Silver Went Parabolic — The Next Trade Is Still Forgotten — Jordi Visser (25 enero 2026)

[0:00] uh in Miami this week for a Bitcoin
[0:03] conference. So,
[0:06] let's go through this. Um
[0:12] week started off uh as a lot of these
[0:15] weeks seem to start off with uh uh some
[0:19] issues involving Davos and Greenland and
[0:22] that lasted for a day. Uh I am going to
[0:25] finally spend a lot more time on
[0:29] silver, copper, minerals. We've talked
[0:32] about uh this being a transition point
[0:34] for the physical world, but I am going
[0:36] to really start to focus the attention.
[0:38] I'm writing a piece on silver for next
[0:40] week. And most of that is inspired by
[0:41] what I've heard at this event as people
[0:44] are trying to pick a top in uh
[0:48] commodities or at least worried about
[0:49] the parabolic moves. At the same time,
[0:52] they're looking to short semiconductors,
[0:55] which have had a parabolic move, and buy
[0:57] software. That seems to be what
[0:59] everyone's doing. I'll cover all of that
[1:01] this week and kind of go through my
[1:03] thoughts uh on where we are in the AI
[1:06] trade and especially for this year on
[1:08] the themes of making money. So, quickly,
[1:10] S&P flat for the week despite the uh
[1:14] horrible start. flat for the week for
[1:15] the Q's, flat for the week for small
[1:19] caps, and the Mag 7 up about 1%. Heading
[1:22] into earnings, uh we did have uh the
[1:26] second largest sigma fall
[1:30] uh in the market in the last, you know,
[1:33] this is uh over the course of the last
[1:36] 18 months uh or sorry, last year. But
[1:39] the main point I think to say is that
[1:41] most of this occurred. It was a big move
[1:43] down but it was also coming off of a uh
[1:46] a low V realized V time. So year to date
[1:50] now we're up 1% but I just want to again
[1:52] highlight that we are being led by
[1:54] energy and materials. So you're getting
[1:56] a lot of alpha there and you can see
[1:58] tech near the bottom. Uh you can see
[2:01] industrials, materials and energy. Three
[2:03] of the best performing stocks. We are in
[2:05] the physical upgrade cycle. I expect
[2:07] this to continue. Um, this to me is a
[2:10] preview for the year. Regardless of any
[2:12] bounce we get in tech or any bounce in
[2:14] software, I think you need to be
[2:16] focusing on energy, materials, and
[2:18] industrials in different components of
[2:20] it. And I'll go through the areas. Um,
[2:23] last 60 days in the S&P, uh, lowest rate
[2:26] of change since we broke out in May,
[2:29] which basically says the S&P is kind of
[2:31] spinning its wheels here for the last 60
[2:33] days, 60 trading days, which is three
[2:35] months. Um when we did get the move
[2:38] lower on Tuesday, uh the turbulence
[2:42] model basically showed more signs of a
[2:45] bottom than than anything. Uh but most
[2:48] importantly, I did get a few reachouts
[2:50] from people on what the turbulence model
[2:53] did and if there was anything there. And
[2:55] effectively from the analysis, it was
[2:57] very consistent with you had no movement
[3:00] in credit whatsoever measured on real
[3:02] time with HYG versus IF. Here are the
[3:05] junk spreads still sitting down there.
[3:07] Uh, as long as anything that happens
[3:10] with Trump is not economically important
[3:13] or earnings important and nothing that
[3:16] was going on with Greenland to me was
[3:17] more than a directional shock to people.
[3:20] Uh, I think I would be on the side of
[3:22] using those events to find the names
[3:24] that you haven't been able to get into
[3:26] yet because they've moved against you.
[3:28] Again, we had more people talking about
[3:30] rates. This will be a theme for me all
[3:32] year. um you know since 2022 when chatpt
[3:37] was launched basically tenure rates in
[3:39] the US have stayed between four and four
[3:41] and a half for the majority of the time.
[3:43] So we're at 423 uh as of Friday. Anyone
[3:47] focused at this being the big trade when
[3:49] you're seeing things like micron
[3:50] technology and silver and all of this
[3:54] stuff is just a waste of time to me. And
[3:56] again I have no care whether tenure
[3:58] rates go to 475 or 375 uh breaking out
[4:01] of this range. I think they're going to
[4:03] be stuck around this level. If I had to
[4:04] guess, we're going to see an upward move
[4:07] in rates, but I just don't think
[4:08] shorting bonds is a trade. Michael
[4:10] Hartnett had a lot of good things out
[4:12] this week in terms of kind of the charts
[4:14] and things that I agree with. Uh, so I
[4:17] just wanted to highlight them, not
[4:18] because we're in agreement, but because
[4:20] I think these are important charts to
[4:21] keep in the back of your mind for what I
[4:22] think is going on. First thing is is
[4:24] just the returns between small cap and
[4:26] large cap. As I've talked about, I
[4:27] expect small caps to be up significantly
[4:29] this year relative to large caps. You
[4:31] can see the inflows. You can see all
[4:33] this. This is the year where I think
[4:34] people are basically going to have a
[4:36] hard time making alpha the way they have
[4:38] in the past.
[4:40] And this is all related to the fact that
[4:42] I think we're at an inflection point in
[4:44] the AI trade. And I really think people
[4:46] are going to have to uh adapt uh in
[4:49] understanding what's happening as the
[4:50] MAG 7 and the deconentration takes hold.
[4:54] Uh just again AI is an important thing
[4:57] and now we're getting kind of the first
[5:00] fear factor I guess uh with where storms
[5:04] connect back to the grid. We'll see what
[5:06] happens but I wanted to bring this up in
[5:07] case something goes on. So this is a
[5:09] visual that you're going to start seeing
[5:11] me spend more time on the entire year
[5:14] because I believe to make money from
[5:16] this point. This is the software boom.
[5:18] We have abundance. Um everyone can try
[5:20] to pick a bottom in software. I view
[5:22] this as the same thing as people trying
[5:24] to predict when the equity market is
[5:26] going to collapse on the back of the
[5:27] debt. Uh the ego that sets in with
[5:30] trying to go against something that has
[5:32] worked for 15 years that now has an
[5:35] insane amount of abundance in it. Uh
[5:38] coding is abundant. Uh apps are now
[5:40] being built in fast time. All of these
[5:42] things which I'll go through as it goes.
[5:44] But just keep this in mind. This is
[5:46] about scarcity. Now I every investment
[5:49] you have to go through and look for
[5:50] bottlenecks and decide whether this is
[5:52] going to be there over the next 5 years
[5:54] and if it is how much of has already
[5:56] been discounted I think the data center
[5:57] scarcity side is known um I think the
[6:01] next stage is really as I'll go through
[6:04] something that fits more into a
[6:06] different visual which I'll highlight
[6:08] which is the conversion from raw p raw
[6:11] energy so natural gas solar coal
[6:15] whatever you want into intelligence
[6:17] There are two two tracks of that of that
[6:20] uh bottleneck and I'll break them down.
[6:21] I've talked about them separately but I
[6:23] have a visual to go through. Uh again we
[6:26] start the economist finds a way to make
[6:29] sure that there's a bottom in the market
[6:30] at least for a period of time but the
[6:32] Donald Trump bare thing to start off the
[6:34] week. I do want to highlight a new drug
[6:36] for each patient. I'm going to highlight
[6:37] this now. This was a story in there.
