Bitcoin’s Next Move Depends On One Fed Decision

[0:00] But the AI thematic side or the the
[0:02] factor side, the vol is up towards 100.
[0:04] Bitcoin vol is still at 30. So, if
[0:07] you're putting an asset in a portfolio,
[0:09] right now you can have three times as
[0:10] much uh Bitcoin on a vol adjusted basis
[0:13] as you can AI. And that means that we're
[0:15] at a point where I think you should
[0:17] start seeing more and more people as
[0:19] they get more focused on Ethereum uh get
[0:22] in. Once we get above the 200-day moving
[0:24] average, I believe we're at the start
[0:26] [music] of something new.
[0:27] >> What's going on, guys? Today, we got a
[0:29] great conversation with Jordie Visser.
[0:30] In this conversation, we talk about all
[0:32] of the deleveraging happening in the
[0:33] stock market, why the mid-cycle slowdown
[0:36] in AI may be impacting your portfolio.
[0:39] We also get into what's going on with
[0:40] inflation and Kevin Warsh's recent
[0:42] comments. And then, last but not least,
[0:44] we go and we cover what's happening with
[0:45] Bitcoin, Ethereum, and the entire crypto
[0:47] industry. Here's my latest conversation
[0:49] with Jordie Visser.
[0:51] All right, Jordie, I thought we could
[0:51] start the conversation. It seems like
[0:53] there's a big unwind happening in the
[0:54] public market, specifically around a lot
[0:56] of the AI names. You started talking
[0:58] quite a bit about uh this like summer
[1:00] slowdown and how that may affect things.
[1:02] What what are you seeing?
[1:05] >> Yeah, it's it's
[1:06] uh
[1:07] this is probably going to be uh a fairly
[1:10] high-level um discussion. I'm going to
[1:12] get a little wonky with the leverage
[1:14] side, but
[1:16] I think
[1:18] um
[1:18] people are going to have to get used to
[1:19] this. So, what's happening is
[1:22] the S&P is making new all-time highs or
[1:25] equal weight S&P is making new all-time
[1:27] highs.
[1:28] Breadth is making new all-time highs.
[1:30] The AI names are going down, so a lot of
[1:32] people are sitting there seeing their
[1:35] Micron go down and their Marvell go down
[1:38] and their go through the list of all of
[1:40] the names. Every single name,
[1:42] Caterpillar, Modine, everyone that's
[1:44] been
[1:45] let's say going up fairly fast
[1:48] is giving back some of that. Now, some
[1:50] of that is related to just what I talked
[1:52] about, which is the AI mid-cycle
[1:54] slowdown is a second derivative change,
[1:56] and Nvidia went through this back in
[2:00] 2024. And I highly recommend everyone go
[2:03] back and look, because from chat GPT to
[2:07] about June of '24,
[2:09] Nvidia went up 12 times.
[2:13] So, very similar to let's say Micron
[2:14] going from 100 to 1,200.
[2:16] Since that period in June of '24,
[2:19] Nvidia's earnings have continued to
[2:22] grow. They've continued to dominate the
[2:24] AI trade, but in 2 years, they've now
[2:27] gone up 48%.
[2:30] Good returns. I mean, you're you're
[2:32] getting 20-plus percent returns, you're
[2:33] outperforming the S&P, but it's not the
[2:36] 12 times that happened. And that's where
[2:38] we are in AI right now for all of these
[2:40] names. Um meaning, most of them had
[2:44] three to 10 times type moves, especially
[2:47] the semiconductor names, which mean
[2:49] they've built in a lot of the next 2
[2:50] years. And I think what the market is
[2:52] going through is this second derivative,
[2:54] where when you look at the earnings of
[2:56] these companies, you're like, "Oh my
[2:57] god, they're up 400%, 500%." But then,
[3:01] a year from now, they're only going to
[3:02] be up
[3:03] 30%, 40%. So, you're decelerating. So,
[3:06] that's the first part is I think the
[3:07] market is coming to term with the AI
[3:09] mid-cycle slowdown. And the other thing
[3:11] I wrote about, which is the fireworks
[3:13] show.
[3:13] The The under-the-hood story is that
[3:16] there is absolutely a deleveraging
[3:18] happening. And I want to use, again,
[3:21] a bunch of um comparisons. So,
[3:24] silver and gold were up massively last
[3:26] year.
[3:27] Uh they had a drawdown the the biggest
[3:29] drawdowns they've ever seen in 30 years
[3:32] on different time horizons that were all
[3:34] within 20 days.
[3:36] The
[3:37] momentum factor for technology has seen
[3:40] an unprecedented move lower. I've posted
[3:42] an X about it. And what that does is the
[3:44] volatility rises, and whether it's a
[3:46] single-name retail person that is taking
[3:49] out leverage, whether it's Korean
[3:50] investors whose margins accounts have
[3:52] been closed or whether it's a long short
[3:56] multi-strap market neutral hedge fund or
[3:59] it's a systematic quant strategy. And
[4:01] we've seen reports and numbers released
[4:04] from the prime brokers of just very
[4:05] large losses in a short amount of time.
[4:08] And because the volatility has gone
[4:10] higher, whether it's risk parity,
[4:12] whether it's any kind of quant strategy,
[4:14] when the volatility goes higher, you not
[4:16] only have to reduce your risk at that
[4:18] point just from volatility beginning,
[4:21] but it also means that that position
[4:22] when it starts to come back cannot be as
[4:24] big. So, I think the best way for people
[4:26] to understand this, this was going to
[4:28] happen in a case like Micron by going
[4:30] from 100 to 1200 and back down to 800 as
[4:33] of early this morning,
[4:35] you've given back now depending on the
[4:37] name 30 to 50%. We've done enough based
[4:41] on what I think is there and I think
[4:42] we'll probably start to form some sort
[4:45] of a bottom here.
