Valuations Are Falling for a Reason: AI Is Repricing the Future — Jordi Visser (22 febrero 2026)

[0:00] Um,
[0:01] short and weak, but as has been the case
[0:04] all year, uh, a lot going on
[0:06] particularly, uh, with regards to
[0:09] AI in the software side. So, um,
[0:13] I just want to bring this up at the
[0:14] beginning. I am,
[0:16] uh, I want to make sure people hear
[0:17] this. My My overall views for the year
[0:20] are that somehow or another the market
[0:21] will finish the year up. Nowhere near as
[0:25] much as you would expect if you thought
[0:27] GDP would be good, profit margins would
[0:29] expand, and and earnings would be good.
[0:32] Uh, this has a lot to do with my views,
[0:34] which I'll go through today, on multiple
[0:35] compression for anything related to
[0:37] software, and then the difficulties that
[0:40] come with people just dealing with
[0:43] cyclical items, uh, on the hardware
[0:45] side. So,
[0:47] I'm going to spend a lot of time on SAS
[0:49] again, um, because I want this to be
[0:51] kind of the last time. I I don't like
[0:52] getting involved in this. I think this
[0:54] is, uh, hurting people by trying to buy
[0:57] into something that is clearly going
[0:59] through an AI disruption that I
[1:01] understand most people don't spend
[1:02] enough time on AI to fully grasp the
[1:04] concept of bespoke. Um, and that is
[1:08] what's happening. Entrepreneurs are
[1:09] rising rapidly, and enterprises are what
[1:12] people are focused on for revenues, but
[1:14] that is not the right way to go through
[1:16] it, in my opinion. Uh, I'll go through
[1:18] that. I'm going to go through the
[1:19] dispersion side again.
[1:21] Uh, I'll talk a little bit about the
[1:23] videos and the stuff that is on the
[1:25] paywall. Uh, I'm going to try to help
[1:27] all of you, uh, but particularly people
[1:29] that have reached out on the trading
[1:31] side. This is going to be a different
[1:32] environment. Uh, Covar, as you guys are
[1:35] learning in terms of covariance matrix,
[1:37] I do not see any change happening. I
[1:39] think this is a new regime. Uh, not only
[1:41] do I think it is, I think as I go
[1:43] through this, hopefully some of you will
[1:45] will see why. Um, we'll go through
[1:48] Bitcoin, where things stand with that.
[1:49] When is it going to break away from the
[1:51] SAS side? The recap, the SAS debate, all
[1:54] that. You see how much is related to
[1:56] SAS. You just can't get away from it.
[1:58] It's the gift that keeps on giving. So,
[2:01] um I did put this out uh on Friday uh to
[2:05] literally go through. I just want to
[2:06] read you something because this, you
[2:08] know, I I write things, especially now
[2:11] with the paywall, a lot of this is for
[2:12] the people uh the individuals who are
[2:14] out there who are trading the market and
[2:16] are are curious as to what's going on,
[2:18] but obviously for asset managers as
[2:19] well. And in most the institutional
[2:21] conversations I've had this week,
[2:24] the question has been when does this
[2:25] stop? My vol has gone up. When are we
[2:28] going to get back to a more normal
[2:29] point? And what I wanted to basically
[2:32] get through here is that the dispersion
[2:35] is massive.
[2:36] Uh as of uh yesterday on Thursday, there
[2:40] were a total of 185 stocks in the S&P
[2:42] with a greater than 15% absolute
[2:44] performance year-to-date.
[2:47] Last year on this day, there were only
[2:48] 81.
[2:49] Um basically in both directions, we're
[2:51] getting massive movements. I go through
[2:53] in that paper why, in my opinion, this
[2:56] is happening and why
[2:57] this is it. Um for the next
[3:02] 5 years, 10 years, whatever case you
[3:03] guys want, uh this is it. So, if you're
[3:06] trying to pick a bottom in software and
[3:08] you think that all of a sudden vol's
[3:09] going to come down within inside the
[3:11] market,
[3:13] just know my opinion is
[3:15] no chance.
[3:16] Uh and I don't like saying no chance on
[3:18] anything, but basically AI will be a
[3:20] disruptive force and behind AI is
[3:23] humanoids. So, I I don't know how people
[3:26] don't understand that. You can't
[3:27] forecast things out 3 years. I've talked
[3:29] about this. The speed is there. So,
[3:32] the software side, this is basically a
[3:34] chart of the
[3:37] S&P overlaid with the PE of the software
[3:42] index just to show
[3:44] how much we're breaking away. The
[3:46] software side has only gone down when
[3:50] the S&P has gone down in terms of
[3:51] people's viewpoint of it. This is a
[3:53] structural shift. Um and as I'm making
[3:56] the argument and I wrote in my Substack
[3:57] this week, this is just the beginning of
[4:00] a bigger story. So, if you're trying to
[4:02] pick the bottom of SaaS, in my opinion,
[4:04] like I've said before, you are an AI
[4:06] bubble-ista. You still believe that this
[4:09] is a bubble. You do not understand that
[4:11] this is the year that there will be
[4:12] billions of Einsteins doing work while
[4:15] you're sitting there not believing in
[4:17] AI. If that is who you are,
[4:19] you're going to have a hard time
[4:20] adapting. Um
[4:23] and here's what you're going to be, if
[4:25] you have a job, in 5 years. This is what
[4:28] you're going to be in the elevator with
[4:29] is effectively humanoids in with you.
[4:32] This is not going to stop and that's why
[4:35] I will just keep showing this over and
[4:37] over again. This is a supersonic
[4:39] tsunami. It is speeding up. It is going
[4:42] at light speed and it is disrupting
[4:44] anything in its path that is not based
[4:46] on scarcity. So, I'm going to go remind
[4:48] you
[4:49] that the physical world upgrade, the
[4:51] paper I wrote at the beginning of the
[4:52] year,
[4:54] the first phase of AI investment 2022 to
[4:57] 2025 was narrowly concentrated in the
[4:59] digital brain builders, hyperscalers,
[5:02] GPUs, and core networking producing
[5:04] enormous capital intensity. I could have
[5:06] included SaaS in there,
[5:08] but limited economic diffusion. As a
[5:10] result, traditional manufacturing
[5:12] indicators like PMI remained subdued
[5:14] despite record AI CapEx.
[5:17] The timeline for this transition
[5:18] stretches not quarters or even years,
[5:19] but decades. The investments tied to
[5:21] this transition are not a passing trade.
[5:23] The road to artificial intelligence is
[5:24] expected to take at least
[5:26] 4 to 5 years and what follows is an era
[5:29] of embodied intelligence
[5:31] driven by robots and blah blah blah.
[5:33] Um not a speculative bubble.
