace, especially the latest part makes sense, but the fact that a 4 and 1/2year-old chip is still growing at this rate of this blue line here, remarkable. This is in my opinion
because AI is truly inflecting to actually become more productive and more functional and this is going to in my opinion change where some of the focus goes for these investment companies and and really where you know ultimately you
might want to invest your money. I think that the true beneficiaries of AI or the companies that utilize the tokens token costs mind you I forgot to mention it but token costs are the white line right here. Token costs have absolutely
cratered. This is the token index and you can see the token index went from a peak of about $2 all the way down to about 90. So you've more than 50% excuse me more than 50% collapsed and
I'm I'm using uh the um this this indication over here this 91 line where they sit right now. So you've seen this collapse in token prices who this is beneficial obviously for underlying companies. So for example, my little
startup reinvest, we can use cheaper token prices and we run a lot of machinery ourselves to do more to provide more value to our customers to signing up for the reinvest terminal or reinvest AI for homes or stocks,
whatever it is, the research. It's great and we can't wait to share all this with you over time. But forget that for a moment. Focus on what we heard this morning or what we reviewed this morning. This morning we went through
going to go through all of this because some of this is just for course members. But one of the things that we realized is that Karp over at Palanteer is starting to get a little tight with who he's letting onto his earnings call. And
I think this is a little sussy because I think a transition is happening. Now I'm not here to bag on Pounder. I'm going to be clear. I've got exposure to Poundier. I like Palanteer's cash flow. Although their cash flow yield's only about, you
know, 1.3ish%. It's not that great. Apple's got about twice the cash flow yield as Palanteer. But this is it. These are the analysts Palanteer put on their earnings call. And I want to grab Karp and shake him and go, "What are you
doing? Why don't you have some people with balls on the earnings call?" They literally use their earnings call as a soapbox for people inside of their company or memers. They literally have Dan Ies who is a bootlicking permutable.
Then you have Brad, this analyst that I call the rocket ship memer. You want to know why I would call the guy a rocket ship me? Congrats on the rocket ship quarter. This was the only question he got to
ask. And he goes, "Congrats on the rocket ship quarter, guys." And then you ready for the really hard question, Alex? You said in your shareholder letter that you see Palunteer being a pioneer of genuine and viable anti-war
movement. What? This is worthless. These analysts are crap. The only analyst that actually crap. The only analyst that actually asked a decent question was one person
and it was Jill. That's it. Jill comes in and says, "Hey, you guys are basically competing directly with Anthropic." And that's where actually things get really interesting because people might not like to hear this,
especially the Palanteer folk, but there is something interesting that happened that's going on and I think I figured it out. So, if I go look at there's an information piece right here that says Anthropic is now working on
using Claw to design its own chips. The company could already be valued as much company could already be valued as much as $1.5 trillion, and their ARR is probably somewhere around $80 billion a year, which is, you know, obviously way
magnitudes more than what Palanteers is. But Palanteer is actually income positive and Anthropic's not, right? Palanteer actually makes money. So you That said, this is very interesting. They're now talking about competing with
chip designers like AMD and Nvidia. It's a high margin product. Now, why do you Well, a because you can make money selling chips. But who are you really trying to sell chips to? Well, in my opinion, I think the writing is on the
opinion, I think the writing is on the wall. Selling tokens. Okay, sell tokens. That's going to go away. That is going to be a very bad business model in the writing on the wall for this. This is, by the way, what Karp harps on. He says,
"Don't give all your alpha and all your data to the big people like Claude or Open AAI." And that's fine. He can harp on that all he wants. But what ultimately happens is Claude and Open AI realize that's what's what's going to
end up happening is people are going to want to run their models locally. What's Mark Zuckerberg at Meta already doing? Meta is already saying, "Hey, you want to run our watermelon model, which is our latest and greatest Frontier AI, no
problem. Daisy chain together two Mac Studios, which in fairness, Apple just had their earnings and their pricing power is actually expanding during a power is actually expanding during a time that memory prices are going to the
moon." And then it makes you wonder, wait, how is Apple expanding their say, I don't know, but they're pulling it off. Uh, look at their income statement. Uh, right here you can see that Apple's gross margins on products
sit at 39.3%. Last year, they sat at 36.8. So, you actually have substantial margin improvement despite or I should say in spite of memory prices skyrocketing. Even their services margins went up.