[4:47] But like silver, like gold, and another
[4:49] one like Bitcoin after October 10th,
[4:52] these things have not come back yet. So,
[4:55] their vol may have come down, but those
[4:57] things have been on the downside and
[4:59] they haven't gone back. The difference
[5:01] is with the AI trade, they were
[5:02] significantly above their 200-day moving
[5:04] average. And so, everyone hears this.
[5:07] Many of these names like Micron, they
[5:09] were unchanged for years. Micron was the
[5:11] same price it at liberation day lows
[5:14] that it was in 2017.
[5:16] So, a lot of these were breaking out of
[5:18] long-term ranges. They're way above
[5:20] their 200-day moving average.
[5:22] I think somewhere between this 30 and
[5:24] 60% retracement, they're going to hang
[5:27] in there and they'll start to move
[5:28] higher again.
[5:30] But I think the fireworks show is over
[5:32] and I think from this point on, some
[5:33] names will do well, some names won't do
[5:35] well, and you'll see volatility and
[5:37] gradually come down, but not quickly and
[5:39] not something that's going to just turn
[5:41] into a oh, that was no problem. It was
[5:43] something real and it's about leverage.
[5:45] >> Do you have a framework for identifying
[5:47] which are the ones that will do well and
[5:48] which ones won't?
[5:50] >> Well, what I'm going through and I I
[5:52] wrote a paper about memory. I think
[5:54] memory is the most important part of the
[5:57] AI trade.
[5:58] I am focused on companies that are have
[6:01] not yet see their step up function in
[6:03] earnings. So, in the case of Nvidia,
[6:06] when it peaked in 2024 from a 12 bagger,
[6:10] one thing started to happen, which is
[6:12] they started to only beat earnings by a
[6:13] little bit. So, the names that I'm
[6:15] focused on are the ones that I still
[6:17] think are going to beat earnings by a
[6:18] significant amount quarters to come, not
[6:21] just this quarter. Memory obviously blew
[6:24] away numbers. So far, we've seen Micron
[6:26] blow away numbers. We've seen Samsung
[6:28] blow away numbers. We've seen ASML blow
[6:30] away numbers.
[6:31] And all the stocks traded lower.
[6:33] So, the semiconductors are going to be
[6:35] fine, but memory is the place that I'm
[6:36] going to be focused on. And then I'm
[6:38] writing a a very long piece, or I did
[6:42] write it, which I'll release on my
[6:43] paywall this weekend, about Vera Rubin.
[6:45] And Vera Rubin is another area where
[6:47] there's certain names in there which
[6:49] haven't seen the step up function in
[6:51] earnings yet. That is going to be the
[6:53] things that drives it is you're going to
[6:54] need to beat earnings significantly. And
[6:57] if you're just beating them by a little
[6:58] bit, you can still produce 25% a year
[7:01] returns like Nvidia has, but you're not
[7:03] going to get back up into the 60s and
[7:05] 70s. So, I think everyone's going to
[7:06] have to now do some homework.
[7:08] That's the reason why I have my 100 name
[7:10] portfolio and I'll be spending the time
[7:11] on which names are better than others.
[7:14] >> Now, a lot of what has been driving the
[7:16] AI trade has been all of these
[7:18] bottlenecks. And a huge piece of the
[7:19] bottleneck was that there was two major
[7:22] American companies, maybe a third was
[7:24] trying to enter the fray in terms of
[7:26] Grok.
[7:27] But, we just got the Kimmy K3 release.
[7:30] And this is a Chinese open source model.
[7:33] It seems to be performing as well, if
[7:35] not better, than Fable or any of the
[7:37] latest closed source American models.
[7:40] Does that change any of your analysis if
[7:43] all of a sudden the winning models or
[7:45] these open-source models versus them
[7:46] being the American closed-sourced?
[7:49] >> No, and this is this is a good time to
[7:51] to bring this up. Um
[7:53] so I I wrote a paper this week on Brad
[7:55] Gerstner's interview uh on well, it
[7:58] wasn't an interview. He he was a guest
[8:00] on the All-In pod.
[8:02] And on that
[8:04] every answer he gave on this to me was
[8:07] very balanced. It kind of went through
[8:10] the pros and cons of all these stories.
[8:12] But the comment that I really focused on
[8:14] the most, which I think people have to
[8:16] understand
[8:18] this is the largest TAM
[8:21] that will ever exist, at least on planet
[8:24] Earth. We'll see if the the TAM in space
[8:26] will actually be something that happens,
[8:28] but that TAM is about intelligence. And
[8:31] intelligence will drive revenues that
[8:33] will drive compute demand. So here's the
[8:35] thing with open-source and and and with
[8:37] the frontier models.
[8:39] Right now, if I wanted to use Kimiko
[8:43] 3.0, I couldn't do it. Neither could
[8:45] you, unless you had the hardware.
[8:47] There's no way to use it. Meaning, could
[8:49] an enterprise use it? Yeah. So it's
[8:51] isolated right now to
[8:54] people that have enough hardware to use
[8:56] it. There's no way to even get the hard
[8:58] You have to have GPUs to be able to run
[9:00] it. Like this is a big model. That's the
[9:01] only way it competes with the frontier.
[9:04] If you want to use Anthropic, you just
[9:07] go onto the cloud and you go use it. So
[9:10] there is this element of well, who's
[9:12] going to be the people that use it? Will
[9:14] developers use it? Yeah, they probably
[9:16] will. Um they're trying to reduce costs.