[5:37] Positioning has been heavily skewed
[5:39] towards services and software for over a
[5:41] decade, but the AI narrative expands
[5:43] beyond code into physical
[5:45] infrastructure, commodities,
[5:46] industrials, and hardware. They are
[5:47] poised to become the structural winners
[5:49] of this new era. I'm just reminding you
[5:52] to go back and read the paper. For those
[5:54] of you just joined the paywall,
[5:56] even the old papers that I put up there,
[5:58] I would go read. They have shelf life.
[6:01] These are not things that I'm writing to
[6:03] say the S&P is here. These are all
[6:05] related to AI, and I'm telling you out
[6:08] of the many, many people I talk to on a
[6:10] weekly basis, which seems to grow, I
[6:12] still put it at about 20% I can have a
[6:15] conversation with at all regarding AI
[6:18] where they know enough, in my opinion,
[6:21] to be making decisions on what's going
[6:22] to happen over the course of the next 3
[6:24] years. It's not a slight against them.
[6:26] They don't have the time, or maybe they
[6:28] don't have the access. This is going at
[6:31] light speed. So, I I'll just leave it
[6:33] there. The regime shift from
[6:34] concentration to dispersion. So, number
[6:37] one, this is a regime shift. This is not
[6:39] a cycle. Number two, asset light is
[6:41] giving way to asset heavy.
[6:44] Buying the losers, meaning the SaaS
[6:46] names, that is not adaptation. You are
[6:48] basically trying to play the game of the
[6:50] prior decade believing that software is
[6:53] going to win in a world where you can
[6:55] create software in seconds. I I just
[6:58] don't understand how people want to have
[7:00] this argument. This is a debate that is
[7:02] futile. It is ridiculous. It's like
[7:04] debating religion. I think it is wasting
[7:07] people's time. Just get the hell out of
[7:09] the way of AI. If SaaS does have a
[7:12] bottom, it's not going to be all of
[7:13] them. If you can pick some, I wrote
[7:15] something positive on Palantir relative
[7:18] to Microsoft cuz I don't want to be
[7:19] involved in it. I have no interest
[7:22] whatsoever. Expect persistent internal
[7:25] volatility.
[7:26] No systemic risk, but structural
[7:28] messiness. This is the whole point. I
[7:30] think there could be multiple 10-plus
[7:32] percent corrections in the S&P um driven
[7:35] by hackings that are going to happen uh
[7:37] because of the agent swarms, driven by
[7:39] deleveraging as I've talked about as
[7:41] asset uh managers have to deal with
[7:43] leverage coming down because CoVaR stays
[7:45] at higher levels, I think it's going to
[7:47] be messy and I think you have to be
[7:48] ready for it. Um I wrote this paper uh a
[7:51] while ago. I want to go into this uh a
[7:53] little bit for Bitcoin at the beginning
[7:55] here because I want to remind people
[7:57] that I I'm
[7:58] again, I didn't expect SAS to go down
[8:00] this fast,
[8:02] but Bitcoin is related to SAS and I
[8:04] don't think it can break away until we
[8:06] get the Mag 7, the hyperscalers, and
[8:10] anything built on code to go down so
[8:12] that there's nothing that people can
[8:13] invest in with certainty. Bitcoin does
[8:16] not get disrupted by AI. AI agents help
[8:18] move the transition from the physical
[8:21] economy to the digital economy. So, I'm
[8:23] going to tell everyone out there that's
[8:24] looking for things to comb in SAS, this
[8:27] is where you go. When SAS finds a bottom
[8:29] eventually, it will, in my opinion, be
[8:32] Bitcoin that leads the way higher and
[8:34] when it's the only growth asset that's
[8:36] working in the forever period of AI
[8:39] disrupting our lives,
[8:41] you're going to jump on board. That's
[8:43] the thing I know that will happen. Uh
[8:45] it's just not happening now, but I would
[8:47] be spending your time on Bitcoin. I'm
[8:48] happy to go through that. Raoul Pal and
[8:50] Andreas Steno Larsen had a conversation.
[8:54] Um obviously, I know these guys. I I I
[8:57] was at an event with them at Raoul's
[8:58] event recently.
[9:00] Uh I have tremendous respect for for
[9:02] Raoul and and don't know Andreas well,
[9:04] but I've read a lot of his stuff. Uh and
[9:06] I think overall, we're on the same page
[9:09] except for something important. So, if
[9:11] you get a chance, go listen to this. I
[9:14] am a defl- deflationista, which I think
[9:17] Raoul is too, but this headline, "Why
[9:19] the Nasdaq Could Surge Again",
[9:21] I do not believe that. And this is
[9:23] really important in this side. I don't
[9:26] think when I hear what people are saying
[9:28] and what these guys said of where the
[9:29] threat is,
[9:31] yes, there's no bubble. I don't believe
[9:33] there's a bubble. Yes, you're spending
[9:35] money out of free cash flow
[9:38] for CapEx in the hope and the belief
[9:42] that you'll get the revenues. I do not
[9:44] believe the revenues will be there at
[9:46] the level that both of they think.
[9:49] Um and I think they need to think that
[9:51] because they've made this point in their
[9:52] head that you need liquidity and to some
[9:55] degree risk assets. And I don't want to
[9:57] speak for them,
[9:58] but I have a very different view. I
[10:00] think the Nasdaq and technology has to
[10:02] underperform for Bitcoin to do well. And
[10:05] I do not believe that people are
[10:07] spending enough time on the deflationary
[10:09] situation that is going on with the AI
[10:11] models, and I will get through that in
[10:13] this. It's very important to me to go
[10:15] through, but this is the way I've broken
[10:17] down what happens
[10:19] for Bitcoin. Let me just move this so
[10:22] that you can see all of them. Phase one,
[10:25] that's where we left. This is the
[10:26] acceleration in the liquidity offset.