That's pricing power. That's impressive. It's kind of like the pricing power you're seeing over at AMD. We can look at those valuations in just a moment, but let's jump into the AMD numbers over here. What do we see here? First of all,
they have the leadest revenue I have ever seen. 1337. That's freaking hot. I like that a lot. But their costs of sales, so basically, and we're going to figure out their margin over here. Their gross margin has grown to 54%. Last
gross margin has grown to 54%. Last year, they sat at 51.7% gross margin. So pricing power is going up at these chip designers and hardware sellers. Apple, designers and hardware sellers. Apple, AMD, Super Microbe earnings, big beaten
raise on their forecast for how many racks they're going to install. And so what's happening? Well, people are realizing that selling tokens is the loser business where you're going to win, at least for now, you're going to
win in two places. You're going to win in selling people the hardware or the in selling people the hardware or the chips or you're going to win by being chips or you're going to win by being the SAS 2.0 that people go to. SAS 2.0
is being the lockin platform. This is like being the palunteer. Okay? So, you're the palunteer. You could go use your own hardware. Palanteer sits on the hardware if you don't want to sell it to the cloud. SAS 2.0 is the platform. It
is like a palunteer, but Palanteer is not going to be alone here for long. They're going to have competitors that help basically use this onlogical process to break down big data into digestible bits that employees at
companies can actually use. Now, a benefit that Poundier has is they're really good with their forward deployment engineers to integrate their AI platform, essentially their sovereign AI into companies. existing customers
are increasing their spending at Palanteer by like 95%. Really this to come across as dunking on Palanteer. Like they are crushing it. Their numbers get better and better and better and better. So what? They have
shills on their earnings call. Their numbers are very good, but the writing is on the wall that competition is coming. And so what's going to happen is you're going to end up getting SAS 2.0. And in my opinion, this is my
prediction. SAS 2.0 I know is going to be driven by a bunch of companies that be driven by a bunch of companies that say, "Hey, you want to use the clawed software? You want to use the open platform? Hey, you want to use the
platform? Hey, you want to use the Palunteer platform or the Salesforce platform, whatever. We'll be the platform for you. We will license it to
you." That's what we're going to do is the bottom line is we will license you our models. We'll license you our software. But guess what? All this crap software. But guess what? All this crap is going to get run on AMD chips, on
is going to get run on AMD chips, on Nvidia chips, on Apple computers, on AS6 that are manu uh that are designed by companies like Marll and Broadcom. Uh or you know obviously they're buying Lumenum which I think is a great deal.
uh they are manufactured by Taiwan Semiconductor uh as well as the Intel Corporation and then of course you've got ASML which is your core for their
backlog is going to be insane when they reported in January because more people want these chip manufacturing machines pretty excited about uh the growth that they end up reporting in January but all this SAS is just going to be an option
for entrepreneurs small medium and big businesses to essentially use this talk about, oh my gosh, it's like circular financing and the buildout and all this, I think the buildout is honestly just getting started because AI
has really only gotten useful in my opinion, like really useful in the last 6 months. And Jensen agrees with that of Nvidia. Now, keep this in mind, the number of spending or the level of spending is not going down. It is not
rolling over. So, understand this. Big tech spending right now offbalance sheet commitments 4.11 trillion. Guess where we were in August? 3.1. Guess where we we were in August? 3.1. Guess where we were in June? 1.8. We have more than
doubled our commitments in just the last five months. Because all of these big If you don't see the writing on the wall, I think you're missing the boat. these companies, we're like, "All right,
so where are the really cheap companies right now?" Well, I'll tell you. AMD and Apple, I love them. I've got exposure to them. They are not the cheapest options them. They are not the cheapest options right now. Uh Apple sits at a valuation
estimates, although I think their numbers are going to be wrong. Wall Street only projects 9% growth for Apple. I think that'll be a lot higher. The last earnings report they had only had one week of new Mac Studio sales.