[9:18] So AI-native businesses are going to
[9:20] have a huge advantage. Will an
[9:22] enterprise company in the United States
[9:24] of America use a Chinese open-source
[9:25] model?
[9:26] Maybe over time. Um but they're trying
[9:30] to get adoption from the humans plus the
[9:33] agents. Humans need co-work. Humans need
[9:37] all of these things and they get used to
[9:38] using a model. So, you just can't keep
[9:41] swapping out every time there's a new
[9:43] model. It's not that easy. And so, I
[9:46] think if they're, you know, if they're
[9:47] going to bet on something from an
[9:49] enterprise perspective, which is where
[9:50] the revenue is being driven, I don't buy
[9:52] into the fact that these Fortune 500
[9:54] companies are going to all of a sudden
[9:56] go, "You know what? I'm going to get rid
[9:57] of Fable 5 and Anthropic and I'm going
[10:00] to move to this Chinese model. We're
[10:01] going to download it and we're going to
[10:03] teach our employees how to use something
[10:04] which doesn't have the tools. So, the
[10:07] intelligence and the benchmarks may be
[10:09] there, which is fine for someone who
[10:12] knows how to code themselves and someone
[10:14] who's an engineer, but it's not that
[10:16] easy for the employees to actually go
[10:17] through. So, I think people have to be
[10:19] very careful about this. I hear this
[10:21] repeatedly. And go back to Brad
[10:23] Gerstner's side. This is the largest
[10:25] individual TAM. Will people use open
[10:27] source? 100%.
[10:29] Will most of the token usage be open
[10:32] open source? Yeah, it already is. But
[10:35] what we've seen so far over the last 6
[10:37] months, and they talked about this on
[10:38] the All-In, the actual
[10:42] enterprises are reducing their open
[10:44] source and going more towards this
[10:45] because that's the way their bureaucracy
[10:47] and everything works. And I think people
[10:49] have who haven't worked for a Fortune
[10:51] 500 company, which I have done,
[10:53] >> [gasps]
[10:54] >> the bureaucracy of getting decisions
[10:56] made, how long it takes to decide on a
[10:59] model, how long it takes to decide on
[11:01] software, and then once you do, you have
[11:02] a long-term plan. I don't see the
[11:04] Fortune 500 companies in the United
[11:06] States switching to a Chinese model
[11:08] anytime soon.
[11:10] >> So, there's two parts of this that I
[11:11] think are are interesting. Um, and these
[11:13] are just things that we're actively
[11:15] trying to figure out at Silvian, so I'm
[11:17] going to use that experience extrapolate
[11:19] where I think a lot of these companies
[11:20] are probably doing. The first is, um, we
[11:23] recently built and released uh, in the
[11:25] product a model router. So, the average
[11:28] consumer that's using Silvian has no
[11:29] clue that this is happening, but um,
[11:32] they can think of uh each query
[11:34] previously was going to kind of the
[11:35] highest intelligence, which also is the
[11:37] most expensive model to be able to go
[11:39] compute it. Which is great when you ask
[11:41] a really complex question, but if you
[11:43] ask what is the date, we probably don't
[11:45] need, you know, superhuman intelligence
[11:46] to be able to go and answer that
[11:48] question. And so the way the model
[11:49] router works is basically it can use the
[11:52] highest intelligence, most powerful
[11:54] model for really complex things, but it
[11:56] can direct queries to whether it's open
[11:58] source or kind of, you know, lower cost,
[12:00] lower compute uh intense uh models for
[12:03] simpler questions. And what I find
[12:06] interesting about the model router is
[12:08] one, we already are seeing a positive
[12:09] impact from it and it's, you know,
[12:10] putting some questions over to the to
[12:12] the uh less powerful model. What I don't
[12:14] know and we're actively trying to figure
[12:16] out how many models could you put into
[12:19] the model router? Does it make sense to
[12:20] do it with two or three? Would you do 20
[12:22] or 30? And I saw Sierra, uh Bret
[12:25] Taylor's uh company, I think that they
[12:27] have a blog post from their engineers
[12:29] where they have about 20 different
[12:31] models. It doesn't mean that's the right
[12:32] way to do it, but it does feel like in
[12:35] in a weird way um I agree that the open
[12:38] source model for the average employee is
[12:39] not going to be the thing that they're
[12:41] going to just, you know, put on their
[12:42] computer and all of a sudden start
[12:43] using. But I wonder if some of these
[12:45] companies start to create this, you
[12:47] know, harness and environment and the
[12:48] tools, and then they basically just use
[12:50] the models on the back end, and the end
[12:52] user doesn't even know what they're
[12:53] using. They just know that, hey, this is
[12:54] faster, this is, you know, more
[12:55] accurate. What do you think about that?
[12:57] >> I think that's the way it'll go. Um I
[13:00] don't think there's any question. I
[13:01] think the barriers to doing it right now
[13:03] are are hard. Um there will be companies
[13:06] that make it easier. So you're getting
[13:08] to the point of let's think of the
[13:10] models as outsourced intelligence.