[10:28] This is the Jeff Booth theme. This is
[10:30] the Lyn Alden and Sam Callahan theme, uh
[10:33] which is that Bitcoin trades like a risk
[10:35] asset. It is highly correlated to
[10:37] liquidity, which Raoul believes in as
[10:39] well. Um he also believes that there's a
[10:41] deflationary shock from exponential
[10:43] innovation. The difference is I don't
[10:46] think the government can now spend
[10:47] enough money to get us out of this. I
[10:49] don't think it's possible. I think this
[10:51] is incredibly deflationary when you get
[10:53] to the intelligence level and you have
[10:55] hyper competition. I have always
[10:57] believed the way that Bitcoin works is
[10:59] for the bottom end of the K of 8 billion
[11:02] people versus the 10,000 people with all
[11:04] the money in the world, there needs to
[11:06] be disruption to public equities. I've
[11:08] done this on this video. I've done it in
[11:10] writings. I've done it in interviews. I
[11:12] am a believer that all public companies
[11:14] filled with bloat, filled with software
[11:16] stacks that everyone is focused on is a
[11:19] game that doesn't exist 10 years from
[11:21] now, or at least if it does, it's in the
[11:22] last legs of it. And that during this
[11:25] time of the capital structure changing
[11:26] and everything else, this is a very
[11:28] dystopian period, which is why it's very
[11:30] dark. The scarcity repricing, which is
[11:33] gold, which is copper, which is all of
[11:36] these things that are scarce. That is
[11:38] where we're going to end up. But right
[11:40] now, we're in this one with people
[11:42] leaning towards the more obvious ones
[11:44] here and the ones that are driven by
[11:47] You guys don't want to see that. Um
[11:50] driven by the money going in spend by
[11:53] the hyperscalers. So, I believe that
[11:55] deflation is going to cost the
[11:57] hyperscalers. Now,
[11:59] to get your mindset in this, I'm going
[12:01] to again say, "So, I'm doing a video
[12:02] series on AI. I think one of the things
[12:06] you guys have to understand, the reason
[12:07] that I do the YouTubes is to hopefully
[12:09] give you some stuff that you can't get
[12:11] from the banks, that you can't get from
[12:12] X. Hopefully, I'm a stable-minded
[12:14] person, but I'm spending my time in AI
[12:17] all day long. I'm
[12:19] talking about what I'm doing, and these
[12:21] videos are meant to help not only you,
[12:23] your employees, but also
[12:26] your kids. I don't think unless you
[12:28] develop an AI mindset, you put in the
[12:30] reps like golf, you think in bets, and
[12:32] you build a relationship with AI that
[12:34] you can possibly understand how
[12:36] dystopian this is going to be. That is
[12:38] the reason why the research is on there,
[12:40] the videos are on there. Thank you for
[12:42] all of you who've joined on. I continue
[12:44] to believe if you go on, I'm going to
[12:45] help you from the investment side in
[12:46] terms of giving you ideas, but I'm also
[12:48] going to help you in terms of the
[12:49] regimes that we're going to remain in as
[12:51] AI dominates, and then hopefully give
[12:54] some things
[12:55] to where you can find some places to
[12:56] make money when we come out of this in
[12:58] the scarcity side. Bill Gurley had a
[13:00] great post that led to something
[13:01] important. Some shocked that Walmart and
[13:03] Costco have higher multiples than
[13:05] software companies. When so much value
[13:07] is in terminal value, the haunting
[13:08] question in will this company be around
[13:10] in 30 years? So, he's referencing
[13:13] Michael Mauboussin.
[13:16] So, I basically went to Mauboussin's
[13:18] paper that he showed in the links and
[13:21] basically brought up, "Give me a summary
[13:23] of this and connect it back to the
[13:24] uncertainty AI brings to long-duration
[13:26] assets and what should happen to
[13:27] valuations.
[13:29] And this is uh uh Michael Mauboussin
[13:30] back in 1997. A stock price is
[13:33] determined by three dynamic
[13:34] expectations, cash flow,
[13:37] risk, and time. If a company's cap
[13:39] remains consistent as years pass, it
[13:41] effectively creates bonus excess returns
[13:44] for shareholder.
[13:45] How does AI disrupt this?
[13:48] Mauboussin notes that cap is determined
[13:50] by the rate of industry change and the
[13:52] barriers to entry. AI acts as a
[13:55] double-edged sword.
[13:57] When uncertainty rises, volatility
[13:59] follows because the market quickly
[14:01] prices in worst-case scenarios. What
[14:03] should happen to valuations? Based on
[14:05] the cap framework, we should expect a
[14:07] bifurcation in valuations.
[14:10] Guys, that's what we're seeing with the
[14:11] dispersion. So, you can disagree with
[14:13] it, you can bet against it, but what
[14:15] this is saying is if you connect the
[14:17] fact that AI is disrupting whether a
[14:19] business will be in place 3 years from
[14:21] now. Forget 30 years ago or 30 years
[14:24] from now. For companies that are based
[14:26] on future cash flows, this is a really
[14:28] important part. I would go read the
[14:30] paper, go spend time on cap, and go
[14:31] through it. Now, I also talked about
[14:33] concentration. I just want to make sure
[14:35] people see this. This is the reason why
[14:37] I think people should be very worried
[14:39] and a deleveraging thing should be a
[14:41] major issue. All of these periods in
[14:43] time,
[14:44] the dot-com bubble is this. The one
[14:47] thing about both of them, they saw
[14:48] deleveraging and they saw a massive
[14:50] rotation.
[14:52] I think you have to be very careful when
[14:53] concentration is this big. I don't think
[14:56] this ends in a horrible scenario for the
[14:58] economy. I think it could if the
[15:01] deleveraging happens too fast,
[15:03] but I still think
[15:05] this is not a systemic issue, but as I
[15:07] go through this for credit and some of
[15:09] the other things that I mentioned last
[15:10] week, there is a risk of it if uh
[15:12] regulators and governments uh don't get
[15:14] involved if it goes too sloppy. I think
[15:17] if software can stabilize, which is not
[15:19] happened yet, uh it can. Uh this is uh
[15:22] another thing that went through and I
[15:24] think you should look at this. This fits
[15:26] in again and I want you to look at these
[15:28] periods. This is a new NS I'm sorry, a
[15:31] high low logic index which is something
[15:33] created in the 1970s
[15:35] which is basically taking the the ratio
[15:37] of these two to some degree and when it
[15:38] gets to very high levels
[15:41] you end up in a scenario. This is not
[15:42] just high low. This is high low relative
[15:45] to the amount of names making high low.
[15:47] You end up with a scenario that
[15:49] it means instability. And the max
[15:51] drawdown from top to bottom on this
[15:54] can be fairly extensive. The main point
[15:56] is I think dispersion creates
[15:58] instability. Instability eventually
[16:00] leads to a drawdown. I still think
[16:02] that's a likely scenario a couple times
[16:04] this year at a minimum. I think it is
[16:06] going to be very hard for stocks to go
[16:08] higher if we're seeing multiple
[16:10] compression in tech names which I do
[16:13] believe is going to be the case anything
[16:14] built on code.
[16:16] Crazy market stat. The S&P 500 is flat
[16:18] year to date but there are just 94
[16:19] stocks in the index up or down less than
[16:21] five
[16:22] percent. So, you've just been all over
[16:25] the place in terms of the the movements.
[16:28] Here's Scott Rubner
[16:30] put this out from Citadel. This is
[16:32] another way to go through it.
[16:35] Just a dispersion side and what I want
[16:37] you to do is just look at the time
[16:39] periods when we've had this kind of
[16:40] dispersion. We didn't even get this in
[16:42] 2020 and I think it's cuz it wasn't a
[16:44] rotation. It was a synchronized just
[16:46] unwind except for some of the software
[16:47] names. You really only have two times
[16:49] going back there in the last 30 years.