They're going to blow this growth out of the water. Uh so I put them closer to 16 to maybe even 20% growth, which puts them close to a two peg. But they're also very strong balance sheet. Fortress cash flow, you know, twice the free cash
flow as Palanteer. Pretty good. Palanteer itself uh has a PEG ratio Palanteer itself uh has a PEG ratio that's also somewhere around 2.1. So why do I bring this up? Well, because if Apple's around 2.1 to, you know, almost
four depending on whose estimates you use. If AMD is sitting around 1.25, Kevin, where are the bargains? Nvidia is a bargain at about a 67. Broadcom is a
a bargain at about a 67. Broadcom is a bargain at a 63. Micron is a bargain at bargain at a 63. Micron is a bargain at like a 33. SanDisk, SKHix, these guys like a 33. SanDisk, SKHix, these guys are dirt cheap right now because all the
leveraged homies who rode the wave of the we meme wave of memory and RAM getting expensive. Next is NAND. By the way, I really think NAND storage like flash that has not gotten built out the way we expect memory to come online. I
expect a lot of memory to come online. NAND, Seagate, uh SKH Highix, that's the next level. That's the next thing to happen in my opinion. Anyway, really a topic for a different video, but the meme wave that happened in memory has
leverage. And so, you've seen these insane swings in hardware stocks. But in my opinion, so many of them have gotten commitments are just skyrocketing and earnings are really good. People are
like, "Oh, but Kevin, you know, the Intel earnings were AWS. Who cares? Did you actually read the earnings? They missed because PCs missed. Bro, AMD's earnings were fantastic. Guess what happened to PCs?
They went to crap. I kid you not. It's right here. client and gaming segment was down 8% year-over-year. Yeah, that's also how Intel got cooked. It has nothing to do with AI. In my opinion, it's a sign of AI strength. AI is so
good, it's driving up the prices of RAM, and as a result of that, what? Well, and as a result of that, what? Well, here you go. You're now kicking butt uh with everything absent, consumers buying data centers basically double over here
at AMD, which is great. Uh so, beyond that, then you've got to think, okay, fine. So where you know where does AMD fit in the stack? AMD fits in with their Halo platform which is basically enterprise artificial intelligence their
Halo platform like their AI max pros you know these can go in mini PCs or larger mini PCs. Big big fan of what AMD can these. Now some of the mini PCs the AMD chips are only like 25 bucks so the
That segment has a lower margin than obviously data center. But I wouldn't be surprised if you keep seeing some of the juice on frontier AI spending over time juice on frontier AI spending over time because it's where the dollars are. The
frontiers want to stay the frontier. The metas want to train you, the watermelon frontier, so they can get you on their platform. Once they get you on their platform, then they jack up the rates. That's literally what Elon Musk just
did. Elon Musk got everybody on Grock with really generous limits. And then guess what happened? Oh, uh, we're spending too much money on Grock. We're bunch of people are like, "Yo, why am I hitting my limit so fast? What'd you
guys change?" They get you in, they jack up the price. That's the whole point. They want to sucker you into the platform. That's what Zuck's doing. Zuck gave me the Max plan for free for a year. You think he's doing that out of
thinks I'm going to become a paying customer in January of 2028, baby. And I might. So, he's not wrong. Good for him. You know, Claude is is probably the weapon
of choice I would choose right now for a lot of different work. Uh especially, you know, design work uh or agentic work, but OpenAI is fantastic. I'd say OpenAI is even better at some uh research tasks for or contract tasks for
like actual documentation review. Really good. problem is the market right now has to deal with the fact that we got a crappy election where Trump is probably are probably going to get hosed. I'm kind of face paling a little bit about
how freaking stupid it was that they didn't figure out how to get a deal beforehand, but you know, I guess Trump didn't put enough effort in. You hear the Cali predictions right now. 74% leaning Dem for the Senate, 93% leaning
to be a landslide for the Dems unless all the prediction markets are wrong. We'll see. But, uh, the next thing you've got to know is the five hyperscalers combined, the five big ones per Goldman Sachs burned 31.4 billion of
free cash flow in Q3 and their capital expenditures were up double 118%. The fact that their capex was at $230 billion and they were only negative 31.4
4 billion is actually kind of crazy. It shows that they've taken all the money they make from selling you, you know, Microsoft, uh, Word, and they're turning all of that money into AI spending and investment because all of them see the
payback period for how good AI is getting. And that's why the commitments are exploding. That's why I'm, you know, like Mr. Buy the dip over here. I'm There's so many good cheap stocks in the hardware sector. This is a steal because
a bunch of weeny babies are getting, you know, weanied out on leverage or because there's an election or some nervousness around Iran, whatever. Labor market's holding up and the juice is just getting swosen over at AMD. Let's take another
some of the other stats here. So, if I grab the uh AMD earnings report that just came out moments ago, another thing that I could see over here is that they want to get into space chips that could sit in space for 15 years. I wrote a
little note that of course because they want Elon to spend on their chips, which partnership, they only want to use SpaceX or they only want to use Nvidia. But AMD knows Elon's going to spend like a drunken sailor. So, why not, baby? Why