[13:13] So let's just pretend like these are
[13:15] people. And you're like, okay, so I have
[13:17] access to these 20 people. Well, you're
[13:19] going to pay the highest price for the
[13:21] most sophisticated things. Where this
[13:23] shows up
[13:25] is that we're in the early stages, the
[13:28] very early stages of
[13:31] the Jevons paradox situation, where
[13:33] adoption is still extremely low. And we
[13:36] know that because there hasn't really
[13:37] been any job losses for any of the
[13:39] companies, and so far they just haven't
[13:41] been hiring people. I think when we get
[13:43] to that point, the downside to me, which
[13:46] I've always thought, is when people
[13:47] extrapolate Anthropic's ARR, and they're
[13:50] like, it'll be a trillion dollar what
[13:52] we're saying, which I agree with, and I
[13:53] don't buy into, they're going to be
[13:55] there. I believe AI will disrupt all
[13:59] businesses because intelligence will be
[14:01] commoditized. Now,
[14:03] that is not going to hurt Anthropic in
[14:05] terms of growing their ARR, but I have
[14:08] believed, and I still believe, that
[14:10] taking this pace, and just because they
[14:14] did it the last 3 years, 10 times, 10
[14:17] times, and they're on pace for 10 times
[14:18] again to get to a hundred billion ARR.
[14:22] People are extrapolating saying, well,
[14:23] next year they'll be a trillion. And I
[14:25] do think there's an element of Jevons
[14:26] paradox, but what you're talking about,
[14:27] and I think what people should see and
[14:30] and get,
[14:32] I don't think they're going to get to
[14:33] those numbers. Do I think they could get
[14:35] to 200 next year? Yeah. Well, that's a
[14:38] doubling, that's not 10 times. So, with
[14:41] everything in AI, I think what people
[14:42] have to realize, what I said with Micron
[14:45] and with all these places, I think it
[14:46] goes with Anthropic as well. Eventually,
[14:49] you reach a point where this was a surge
[14:51] out of nowhere, but eventually
[14:53] competition shows up. And you and I have
[14:55] talked about this, and I've said it
[14:56] before, and I'm saying it more actively
[14:58] as time goes on.
[15:00] Demis Hassabis wrote a piece on AGI.
[15:04] Demis Hassabis is the person that I
[15:06] think is the most level-headed of the
[15:08] people who I also think wouldn't say
[15:10] anything unless he was confident that we
[15:12] were getting there. I think Dario
[15:13] Amodei, I think Sam Altman, they run
[15:15] businesses where they're talking about
[15:18] it. Sundar Pichai, they Demis Hassabis
[15:20] doesn't run Google. He has talked about
[15:23] AGI since 2010 and forecasted that it
[15:26] would be around 2030. So, he's right on
[15:28] the number. Well, he's now convinced and
[15:30] saying
[15:31] the world is going to change over the
[15:33] next 3 years. And what it means for
[15:34] public companies and private companies
[15:36] and any company that exists,
[15:39] you don't know if your business will be
[15:42] disrupted by AI, but the probability of
[15:44] that happening rises every day that we
[15:46] get closer to AGI. So, we've talked
[15:48] about terminal value. We've talked about
[15:50] how you can't view Adobe and Salesforce
[15:53] past the next 3 years.
[15:55] I think you're starting to get in that
[15:56] with all companies, and especially with
[15:59] AGI, it even hits the physical
[16:00] companies, the energy companies, the
[16:02] Microns, the stuff like that, 3 years
[16:04] from now. So, I think that's all coming
[16:05] and that gets us back to This is the
[16:07] reason why I say Bitcoin is the only
[16:09] thing that has a moat, which is what's
[16:11] an asset that I don't have to worry
[16:13] about it being disrupted by AI,
[16:16] but I still have to invest. I think
[16:18] every single asset is going to run
[16:20] through a problem where AI is going to
[16:21] run into its moat.
[16:24] >> All right, boys. I need you to pay
[16:26] attention and listen. Most of you don't
[16:28] realize, after the age of 30,
[16:30] testosterone drops about 1% a year. And
[16:32] the bigger issue isn't just total
[16:34] levels. There's a protein called SHBG.
[16:37] Yeah, I'm in my science bag with that
[16:38] [music] one. SHBG.
[16:40] Then it actually binds your testosterone
[16:42] and it locks it up. So, even if your
[16:44] number looks fine, a lot of it isn't
[16:46] doing anything for you. That's where
[16:48] Mars men comes in. Mars men is designed
[16:50] to fix that. Clinically [music] dosed
[16:52] ingredients that help free up bound
[16:54] testosterone and support natural
[16:56] production, energy, and recovery. It
[16:58] works with what your body already makes,
[17:00] not against it. And the feedback that
[17:02] they're getting is consistent. People
[17:04] are reporting more energy throughout the
[17:05] day, sharper focus, better gym recovery.
[17:07] It's not one-time spike. [music] It's
[17:09] just simply a higher floor. Now, this
[17:11] product is made in the United States.
[17:13] It's [music] third-party tested and it's
[17:15] backed by a 90-day money-back guarantee.
[17:18] For a limited time, our listeners can
[17:19] get 50% off for life plus free shipping
[17:22] and they'll get three free gifts, too.
[17:24] You can go to men go to mars.com. Again,
[17:28] that's men go to mars.com for 50% off
[17:31] and three free gifts when you check out.
[17:33] After you purchase, they will ask you
[17:35] where you heard about them. Please
[17:37] support our show and tell them our show
[17:40] sent you. Again, men go to mars.com, go
[17:43] check out your testosterone and fix it
[17:45] if it's broke.
[17:46] >> What I also find interesting about the
[17:48] open-source models versus uh the
[17:50] American closed-source models is
[17:53] everyone is I I think excited about
[17:55] open-source allowing you to self-host
[17:57] and be able to actually use your data in
[17:59] your environment without having to worry
[18:01] about the zero data retention policies
[18:03] or, you know, the big model labs going
[18:05] to take your information and train their
[18:07] model or eventually compete with you.