[16:51] One was during the Great Financial
[16:53] Crisis. The other one was during the dot
[16:54] com bubble and during LTCM. What I want
[16:57] to highlight is if you match up these
[16:59] periods in every case credit spreads
[17:02] went higher. That's the only thing we're
[17:04] missing at this point. You're going to
[17:05] see me focus a lot more on credit as
[17:07] part of the deleveraging cuz that's the
[17:09] thing I'm watching.
[17:10] You want convexity in your portfolio if
[17:13] you agree that COVAR is going to stay
[17:15] high which means you want vol uh VIX
[17:17] calls, you want um anything on the high
[17:20] yield in terms of CDX side. You want to
[17:21] find convex trades where you're making
[17:24] money not if the market falls, cuz I
[17:26] don't think the S&P can fall with strong
[17:28] earnings and GDP I with strong GDP and
[17:31] profit margins sitting up, but I think
[17:33] it can fall 15% in a week and then rally
[17:36] back over the course of the next month.
[17:38] Um AI disruptions in earnings. So, we're
[17:41] coming out of the earning season.
[17:43] This is not going away. AI will be a
[17:45] disruptive force. Um it's compressing
[17:47] timelines, changing the nature of work.
[17:50] Uh Opus 4.5's debut that changed the
[17:52] world.
[17:54] This is from Jeff Seabert and I'm just
[17:56] bringing this up again because
[17:58] he's literally talking about the fact
[18:00] that the world has changed forever. And
[18:02] so, if you take what I've written and
[18:04] you keep going through,
[18:06] um this is why this is happening now.
[18:08] This is why this started in October for
[18:11] the most part. Uh but then for SAS, I
[18:14] would say it really got bad post the
[18:16] November period. Here's the S&P. We are
[18:18] just going sideways.
[18:20] Uh it looks a little roundy to me. I
[18:22] think it needs to break out, but even if
[18:23] it does break out, I I think we're going
[18:25] to continue to see multiple compression
[18:28] uh in the better part. Now,
[18:30] for the month, again, tremendous
[18:33] dispersion. Industrials, materials,
[18:34] utilities, energy. See this?
[18:36] These are all the sectors I would say
[18:38] that are the data center buildout and
[18:40] the things necessary for the $650
[18:42] billion.
[18:43] What do you have down here? Tech,
[18:46] consumer discretionary with Amazon,
[18:47] Tesla, and this. So, you got the Mag 7
[18:49] down here.
[18:50] This is the bad one. This is the one
[18:52] that worries me. Financials down.
[18:54] Financials should be up here if this is
[18:56] a PMI related boom. What is going on in
[18:58] financials should be worrisome to
[19:00] people, particularly when you have
[19:01] utilities here. There is a te- 9%
[19:04] spread. Utilities are outperforming
[19:06] financials
[19:08] in in a month where or so far where you
[19:12] would expect uh it to be the other way
[19:14] when you see those other sectors up. I
[19:16] showed you before the software side
[19:18] we're seeing multiple compression in the
[19:19] S&P as we go sideways earnings are good.
[19:21] You are starting to see this compression
[19:24] go.
[19:25] IGV
[19:27] Here it [clears throat] is quarter to
[19:28] date now. It's down again this week.
[19:31] We are now at the worst quarter. There's
[19:32] only two two three quarters all together
[19:35] but two periods. It was worse during
[19:38] the great financial crisis and it was
[19:40] worse during 02 into 03. Now in both of
[19:43] these cases
[19:46] the market was down. Right now the S&P
[19:47] is flat for the year. So this is not a
[19:49] relative thing. This is absolute. So if
[19:51] you did an absolute we're obviously in
[19:53] worst case than that. Software just
[19:55] cannot catch a bid. It did not catch a
[19:56] bid this week and we had a new gutting
[19:59] on Friday. Cyber stocks slide as they
[20:01] So now now we're hurting something where
[20:04] the multiples are still very high
[20:05] because it seemed like AI and agent
[20:08] swarms would be the perfect place
[20:10] to have
[20:12] these names do well. Well, they were
[20:14] thrown under the bus.
[20:16] IT services this is the consultants.
[20:18] This is Gartner. This is Accenture.
[20:22] Massive fall so far this month down 17%
[20:26] only one month worse in the last 22
[20:29] years. So even the IT consultants are in
[20:32] there.
[20:33] Here's a 200 week moving average the
[20:34] yellow line of the S&P 1500 software
[20:39] index. I don't know why anyone wants to
[20:41] pick a bottom of this. This is not a
[20:42] friendly chart but I will tell you this.
[20:44] We have not closed below there except
[20:46] for
[20:48] one day here one week sorry one month or
[20:51] one week there one week there one week
[20:53] there and then we bounced immediately
[20:54] back up. This goes all the way back
[20:56] almost until the iPhone launch which is
[20:58] when this went up. 200 week moving
[21:00] average go back and look at it over
[21:01] time. It is a good judge for whether
[21:04] you're in a real true bull bear market
[21:06] or not. So software at best is going to
[21:09] be a value trap, but I think we are
[21:10] going to break.
[21:12] Another group within the financials got
[21:16] hurt again this week. Blue Owl blew up.
[21:19] Um, this is the this is my equal weight
[21:23] index of the private equity firms
[21:25] including Apollo and Blackstone.
[21:28] Uh, Ares you can just see. I mean it's
[21:31] been down six weeks in a row right now
[21:34] as we are spreading into private credit
[21:36] and private equity.
[21:38] I mentioned utilities and financials.
[21:40] This is a chart of utilities
[21:42] outperforming financials for the quarter
[21:45] and you can see aside from these
[21:47] horrible periods when almost every time
[21:50] we've seen a number this big, you have
[21:52] credit spreads, junk spreads. So, this
[21:54] is junk spreads widening out. Uh, right
[21:57] now they're tight. Again, I think if you
[21:59] want convexity with what's happening, I
[22:01] think you want that stuff. Um, a place
[22:04] again since everyone is trying to pick
[22:06] the top top in semis and trying to pick
[22:09] the bottom in IGV, I will reiterate this
[22:12] package thing that I created which is an
[22:14] equal weight of the names that benefit
[22:16] from the move to edge devices that are
[22:19] more sensitive. A lot of analog names.
[22:22] Um, if you want the names
[22:24] for the people on the paywall, you have
[22:25] access to them.
[22:27] Uh, for the clients they've already
[22:28] gotten this, but um, this looks like
[22:31] it's going to break out and here is the
[22:33] report that I wrote on it. At the end of
[22:35] the year, I still think these are the
[22:37] names that you want to be involved in.
