Desaceleración de mitad de ciclo en IA — Jordi Visser (llamada exclusiva de junio 2026)

[0:20] All right, for those who are already on, we're going to wait another minute until everyone's
[0:51] let in.
[0:52] We just open everything.
[1:43] There's still people coming in, but we're at a quorum.
[1:47] So all right, what I want to do today is really go through a whole bunch of slides that kind
[1:56] of take you through what I've been writing about for the past, it's been about three
[2:00] or four weeks that I've been talking about this, but I'm now calling this, and I think
[2:05] this is the best macro analogy to use, which is navigating the AI mid-cycle slowdown.
[2:13] And we're just at a different stage right now.
[2:15] So we came out of the first quarter.
[2:18] That was catch-up quarter.
[2:20] And as I've been writing about, and especially with two weeks ago's piece titled Fireworks,
[2:25] I think the fireworks show is over.
[2:27] I really do believe that.
[2:29] I believe now investors and especially retail
[2:35] on the leaders of this, but then the laggards,
[2:38] which would be more the investment professionals.
[2:40] I think now everyone is fully on board.
[2:42] And like I said, I can usually tell
[2:46] from the pushback I'm getting on some of the things
[2:48] I'm writing and some of the conversations.
[2:50] And it's impossible right now for people to envision
[2:53] that the CAPEX numbers and the token usage
[2:57] and all of this stuff isn't gonna happen.
[2:59] So we've got into the linear math game.
[3:01] And I think we're, if you take the mid-cycle slowdown
[3:05] and you go back and think about what that is
[3:08] from a macro perspective,
[3:09] it's really a second derivative change.
[3:11] We're just past the major point of the growth rate.
[3:15] And usually what happens after you've gone
[3:17] through that first pace is the bar gets higher
[3:21] and at the same time you start hitting all of the headwinds.
[3:24] And that's what I'm gonna go through
[3:25] is just highlight the headwinds
[3:26] and also highlight the internals of the market.
[3:29] So every paper I've written,
[3:32] this is three weeks ago on running in the scarcity.
[3:35] This is the Firewix show last week,
[3:38] running in the hot scarcity.
[3:40] These are the videos on this.
[3:42] This one in particular, the bandwidth wars.
[3:44] I really spent a lot of time on Fujikura,
[3:46] which was on the optical side yesterday.
[3:48] We had fears on the optical side.
[3:50] So all of these were really more that
[3:53] I think we're running into a different risk factor.
[3:55] Here's the chart of the optical index I have
[3:59] versus Fuji Cora before Fuji Cora came out
[4:02] and said they were having bottleneck problems
[4:04] and the stock has traded off
[4:05] and continued to trade off now trading at the lows.
[4:09] And the optical stuff yesterday was hit
[4:11] on the back of a report from Semi-analysis.
[4:16] So on the one side, you've got the Semi-analysis analysts,
[4:20] I like them a lot, they're walking around Computex
[4:23] and they basically come to the decision
[4:25] their channel checks are saying that things are going to be pushed out. It may not be 2027,
[4:30] blah, blah, blah, maybe 2028. Then you get Morgan Stanley, which comes on the other side
[4:34] and disagrees and says their supply checks contrary to what's been reported. Even Nvidia
[4:40] denies the reports. The main point of this, and again, here's the Fuji core chart,
[4:47] we're just at a point where there's headwinds in the space. The optical side is going to be built
[4:52] out neither one of those pieces changes that the only thing it says is they're
[4:56] arguing about when it's actually gonna hit and this is after Jensen Yuan just
[5:01] spoke and kind of talked it up and Nvidia talked it up so if you take Fujikura
[5:06] and now you you don't you know it's not part of the basket but you take all of
[5:11] the Japan names that I have which is the white line equal weight the whole
[5:16] market over there has been weakening and that is the thematic portfolio over the
[5:19] top of it. Now the Asian stuff including Korea, I've shown the machinery side, I've shown the
[5:25] construction side. All of the bottlenecks are starting to show up in things and investors are
[5:30] now long and the correlation amongst all of these things is extremely high. So when you get the
[5:36] machinery and construction sector breaking down in Korea, I've shown, I showed this before, this
[5:42] This is it now updated and you've got over here SK Hynex which hasn't fallen yet, but
[5:49] again the reason I started to get a little bit more negative on memory is I think it's
[5:53] a little bit more vulnerable to this side.