[6:39] This is again going to be important to
[6:41] me when we get to the software side and
[6:43] being in an age where now you're
[6:45] customizing. Uh think of a new drug for
[6:48] each patient as different than Advil for
[6:51] everyone or drugs that are made for
[6:54] everyone goes through an FD everyone
[6:56] takes it. um customizing things
[6:59] specifically for a patient is going to
[7:00] be the trend, but customizing workflows
[7:03] specifically for businesses is going to
[7:05] be the solution that will disrupt
[7:07] software over as time goes on. Uh so the
[7:09] Greenland side, I talked about this a
[7:12] year ago, you're going to continue to
[7:14] hear me talk about this. We have a
[7:16] shortage of critical minerals. We are in
[7:18] an arms race around the world for AI.
[7:20] We're in an arms race for military
[7:22] supremacy. All of these things require
[7:25] critical minerals. In the past, we have
[7:26] talked about rare earth as a major
[7:29] factor in this. So, anyone that doesn't
[7:32] want to connect Greenland to critical
[7:33] minerals in Greenland to the strategic
[7:36] implications of making sure that the
[7:38] other AI uh components with the
[7:40] military, China, Russia, Iran, they are
[7:43] involved in all this decision making.
[7:45] So, it's not just about the rare earth,
[7:47] it's also about security. Same thing
[7:49] went for Venezuela. I had some people
[7:51] reach out to me here and say that people
[7:53] had said that it was ridiculous for me
[7:54] to post something about Venezuela's
[7:56] importance with crit critical minerals.
[7:59] I don't think people fully grasp this.
[8:01] 100% everything in the world is about
[8:03] securing the supply chains right now in
[8:05] the military making sure that Iran and
[8:07] Russia in particular when it comes to
[8:09] the US along with China are in their
[8:11] hemisphere. Um we are in a different war
[8:14] type situation with drones over tanks.
[8:16] We're in a different war situation with
[8:18] drones over ships and making sure that
[8:20] bases are at a particular pace and also
[8:22] that we have access to critical minerals
[8:24] because these will be in a shortage for
[8:26] a long long time. And Howard Luck Lutnik
[8:29] confirmed it. He told reporters
[8:30] Venezuela possessed significant amounts
[8:32] of critical minerals and a great mining
[8:34] history that's gone rusty. He believed
[8:36] the Trump administration could fix. So
[8:38] again, the administration talks about
[8:39] it. So for people that are not fully
[8:41] grasping this and talking about oil, we
[8:43] have abundance of energy. We have more
[8:45] fossil fuels than we need for the
[8:47] environment we're in. But we have
[8:48] shortages of and this was announced by
[8:51] the White House um on January 14th
[8:55] again around critical minerals. This is
[8:57] a focal point. This will remain a focal
[8:59] point. Don't forget it. It's what
[9:00] stopped the tariff situation with the
[9:02] rare earths and China showing that they
[9:04] had this monopoly in exchange for the
[9:06] chips. That is our monopoly.
[9:09] So you also had Riotinto and Amazon
[9:11] collaborate to bring copper. So
[9:13] basically, Amazon's in the copper mining
[9:15] business as of this week, the silver
[9:18] chart. And yes, there will be a
[9:20] pullback. And it makes sense that we're
[9:23] just above 100. Uh this was before the
[9:25] close on Friday, but I wanted to
[9:27] highlight the parabolic nature of this,
[9:29] which is these are impossible charts to
[9:32] buy. Where do you buy something that has
[9:34] already broken out? you had basically 14
[9:36] years to buy silver and in particular up
[9:39] until about September October it really
[9:42] wasn't moving and then this move
[9:44] occurred. So the question is that being
[9:46] at this conference and hearing people
[9:48] talk about silver and thinking about it
[9:49] as a precious metal. I just want to make
[9:52] the connection. It is not a precious
[9:53] metal is directly related to the AI
[9:55] trade. It is an industrial metal and
[9:57] more importantly the fact that people
[9:59] think of this as something that is just
[10:01] representative of what gold has done. I
[10:03] think we have to dig deeper in this.
[10:05] This is the overlay of micron with
[10:06] silver. These are parabolic moves. These
[10:09] are the types of things that if you
[10:10] don't get in early you have a hard time
[10:12] getting into. I posted this because Elon
[10:15] Musk during the holiday season posted
[10:17] about silver. This is not good. Silver
[10:19] is needed in many industrial processes.
[10:21] That's an understatement. It is huge in
[10:24] solar, in NPUs, in switch gear, in
[10:27] circuit breakers, data center, advanced
[10:29] packaging at chips, EVs, FSD, edge
[10:31] devices, drones, radar, electronic
[10:34] warfare, robotics, humanoids, and
[10:36] sensors. All of these are in S-curves.
[10:39] They're all happening at the same time,
[10:41] which is why I'm writing a paper because
[10:42] I want to make sure people realize
[10:44] regardless of your belief in where
[10:45] silver should be, whether it should be
[10:46] 100 or 75, the mining stocks are going
[10:49] to continue to have a bid underneath
[10:50] these. This is not a cyclical move in
[10:52] silver. And that's the point of this.
[10:53] Forget buying into it at this point. It
[10:56] is price inelastic. It does not matter
[10:59] whether it's $300, $400, $500. It is
[11:02] such a small portion of
[11:04] data centers that it's a small portion
[11:06] of AIS. It is not huge in this, but it
[11:10] is necessary in this. So, it has become
[11:12] a critical part in drones. Uh, it is a
[11:15] critical part in a lot of different
[11:16] things. But when you're talking about
[11:18] drones versus tanks, massive amounts of
[11:20] drones versus very limited dranks,
[11:22] extremely high replacement rate. This is
[11:24] just showing the importance or the
[11:26] transition in silver for the military
[11:28] side. And this is just with drones. This
[11:31] doesn't include the other side of it. It
[11:32] does include the data centers and the
[11:34] places that it's necessary uh for NPUs
[11:37] and edge devices for humanoids. All of
[11:39] that's there. Uh I'm just bringing this
[11:41] up because Cuba would be the I guess the
[11:44] last place where there's a Russia China
[11:48] situation and where there's critical
[11:49] minerals. So don't be surprised if Cuba
[11:52] becomes uh a story at some point this
[11:54] year as well. Now we did have Davos
[11:58] going on and Nvidia and Jensen Yuang
[12:01] spoke many many times. I just wanted to
[12:03] highlight this thing that he said on
[12:05] Friday which is the largest
[12:07] infrastructure buildout in US history or
[12:10] human history with estimates pointing to
[12:12] $85 trillion of investment over the next
[12:14] 15 years spanning energy chips, data
[12:15] centers and AI factories. The buildout
[12:18] is still in the early innings. The
[12:19] reason I bring up the silver thing, the
[12:20] reason I bring up the copper thing,
[12:23] we're at the very early stages. So when
[12:24] I wrote the paper this year about us
[12:26] being at the transition point away from
[12:28] software and into hardware, we're at the
[12:30] very beginning stages of building AI
[12:32] factories. We're at the very beginning
[12:33] stages of having a shortage in energy of
[12:36] needing the micron. That is why you're
[12:37] seeing these parabolic moves because
[12:39] we're in a commodity bull market and
[12:41] where there's shortages you buy as much
[12:43] as you can particularly when they're
[12:45] price insensitive uh for the demand. So
[12:48] this came out uh Friday. Bernstein is
[12:51] showing what the copper shortage will
[12:53] look like. So I want you to think back
[12:55] to Jensen Yuang's thing and he basically
[12:57] said for the next 15 years. So I wrote
[12:59] my paper. This is 2026. Here's 15 years.