[18:09] The downside, though, is a lot of what
[18:12] we're seeing with Soviet and the Chinese
[18:14] open-source models is it is not well
[18:17] understood how some of the Eastern world
[18:20] values are incorporated into the
[18:22] weights. And so, one aspect that I think
[18:25] is not a conversation really at all, but
[18:27] will become a conversation is this idea
[18:29] of like cultural weights. So, if you
[18:32] think of finance maybe is the best
[18:33] example, if you use the generalization
[18:35] of America and the West is capitalist
[18:37] and China and the East is more
[18:39] socialist, well, you obviously want the
[18:42] capitalist responses or that kind of
[18:44] capitalist weights into the model if
[18:46] you're in the West.
[18:48] We [snorts] just don't know, right? It's
[18:50] really hard. I don't see a lot of
[18:51] research there yet, but I do think that
[18:53] just as much as people are worried
[18:55] about, you know, what is the Google
[18:56] algorithm or what is the TikTok
[18:58] algorithm feeding the people, this idea
[19:00] of like these cultural weights is going
[19:02] to become a huge part and it goes back
[19:04] to your point about these large
[19:05] companies. There's like a technical and
[19:07] user experience barrier that they've got
[19:10] to overcome, but then also, you know, if
[19:12] you're I don't know, a customer service
[19:14] uh uh
[19:15] product and all of a sudden you start
[19:18] sharing, you know, kind of cultural
[19:20] values in response to customers,
[19:23] companies are probably going to get
[19:24] pretty uncomfortable very quickly. And
[19:26] so just feels like it's not only
[19:27] technical, there's also this other
[19:29] element that um is very uh
[19:32] not understood at the moment that we're
[19:33] going to have to do a lot of work on.
[19:35] >> Yeah, and this is why in the end I you
[19:37] you left one thing out uh or I don't
[19:39] know if it's intentional or you guys
[19:42] haven't thought about it with Silvia,
[19:43] but
[19:44] there is no way that we know for sure
[19:48] that if you host your own model on your
[19:50] own computer
[19:51] that you're safe from people having
[19:54] information from you. And let me let me
[19:57] let me let me make sure people
[19:58] understand why.
[19:59] Once you start going on the internet,
[20:01] you're exposed. So if you have this
[20:03] thing sitting someplace and it's
[20:05] contained, but you're never accessing it
[20:08] to the internet, we don't know what back
[20:09] doors or things could be in any of these
[20:11] models to get to get through. So I I've
[20:13] I've heard people talk about it. I I
[20:16] think the naive part is when people get
[20:18] into this, "Well, I'm not exposed." If I
[20:20] use Chem I K 2.5, which I have on one of
[20:23] my hardware,
[20:25] I'm not worried about it in the same way
[20:27] only because I don't have any of my
[20:29] important information on my business in
[20:32] there.
[20:33] But on me going into the internet or
[20:35] anything like that, if I use agents and
[20:37] they've got credit card information or
[20:39] wallet information, I just don't think
[20:41] we're there. So I everything that you
[20:43] said, I think it leads back to the fact
[20:46] that
[20:47] it is very difficult for a US Fortune
[20:49] 500 company to not use US frontier
[20:53] models and probably
[20:55] Google, Amazon, Microsoft as the cloud
[20:59] provider in some kind of
[21:02] you know, uh
[21:03] cyber related way that they can think
[21:06] about things along with their own cyber.
[21:08] And that's why I think the adoption will
[21:10] continue to go.
[21:11] Um the reason I said that on the AI
[21:13] side, I'm not worried about where we are
[21:15] and what stocks I want to I want to be
[21:17] involved in. I said it last week, I'll
[21:19] say it now, and I'm going to show it on
[21:20] the weekend video.
[21:22] We have barely entered the agentic
[21:25] world, barely. Consumer agents are
[21:27] coming. Right now, we're at enterprise
[21:29] agents, and they're barely being used.
[21:31] And that's why when you go back to Brad
[21:32] Gerstner's line, this is the largest TAM
[21:36] of intelligence that it could be. So, if
[21:39] we knew that the number was $20
[21:42] trillion, and that there were six model
[21:44] companies,
[21:46] they're all going to make money.
[21:48] That like in the end, they're all going
[21:49] to make money. So, getting worried
[21:51] about, you know, will they make any
[21:53] money? Will open-source take everything?
[21:55] It's not going to work like that. It's
[21:57] going to be spread out. There'll be lots
[21:58] of different players. Jevons paradox is
[22:00] in play. Compute demands are everywhere.
[22:03] Everyone's going to get a little piece
[22:05] of the pie. The question is, if you're
[22:06] pricing this stuff really low, like the
[22:09] Chinese models are, like Meta is,
[22:12] do you have other ways to make money?
[22:13] Meta's trying to get it through the
[22:14] consumer agent side, through
[22:16] advertising, through a bunch of things
[22:17] within side their world. Apple's trying
[22:20] to do this with the Gemini model, to get
[22:22] Siri in, to get volume going that way on
[22:24] the hardware side, as well as dominating
[22:27] in in terms of the personal intelligence
[22:29] side. Everyone's in a race to monetize
[22:31] this. I think it's a big pie. I just
[22:33] think the model providers are going to
[22:35] have an uphill battle from here,
[22:36] compared to the last 2 years.
[22:38] >> Another thing that happened this week
[22:39] that I think is having a big impact on
[22:40] the market is obviously the inflation
[22:43] report. It came in much cooler than I
[22:45] think people were expecting.
[22:47] Um we also got some comments from Warsh.