[22:40] Um, one data point that matters. This is
[22:41] the durable goods. I've shown this
[22:43] before. It ended up being right. I
[22:45] showed you guys this all starting from
[22:48] the summer time that the capital goods
[22:50] were already breaking higher. This is
[22:52] what measures the data center
[22:54] expenditures. So, we got our PMI lift up
[22:57] catch playing catch up. We are headed
[22:59] towards 60 guys. It is going to happen.
[23:03] 30-year JGB yields, remember those?
[23:05] Panicked move. Um
[23:07] largest monthly decline now in 30-year
[23:09] JGB yields since 2016.
[23:12] We are headed for a deflationary
[23:14] spiral, guys.
[23:16] Uh I do not know how this guy continues
[23:19] to
[23:19] to put stuff in X. Um
[23:22] I don't believe in software narrative
[23:24] and I think there will be a bounce. High
[23:26] gross leverage in semis versus software
[23:28] will be reduced soon. That's all I need
[23:30] to know to be long semis and short
[23:32] software still.
[23:34] Um you're still going back. Everyone
[23:36] looks for a bullish argument. Here's one
[23:38] that came out on software. Is it
[23:40] completely dead? At least it's logical.
[23:42] They go through the things that have
[23:44] killed it. There's 10 moats. These are
[23:46] the ones that are there. Blah blah blah.
[23:48] don't care. You guys can go read it on
[23:50] your own. I bring that to you so you can
[23:51] look. Um Nick Evans, who I've met, um
[23:55] fun beating 99% of its peers. He sees
[23:57] few software firms surviving AI.
[24:00] I'm with Nick on this. I see few
[24:02] surviving in the long term for the next
[24:03] 3 years. Yes, uh but that has more to do
[24:06] with the um the addiction that the
[24:09] enterprises have with this stuff.
[24:10] Gradually as the agent swarms destroy
[24:12] everything or replace everything, uh you
[24:15] will be dealing with uh companies that
[24:17] will have access to a billion Einsteins
[24:19] to do their work and they won't need to
[24:20] have this stuff that uh people call
[24:22] software. Um Sol Tan, who I know very
[24:25] well, wrote this piece saying software
[24:27] will be free. Again, I'm giving you both
[24:28] sides of the debate. Um Dennis and the
[24:31] team put out a great thing here. If if
[24:34] you're still buying software, um I
[24:36] wouldn't. Um if you want something
[24:38] quantitatively to show it and you guys
[24:40] should be getting these guys
[24:42] are all over the AI side from a data
[24:44] perspective. Uh
[24:46] software has seen the largest decline in
[24:48] positive EPS revisions.
[24:51] That's a bad sign.
[24:53] Uh again, back to the Jeff Seibert in
[24:55] terms of what's going on, the world
[24:56] changed. Fiverr, um um
[25:00] this is a note they put out in
[25:02] September. I remember it. If you guys
[25:03] haven't used Fiverr, it's an outsourced
[25:06] place. They have employees in Pakistan
[25:08] and India and at various places around
[25:10] the world where you can basically pay
[25:13] them local money
[25:15] to do work for you tech-wise. And so,
[25:18] revealed the restructure allowed the
[25:19] company to go AI first. So, here's a
[25:21] company that I fully expected was being
[25:23] quick to this. They'd be fine. And this
[25:26] week their stock price collapsed as the
[25:28] freelancer platform tries to put a
[25:30] positive spin on AI disruption. Nobody
[25:33] can hide even if they're admitting it
[25:35] and trying to get ahead of it. The
[25:37] uncertainty of long duration assets,
[25:39] again, this gets into the Mobius and
[25:41] thing and I just want to bring this up
[25:43] as we go through it because
[25:45] the valuations fit in with it as to
[25:47] what's happening. Private software
[25:48] companies release earnings early to calm
[25:51] AI fears.
[25:52] You didn't calm them. This was earlier
[25:53] in the week, but at least they're trying
[25:55] everything. Carlyle and Blackrock,
[25:58] private They tried to do stuff, too.
[25:59] They're buying software loans to boost
[26:01] CLO profits.
[26:03] Not going to work.
[26:05] This was story about Evercore saying
[26:07] that so far there have been exactly
[26:09] three insider transactions in large cap
[26:12] software. If they're not buying, why
[26:13] should anyone else buy? The SAS
[26:15] apocalypse is a credit event, not a tech
[26:17] story.
[26:19] This is where everyone should be paying
[26:21] more attention
[26:23] cuz it has destroyed a lot of equity,
[26:25] but it is also testing the debt world.
[26:29] And this is where the Blue Owl
[26:30] permanently halts redemptions. They came
[26:33] out and said they're not permanently
[26:34] halting anything, but whatever the case
[26:36] is,
[26:37] there's a whole big story here that
[26:39] continues to happen. I think this has to
[26:41] do more with we're getting credit
[26:44] problems that are just spreading. They
[26:45] didn't just start here. Here's the
[26:47] private equity companies. This fall here
[26:50] was related to the auto subprime auto
[26:52] lending, which was Tricolor.
[26:54] Um
[26:55] this has been a problem. So, private
[26:57] equity, private care credit, VC, as I
[27:00] did with the Mobius and things, I am
[27:02] incredibly negative on all long duration
[27:04] assets. If it's illiquid, I'm even more
[27:07] negative. These companies are trapped in
[27:09] things that are longer that take longer
[27:12] to get through, and they've got money
[27:13] trapped up. I think they should be
[27:15] trading at a discount, not necessarily
[27:17] the companies, but at least the way
[27:19] people view them for endowments, for
[27:21] foundations, for pension funds, for
[27:24] sovereign wealth funds, anyone that has
[27:25] a lot of long duration assets, the
[27:28] hyper-competitiveness, regardless of
[27:30] whether it's VC or anything else, I just
[27:33] don't know how you can value things when
[27:35] you're talking about humanoids in an
[27:36] elevator 5 years from now, when you're
[27:38] talking about 3 years from now having
[27:40] certainly recursive learning and AGI,
[27:42] we're already at a point where it's
[27:44] moving too fast. So,
[27:46] um private credit should worry about a
[27:47] singularity in software debt.
[27:51] If losses spike too quickly, they could
[27:53] lead to a widespread crunch. In a tail
[27:55] outcome, the knock-on effects for all
[27:56] issuers attempting to access credit,
[27:59] including hyperscalers,
[28:01] could be significant, potentially
[28:02] undercutting capital spending investment
[28:04] plans. This is where the negative thing
[28:06] that I think will be a story this year.