[5:57] That's where a lot of the pushback has been from the buy side at this point in terms of
[6:02] what they think.
[6:03] So the navigation part, I update all, I put all these tools into the paywall every week
[6:11] for a reason, which you can go back and then look at the deterioration that's happening
[6:15] from a technical basis underneath.
[6:18] It's not just to look at and say, what's trending well?
[6:21] It's really to see, how's the deterioration happening?
[6:24] And basically, since that first paper, here are the percentage of the names that are above
[6:29] the 20-day moving average each week.
[6:32] So I uploaded it, and as of last week, it had fallen from 59 to 39, the 50-day.
[6:39] you can see that gradually we're breaking. So if your long names that are sitting at
[6:43] all-time highs or near all-time highs while this is going on, it's not just Korea, it's
[6:47] not just Japan, it is resetting. And to me, you know, this is the reason that you should
[6:55] be using this stuff. The update as of today, I did it for yesterday, I did it for today.
[7:01] The 20-day as of now has stabilized, we're below 50%, but the 50-day keeps deteriorating
[7:07] and even the 200 day we're losing names kind of one by one.
[7:10] So there's no doubt at this point
[7:12] because of how highly correlated these names are
[7:16] that you're starting to see more weakness.
[7:18] Now, I wanted to, I uploaded this for the subscribers today.
[7:24] So as long as you're in, you know, you can go check
[7:29] and see if this is something you have access to.
[7:31] A lot of people have asked me about my knowledge brains
[7:34] that I've talked about on some of the videos.
[7:36] So I wanted to give you guys a simplified way
[7:39] of doing it yourself.
[7:41] This is a way to do it using just co-work
[7:46] and creating a Python script.
[7:49] It'll create it for you.
[7:50] And then all you have to do is run it.
[7:54] You'd be shocked at how easy it is to do.
[7:56] It took me less than 30 minutes.
[7:58] This is it's time, but 30 to 45 minutes
[8:02] and you can download every one of
[8:03] Jensen, you won speeches this year, or interviews,
[8:06] starting from CES all the way down.
[8:08] You can do it for whatever you want.
[8:10] You can pick a different one.
[8:12] You just copy and take this word doc off the subscriber
[8:17] paywall, upload it into an LLM,
[8:19] and it'll take you through how to do it.
[8:21] And this way you can go in,
[8:23] and this is what I did with basically everything
[8:25] that's been going on.
[8:26] Go through Jensen's comments on Vera Rubin
[8:28] and the importance of opticals.
[8:30] Give me a summary on his thoughts on opticals
[8:31] through a progression lens
[8:33] from the beginning of the year until today.
[8:35] And the reason I wanted to do that is basically
[8:37] you've got these people debating
[8:38] what's happening in opticals
[8:40] and they're saying this won't happen, this won't happen.
[8:43] Let's go back and see what Jensen Yuan has said,
[8:45] but let's also see how early we are into the optical trade,
[8:48] which is the reason why I've stayed long Marvell
[8:51] but gotten out of micron.
[8:52] Marvell to me, we're still in the early stages of the optics.
[8:55] I don't really care that much
[8:57] whether we get the full build out in 2027, early or late.
[9:02] What I care about is where the earnings are gonna go
[9:05] and how easy it's gonna be to see,
[9:07] even if there's a slight delay,
[9:08] because we're further ahead in the curve here
[9:11] in terms of the demand side.