[22:48] What was your take on the data, and then
[22:50] his commentary afterwards?
[22:53] >> Well, I I think the most important part,
[22:55] um
[22:56] you know, what you and I talked about
[22:57] the last couple weeks, and I said
[22:59] probably about 3 weeks ago that
[23:02] one of the biggest surprises to me with
[23:04] this whole thing in the Strait of Hormuz
[23:06] is how people could literally be this
[23:09] wrong on something they had studied for
[23:12] I don't know, 20, 30 years. If this ever
[23:14] got shot down and we took down 20% of
[23:17] this, what would happen? And the fact
[23:19] that not only did crude oil trade down,
[23:21] and obviously we're trading back up now
[23:23] with rekindling and bombing and
[23:25] everything, but
[23:27] the second you stop bombing, we have the
[23:28] next ceasefire, what's going to the
[23:30] crude? I can't imagine it doesn't go
[23:31] straight back down. I mean, I I don't
[23:33] fall for the trap again, especially when
[23:37] inflation swaps this week actually went
[23:39] down. So, 2-year inflation swaps as of
[23:42] yesterday were actually lower than where
[23:44] we started the week. So, we're at the
[23:46] lows. So, I think when you go through
[23:48] the inflation data, number one,
[23:51] there's just no way to read it. Unless
[23:52] this month was a fluke on the core
[23:55] services side,
[23:57] every single inflation data but one, PCE
[24:00] core is the only one that's pointing
[24:02] upward. Everything else, whether it's
[24:04] median
[24:06] inflation,
[24:08] trimmed mean, whether it's the CPI core,
[24:11] whether it's sticky inflation, or
[24:13] whether it's trueflation, they've all
[24:15] come down while the PCE core is at the
[24:18] highs of the year. And so, I'm leaning
[24:20] towards the fact that the inflation side
[24:22] is a non-story for the rest of the year.
[24:25] Now, that being said, with the change in
[24:28] inflation, we reduced the Fed rate hike
[24:30] significantly. And I've talked about how
[24:32] I think this is a very big positive for
[24:34] the debasement trade, very big positive
[24:36] for crypto. We haven't seen it play out
[24:38] yet. We've seen Bitcoin act much better
[24:40] during a momentum unwind, which is not
[24:42] normal. But what Warsh said again, he's
[24:44] reiterated,
[24:46] he Basically, the headline for me is
[24:48] this is not a hawkish-dovish thing. This
[24:50] is a reform thing.
[24:52] He is fixated on a point that you, you
[24:55] know, brought up at the beginning of the
[24:56] year repeatedly, which is we can't trust
[24:59] any data from the government. He
[25:01] believes that the Fed has been very
[25:03] backward-looking, which I know I agree
[25:05] with. I think looking at anything at a
[25:07] time of AI, you know, 6 months later
[25:10] when the data gets revised for, you
[25:12] know, a long time, how can that be the
[25:15] types of things that you make
[25:17] decisions on? So, I think his message to
[25:20] everyone there is we're in a completely
[25:21] different time, and I think that aligns
[25:23] with AI. He has talked about the fact
[25:25] that AI is very disruptive. He's a big
[25:27] believer in AI. He's also a big believer
[25:29] in digital assets, and I don't think
[25:30] people should forget that. So, I think
[25:32] monetary policy, the old traditional
[25:34] way, is gone, and the most important
[25:37] thing for people is we've now taken the
[25:38] July rate hike to a 10% chance.
[25:42] So, if you believe what the um
[25:45] uh where the expectations are, which is
[25:47] one rate hike before the end of the
[25:48] year,
[25:49] you're kind of saying he's going to
[25:51] raise rates before the midterm
[25:53] elections, and
[25:54] if he was going to do it, I think July
[25:56] would be the time. I don't see him doing
[25:58] it in September, October.
[26:00] >> One of the aspects to me that
[26:03] um
[26:03] fuels uh
[26:05] like it has been exasperated
[26:07] in my understanding of all of this. Last
[26:10] Sunday, I went to the Jay-Z concert, and
[26:12] while we were there, we walked around
[26:14] and I interviewed a bunch of people,
[26:15] which basically just meant I put a
[26:16] microphone in their face and said, you
[26:17] know, why is everything so expensive?
[26:20] If you did knew nothing about economics,
[26:22] you'd never seen a data point in your
[26:23] life, you would think that every single
[26:25] person is just completely unable to
[26:29] afford anything. Groceries are too
[26:30] expensive, gas is too expensive. We did
[26:32] it the day after the news came out that
[26:34] rent in New York City uh the median rent
[26:36] just hit a brand new all-time high of
[26:38] like 5,250
[26:40] bucks or whatever, right?
[26:43] At the same time,
[26:44] they're at a concert where the tickets
[26:46] are, you know, 500 bucks, 700 bucks,
[26:48] whatever. So, like there's a little bit
[26:49] of irony in people who are at this event
[26:51] where the ticket prices are really high,
[26:52] talking about everything so inexpensive.
[26:54] So obviously they got money from
[26:55] somewhere they know somebody who's got
[26:56] money.
[26:57] But the second aspect [clears throat] of
[26:59] it was almost nobody was actually
[27:01] referring to inflation. They were just
[27:03] referring to the aggregate increase in
[27:05] price over the last five or six years.
[27:07] And so what I do wonder is, you know, is
[27:10] there good news on the horizon? It's
[27:12] just going to take a couple of years.