[28:08] I would be very surprised if at some
[28:10] point we don't have credit fears that
[28:13] rise to the level that the hyperscalers,
[28:16] quote-unquote, would have to cancel
[28:17] CapEx, or this would be an issue. So, if
[28:19] you think about what's happening with
[28:20] Blue Owl, with uh their fact, and again,
[28:24] it's just a fear side of AI. But, the
[28:27] reality is it's because you have no
[28:29] certainty in the future. This is not a
[28:31] linear economy anymore. This is not a
[28:34] one with cycles. This is something where
[28:36] the uncertainty will be here forever. It
[28:39] will never get better, which is why I
[28:40] show it as a supersonic tsunami. The
[28:43] speed is going,
[28:44] the competition is coming every day from
[28:46] entrepreneurs that are eating away at
[28:48] revenues, they're eating away at growth
[28:50] rates, and it's bespoke. You customize
[28:52] things for what you want, which is
[28:54] really the way software should be.
[28:56] Here is a chart. This is the tech sector
[29:01] OAS, the option adjusted spread
[29:04] in the bond market, the high yield side.
[29:06] So, this chart goes back 14 years, and
[29:10] it's overlaid with CDX for high yield.
[29:14] So,
[29:15] CDX is still sitting down here, guys. Um
[29:18] the tech side is up here, and the reason
[29:19] CDX is down here is cuz the other
[29:21] industries are not growing.
[29:23] I think tech matters, and I think
[29:24] contagion is an issue. Now, here is that
[29:28] same chart in terms of the tech side.
[29:29] Here is again, you're starting to get
[29:32] the OAS on the overall side working out.
[29:35] It was still at lows here, but this next
[29:38] leg in 2026 has got it going.
[29:41] And again, this is junk spreads down
[29:43] here. You can already do them the
[29:45] correlations in your head. This is what
[29:47] I did at Morgan Stanley. I traded
[29:49] emerging markets throughout the '90s. I
[29:50] would look at these, and I believe once
[29:52] credit starts to go, you end up in an
[29:54] issue very quickly because it's the fear
[29:56] of getting paid out. Now,
[29:58] these charts matter all of a sudden
[30:00] because we didn't have anything
[30:01] happening at all in high yield, but now
[30:03] we have something on a big sector.
[30:06] We've got credit card delinquencies up
[30:09] here.
[30:10] We've got auto loan delinquencies, the
[30:12] subprime side. Remember, this is what
[30:14] got us into trouble with Tricolor and
[30:18] everything else, Carvana and everything
[30:19] else. Uh student loan delinquencies.
[30:24] Walmart had their earnings. They cited a
[30:27] hiring recession. How are things going
[30:30] to get better if AI is now going to
[30:31] disrupt the job market more than it
[30:33] already have? Several indicators are
[30:35] showing warning signs, including a rise
[30:37] in US consumer debt in delinquency.
[30:40] Americans are growing more concerned
[30:41] about their employment prospects.
[30:44] Unprecedented jobless boom. And again,
[30:47] we had 130,000 job creation last month
[30:50] and 137 positive health care jobs. So,
[30:54] again, that's not really helping disrupt
[30:57] or argue against the AI side. Uh CMBS,
[31:00] remember uh Silicon Valley Bank when
[31:02] this whole kicked in right here? This
[31:04] isn't fixed yet. Um
[31:07] all-time high. Is AI going to help this
[31:09] or hurt it? I see things being printed
[31:11] about how AI is going to create more
[31:13] jobs and people are going to go into
[31:14] offices.
[31:16] God.
[31:17] Good luck with that one. Lenders are not
[31:19] paying.
[31:21] Or lenders to the owners,
[31:24] they're not paying. Um
[31:26] money managers are now worried that
[31:27] companies are over investing. This is
[31:29] where you start to get into the hyper
[31:30] scalers and I just want to bring this
[31:31] up.
[31:32] And this these next three should at
[31:36] least remind you of why this will be a
[31:39] story.
[31:40] Dario Amodei perfectly explains why
[31:41] Oracle and Open AI are in a very risky
[31:43] position. Oracle and Open AI. He says
[31:46] that you can commit to buying 5 trillion
[31:47] worth of compute if you're projecting 1
[31:50] trillion in revenue rate. If you
[31:52] retrieve 800 billion, it's still going
[31:54] to it'll still be a staggering number,
[31:55] but the company will go bankrupt. This
[31:57] is the risk that they are facing and
[32:00] Oracle's involved in. Now,
[32:03] I want to bring this up because
[32:04] Anthropic is winning.
[32:06] And on January 22nd,
[32:09] this was the news story. Anthropic's
[32:11] gross margin guidance
[32:13] as revenue surging fell
[32:19] by a 23% spike in inference costs.
[32:22] These figures hold significant
[32:24] implications for investors. I just want
[32:27] to remind you Anthropic gross margins
[32:30] are now being forecast down to 40%. This
[32:33] is the winner.
[32:35] I'm only bringing this up because the
[32:37] cost of the compute is going higher,
[32:39] which makes the cloud cost up there.
[32:42] We've already got delays on the cloud
[32:43] because the data centers are being
[32:45] delayed through bottlenecks and through
[32:47] the government, which I'll show more of.
[32:49] The spending is happening.
[32:51] And here's the issue.
[32:54] China is competing with models that are
[32:58] anywhere from 80 to 90% cheaper. We are
[33:01] in a deflationary spiral. So, when you
[33:03] spend money on CapEx
[33:06] for out of free cash flow or using debt
[33:09] as OpenAI is or borrowing go out and
[33:11] taking tons of money,
[33:13] it's about the revenues. And there might
[33:14] be orders, but if you can't get the
[33:16] orders or if they're delayed, that's a
[33:18] problem. But if the cost of your compute
[33:21] can't keep up with the inference cost or
[33:23] the build-out cost, then you're really
[33:24] in trouble.
[33:26] You got to do a hell of a lot more
[33:27] volume, and that's really hard to get
[33:29] when there isn't enough token usage. So,
[33:31] you got to charge more money, and if you
[33:33] charge too much money, there's people
[33:35] offering stuff already at a 90%
[33:37] discount. I don't know how you get out
[33:39] of this. Oracle,
[33:41] the
[33:42] lawsuit,
[33:43] defendants failed to disclose to
[33:44] investors that one, Oracle's AI
[33:48] infrastructure strategy would result in
[33:50] massive increase in CapEx without
[33:52] equivalent near-term growth in revenue.
[33:54] There's a lawsuit going on specifically
[33:56] on this. Oracle stock was down big on
[33:59] Friday, and remember, this is their CDS.
[34:04] Here's the chart of the hyperscalers
[34:06] relative to the S&P overlaid with the
[34:09] private equity companies.
[34:14] All right.
[34:15] Data center delays, 125% surge in data
[34:18] center opposition.
[34:20] Moratorium nation, getting
[34:22] cancellations.
[34:25] Hey, I
[34:26] even in XAI's Colossus, you can go
[34:28] through and see a notice of intent to
[34:30] sue for violations and folding
[34:32] there on
[34:34] costs, on pollution, on everything. And
[34:38] what happens if we already have the
[34:40] Democrats at 83% for the house?