[9:13] So there could be buying in this way before it's built out.
[9:16] In a lot of the case with memory, that's what's happened.
[9:18] So I'm not gonna read all this.
[9:20] You guys can either pause it, read it here,
[9:21] go take the slide from the paywall,
[9:24] but you can go through and you just see
[9:26] that it takes you through every single month.
[9:28] And then at the end, we'll give you a summary
[9:30] on all of these pieces.
[9:31] I also asked it for the next thing
[9:34] to combine this with the Marvell announcement
[9:36] that happened in March when I wrote the paper,
[9:41] just to get a sense as to,
[9:43] okay, was he talking about Marvell?
[9:45] That is the reason why I got lucky
[9:47] in the fact that I wrote a paper a day before
[9:49] they made a big investment into Marvell.
[9:51] But the reason was because every time he talked
[9:54] in the first three months,
[9:55] the knowledge brain was telling me
[9:57] that he was mentioning Marvell without saying Marvell.
[10:00] I can't emphasize enough how important it is
[10:02] to basically use these and I just gave you a 30 minute way
[10:05] to set it up and to replicate it
[10:07] for whatever people you want, whatever things you want.
[10:10] It's taking the transcripts from YouTube.
[10:12] It's using the API from YouTube.
[10:14] Everything is set up there
[10:15] so you don't have to go through the process
[10:17] of copying, pasting, you can just run it.
[10:19] But it just gives you a sense as to how quickly
[10:21] you can get up on creating not only an idea
[10:23] but stay on top of the idea as it happens.
[10:27] In the 16th, Jensen never actually puts a date
[10:30] on the 800 volt DC rollout.
[10:31] So this is really me asking the question
[10:34] about the 800 volt rollout
[10:37] and whether there's a delay or not
[10:38] and what he has said at each of the different times
[10:40] in terms of expectations to see whether they're met.
[10:43] And basically everything that I've seen
[10:46] is that there's nothing out there
[10:48] truly on the delay side,
[10:50] even though I do believe that the risk
[10:53] of what people are talking about.
[10:55] It doesn't really matter whether it's late 26,
[10:58] the first rollout, late 27,
[11:00] it only matters from a trading perspective,
[11:02] but it may not matter from the Marvell side of the purchases.
[11:06] So just remember that, and again,
[11:08] this whole thing gets back to,
[11:10] why are we in the mid-cycle slowdown?
[11:12] Aside from that one thing that happened
[11:15] in the last two days, let's go through it.
[11:17] First of all, the pushback on AI just continues
[11:20] to intensify, you've got Eric Schmidt being booed,
[11:23] You've got the data centers in terms of the midterms and popularity.
[11:27] You've got Trump and Bernie Sanders agreeing on maybe something that the people should
[11:32] own the AI boom.
[11:33] I think anything involved with owning hyperscalers and Anthropic and open AI, I'd be very, very
[11:39] wary.
[11:40] And then you got the 1.3 billion followers of the Pope, the Pope stepping into the party
[11:46] to talk about the negativity.
[11:48] You have the capital raises.
[11:49] So we're finally at the stage where we need to tap these markets.
[11:53] And so you got Alphabet, and then on Friday you had Rumors of Meta trying to get out there
[11:58] in front of the $3.6 trillion open AI.
[12:01] You've got all these companies coming out to raise money.
[12:05] Anything on the bottleneck side or anything on the delay to Verreroobin, it's obviously
[12:09] bad for everybody involved in this, but they're all racing to try and get this money, again,
[12:14] to build things out.
[12:16] On the bottleneck side and from the delays, this is a good site in terms of just highlighting
[12:21] all of the different components.
[12:22] And again, if one of these is delayed, the photonics, that has an impact back to people
[12:26] using these things, every single one of them is interconnected.
[12:29] It's a very complex thing.
[12:31] Seeing this gets broken down into the physical layers and each component piece, it is a complex
[12:36] web and anything can slow it down.