[13:03] This is the predicted shortage. This is
[13:06] not cyclical. This is not okay copper
[13:09] goes to a certain level and it's going
[13:10] to collapse. You have to think about
[13:12] that because that's the way these stocks
[13:13] have been priced and that's the way they
[13:15] show up. So right now to show how much
[13:17] people doubt the non-yclicality of
[13:19] miners here is the move in basic
[13:21] resources in the uh euro stocks I want
[13:25] you to highlight we're at the same level
[13:27] we were in the great financial crisis.
[13:29] This is going much higher and this will
[13:31] be as I said materials and energy will
[13:33] be the two best performing stocks this
[13:34] year. I said this the first week, it
[13:36] continues to be the case. And right now,
[13:38] earnings estimates, the analysts are
[13:41] still freaking out about cyclicality.
[13:45] This will be a long-term play. From a
[13:47] valuation basis, it's they're not only
[13:49] rejecting kind of earnings estimates,
[13:52] we're at valuations which say you got a
[13:54] long ramp run rate into here. Even if
[13:57] nothing else happens, as long as silver
[13:59] stays around the same level, you can see
[14:01] where the money's going to go. Now, I do
[14:03] believe this is the 1970s. Now, before
[14:05] you get into the inflation stuff, I'll
[14:07] get into that. But these are the types
[14:08] of things that Michael Hartnett put out
[14:10] that you want to be invested in. Look
[14:11] where tech is. So, we just entered a
[14:13] point where you wanted to be the
[14:14] opposite. You didn't want value. You
[14:16] wanted growth. Growth is here. It's
[14:18] still going to work, but it is way
[14:19] behind a lot of these. I completely
[14:21] agree with small cap, the value side,
[14:23] commodities, all of that. So, if you're
[14:25] a mutual fund sitting out there and your
[14:26] portfolio is weighted heavily towards
[14:28] growth, remember your current waitings.
[14:30] Yes. technology sector 33, but when you
[14:32] add in the communication services,
[14:35] you're up at 43. When you add Amazon and
[14:37] other things from the consumer
[14:38] discretionary side, you're basically at
[14:39] 50%. Energy, materials, 5%.
[14:44] Nobrainer.
[14:46] Get your money in these things. We need
[14:48] and we cannot convert raw energy into
[14:51] intelligence. So, we were using GPUs for
[14:54] the data center. That was for the
[14:55] training models, but we are now entering
[14:57] the AI factory world. that
[14:58] infrastructure buildout that Jensen Yu
[15:00] Wong talks about, who better to kind of
[15:02] say what's going on. Now, from a
[15:03] charting basis, you've got healthc care
[15:06] ready to break out. I've talked about
[15:07] the drug and the pharma side. You've got
[15:08] energy ready to break out of a big base,
[15:10] John Rog style. You have materials ready
[15:12] to break out of a John Rog big base
[15:14] style. And I highlighted the euro stocks
[15:16] are even better. So, technically, this
[15:18] is not just an AI story. This is also
[15:20] the technical picture, the underweight
[15:22] picture, everything. DRAM, another super
[15:24] spike. So, this is the old DRAM prices.
[15:27] This is overlaid with PMIs. Commodities,
[15:30] commodities, commodities, silver,
[15:31] copper, DAMS. You're going to have
[15:33] everything along those lines going
[15:35] higher. EM a natural play on this. I'm
[15:38] going to spend a lot of time on EM. I'm
[15:39] going to spend a lot of time on Brazil
[15:40] today. EM is a natural place for this to
[15:43] be. They benefit a lot from what we're
[15:45] talking about, but in particular
[15:47] commodities. I want to be long emerging
[15:49] markets. I wrote this in Brazil in June
[15:51] of last year.
[15:53] The dollar to me will be weakening on
[15:55] the back of the capital s uh the capital
[15:57] account surplus reversing because the
[15:59] mag 7 will be underperforming. If
[16:01] materials do well look at the material
[16:03] side in Europe versus material side in
[16:05] the US compare them as a waiting in the
[16:07] portfolio and you get very very
[16:09] different things. Technology in the US
[16:10] and particularly the mag 7 if they're an
[16:13] underperformer and large cap is an
[16:14] underperformer you're going to have a
[16:15] massive move into the manufacturing and
[16:17] commodity stuff that's better for
[16:18] Europe. It leads to capital outflows. uh
[16:21] the next phase of artificial
[16:22] intelligence will be defined by
[16:24] hardware, power, materials, not
[16:25] software. I wrote this in June of last
[16:27] year. If you would have followed this
[16:28] with silver and everything and
[16:30] everything along those lines and semis
[16:32] and micron and all all of this stuff as
[16:34] opposed to software, you'd be doing very
[16:37] well since there. And finally, a global
[16:39] push to secure critical minerals due to
[16:41] the trade war between China and the US
[16:42] where Brazil is rising fast as a
[16:44] strategic supplier. Just think about the
[16:46] countries we're seeing and how much
[16:47] critical minerals are bringing up. This
[16:49] is a new story. Brazil breaking out.
[16:52] This is what it's done since I wrote
[16:53] that paper. It's up 30
[16:56] 30% since there. I think it has a lot
[16:59] further to go. This is the relationship
[17:01] historically over the last 14 years
[17:03] between silver and EWZ. Obviously,
[17:06] there's a dollar component. There's a
[17:07] Brazil component, which gets back to
[17:09] commodities. Well, here's EWZ with the
[17:12] most recent thing of silver. If you're
[17:13] looking for lagards, get into Brazil. We
[17:16] just had one of the best performances on
[17:19] a weekly basis in EWZ relative to the
[17:21] S&P. Brazil is up in one of the best
[17:24] performing markets in the world right
[17:26] now. Now, this is the negative side and
[17:28] why I said earlier, do not worry about
[17:30] the inflation component. So, I just want
[17:33] to highlight he put this as well. This
[17:35] is not going to happen. And the reason I
[17:37] can say this with certainty or at least
[17:39] where I'm willing to put my uh
[17:43] uh my uh my money on the line here, I
[17:46] highlighted this because I do expect a
[17:48] possibility that this could occur. This
[17:50] would take us back up to 4%. This to me
[17:52] is not going to happen because what we
[17:54] have right now is an employment
[17:55] situation which is weakening. So wages
[17:58] are not growing. We have a housing
[17:59] affordability issue. So house prices are
[18:01] going to be moving lower. They have been
[18:04] moving lower and I think the Trump
[18:05] administration is focused on that. But
[18:07] most importantly, this happened because
[18:09] of oil and this is the difference right
[18:11] now. We have an abundance of fossil
[18:13] fuels. So in 1974 when everyone scared
[18:17] the hell and said we're going to have
[18:18] another rise of inflation. This is the
[18:20] price of of oil per barrel. So we went
[18:23] from 10 to 35 to get that second wave of
[18:27] inflation. So you had the first wave
[18:28] here and then you had the second wave.
[18:30] We are not getting that this time. So if
[18:32] you think oil is going to go effectively
[18:34] from 60 to 180, then yeah, we're going
[18:36] to get inflation at those levels. If you
[18:38] believe that we have a surplus, it's
[18:41] fine. This is the overlay of CPI
[18:42] year-over-year with that oil chart in
[18:45] those years. So that was an oil thing.
[18:47] We have a very unique situation that is
[18:49] not about oil, but I do want to be long
[18:51] energy because we do need power. I do
[18:53] think oil has a floor because of how
[18:56] much the commodity stuff is booming and
[18:58] how much PMIs are going down. But if oil
[19:00] only goes to $70, you're still going to
[19:02] see XLE outperform. And again, I include
[19:05] this as software. Everyone is trying to
[19:07] pick the bottom in software, forget it.