not make a chip that at least is going to attract Jeff Bezos? Uh, okay. So, makes sense. They're going to spend money like crazy. Then, uh, you've got, uh, we talked about client data
centers. We talked about data centers being up double. That was impressive. We talked about, let's see here, third quarter revenue. Ah, yes. Okay. So, let's look at the revenue numbers. And we could see that if we jump onto the
income statement over here, this was very impressive. So their sequential growth is 15.9%. They're forecasting, so that's from the between this quarter and next quarter, it's going to almost double from 15.5
uh 15.9% to 27%. So in other words, you're going to see a growth acceleration at AMD. And this is beating estimates. This is one of the reasons why you're seeing AMD rush back
reasons why you're seeing AMD rush back to that 520 legendary line that we have. Every time in my opinion AMD is under 520, you should just buy it. Not personalized financial advice, but that line right there is very juicy because I
think we're going to stay on top of that line in the long term. Now, AMD is not the cheapest out of these. We have the valuation for it right here. Uh we've valuation for it right here. Uh we've got AM Dizzle. Let's see here. AMD, AMD,
where did I put my valuation? I got to find it. So, Apple was two uh based on a find it. So, Apple was two uh based on a higher growth rate. AMD's forecast or their PEG ratio was 1.2. That makes them twice as expensive as
Nvidia. So, that's their PEG ratio based on their forward growth of 56% on EPS. They're twice as expensive as Nvidia, but they're obviously a fraction of the size. So, the argument is that maybe they have more potentially to grow. I'm
a big fan of this. I I think the whole stack is how you kind of diversify. Again, not personalized advice, but I look at it as you kind of want exposure to AI wherever you can get it. You don't need to go all in on one Neo Cloud or
one stock because the volatility will crush you or break your heart or both. Instead, put a little bit into Marll, a little into Broadcom, a little into AMD, a little Apple, a little Nvidia. Get a little Microsoft in there. I don't like
don't mind a little Amazon, a little SpaceX and Tesla, you know, to me that's a bucket that's going to be deliciously gor glorious over over time. Uh or if of venture capital, you know, a little bit of AI startup exposure, whatever. Uh
the the whole bucket is is very exciting to me. Uh so looking some more at these numbers here, we talked about carve margin improvement uh moving up operating income 2.22 times. Really impressive. So really,
really nice growth here. Now, Amazon's purchase orders for AI have doubled since August. So, it's not just these companies that are doubling operating income. Amazon's literally spending twice as much money. Uh, that's crazy.
Uh, so kind of impressive. The account does include guarantees at their full size before any discount. So, in fairness, some of these companies might get a little bit of a discount on some of those uh uh commitments that they're
making. Small little note we've got over here, but this is still very impressive. here, but this is still very impressive. Uh in or how should I put this? Buybacks as a result of this are getting burned. You know, free cash flow over at even at
AMD is on the small side. Uh free cash flow, their balance sheet, they got of money to pay their bills, both short and long term. So, that's not a problem. But free cash flow is only 554 million. This is why they tend to offer their
This is why they tend to offer their stock as opposed to uh actually offering cash to invest in companies like Nvidia does. So they'll go to like OpenAI and say, "Hey, work with us and we'll give you call options on AMD stock at 1
cent." They're called warrants when they're with a company, right? Then uh buybacks over here as well. AMD also took on another $4.7 billion in debt. So even AMD is like, "Oh no, this is the moment, man. We gota we got to spend
more money. Let's borrow more and grow more." Okay. Now, when it comes to least capex, Morgan Stanley is forecasting $195 billion in lease capex in 27, 246
in 2028. Honestly, it'll probably grow even more than that. Lease spending, by like they have free cash flow when, in know, that's not showing up on the plant property and equipment side. That just
advertise out their leases or their lease contract terms uh in the cost of goods sold. So, a little bit of trickery that does go on there. But still to me, these numbers are saying if these big boys are now just doubling their spend
suggests we're more early innings rather than late innings. Kind of wild. Now, obviously recently some fundraising concerns have been present. Open AAI wanted a $30 billion uh at a $1.4 trillion uh valuation in September. That
has now fallen. Uh they're only going to raise $20 billion at a $ 1.25 billion a trillion dollar valuation, $1.4 trillion down to 1.25. 25 trillion. Pretty big numbers. But anyway, smaller check, smaller price. People are a little
nervous right now. Again, partly election, partly Iran. This this all this stuff will be over soon. SpaceX obviously is going to have to borrow somewhere around a trillion dollars. You know what's his name? Um
uh who was it? Uh Bill Aman. Bill Aman's like, "Oh, SpaceX is going to have to You know, I made a video about that a week ago before he tweeted it. I wonder dollar problem for SpaceX. I encourage you to watch it. But that's also a sign
that it's not just Amazon or Google or Neta or whatever that's spending on all this. It's also your boy Elon. But I mean that that is almost the default. Like it's like asking if the Pope is Catholic. Is Elon going to spend money?