[12:39] And that's why I keep saying the probability of this going smoothly is next to zero.
[12:44] the beginning part when the agentic side really turned the corner in October of last
[12:49] year with Opus 4.5. This is still something which has been ordering ahead of time. I believe
[12:55] the ordering will slow. I believe the cost issue has become a problem. And as I go through
[12:59] the rest of this, there's a lot of other stories. And then just yesterday, Crusoe pauses a Wyoming
[13:05] site. And again, whether you buy into any of these negative headlines, there is a truth
[13:11] to the fact that there's a lot of delays happening.
[13:15] I think this is a much bigger risk than people realize,
[13:18] which is the self-improvement risk.
[13:20] This was released on Anthropoc site today
[13:22] when AI builds itself.
[13:25] And they're basically talking in this thing
[13:27] about recursive self-improvement and its implications.
[13:29] We are there.
[13:31] It is as important, if not more important, than OPUS 4.5.
[13:35] It will disrupt the needs.
[13:37] It'll disrupt a whole bunch of things.
[13:39] and it will also be on the positive side for me on the Eli Lilly and all of the biotech and
[13:45] pharmacide, which I'll get at the end from an application side. I remember mythos as you go
[13:51] through wasn't released for the danger as well. A lot of this stuff I did today was on Fable 5 and
[13:57] mythos. The token bill comes due. So basically every podcast I listen to now is about the extreme
[14:06] costs that are going across the industry. Two months ago, we heard Uber basically talk
[14:11] about this in terms of their budgets, but now it's becoming a much, much bigger story.
[14:15] Almost every single podcast I listen to, this has become an issue and it's starting to show
[14:19] up in the data. Sam Waltman talked about it, shocked people last week because he basically
[14:24] said they're still trying to figure out how to deal with the rising costs as we move away
[14:29] away from the unlimited usage and the token side is now starting to go, and this comes
[14:36] at the same time.
[14:38] That DeepSeq V4, which is a real competitor, lowered their price for a 75% price cut.
[14:46] It's forcing in China rivals to slash prices, rattling the cloud providers.
[14:51] This all has impacts to the US.
[14:53] Here's the open router consumption side.
[14:56] So we had this massive rise here when open claw basically started, then it came down,
[15:03] you're getting a rise again.
[15:04] And this is an issue for anyone that's looking to see what's going to happen to Enfropix
[15:08] ARR, how much of the growth rate is going to continue if the costs are going higher
[15:14] and at the same point you have competition of cheaper prices that just slash the prices.
[15:19] Now, open AI last year, May 27th was the peak of usage and then they saw it plummet in the
[15:25] summer when the students were home. I've been talking about this a lot with people over
[15:30] the last week and a half that you have to be careful in thinking about the difference
[15:35] between seat based businesses and usage based businesses. On those usage based businesses,
[15:42] there is seasonality. There is no doubt in the summertime there should be less usage,
[15:47] not just because of school, but because of 20% of people being out of the office every
[15:52] single day during the summertime due to vacation time.
[15:55] So the question is, what happens to usage during there?
[15:58] Is that going to impact the ARR growth rate?
[16:02] Are we going to be at a second derivative slowdown because of the seasonals?
[16:05] Do they happen to come at the same time as where we are now?
[16:08] Last year, May 27 was the peak.
[16:11] Here's token pricing peaked May 28, now falling sharply.
[16:18] this is because of more people moving to deep-seek. It doesn't really matter to me. This is the token
[16:26] price. It's coming down right now. When it came down here, the reason given was because so many
[16:31] people were going to deep-seek at that time as part of the growth that that's what took it down.
[16:37] Regardless, we had been straight up for a while. We've now taken out the 50-day moving average
[16:41] and pricing is back to where it was April. And I remind you that the cost of a data center is
[16:46] increasing significantly. What I'm interested in is that the daily move in this is actually
[16:51] accelerating on the way down. So we've had other periods like there where it's negative,
[16:55] but this one so far looks more like something which is a bigger story. Maybe it has to do
[17:00] with usage coming down again because of school and because of the summertime.