[27:14] Well, that doesn't help politicians at
[27:15] the midterm. But it does feel like so
[27:18] much of the discontent the kind of
[27:21] issues in the actual economy in terms of
[27:24] people's perspective of their real life
[27:26] experience. It's all aggregate price
[27:28] increase. Almost none of it has to do
[27:30] with, you know, the inflation that you
[27:32] and I would think about in terms of the
[27:33] difference between 3.5 or 3.9%.
[27:37] >> Yeah, I so I've thought a lot about this
[27:39] and and
[27:41] here here's my uh
[27:43] here's my take.
[27:46] If something bad happens to you,
[27:48] uh you know, you have a ski accident,
[27:51] you hurt your knee, you get into a car
[27:53] accident.
[27:55] For the next five years,
[27:58] that thing is pretty front and center in
[28:00] your brain.
[28:01] >> [clears throat]
[28:01] >> Um
[28:02] I believe what's happened is we can all
[28:04] remember what things cost before COVID.
[28:07] Like it's in our head. We remember what
[28:10] the car we bought cost. So every day
[28:13] someone has to just make a decision
[28:14] because cars have a some type of
[28:17] lifespan. You're like, "You know what?
[28:18] I'm going to go buy I'm going to I'm
[28:19] going to trade this one in and go buy a
[28:20] new car." And
[28:22] I mean, I've run into this recently
[28:24] where I went to go see because I have a
[28:25] Model S that I bought in 2021. Well,
[28:28] they're they're stopping the Model S.
[28:30] I'm actually using the full self-driving
[28:32] now on my car and it's an old version. I
[28:34] want the newest version.
[28:36] The car the cost is up significantly
[28:38] from when I bought it in 2021. So my
[28:41] brain remembers what I paid on
[28:43] something. And I think this is one of
[28:45] the things with you with with us with
[28:47] COVID, never does inflation go up that
[28:50] much in that short amount of time. Not
[28:52] in all of our lifetime. So, normally
[28:53] when inflation goes higher, it's this
[28:55] gradual process. When it happened in the
[28:58] '70s, it was driven by oil prices, and
[29:00] so oil went up, oil went down. This one
[29:03] wasn't that. This was Here you go,
[29:04] here's trillions of dollars, everything
[29:06] went up in price. Every single thing
[29:08] went up in price. And so, I think we all
[29:10] remember when eggs didn't cost
[29:12] $6, $7 a carton. We all remember when
[29:15] XYZ didn't cost this. And I think it
[29:17] takes a long time for that to wear off,
[29:20] but I think when you add in the
[29:21] polarization and the fact that people
[29:23] need someone to blame for this, and the
[29:24] fact that AI is standing in front of
[29:26] them where they don't feel like their
[29:27] job is ever safe the way it was, I
[29:30] really think I'm I'm getting to the
[29:31] point now I agree with you because
[29:35] I grew up in a house where my
[29:35] grandmother told me about the Great
[29:37] Depression all the time, about not
[29:38] having money for food, being thrown out
[29:39] of her house in in her teens to
[29:42] basically they couldn't afford She had
[29:43] to go out and get a job. She had to go
[29:45] do something. Um, and just remembering
[29:47] what the Great Depression was, this is
[29:49] not a We have the unemployment rate at
[29:51] near all-time lows. You can borrow money
[29:53] for anything right now. So, I I agree
[29:56] with you that I think this is more of a
[29:57] psychological inflation thing, and I
[29:59] think it has a lot to do with just what
[30:01] has happened in the last 5 years.
[30:03] >> Given Warsh's comments, um, you know,
[30:05] the first meeting that he really did a
[30:07] press conference, he didn't say much.
[30:09] Um, this was a little bit more robust in
[30:11] in commentary. Uh,
[30:12] do you change your mind at all about how
[30:15] you're thinking about him as the leader
[30:16] of the Central Bank or actions he may
[30:18] take?
[30:19] >> No, I actually
[30:21] I think he's the right person for the
[30:22] job because I do think the most
[30:24] important thing for this current
[30:27] Fed chair to be
[30:28] is young enough
[30:30] to understand the importance of AI,
[30:33] which I don't think Jerome Powell had.
[30:35] They didn't even pay attention to AI,
[30:37] honestly, until like the fourth third,
[30:39] fourth quarter of last year.
[30:41] Um, at least from what they said
[30:42] publicly. The second thing is, I think
[30:45] it needs to be more of a market person
[30:47] and less of an academic. And I think
[30:49] that's another important part that, you
[30:51] know, Warsh
[30:53] worked at Morgan Stan- Like Warsh has a
[30:55] background in capital markets in the
[30:57] same way that Bessent does. So, you have
[30:59] Bessent and him involved right now in
[31:02] the market at a very important time. For
[31:05] people not focused on digital assets,
[31:06] and again,
[31:08] th- this is you're going to be hearing
[31:10] me talk about this a lot more. I mean,
[31:11] our our lives are crossing over to a
[31:13] very important point starting there
[31:15] because,
[31:16] you know, everything I'm starting to do
[31:17] in my video is related to what's going
[31:19] to happen for me in the second the final
[31:22] four months of the year when I get out
[31:24] of Maine.
[31:25] I mean, I'm very focused on the rise of
[31:27] agents and what Bessent and Warsh are
[31:30] talking about, which is the importance
[31:32] of the United States of America
[31:34] being the leaders of the financial
[31:36] guardrails, not losing the
[31:40] the benefit that they admit has been a
[31:41] huge benefit for the world, which is
[31:44] reserve currency of the world, Swift,
[31:47] all of these different things. Digital
[31:49] assets are major part of it, and the
[31:51] growth that's happening in prediction
[31:52] markets, the growth that's happening in
[31:53] real world assets and tokenization. The
[31:56] news we saw this week on tokenization.