[19:09] You can get a bounce. You can get a
[19:11] trade. You can go. That's an ego trade
[19:12] as far as I'm concerned. You want to buy
[19:14] scarcity. We have problems with
[19:16] commodities. We have problems with
[19:17] power, electricity. There are plenty of
[19:20] ways to make money. The electricity
[19:21] thing will be deals that Chevron, Exxon,
[19:23] all of them will benefit. I don't want
[19:25] to be short this chart. This is a
[19:26] beautiful looking thing. And this is XLE
[19:28] over software or over IGV. That's the
[19:32] trade you want to have. I showed it is
[19:34] Chevron versus um uh Salesforce.com last
[19:38] week. So back into the abundance
[19:39] scarcity, you want to be short abundance
[19:41] software. You want to be long everything
[19:43] associated with a buildout. This is the
[19:45] reason why and I want to make sure that
[19:47] you guys can fully see this. Oh, you
[19:51] don't need to see me. Um so this thing
[19:54] here is very simple. This is showing
[19:57] that we have plenty of energy. We're
[20:00] making it into intelligence. There are
[20:02] two tracks here where there are
[20:03] shortages. One is turning energy, the
[20:06] raw power into electricity. And that's
[20:07] where the turbines show up. That's where
[20:08] the switch gear shows up, the
[20:10] transformers. These are all bottlenecks.
[20:12] That is where the data center has been
[20:14] and this is where trades have worked. On
[20:17] the other side, you have all of the
[20:19] advanced GPU. You have the high
[20:21] bandwidth memory. All of this has also
[20:24] been in a shortage. This is the place
[20:26] where scarcity exists. So you are taking
[20:28] raw energy which we have an abundant
[20:30] amount for but it's the conversion
[20:32] process into intelligence where we have
[20:34] the issue that will remain the issue as
[20:37] we go forward. The place where you want
[20:39] to be this advanced GPUs that's why I
[20:41] showed this packaging thing that is a
[20:43] lot of names there. This right now
[20:46] becomes important as we get into edge
[20:48] devices. This here is really more for
[20:50] the cloud and the because of the data
[20:52] centers. you're going to get into the AI
[20:54] factories which are more for on premise
[20:56] things but I just wanted to bring that
[20:57] up because you want to invest in the
[20:59] shortages so everyone wants to take the
[21:01] other side of this it's in X every day
[21:04] everyone wants to short semis and be
[21:06] long software they think this is a
[21:08] bubble they look at it this way I just
[21:10] wanted to show who wants to be long this
[21:12] chart this is just that inverted
[21:14] whenever I see people kind of doing this
[21:16] they're looking at it and they're more
[21:17] comfortable shorting semis and uh versus
[21:20] and being long software as some trade,
[21:23] but nobody wants to catch a falling
[21:25] knife. And this just continues to go
[21:26] lower, and I think it will continue to
[21:28] go lower over the next five years with
[21:29] episodic shifts. This is software
[21:32] relative to the NDX. Again, who wants to
[21:35] buy stuff like this? Can we get one of
[21:37] these where it reaches a point and then
[21:38] it goes sideways? Sure. If that the best
[21:41] trade you can come up with to try and
[21:42] pick the bottom of software, it's not
[21:44] worth it. So, let me go through the
[21:46] reasons. I wrote a piece, why buying
[21:47] cheap software is now the new AI bubble
[21:49] trade. I fully believe in hearing people
[21:51] talk that people are fighting this,
[21:53] particularly people who are really good
[21:55] at coding trying to tell me that claude
[21:57] code isn't good enough and it won't
[21:59] work. I just want to like bring it out.
[22:01] If this is the best trade you can come
[22:02] up with with all these parabolic charts,
[22:05] I think you're in trouble. Go to energy,
[22:06] go to materials, go to some of the names
[22:08] that haven't yet broke out where the
[22:10] analysts are still not positive and
[22:11] they're treating this as a cyclical
[22:12] thing. A rerating is when stocks that
[22:15] have been priced at a certain multiple
[22:17] are about to go through a dramatic
[22:19] shift. I think that's going to happen in
[22:20] the drug sector, in the health care
[22:22] sector, and I think that's going to
[22:23] happen for sure in the materials and
[22:25] energy and industrial sector where
[22:27] people are scared of the next cyclical
[22:30] um point where this will break down. And
[22:32] if oil doesn't go higher, I don't see
[22:35] anything that will stop copper and
[22:37] silver from being an issue. They are not
[22:39] going to create demand destruction, at
[22:41] least not for the AI trade. They will
[22:43] have impacts on phones. They will have
[22:45] impacts on autos. So will high bandwidth
[22:47] memory. But the issue comes in if it has
[22:50] a uh a 10% increase in the price of
[22:53] those things. It may be bad for
[22:55] inflation, but it's not going to have an
[22:57] impact on the deflationary components
[22:58] without oil going higher. So just
[23:00] another thing on the software side.
[23:02] Claude made the Wall Street Journal uh
[23:04] in terms of taking the world by storm
[23:07] and this was all about the coders
[23:08] agreeing with it. Anthropic revenue is
[23:10] now racing again. They're the ones
[23:13] dominating on the enterprise side from
[23:14] the coding. VCs are piling in. They're
[23:17] going through and I mean they're going
[23:19] to have an IPO. Uh Agentic Coding is
[23:22] accelerating app releases. This was from
[23:24] code two. There were two charts that
[23:27] went around. I just want to highlight
[23:29] how important this is from the
[23:31] competitive side. So, one of the areas
[23:34] on software that people have to
[23:35] recognize is that you can think that
[23:38] enterprise software is bulletproof or
[23:40] that it's gotten cheap. Competition is
[23:42] rising by the day. It may not be able to
[23:45] directly have an impact today, but you
[23:48] are getting so many more apps and as the
[23:50] models get better, this happens. So,
[23:52] this is the coding side of vibe coding.
[23:54] This is when Vibe Coding took over. For
[23:57] three years, we were having no app
[24:00] releases. And now all of a sudden, we're
[24:01] up at 60% year-over-year. It's been
[24:04] increasing. It's just a dramatic shift.
[24:07] And to ignore it would be a mistake. So,
[24:09] I took the other chart, you can see it
[24:12] here in terms of more the smoother one,
[24:14] which gets into the amount of uh apps
[24:16] released every month. I said okay now
[24:19] taking this uh agentic platform
[24:21] competition for SAS combined with
[24:23] reality of vibe coding began what is the
[24:26] summary for the enterprise SAS model and
[24:30] basically it said okay we're agentic AI
[24:32] will directly change the workflow side
[24:34] this is uh an existential tension
[24:40] enterprise SAS platforms whether it's
[24:43] these names here monetized by encoding
[24:45] best practice workflows forcing standard
[24:47] ardization across teams. This is the
[24:50] other point and this is where I'm going
[24:51] to get into this company GenSpark as we
[24:53] go through it which is if you're doing
[24:56] standardization and I had an issue with
[24:59] Salesforce.com when they came in and the
[25:01] reason I had an issue is because a lot
[25:03] of the things that I wanted for what we
[25:05] did for fundraising
[25:07] they could not be done in there. It
[25:09] wasn't part of it. So we had to figure
[25:11] out a way to adapt it and add things
[25:13] that we wanted and get them into the
[25:15] database. So they create a form which is
[25:19] fairly standardized for everyone to fit
[25:22] the majority of people. SAS is built
[25:24] with everyone in mind. Meaning it does
[25:26] not fit 100% for anybody. The problem is
[25:28] enterprise clients want their exact
[25:29] workflow automated. They aren't going to
[25:31] be satisfied with 80%. Because the
[25:34] person you just convinced them to fire
[25:36] was already doing 100%. That is the big
[25:38] issue. So when people look at how is the
[25:41] productivity going to work as I go
[25:42] through this you're going to start to
[25:44] see the success by companies has been
[25:47] not through using the enterprise
[25:49] software it's through agentic AI and
[25:51] it's through having those things
[25:53] effectively be databases where the data
[25:55] is the value but not the software that
[25:58] is an important component of this which
[25:59] is going to become more of a story as
[26:01] time goes on now this is the company
[26:03] GenSpark so it's talking about how
[26:06] GenSpark is competing with software.