[17:04] What did Jensen say just recently about a one gigawatt data center? It started at 30 billion.
[17:09] It's now at 50 to 60 and soon will be 80 to 100. So the cost is going higher. The competition is
[17:15] increasing, the token abilities going down, these companies are racing to continue to
[17:21] borrow money to build out the CapEx and to owe by the way, we still, regardless of the
[17:26] situation in oil, where as of now, as the news has gotten worse this month, oil prices
[17:32] have actually gone down, which just says something to all of the oil doomers.
[17:38] But I will say this, there's no doubt at this point that we're going to have supply shocks
[17:43] from a lot of different things there.
[17:45] And Helium is getting brought up more and more.
[17:47] This is from an article this week on June 9th today.
[17:52] It's by, I can't even read what it is.
[17:54] You guys can copy it and expand it.
[17:56] But there's a lot of stuff going on in there
[17:59] that is causing force measure clauses
[18:02] to have been triggered with inside the Helium side,
[18:04] any slowdown because of that.
[18:06] And this is what the chart looks like
[18:08] of my thematic portfolio.
[18:10] So you're at the point now where not only
[18:13] have you broken through the 20 day moving average,
[18:15] I mentioned that we now have over 60%
[18:19] of the individual names below there.
[18:21] I've shown that each of the other groups
[18:23] that's more related to the industrials has already peaked,
[18:25] but now also the 20 days moving down.
[18:28] I fully expect a reset back to the 50.
[18:31] It might take a month, it might take two months.
[18:33] I expect very much consolidation.
[18:36] I find it hard to believe that we're just gonna zoom higher.
[18:38] And remember, this is an equal weight portfolio,
[18:41] at least it started that way.
[18:43] And that just means that no one name
[18:45] or no one thing can do it.
[18:46] That's both a positive and a negative.
[18:48] The positive is we only broke out
[18:50] of kind of where we were at around the opus point
[18:55] back in April.
[18:56] And that's because of the diversification.
[18:58] A lot of the names, especially like Texas Instruments
[19:00] and Intel, a lot of those names only broke out
[19:04] during the time.
[19:05] If you want, and I say this on almost every video,
[19:08] This is the way that I wanna play the thematic portfolio
[19:11] during the slowdown.
[19:14] I wanna be short the hyperscalers equal weight
[19:16] and I wanna be long the thematic portfolio.
[19:18] That is the trade that to me has been going on.
[19:20] That is the one that is driving this.
[19:22] I've talked about this as benchmark arbitrage.
[19:25] I've talked about it as your CapEx is my opportunity.
[19:28] You're short the spenders.
[19:29] You're short the people raising capital
[19:31] and you are long and you're short.
[19:34] You're not getting hurt by the China stuff.
[19:35] You're not all of those trades over there.
[19:37] the hyperscalers are on the short side
[19:39] from a peg ratio basis.
[19:41] They're all still massively expensive
[19:43] relative to these other names on the infrastructure side
[19:46] with I think structural.
[19:47] Those things are on the coding and the software side.
[19:49] So this is just my play on it
[19:51] is I wanna be long the infrastructure short
[19:53] the hyperscalers and while this consolidates
[19:55] you might have the hyperscalers outperformed
[19:58] but when we finally come out of this
[19:59] the beta will be back on the infrastructure names
[20:01] and it'll happen one by one.
[20:03] The hyperscalers themselves,
[20:05] Amazon broke through the 50-day the other day.
[20:08] Meta, in a declining market with the 200-day down,
[20:11] is below the 50-day.
[20:12] You've got Google, all the good news on Google,
[20:16] all the people talking about how great of a company
[20:19] it's gonna be the winner,
[20:20] and they're right back on the 50-day.