[34:44] Number is falling on for the Senate. And
[34:48] what happens if the
[34:51] AI
[34:52] you can't delay anything.
[34:54] Again, you have orders. You're not
[34:56] allowed to delay things. Altman calls
[34:58] China's AI progress remarkable this
[35:00] week.
[35:03] How China caught up on AI and may now
[35:06] win the future. This should be scaring
[35:09] the hell out of everyone. And again, I
[35:11] realize
[35:12] enterprises are not going to use Chinese
[35:15] open source and I Dario Amodei said
[35:18] that. But
[35:19] if the models are as good and they're
[35:22] 90% cheaper
[35:24] then every entrepreneur is going to use
[35:26] them. And this is where the K-shaped
[35:29] economy and the end of public companies
[35:31] gets to. And again, I'm going to keep
[35:33] saying it. AI is the most disruptive
[35:35] force that has ever hit. It is
[35:37] deflationary and the Chinese are
[35:39] offering models and they are racing to
[35:42] keep up. Deep Seeking Quan just captured
[35:45] 15% of the global AI I market in a
[35:47] single year. They went from 1% to 15.
[35:52] You can go read an MIT article. In fact,
[35:54] I highly recommend if you want to go
[35:55] read and go through it. I'll read you
[35:57] some of it. Found that Chinese open
[35:59] source models have surpassed US models
[36:01] in total downloads. But these models
[36:02] differ in a crucial way from most US
[36:04] models.
[36:05] You actually get the models weights,
[36:07] numerical values that get set when a
[36:09] model is trained. So anyone can
[36:11] download, run, and study, and modify
[36:12] them. And
[36:15] they keep getting better. They will not
[36:17] just offer cheaper models who want
[36:19] access to them. They will change where
[36:20] innovation happens and who sets the
[36:22] standards.
[36:24] They will become infrastructure for
[36:26] global AI builders. The adoption of
[36:28] Chinese models is picking up in Silicon
[36:29] Valley, too.
[36:31] A general partner in Andreessen Horowitz
[36:33] has put a number on it. Among startups
[36:35] pitching with open source stacks,
[36:36] there's about an 80% chance they're
[36:38] running on Chinese open source models.
[36:40] These are the enterprises,
[36:41] theoretically, in the future.
[36:43] They're not going to use salesforce.com.
[36:46] They are not going to use this stuff.
[36:48] They will be AI native building their
[36:49] own stuff. They're not even looking to
[36:51] use US frontier models.
[36:54] The demand is also rising globally.
[36:56] Limited new subscriptions to its coding
[36:59] plan for a new company or new ZAI
[37:02] reports that the system's user base is
[37:04] primarily concentrated in the US and
[37:06] China, followed by India, Japan, Brazil,
[37:08] and the UK.
[37:10] Open Claw. And this is an important
[37:12] thing. If you've heard Open Claw, but
[37:13] you don't know what it is, this is the
[37:14] beginning of the agents. When the agents
[37:16] are running, as someone who has a Mac
[37:18] mini, your costs go up if you're using
[37:20] US models. If all of a sudden now it's
[37:22] costing you $500 a day to run on a bunch
[37:25] of AI agents to do stuff overnight for
[37:27] you, and you can do it for $50 with a
[37:29] Chinese model, and it's protected where
[37:31] you don't have to deal with the
[37:32] security, this is why Open Claw
[37:35] accelerates AI usage for China, and it
[37:39] will hurt the inference ability of
[37:42] people in the US offering it. It's going
[37:46] to happen. It already has happened. I'm
[37:49] using an open source Chinese model, and
[37:51] it's this one. Kimi K2.5
[37:54] had surpassed Claude as it became the
[37:55] most used AI model by token count,
[37:57] meaning meaning it was handling more
[37:59] total text process
[38:01] across user prompts and model responses.
[38:03] China is catching up. You can now deploy
[38:06] Open Claw in seconds on the cloud. You
[38:08] don't even need a Mac mini. They're
[38:09] making this easy. Chinese They just
[38:12] released a desktop automation that agent
[38:14] that runs 100% locally.
[38:17] Record low hallucination for the GLM 5.
[38:22] Genie 3 released a world model released
[38:24] by Google. And then immediately there
[38:26] was a Chinese open source competitor.
[38:29] DeepSeek 4, the rumors have surfaced and
[38:31] the numbers are terrifying. The best
[38:33] model scored 83.7 program test blah blah
[38:35] blah, making it high in this, assuming
[38:38] it's true. We don't know.
[38:40] But it'll be released soon. Why are
[38:41] Chinese AI models dominating open
[38:43] source? This gets through it, but
[38:46] they've captured the space. Research
[38:47] published by Sentinel 1 and Cenith
[38:49] mapping 175 exposed AI hosts across 130
[38:52] countries over almost a year. Alibaba
[38:55] Qwen consistently ranking second only to
[38:58] Meta's Llama in global deployment. More
[39:00] importantly, it's running multiple AI
[39:02] models again becoming the de facto
[39:04] alternative to Llama.
[39:07] Uh token share by region, again you can
[39:10] see how much the red side here, the
[39:12] Chinese continue to move higher, again
[39:15] by region.
[39:16] Uh
[39:17] dominating on open router.
[39:20] You can get another chart here, but more
[39:22] importantly from Shanghai macro
[39:23] strategist,
[39:24] uh
[39:25] top eight AI models now have a combined
[39:27] 47 market share. Isn't this the biggest
[39:29] nightmare for American AI firms that
[39:30] have spent 10 times more than their
[39:32] Chinese counterparts? That is the other
[39:34] thing about this is
[39:36] we're assuming that the CAPEX being
[39:38] spent in the US makes sense.
[39:40] It costs they're doing spending less
[39:42] there, which means they're being more
[39:43] efficient. Uh
[39:45] again, it's just don't even fight the
[39:47] fact that this is happening. It's very
[39:50] hard to measure, but this is where I
[39:51] differ from people that believe that A,
[39:53] the US models are going to win
[39:55] uh in such a big way like they did the
[39:57] mag seven did with everything else. And
[39:59] secondly, the Chinese have always been
[40:02] deflationary for things and I think that
[40:04] is a very scary thing when you're
[40:05] thinking that these models are going to
[40:07] be used in a way that you can monetize
[40:10] when deflation is going down.
[40:13] Netflix threatens ByteDance. If you
[40:15] haven't seen this week in this C dance,
[40:19] so I uh bite dance.
[40:21] These are AI generated movies, videos.
[40:24] If you haven't seen them yet, it is
[40:26] staggering. If you haven't seen this
[40:27] one, go watch it.
[40:29] Uh
[40:30] >> [clears throat]
[40:30] >> It's insane how good everything is.