[26:11] They either next raise after hitting 100
[26:13] million in ARR.
[26:16] It's an agentic AI platform designed to
[26:18] handle complex real world tasks without
[26:20] being rigid and predefined workflows.
[26:24] So less control, more tools. Again, you
[26:27] go back. You want to have more
[26:30] flexibility and adaptability and not
[26:32] have to deal with something which is
[26:34] scripted.
[26:36] You want things that are adaptive that
[26:38] are able to act, fail, recover, learn in
[26:40] real times, make changes, go through it.
[26:42] That's the power of AI agents. And maybe
[26:44] some of the incumbents can actually get
[26:47] there, but it is very challenging and
[26:49] very costly to spend the money to get
[26:52] these agentic flows from everything that
[26:54] I've seen to actually be able to
[26:55] customize to an individual. So, I find
[26:58] it very challenging and I don't know how
[27:00] people are going to pay for it when they
[27:01] can just use the data which they own.
[27:04] Um, I brought this new drug thing up
[27:06] because again, this is from the article
[27:08] with inside the economist which I
[27:10] referenced. Everything is going to be
[27:12] hyperpersonalized. So when you're
[27:14] thinking about the way people talk about
[27:17] software, we are in a new world where
[27:19] everything should be customized to your
[27:20] workflow. It should be customized to
[27:22] your body. You make traditional
[27:25] medicine. You make one drug like Advil
[27:27] for millions. Traditional software, one
[27:29] workflow for many companies. The new
[27:32] world of Agentic, one solution per case.
[27:36] That is a very different thing for
[27:38] incumbents to be able to do. Enterprise
[27:40] software in the age of aentic platforms
[27:42] and vibe coding. Software supply I
[27:45] showed you at the beginning has
[27:46] exploded. Apps are going nuts. I build
[27:48] software. I can go through this. Custom
[27:50] beat standardized. The customization is
[27:53] the critical part. Workflows are no
[27:55] longer the moat. Value shifts from
[27:57] software to results. When you're trying
[27:58] to get revenue per employee higher, the
[28:01] critical thing is to replace people with
[28:05] customized workflows, not software.
[28:08] Software might make people more
[28:10] efficient, but it it does not allow you
[28:13] to not hire as many people. So, the
[28:15] seatbased pricing breaks, revenue
[28:17] pressure precedes churn, discounting,
[28:20] longer sales cycles, and renewed
[28:21] friction appear before customers
[28:23] actually rip out the platforms. This is
[28:25] kind of one of the things I think you
[28:27] have to be all over these software
[28:28] companies and their earnings and really
[28:30] do a deep dive. This is not a valuation
[28:32] thing. Do not buy these for valuation.
[28:34] Buy these because you think the company
[28:35] is able to adapt and they're seeing
[28:37] success in terms of the revenue side
[28:39] that is actually growing because it is a
[28:42] lot of money I think for people to spend
[28:43] on this when their goal is ultimately to
[28:45] have people go through. The winning SAS
[28:47] model becomes agent native. It's forcing
[28:50] enterprise SAS to evolve or accept
[28:52] structural value compression.
[28:55] That's in my opinion what's happening.
[28:56] What do incumbents need to do to
[28:58] survive? It's hard. I I I just don't
[29:02] see. They're basically saying, "Here's a
[29:04] system. Follow the process. Tell me the
[29:06] goal. I'll handle the rest."
[29:09] Again, I think this is a problem because
[29:10] the people don't really use AI that
[29:12] much. This doesn't make a lot of sense
[29:14] to them. But as someone who uses it all
[29:15] day long and watches it evolve, it gets
[29:17] better every single time. I don't see
[29:19] how software companies are going to be
[29:20] able to have update upgrades going
[29:22] constantly. Um, if you want an article
[29:24] to go read about why SAS might not be as
[29:27] safe as investors think, and I think
[29:29] Renee Selman does very good work, um, in
[29:32] terms of putting things out, and I think
[29:34] this was a very, let's say, balanced
[29:38] thing talking about they're selling off
[29:40] hard, investors are looking for places
[29:42] to go, value comes from trust, blah,
[29:45] blah, blah. uh when you go through it
[29:46] the real risk isn't current AI
[29:48] capability it's human misjudgment of
[29:50] exponential process progress this is
[29:52] really important we instinctively model
[29:55] technology linearly
[29:57] while it often involves
[30:00] multi multiplicatively
[30:04] exponential blindness examples the main
[30:06] point of this is as things are changing
[30:08] quickly your brain doesn't allow you to
[30:12] believe that things are changing that
[30:13] quickly and that every time that this
[30:15] happens, you want to go pick again, buy
[30:19] into the weakness. When exponential
[30:22] change happens, like silver and it goes
[30:24] to 300 and it goes to 400. Once it gets
[30:27] to 60, it's overbought. Then at 70, it's
[30:29] overbought. Then 100, it's overbought.
[30:31] Well, this is parabolic. Parabolic stays
[30:33] parabolic until it's not parabolic. But
[30:35] if the demand side is there and the
[30:37] shortage goes through, because we have a
[30:38] bunch of S-curves that didn't exist 15
[30:41] years ago, because drones were not
[30:43] tanks, because there were no EVs that
[30:47] matter, because FSD wasn't a thing,
[30:49] because China and the US weren't in an
[30:51] AI arms race, blah blah blah. You get
[30:53] the picture. All right. Um, you should
[30:55] go listen to these. They're all about 30
[30:57] minutes. So Jensen Yuang spoke,
[30:59] interviewed by Larry Frink. Um, and he
[31:02] basically went through the things that
[31:03] I've already highlighted.
[31:05] I don't think there's anything there
[31:06] that's uh that's new. But I do think uh
[31:09] if you want to sit there and pause it
[31:11] and read it, it gets back into this. So
[31:13] everything he said gets back into this
[31:15] chart. We're going to have shortages for
[31:16] a long time. They're not going to ease
[31:18] up. We cannot build the things we need.
[31:20] To get more silver takes a long time. To
[31:23] get more copper takes a long time. To
[31:25] get more gas turbine takes a long time.
[31:27] High bandwidth memory, we'll see how
[31:29] long that takes. China's already trying
[31:30] to do it. Um I don't know what the right
[31:33] price is but I do know at some point
[31:35] whether it's Micron SKH high Sandisk
[31:37] they'll get to levers will they've built
[31:39] in enough of the next three years the
[31:41] question is where that'll be uh for
[31:43] silver and copper it's going to take a
[31:44] lot longer than that uh just because I
[31:46] don't think it's there the stocks which
[31:48] have rallied are going to rally more but
[31:51] this is also going to spread into
[31:53] chemical names which have started to
[31:54] rally the basic materials thing you have
[31:56] to go there's a cyclical discount in
[31:58] there of which for the next 15 years
[32:00] it's going be like the software side
[32:02] was. It's not going to show up in profit
[32:04] margins. It's just going to show up in
[32:05] the consistency of earnings which are
[32:07] going to be there because the prices are
[32:08] going to be higher. So from energy to
[32:10] intelligence, this is the physical stack
[32:12] underneath. And again, exactly what I
[32:14] highlighted. All of these things have
[32:16] shortages.