[20:22] I would watch this one the most,
[20:25] partly because if you do an overlay of the PEs of Google,
[20:28] relative to the PEs of NVIDIA,
[20:30] the PE of Broadcom, the PE of any of these,
[20:34] you're gonna be surprised at how correlated they were
[20:37] and all of a sudden they're completely broken apart.
[20:39] I'd rather be long in video as a winner in this
[20:42] relative to Google in terms of this part
[20:44] because I still think multiple compression is a story
[20:47] and here's lowly Microsoft,
[20:49] which again, below the 50 days.
[20:50] So only one of the four hyperscalers
[20:53] is actually above the 50 day now.
[20:58] The hyperscalers have led the declines
[21:00] in my thematic portfolio.
[21:02] Basically, anytime we've had a correction that lasted,
[21:05] the hyperscalers broke first.
[21:07] Here's the hyperscalers in there.
[21:08] So as long as the hyperscalers are negative,
[21:11] I think the AI trade will be vulnerable,
[21:13] which is another reason to be long and short at the S&P.
[21:16] Same problem.
[21:17] If the hyperscalers are down
[21:19] and the thematic portfolio is down,
[21:22] there ain't nothing that's gonna hold the market up.
[21:23] Right now we've got a rotation going on.
[21:25] This is not an S&P side.
[21:27] I'm gonna get a little bit into the vol
[21:29] for you dispersion people
[21:30] so you can start to see the rationale behind
[21:33] what I think is going on.
[21:34] If we do get an S&P fall, if we do see a VIX spike,
[21:39] I believe it'll be a one or two day event.
[21:41] I do not think we are in anything more
[21:43] than a slow, mid-cycle slowdown for the AI trade,
[21:48] for the infrastructure trade, and if that's right,
[21:51] as I show you, I think it's gonna be a mistake
[21:54] to get too much on the VIX side, the VIX side,
[21:56] all of that stuff, and it's because I wrote this,
[21:59] I've shown this article before.
[22:01] I've had a lot of conversations with pension funds
[22:04] and endowments on this, mutual funds,
[22:06] I think are already doing this,
[22:08] but the reality is your CapEx is my opportunity,
[22:10] the benchmark arbitrage is the reason for the dispersion.
[22:14] It should be here.
[22:14] This is not some problem in the vol market.
[22:17] You're having a rotation.
[22:19] Here's my thematic portfolio
[22:21] with the 95 equity tickers that are in there.
[22:24] It's $20 trillion in market cap.
[22:26] The Mag7, magically, is 22.
[22:29] We're basically rotating, strip out NVIDIA,
[22:31] because it's in both.
[22:33] We're rotating from six names into 95 names.
[22:39] That's the unwind of the concentration.
[22:41] That is the benchmark arbitrage.
[22:42] You wanna be long, the infrastructure names.
[22:46] And you wanna be short,
[22:47] the market cap of the hyperscalers.
[22:49] So here's your dispersion.
[22:50] This is the equity vol, all the way up here.
[22:53] down here is the VIX now if this is going to spread into something more and
[23:00] here's the hyperscalers inverted if we're going to turn into something more I
[23:04] really do think the thing that matters the most is the IG market this is on CDX
[23:09] we haven't seen a whiff of anything in the credit markets if they become an
[23:14] issue that's where it becomes a bigger story to emphasize the dispersion point
[23:18] even further. Here is Morgan Stanley's Moe. This is the 30-day vol, realized vol of it.
[23:26] We are at the highest level in the last five years. We just took out the 22. If you do
[23:32] the tech version, we are at the highest in their history. That's a 25-year chart. So
[23:38] the 30-day vol on this or the movement, the vol on it is just through the roof. So that's
[23:44] That's 30-day vol. That's 30-day realized vol. All-time highs on the tech mo. Get used
[23:49] to the back and forth. Wanted to include this. Jeff put this out this morning, and I agree
[23:56] with the point, meaning you've had these big falls going on, a lot of volatility and momentum.