Catholic. Is Elon going to spend money? Is water wet? They're all the same, Is water wet? They're all the same, right? So the problem with this is, you that's where some of that nervousness is coming from. Bank of America's HeartNet
argued that capex in total, not just leases, is going to go from $240 billion leases, is going to go from $240 billion in 2024 to $1.2 trillion per year in 2027, or at least for 2027. It's expected to grow even more into 2028,
which is kind of wild to think about. Free cash flow as a result goes from positive240 billion to negative 180. That's why you're seeing some of that credit default swap stress. Uh, and that's why some people are like, "Oh,
Well, eventually it's probably all going to blow up into a bubble, but the valuations we have today are like postbubble valuations. You know, Cisco was trading for like a 10 to 20x valuation after it collapsed. That's
kind of where we sit now on a lot of these tech companies or worse. Anyway, if I look at the credit default swaps for Coreweave, which is a really kind of sensitive analysis for how people have been feeling lately, they have popped up
mostly because the market's been selling off, but they have not been going up as much as they did historically, which is good. I'll pull those up on screen in just a moment. Uh, also worth seeing that you've got
uh Oracle's bond yield, and this is the other part is it gets a little more expensive to feed the debt kitty. Oracle's bond yield for their 40-year debt is at 8.43%. Microsoft is paying 6.08%, which is
wild. These are the credit default swaps for Coree, showing that they've risen, but you know, we're not going to new highs even though the market's selling off. you've had uh quite a bit of a bare market setup so far, four of the mag
seven are in a bare market. More than 60% of stocks in the S&amp;P 500 are more 60% of stocks in the S&amp;P 500 are more than 20% off of their all-time highs. Yardini, you know, Edardini and their research group, they actually think this
"All right, let's just prioritize money where it belongs rather than just everywhere." Okay. Uh Deutsche Bank thinks there's still a little bit of downside, maybe as much as 5% more downside, mostly because
you need to get through this Iran situation. And you know, it doesn't it's election. So, it means we got to deal with the choppiness for a little while with the choppiness for a little while longer, which just sucks. Oh well, we'll
get through it. Uh so, then you've got data centers. uh data centers sit around data centers. uh data centers sit around 29% of mega project starts. Now why do I say that? Because it used to be liqufied natural gas. They're at 23%. Data
centers are now the big play. That is obviously a big part of this election is data centers. Well, prediction markets say the answer to that is no. People aren't going to want them. We'll see. I mean, you know, this is always um this
is well, how should I put it? My my goal is always to be as neutral and unbiased know, between you and me, the boom is just getting started. To me, that just
gives us more time to, you know, you can get maybe, you know, I'm kind of like a 9010 is sort of my mindset. You know, it's sort of like 90% taking the time to buy real estate and that's sort of just where our world is.
And then, you know, maybe maybe it's 5% stocks or it's 10% whatever, right? So, I'm kind of in that world. I'm really bullish on that world though. But I also estate, which makes me really interested
in real estate. Perfect kind of time. Uh but I kind of put hardware in that same bucket right now. People are so bearish on AI hardware and it's nuts. It was just crazy. So anyway, that's uh that's my take on where we sit and uh in the
for you. &gt;&gt; Oh man, your boy Kevin just lost lots of money. Oh, but that's okay. I am with you. With you. Oh man. Good boy. Kevin
just lost lots of money. Oh, but that's okay. I am with you. Here we go. Come on. Push it. Find the dip. Find the dip. Here we go. Come on. Push it. Find the dip. Find the dip. Here we go. Come on. Push it. The dip.
&gt;&gt; If you like that video, check this one out. I think you're going to love it. told us here? I feel like nobody else knows about this. We'll we'll try a &gt;&gt; Congratulations, man. You have done so much. People love you. People look up to
&gt;&gt; Kevin Pra there, financial analyst and YouTuber. Meet Kevin. Always great to YouTuber. Meet Kevin. Always great to get your take.