[32:18] All of these things are issues. That's
[32:19] where you want to focus your attention
[32:21] for the next at least decade. Uh which
[32:24] is very different than the prior decade.
[32:26] Uh all of these different components. So
[32:30] as you go through this just think about
[32:32] the fact that we are exactly in the
[32:33] hardware stage and we're in the very
[32:35] early innings and that's because chatbt
[32:37] and the data center buildout was the
[32:39] main story. Data centers are going to
[32:41] continue to get built but we're now
[32:43] going to be putting intelligence into
[32:44] everything and that's the difference
[32:45] between an AI factory which is more
[32:47] about the inference side and a training
[32:50] model which is more about uh getting
[32:52] better AI models. So uh supply side
[32:56] constraint he's describing the physical
[32:57] bottleneck of intelligence which is
[32:59] where we're at. If the infrastructure is
[33:00] not built, intelligence simply cannot
[33:02] exist at scale. So it has to be built
[33:04] and the money will be spent to have it
[33:06] built. I thought the most important um
[33:09] presentation with the most information
[33:11] was by anthropic head Dario Modi. And
[33:14] basically again to show you I bring in
[33:17] the transcript. It's about 30 minutes. I
[33:19] run the hedge funds analyst skill on the
[33:21] transcript. It gives me about a a 12page
[33:24] output giving me a bunch of different
[33:26] components. I'm not going to show it
[33:27] again for anyone who wants to get a
[33:30] presentation on this. I will show you,
[33:32] but it creates a PM research memo
[33:34] directly from that transcript. Goes
[33:37] through all of the things he says and it
[33:39] leads into
[33:41] cognitive ability is doubling every four
[33:43] to 12 months. Concrete evidence point
[33:45] lead product team for clog code hasn't
[33:47] written code manually in two months.
[33:49] Again, new information all been written
[33:51] by Claude. Competitors have gone
[33:54] consumer oriented. So this is obviously
[33:56] directly related to open AI and I think
[33:59] this was one of the more important
[34:00] things. While anthropic focuses on
[34:02] enterprises which are adopting now right
[34:04] now and they have the money and they
[34:07] have the productivity, the people that
[34:09] are focused on the consumer side are
[34:12] spending more money than the people in
[34:14] this case anthropic on the enterprise
[34:15] side. And what Dario was basically
[34:17] saying is I think enterprise business
[34:21] described as more stable than consumer
[34:22] with better margins. So the heavy
[34:25] spending the aggressive spending is
[34:27] being done by open AI and anthropic is
[34:30] getting more likely of getting business
[34:32] and more stable because the adoption
[34:34] continues to grow. In terms of China
[34:36] competition for his business, I have
[34:39] almost never lost a deal to a Chinese
[34:41] model. So as the enterprise stuff goes
[34:43] out and as coding is gone, I find it
[34:45] very unlikely that anyone is going to
[34:47] compete at this point with with
[34:49] Anthropic. They have made this point and
[34:51] I think that's important because right
[34:53] now the valuation of the two companies,
[34:55] Anthropic is about a 50% discount to
[34:57] OpenAI and I think OpenAI with new lows
[35:00] going on right now for Oracle. I think
[35:02] the Open AI story could be an issue as
[35:04] as as the year goes on. Uh at least for
[35:06] a scare and at least for people
[35:08] questioning whether OpenAI is in
[35:10] trouble. Um, maybe they make some
[35:12] headway uh with their new model. I I I
[35:16] find it hard to believe that it's going
[35:17] to be that big of a differentiator and
[35:19] maybe their product that Johnny Ives is
[35:21] involved in by the end of the year ends
[35:22] up being something. But for right now, I
[35:25] think it's very very difficult. And he
[35:26] talks about it on the bubble risk. He
[35:28] separates the technological trajectory,
[35:30] meaning the models are getting better
[35:31] and better and they're going to continue
[35:33] to get better and better. But on the
[35:35] enterprise deployment bottleneck, the
[35:38] technology is capable of is probably 10
[35:39] times what the enterprises are able to
[35:41] deploy. This is going to be a major
[35:43] story as I finish this up for all of you
[35:44] out there running businesses that are
[35:47] trying to figure out how to incorporate
[35:48] it. Uh you have to think about it from
[35:50] what the bigger picture is and you have
[35:52] to get your employees to start using it
[35:53] immediately and the question is have you
[35:55] provided an environment that allows them
[35:56] to do that. Um that is where I'm
[35:58] spending most of my attention with
[35:59] people on at this point. the second and
[36:01] third order effects of what Dario said.
[36:04] Uh there's an enterprise deployment
[36:05] bottleneck. The labor market disruption
[36:08] is happening and will continue to
[36:10] happen. Consumer AI commoditization.
[36:13] Uh I agree. Most consumers still use it
[36:15] as a chatbot and I don't know what
[36:17] they're going to pay more money for
[36:19] makes it very difficult. the capital
[36:21] allocation stress. Again, companies are
[36:23] spending different amounts of money and
[36:26] the question is if you're getting models
[36:29] that are better, but your end client is
[36:31] a consumer and they don't know how to
[36:33] monetize them. Do we have a gap there
[36:35] between the models which are getting
[36:36] better, the compute which are spending
[36:38] tons of money on for the future? I think
[36:40] there's an air pocket here for some of
[36:42] these. Again, if OpenAI is not doing
[36:44] well, then I think you're going to have
[36:46] the entire ecosystem, which again is
[36:48] just going to have to build in the risk
[36:50] that they're all going to run into
[36:51] overspending. Even though I don't
[36:53] believe it's going to be an issue, and
[36:54] this is not a bubble, I do believe that
[36:57] there will be questions and uncertainty
[36:59] over valuations for these companies
[37:01] unless they're showing the revenue side.
[37:03] And remember, part of the inability of
[37:05] of getting the capacity uh uh the
[37:09] bottlenecks is to be able to get more
[37:12] cloud storage and more cloud dollars and
[37:15] and if you've got capacity issues, which
[37:17] you still have for all of the big
[37:18] hyperscalers, I think this is going to
[37:20] be an issue. Um the economic potential
[37:23] of AI is enormous, many trillions of
[37:25] dollars, blah blah blah. The key
[37:26] uncertainty is adoption speed, not
[37:28] capability. This is becoming a bigger
[37:30] and bigger thing. And I think his
[37:31] statement of AI is 10 times ahead what
[37:34] enterprises are currently able to deploy
[37:36] the ability for the companies to adapt
[37:39] which is where my business is really
[37:40] starting to um blossom has become an
[37:43] issue and you have to if you want to
[37:45] catch up as I get through this I will
[37:47] show that about 12% of companies are
[37:50] being successful with adoption in a good
[37:53] way meaning they're already seeing very
[37:55] very strong positive returns the gap in
[37:57] the numbers in terms of how much more
[38:00] successful they are versus the average
[38:02] is dramatic and it's really hard to
[38:04] catch up guys if you don't use it every
[38:06] day. The best analogy I can give is a
[38:09] golf swing. If you think you can go
[38:11] learn golf by playing once a week, going
[38:13] to the range and hitting one bucket of
[38:15] balls, you if you really want to get
[38:18] your game to a level that goes fast, you
[38:20] have to put in an enormous amount of
[38:21] reps and it has to be daily. Uh
[38:23] especially as you're older. And the
[38:25] problem is very few people are doing
[38:27] that. Uh I won't go through and read all
[38:29] the different uh timestamps here.
[38:32] Uh I also won't go through this again.
[38:34] He's talking about the same thing which
[38:36] is their focus on enterprise is winning.