[24:02] It feels like a factor unwind. It feels like something bigger, but at the same point, the
[24:09] the number of names in the equal weight S and P
[24:13] has broken above their 20 day.
[24:14] Remember what I said about the thematic portfolio
[24:16] breaking down?
[24:17] Well, this is the other side of the equation
[24:19] and that says to me, that is a rotation.
[24:22] Same thing today, just look at where small caps are.
[24:26] So again, how you're navigating this as a subscriber,
[24:29] how you're getting through this.
[24:31] My job here is to give you kind of all of the pieces
[24:34] as to why this will be a worse time
[24:37] in terms of relative to the prior quarter,
[24:39] highlighting every now and then a turbulence side.
[24:42] We did get a turbulence signal finally.
[24:46] We had one back in here, but like I said,
[24:48] there was no cluster.
[24:49] We got one here, no cluster yet.
[24:52] I think when we get a cluster, it's gonna be a bottom,
[24:54] which is what historically has happened.
[24:56] In this case, it just happened as the event went down,
[25:00] so I don't think it's anything to bring to you guys.
[25:02] I am working on something super cool.
[25:04] This is my cursor output.
[25:08] You can see how much I've done.
[25:09] There are thousands of line of code on this.
[25:11] When it's done, I will basically figure away
[25:14] if you guys have access to this real time, if I can.
[25:17] I may put it up on a separate site
[25:19] or it may just update over there.
[25:21] But this is meant to be a dashboard
[25:22] specifically on the agentic side
[25:24] using my names in my thematic portfolio
[25:27] with a PCA analysis, correlation analysis,
[25:31] a PCA spread, I'm adding volatility,
[25:33] I'm adding breadth.
[25:34] So this way you guys have things
[25:36] that you can use signal-wise on your own
[25:38] to kind of navigate from a risk basis.
[25:41] All of the stuff that I give you
[25:42] in terms of the fundamental sheet,
[25:44] the exhaustion sheet, the technicals,
[25:46] every single day I would bring these up
[25:48] or put them in a folder,
[25:49] the same way that you're building a knowledge brain,
[25:51] just download every single week,
[25:53] those Excel files, put them into one folder,
[25:56] and then just go into Claude
[25:58] and ask it a question in co-work.
[25:59] Can you go through?
[26:00] And for the ones that are currently below the 50 day,
[26:03] Give me the list of 10 that are now closest to the 200 day
[26:05] with the 200 day still upward sloping.
[26:08] Whatever you want to do,
[26:09] this was just something I did to kind of get a sense
[26:13] as to the names like Hubble and these others where,
[26:15] okay, the 200 day is still sloping upward.
[26:18] In this case, silver massively upward
[26:22] were below the 50 day.
[26:23] We just broke below the 200 day as well.
[26:26] Like these types of things here
[26:28] are what I'm interested in kind of going through
[26:31] that you guys should be doing as well.
[26:33] You can add in questions on the fundamental sheet
[26:35] if you want, and in this case, I did do that
[26:38] just as an example.
[26:39] On Friday's thing, I uploaded all three
[26:41] and I said you're an experienced hedge fund
[26:42] quantitative analyst.
[26:43] Go through these three files from the close of June 5th
[26:46] and give me the five stocks I should buy and why.
[26:48] You can play around with this however you want.
[26:50] You wanna follow this, go ahead.
[26:52] Number one on the list is Micron
[26:53] for a variety of different reasons,
[26:55] but you can sit there and say what you wanna play around with
[26:58] and just go that route.
[27:00] All right, this is the final couple of minutes.
[27:03] I talked about this last week.
[27:05] So if these aren't working right now,
[27:07] when we're mid-cycle slowdown,
[27:08] but the market is doing fine, where are we going?
[27:11] We're going up here to the application.
[27:13] I'm very much a big proponent of this.