[40:34] Hollywood AI nightmare, the door
[40:35] brothers. This one I watched,
[40:37] unbelievable. Um artificial intelligence
[40:39] has stormed into Hollywood with
[40:41] breathtaking speed and alarming
[40:43] consequences.
[40:45] Um
[40:46] It just is. Here's Netflix stock. If you
[40:49] don't think it's getting impacted, um it
[40:52] is. Everything is being disrupted by
[40:54] artificial intelligence, except for
[40:56] Apple cuz they haven't spent any money
[40:57] on it. Here's Apple spend. So, maybe
[41:00] Apple's going to be fine. I think it'll
[41:01] be interesting to watch. Uh they've
[41:03] avoided this thing. Um we'll see. I can
[41:06] see the argument even though Siri hasn't
[41:08] worked. Uh AI's compressing timelines,
[41:11] changing nature of work. Another article
[41:12] that came out. This is the Jeff Siebert
[41:14] one. If you guys want to go read up on
[41:16] it since I mentioned it in here. If you
[41:18] didn't see this, Open AI hires the Open
[41:21] Cloth founder, Peter Steinberger. Um
[41:25] So, Claude this was created out of kind
[41:27] of Claude code. Now it's uh been he's
[41:30] been hired by Open AI.
[41:32] It's latest hire changes everything
[41:34] about AI's next move. We are in the
[41:37] agentic world, guys. Uh so, get ready
[41:39] for billions of Einsteins. And of
[41:42] course, if you haven't used Grok, now
[41:44] he's got the names poking up on the
[41:46] agent. So, we actually have
[41:50] the agents working. They were watching
[41:52] working on my chemical uh thing for me,
[41:54] which will come out next week. I got too
[41:56] busy. I put two papers out this week. Uh
[41:59] one of them uh
[42:01] you definitely should look into uh in
[42:03] terms of the uh the the
[42:06] nan side. Uh I'll get into that in a
[42:08] second. Google releases Gemini 3.1.
[42:12] Grok 4.20 came out. So, the model
[42:15] leapfrogging continues. Jensen Huang
[42:18] teases upcoming surprise chip. A chip
[42:21] that will surprise the world will be
[42:23] unveiled next month. I still believe
[42:25] Nvidia is the key to the market as long
[42:28] as their stock holds up, and you'll get
[42:30] earnings from them next week.
[42:32] As long as their their stock holds up, I
[42:34] think the S&P can survive until it turns
[42:38] until credit starts to make things ugly.
[42:40] If Nvidia for any reason
[42:43] uh
[42:44] it doesn't matter what the earnings are.
[42:45] If they sell off, then I think you'll
[42:47] probably get a get a correction at that
[42:49] point. Um Nvidia and Meta did a deal
[42:53] again for a lot of chips.
[42:55] Uh rapid AI demand for memory is fueling
[42:57] a growing chip crisis. Everyone wants to
[43:00] short memory. They look at the charts
[43:02] and say it's got to be in there, which
[43:03] is why I wrote this.
[43:05] Uh for those of you who haven't signed
[43:07] up, this is just again, if you want
[43:09] ideas, there's a bunch of stocks in
[43:11] here, some that I love uh in particular
[43:14] that have gone sideways for
[43:16] really the last couple years.
[43:18] Uh
[43:19] we're about to hit a flash wall. So,
[43:20] this goes through, and this was
[43:23] after an interview with the CEO of
[43:25] Phison Electronics out of uh Taiwan.
[43:29] Great interview. Had it translated, went
[43:31] through it. I think there's a tremendous
[43:32] amount of ideas in it. Um in particular,
[43:35] the demand drivers as to what's going
[43:37] on.
[43:38] This is just going to continue. So,
[43:40] you've got the US cloud, the build-out.
[43:42] You the China cloud build-out. But then
[43:44] you're getting into the edge device that
[43:45] is starting in the second half of this
[43:47] year, the consumer roll-out. And then
[43:49] you get into the institutional
[43:51] on-premise, education, 2026. There are
[43:54] so many levels to this memory side, and
[43:58] there just isn't enough. Um JP Morgan
[44:00] launched its Halo acronym
[44:03] for heavy assets, low obsolescence. Um
[44:06] again, I think this fits in with my
[44:08] abundance scarcity thing, but everyone's
[44:11] got to come up with their own ideas.
[44:12] Stan has jumped on board the EWZ trade.
[44:15] One of his largest new positions.
[44:19] All right, this is where I want to bring
[44:20] it up. So, this is Bobby Maguire, for
[44:22] those of you who don't know, one of the
[44:23] traders at 22V, and this is him
[44:27] with his Mac Mini
[44:29] Open Claw Agent setup. I asked him to
[44:32] send me a photo because
[44:34] you guys should be doing this. Bobby has
[44:37] embraced everything with AI. He's run
[44:40] with it. We started talking about it I
[44:42] guess a year ago. He mentioned and I
[44:43] noticed that every week that he came in,
[44:45] he was more amazed by it. So, when I'm
[44:47] doing these videos and I'm saying you
[44:49] have to become AI native, you have to
[44:51] think and have an AI mindset,
[44:54] here's Bob Maguire,
[44:56] here's me.
[44:57] We're not two young guys. We did it. We
[44:59] have no coding skills.
[45:01] If you guys know Bobby, call him up,
[45:03] talk about his experience. For you guys
[45:05] in Boston, I'll be in Boston this week
[45:07] meeting some of you guys on the asset
[45:09] manager side.
[45:11] Mike Walsh, who's hosting or at least
[45:14] putting the thing together, he reached
[45:16] out to me after a call and said, "Hey,
[45:18] is it possible that I could build an
[45:20] app?" And I said, "You can build an app.
[45:22] You just need to go onto Cloud Code, and
[45:25] I think you can do it with just these
[45:27] simple instructions." So,
[45:29] over the weekend,
[45:31] he ended up building
[45:33] his own
[45:34] app on the website. He did this,
[45:37] lifestyle change one. Give kudos to Mike
[45:40] Walsh. I bring this up because the
[45:41] leaders of companies, if you're
[45:42] watching,
[45:43] and I've met some of you over the course
[45:45] of the last year and you've had me come
[45:46] in from a consulting basis,
[45:48] the leadership of the companies has to
[45:50] do this. If they do it, then everyone
[45:52] else will follow.
[45:54] I think these guys have done an awesome
[45:55] job.
[45:56] I'm going to keep promoting this site.
[45:59] I'm trying to help people with it. The
[46:01] videos are going to be there to help you
[46:02] develop an AI mindset. The investment
[46:05] side is going to help you hopefully
[46:06] avoid some things, think about things,
[46:08] proactively give you risk management
[46:10] side.
[46:11] Uh please subscribe. Please send it to
[46:13] your family and friends, and let me help
[46:15] as many people as possible. Uh and I'll
[46:17] see you guys next week.