[38:38] So the synthesis between Jensen and
[38:41] Daario. Jensen explains why intelligence
[38:43] is scarce. Dario explains why revenue is
[38:45] delayed. I think if you listen to these
[38:48] two, it's very difficult that we've got
[38:50] a hard ceiling on AI capabilities based
[38:53] on how much infrastructure we need to
[38:55] build to have it go forward. This is not
[38:57] just for better models because the
[38:58] models are going to get better from
[38:59] Blackwell. This is mainly for edge
[39:01] devices, for autos, uh for getting uh AI
[39:05] factories built for for these companies
[39:07] that have massive amounts of data, but
[39:09] it's all siloed and unstructured. Uh and
[39:12] then adoption is the bottleneck on the
[39:13] AI monetization. And so I think you're
[39:15] going to have massive spending, which is
[39:17] good for commodities, hoarding and
[39:18] everything. But I think in terms of
[39:20] seeing the companies pay lots of money
[39:21] to software companies or pay lots of
[39:23] money to the hype to the AI model
[39:25] companies, I just I think we're in an
[39:27] air gap here. And I think these
[39:28] companies are they're priced
[39:30] differently. Uh 30 PE companies for
[39:33] software and for things related to AI
[39:35] versus
[39:37] uh materials stocks as I showed, they're
[39:40] just mispriced relative to each other
[39:42] for the stability of the earnings going
[39:44] forward. uh Satcha Nadella spoke and he
[39:46] basically continues to talk about the
[39:48] fact that there's going to have to be an
[39:49] agentic world. You can go listen to him.
[39:51] He did one with Allin. And while all
[39:53] this was going on, while all of the big
[39:55] leaders uh for the AI models,
[39:59] let's say the bigger companies, I'm not
[40:00] throwing Zuckerberg in there, but where
[40:03] was Sam Alman? Um sending another
[40:05] signal. He needs more money. Uh he also
[40:08] announced that uh they'll be having ads
[40:12] which I agree this is just another sign
[40:15] that they are desperate for money at
[40:16] this point and even Demis who spoke at
[40:20] Davos said he was surprised that they've
[40:22] moved on ads. It definitely seems like a
[40:25] a money grab and Deep Mind said it has
[40:28] no plans for that. So the company that
[40:31] knows the value of ads better than chat
[40:32] GPT is not going to follow their lead.
[40:35] Elon
[40:36] spoke as well. Great interview again
[40:39] with Larry Frink. Um key points he
[40:42] brought up uh Tesla, SpaceX, XAI and
[40:46] robotics around a single mission
[40:49] maximizing the probability of future
[40:50] civilization. Uh he talked about
[40:52] abundance. He talked about aging. He
[40:56] talked again about the true bottleneck
[40:58] being electrical power. I think the most
[41:00] important thing that Elon talked about
[41:02] and the reason you have to keep
[41:03] listening to him is we are in an age
[41:05] where we need the engineers to basically
[41:07] figure out how to deal with scale and
[41:09] vertical integration which he's done. At
[41:11] the same time, I talked about this last
[41:14] year. This is the trigger point for me.
[41:15] I believe this is a major event. Tesla
[41:18] robo taxi is now driving in Austin with
[41:21] no safety monitor in the car.
[41:24] That is basically a humanoid on wheels.
[41:27] We have entered now this. We've entered
[41:30] the point. And if you're wondering,
[41:32] you're starting to get more and more of
[41:34] these. The first trip
[41:37] $4.31. I saw one there where it was
[41:40] basically, you know, 20% of the cost of
[41:43] a Whimo regard or of an Uber. Regardless
[41:46] of what your thought process is, the
[41:49] reality is setting in. And it's not just
[41:51] the state of Texas with Austin and Elon
[41:54] basically believing that it's time. But
[41:57] then Lemonade announced that it will
[41:59] offer a 50% rate cut for drivers of
[42:01] Tesla vehicles when FSD is steering
[42:04] because it had data showed red. This is
[42:06] a really big deal because it means
[42:08] they're effectively allowing you
[42:10] basically to pay for the FSD yourself by
[42:14] having your insurance cut. So, just a a
[42:17] historic time for robo taxis and anyone
[42:20] who doubted this, of which the majority
[42:22] of people on the on the investment world
[42:24] that I've talked to over the course of
[42:26] the last year, they doubt Elon. They're
[42:28] not ready for what's happening. We're in
[42:29] the hardware stage. I think Tesla's
[42:31] stock will have a phenomenal year. Now,
[42:33] Ray Kurszwell, the father of singularity
[42:36] in the modern era, uh gave an interview
[42:39] on moonshots. I highly recommend it. If
[42:41] you take what Ray Kurszswwell said about
[42:43] the future and Elon Musk, Ray Kurszwell
[42:45] is known for his predictions and first
[42:49] said AGI would arrive by 2029 and he
[42:52] said that in the 1980s so he's had a
[42:54] good track record of approximately when
[42:56] everything was happening. He's had some
[42:57] misses but when you're trying to
[42:59] forecast 10 to 20 years out um he has
[43:01] been amazing. So exponential AI
[43:04] acceleration is real and it's underway.
[43:06] AGI is imminent. Uh abundance is the
[43:09] dominant outcome. This gets into a lot
[43:10] of disruption which I've talked about.
[43:12] Abundance is the major theme. You want
[43:14] to right now short any company that
[43:15] suffers from abundance and to get to
[43:17] abundance we have a scarcity side. You
[43:19] want to invest in scarcity. Anything
[43:21] that you invest in is it scarce or is it
[43:24] abundant? Um energy not algorithms is
[43:26] the binding constraint. They agree on
[43:28] human AI integration is in inevitable.
[43:30] Robotics is the next major phase and
[43:32] it's beginning in 2026 or 2027. Think
[43:35] robo taxis. Think optimists. optimism
[43:37] not fatalism is the way to think about
[43:39] this. Uh so again I would just focus on
[43:43] the fact that two of the smarter minds
[43:44] on the future are in agreement. Now data
[43:46] center projects this gets back again to
[43:48] another risk with the oracles the open
[43:50] AIS and all this stuff in terms of the
[43:51] data centers. You're going to be seeing
[43:52] this news happening more and more. 64
[43:54] billion dollars of data center projects
[43:55] have been blocked or delayed amid local
[43:57] opposition. Uh
[44:00] this data center thing for people who
[44:03] are not paying attention to it is
[44:04] definitely becoming a much bigger story.
[44:06] Google warns grid connection delays are
[44:08] now the biggest threat to the data
[44:09] center expansion. Uh in Memphis, Elon
[44:13] Musk's supercomput is getting push back
[44:16] from the pollution that is happening
[44:19] nearby. This is a big one and amid
[44:21] rising local push back US status
[44:23] counteration. These are all stories
[44:24] within the last couple weeks. Um, CEOs
[44:28] say AI is making work more efficient.
[44:31] This is from the AI daily brief. Uh,
[44:34] I've never shown this one. I listen to
[44:36] this podcast usually a few times a
[44:38] month. I think if you are a business
[44:41] leader in particular looking for how
[44:44] you're doing on adoption. I think this
[44:46] is really important um podcast to listen
[44:49] to. Uh here's the highlights. The uppart
[44:52] reviews new enterprise surveys from PI
[44:55] uh from PWC, workday and section and
[44:58] basically says that
[45:01] there's a widening gap between AI
[45:03] leaders and laggers. So the adoption
[45:04] thing is getting a headline that no
[45:06] one's adopting to it because if only 12%
[45:08] are adopting that means nobody's
[45:10] adopting. But when you go through the
[45:12] data and you realize that if you take
[45:14] the top 12% the gains they are seeing
[45:17] are dramatic.