[27:16] If you haven't read the Eli Lilly piece,
[27:18] go read it in detail, go put it into an LLM,
[27:21] start uploading things.
[27:23] This is from Mythos 5 Release.
[27:26] These are all of the quotes of people
[27:30] that have been involved in it.
[27:32] Our internal protein design experts
[27:34] accelerated aspects of the drug design process
[27:36] by around 10 times.
[27:37] This is a striped thing, not having to do with biotech.
[27:40] Mythos 5 is our first model,
[27:41] consistently produced novel compelling scientific hypothesis
[27:45] in blinded head-to-head comparisons, blah, blah, blah.
[27:49] Mythos conducted novel genomics research.
[27:51] I'm listening to so many podcasts right now
[27:54] and what is happening in biotech
[27:55] in that IP side that I talked about.
[27:57] Here's the reason why it's important
[27:59] from a chart perspective.
[28:00] So we got Eli Lilly here making new all-time highs.
[28:03] The white line here is the XBI, the equal weight biotech index.
[28:10] These are correlated, obviously, right now.
[28:13] The reason that's important is because, historically,
[28:16] they're not correlated.
[28:18] So we're at a point now where, since this red line, which
[28:24] is really, in my opinion, when we started to see the change in the market.
[28:27] So when you hear people talk about, oh, junk is killing me and NVIDIA is not working,
[28:31] all of that shift started to happen here.
[28:33] So the agentic side from Opus 4.5 was here.
[28:36] But to me, the agentic side really started to work in August of last year,
[28:40] which is why a lot of the junk first quality started to shift.
[28:44] Small caps started to work.
[28:46] You ended up having the PMIs.
[28:49] I started writing papers on that.
[28:50] I was already seeing this shift happening.
[28:53] Well, here it is now.
[28:54] They're correlated and they're moving together.
[28:56] And whenever I see a change in correlations
[28:58] in something that's aligned,
[29:00] Eli Lilly so far this year,
[29:02] M&A has done 10 or 11 deals.
[29:05] That is almost as many as it had done
[29:08] the prior five years combined.
[29:10] They're buying everything in site.
[29:12] And I'm telling you, the podcast that I'm listening to
[29:14] are mind blowing in terms of what's coming on this side.
[29:17] If you want to take it from a candlestick basis
[29:19] to finish this off, here is the XBI.
[29:22] beginning in August, every month.
[29:26] Not yet this month, we'll see what happens,
[29:27] but all of these white candles means
[29:30] it closed the month above where it opened.
[29:34] That is a chart over 20 years.
[29:38] It's a 12-month streak now.
[29:40] These are the prior streaks that had ever happened.
[29:42] We've never seen anything like this.
[29:44] So when I can find places where a theme is happening
[29:47] on the application side, where the TAM is insane,
[29:51] and where one company in particular, Eli Lilly,
[29:54] has all the cash, has been spending it
[29:56] specifically on data centers,
[29:58] while their competitors were handing out dividends
[30:01] in Denmark.
[30:02] You end up in a situation where I really do believe
[30:05] that people are missing the boat on what's happening.
[30:08] And I love it when a stock is making new all-time highs
[30:10] as the rest of the market is going down.
[30:12] And Biotech, which is the part about the IP,
[30:16] they have IP, they have ideas.
[30:17] And if you go read the paper on Lilly
[30:19] and you start understanding what TUNE lab
[30:21] and what the Co-innovation lab is, what Lillipot is,
[30:24] and you start going through who actually has this
[30:26] on the farmer side, you start getting excited
[30:28] about the next wave of AI.
[30:30] That's it for the subscriber list.
[30:32] There's a lot of things going up.
[30:33] The slides for this will go up on the paywall.
[30:35] They may already be up there,
[30:36] but they will be up there by the end of the day.
[30:39] And hopefully this is gonna help you navigate
[30:41] the AI mid-cycle slowdown.
[30:42] I'll see you next month.