[45:19] So if you're not using it, you're
[45:22] suffering and you're falling way behind.
[45:23] The biggest drivers of AI proficiency
[45:25] are leadership expectations followed by
[45:27] access to tools and a coherent AI
[45:29] strategy. ROI compounds with
[45:31] proficiency, reinforcing that enterprise
[45:35] integration problem, not an AI hype
[45:37] problem. And I could not agree more. It
[45:40] is staggering to me how little
[45:42] leadership allows their employees to use
[45:44] it, but they want the benefits that
[45:46] come. That is what the main story was in
[45:49] this. So
[45:51] the top 12% deploy AI the way previous
[45:54] generations deployed. AI is embedded
[45:56] directly into core workflows not layered
[45:58] on top of them. Again this is really
[46:00] important in terms of going through. AI
[46:02] sits inside operational systems. It has
[46:05] access to it. Firms with strong AI
[46:08] foundations are three times more likely.
[46:10] So basically you need to have an AI
[46:13] plan. You need to have an agentic plan.
[46:16] How does this I just copied what I just
[46:18] showed you and brought it back into
[46:20] GenSpark. How does what I just got from
[46:23] the AI daily brief in terms of the
[46:25] outcome from results from surveys align
[46:29] with what GenSpark which I showed you
[46:31] earlier which is an agentic thing to
[46:33] compete with
[46:35] standardized software flows SAS flows.
[46:39] It aligns almost perfectly. The
[46:40] description shared is essentially a
[46:42] clean articulation of GenSpark's core
[46:44] philosophy, not marketing fluff
[46:47] here are the way that they line up
[46:49] and this is the most important thing.
[46:51] Loser forms, AI as tool, rigid workflows
[46:54] again gets into enterprise SAS, human
[46:56] rework, static models, efficiency tools.
[46:59] Everything here is about agents and it's
[47:02] about what GenSpark is basically saying.
[47:04] saying GenSpark is building a type of AI
[47:06] system that only the top 12% of
[47:07] companies have figured out how to deploy
[47:09] internally and trying to product size
[47:11] it.
[47:13] Uh employees are three time 3.1 more
[47:16] likely to hire AI ready talent. Okay,
[47:18] this is why when the when my video
[47:21] series is launched and in particular
[47:23] most of the feedback I'm getting is
[47:25] starting to come from the parents and
[47:28] from college kids. Uh you have to be I
[47:30] native. The only way you can be IIA
[47:32] native is to use it all day. If you work
[47:34] for a firm and they don't let you use it
[47:35] all day, you're becoming less relevant
[47:37] and it is hard to catch up. You are
[47:38] trying to compete with golfers who are
[47:41] practicing every day for four hours, 5
[47:43] hours, 6 hours a day. You're trying to
[47:45] get good at golf by not having access to
[47:47] a golf club except when you're in your
[47:48] free time, which is much harder to do.
[47:50] That is why it is critical for people to
[47:52] be using this all day long. Schools, not
[47:54] letting kids use it all day long or
[47:56] telling them feing the fear of God into
[47:59] them, it doesn't help. So parents,
[48:01] students, continue to watch this, but
[48:03] also when the videos go up, I'll let you
[48:05] know. Uh I'm in the process hopefully
[48:07] very very soon. I say this every week,
[48:09] but every week that passes, we're one
[48:11] week closer to when this launch will
[48:12] happen. Final part on Bitcoin. Bitcoin
[48:15] and crypto market structure bill will
[48:16] pass very soon. The question is not is
[48:19] the bill going to pass. It's still the
[48:21] question of what's going to be in it. Um
[48:23] I wanted to just bring up the fact that
[48:26] this has been a big theme at the crypto
[48:28] event. People have talked about why
[48:30] Bitcoin's not moving like silver.
[48:32] Bitcoin is a$ 1.8 trillion asset. Silver
[48:34] has added$ 1.5 trillion in market cap in
[48:37] 23 days. I bring this up because when
[48:38] this stuff moves, it moves. And if you
[48:41] wait for it, you end up not getting
[48:43] involved. Um Brian Armstrong spoke about
[48:45] the Clarity Act. I'm not going to go
[48:47] through all the details, but he's saying
[48:49] the Genius Act was a clean bill that
[48:51] fostered uh innovation. This is the one
[48:53] that passed over the summer. The Clarity
[48:55] Act has become a vehicle for traditional
[48:57] banks to try and get out of some of the
[48:59] things that were in the Genius Act.
[49:01] I think the most important thing is just
[49:02] to get the regulatory framework. So,
[49:04] here's Bitcoin. Uh, now we've got a
[49:06] negative signal on the short-term side.
[49:08] I had said once we broke above 92, to me
[49:11] it was the buy side. Well, I've already
[49:14] been buying for all of this period here,
[49:16] including a little last week. It's gone
[49:17] down. The chart still looks horrible.
[49:19] Now, we've got a MACD sell signal on the
[49:21] daily. The weekly is about to cross. But
[49:23] the main point is in being here,
[49:26] we're just not getting any kind of lift
[49:29] that lasts. That will eventually change.
[49:32] And I just want to highlight in these
[49:33] final slides that number one, this is
[49:36] Ethereum versus Bitcoin. Ethereum to me
[49:38] is utility. Bitcoin is store value. The
[49:41] utility side is still doing well. As
[49:43] long as this chart looks this good, and
[49:45] again, we had a big rise up. Let's take
[49:47] this as five waves. We've got a 200 day
[49:49] moving average, which is pointed up.
[49:51] This is a good sign for the utility side
[49:54] which is the network effects which is
[49:55] stable coin. So the fact that Ethereum
[49:57] is doing well is good. Now if I want to
[49:59] have a comparison silver utility gold
[50:02] store of value. So it's not surprising
[50:04] that if we look back over the last year
[50:07] and a half until November silver verse
[50:10] gold was very correlated with Ethereum
[50:12] verse Bitcoin. They bottomed in April.
[50:15] Here's what's happened over the course
[50:17] of the last month. So, since November,
[50:20] silver's gone straight up. I bring this
[50:22] up because I do think we're going to
[50:23] have this type of move in crypto soon.
[50:26] Remember in Micron,
[50:28] I was able to buy a lot of Micron during
[50:33] this period in here. All of it. I just
[50:36] was buying it buying it because I
[50:37] believe that high bandwidth memory was
[50:39] going to be an issue. I feel the exact
[50:40] same way with Bitcoin. I actually bought
[50:43] some stock at the same level it was in
[50:45] 2018 before Micron went up from 70 to
[50:50] 400 in a period of less than 9 months.
[50:55] So you're left with another one. This is
[50:58] the basic resources stocks. Here's where
[51:01] we were in the summertime. Basically the
[51:03] same level we were in 2018 and now we've
[51:05] gone parabolic. Don't wait for the
[51:07] parabolic moves. Bitcoin and Ethereum
[51:10] and all of the crypto space will be
[51:11] there as well. And the reason for me is
[51:15] this is my proxy for Bitcoin. The orange
[51:17] line is Bitcoin. The white line is
[51:20] something I created as a proxy which
[51:21] involves all of the pieces to me that
[51:23] Bitcoin is involved with. It involves
[51:25] copper as the reflation side. It
[51:28] involves Q's as the innovation side and
[51:30] involves gold as the uh store value
[51:34] side. And you can see how strong a
[51:36] waiting of those is relative to Bitcoin.
[51:39] It will always move away from it, but
[51:41] eventually it follows the same path. And
[51:43] since I think gold, copper, and stocks
[51:46] are going to continue to move higher, I
[51:48] want to be in there. That's it for this
[51:50] week. Um, I will see you guys next