[31:58] All of this stuff is happening, and so I
[32:00] think Kevin Warsh is the right person at
[32:02] a major inflection point that very few
[32:05] people on Wall Street
[32:07] that are on the hedge fund world are
[32:10] fully embracing yet. I think they're
[32:12] realizing they have to, and with AI now
[32:15] being in a massive volatility spike, I
[32:18] think now we're going to start to see
[32:20] some money looking for new beta, and I
[32:21] think that's where Kevin Warsh is going
[32:23] to be a a big asset. So, I'm very happy
[32:25] with the reform that's going to happen
[32:26] on all these fronts related to AI.
[32:29] >> The second that inflation starts to cool
[32:31] and interest rate hike odds start to
[32:33] come down, I think a lot of people look
[32:35] at something like Bitcoin and try to
[32:36] understand how is that reacting and and
[32:39] what is that telling us about the
[32:40] future. Seems like Bitcoin had a pretty
[32:42] strong bid this week. Um what was your
[32:45] read on that?
[32:47] >> Yeah, I'm actually going to take it a
[32:48] different direction. I don't I mean, I
[32:50] know Bitcoin is is is more of your focal
[32:53] point. Um Ethereum's had the big move.
[32:56] Um
[32:57] it's up as of yesterday's close or let's
[32:59] say yes since it doesn't close. As of
[33:01] yesterday's equity close,
[33:03] uh it was up close to 20% month to date.
[33:06] Um and just so people understand,
[33:08] that would be if we finish the month
[33:10] with 20% the largest month or the best
[33:13] month since August of last year. Um
[33:15] Ethereum has outperformed Bitcoin. It's
[33:17] been the largest on a cross versus
[33:19] Bitcoin since then. The reason I care so
[33:21] much about Ethereum is because I do
[33:23] believe if AI agents are coming, if
[33:25] tokenization is coming, if all of these
[33:27] things are happening, this is an energy
[33:28] within side of the revenue side of
[33:30] crypto
[33:31] and we there's
[33:33] there's no there's no way to get around
[33:34] what's happening volume-wise. I think
[33:36] the Stripe
[33:38] bid for PayPal
[33:40] was a big deal, too. I have my crypto 40
[33:43] name equal weight basket, similar to my
[33:46] AI thematic one, meant to deal with the
[33:48] agentic world.
[33:50] And PayPal was one of the names in it.
[33:53] And so
[33:55] number one, I was happy I've done my
[33:56] homework and figured out which public
[33:58] companies cuz there's six of them that
[33:59] are in there including Robinhood,
[34:01] how they all fit in with this crossover
[34:03] with the AI agents. But I think Bitcoin
[34:06] is just hanging in there very well on a
[34:08] relative basis. It's obviously done well
[34:10] versus um AI. But the AI thematic side
[34:14] or the the factor side, the vol's up
[34:16] towards 100. Bitcoin vol is still at 30.
[34:19] So if you're putting an asset in a
[34:21] portfolio right now, you can have three
[34:22] times as much uh Bitcoin on a
[34:24] vol-adjusted basis as you can AI. And
[34:28] that means that we're at a point where I
[34:29] think you should start seeing more and
[34:31] more people as they get more focused on
[34:33] Ethereum
[34:34] getting. Once we get above the 200-day
[34:36] moving average, I believe we're at the
[34:39] start of something new. Until we are
[34:41] above there, I'm trading it actively.
[34:42] I'm trying to make sure that if a bottom
[34:44] is made here,
[34:46] it's there. I even put little started to
[34:48] buy a little bit of Micron this week in
[34:51] some of the semis. Micron, I have a
[34:53] position again. It's very small relative
[34:55] to what it was back then, and I'm buying
[34:57] it at higher prices than where I ended
[34:59] up selling it on average.
[35:01] But that's because the new news that has
[35:03] come out over the course of the last 6
[35:05] weeks and the consolidation and the the
[35:07] deleveraging that's happened,
[35:10] I feel more confident in terms of
[35:12] putting a little bit little bit of money
[35:13] in there, but I'm still much more
[35:15] heavily weighted towards crypto.
[35:17] >> When you think of the crypto allocation,
[35:19] is it just Bitcoin and Ethereum?
[35:22] >> For me at this point, it's Bitcoin,
[35:24] Ethereum, and and MicroStrategy. So I'm
[35:26] not
[35:27] I I'm I I don't think from
[35:30] I'm sure this will change as I spend
[35:31] more time in the space.
[35:33] Uh
[35:36] I've never been a big stock person as a
[35:38] macro person. It's been thematic. So
[35:40] even though Micron is in the portfolio,
[35:42] if you ask me why, it's because of AI.
[35:44] Everything I have in the portfolio to me
[35:47] is either an AI trade or it's a crypto
[35:50] trade. And so that's it. I I really Eli
[35:53] Lilly, it's an AI trade.
[35:56] Silver, it's an AI trade. Like anything
[35:59] I have in the port- folio is related to
[36:01] the stuff that I write about and I talk
[36:03] about. I believe we are at the
[36:04] convergence between AI and crypto and
[36:07] that all public companies will be
[36:10] disrupted the same way Adobe and
[36:12] Salesforce have been. Same way the
[36:14] hyperscalers have. They're they just
[36:16] rallied off the bottom. That's why the
[36:17] S&P's hanging around the all-time highs
[36:19] while AI trades lower.
[36:21] But over the last 9 months, they haven't
[36:23] done anything. So again, when you go
[36:25] through things, I'm just more focused on
[36:27] the AI infrastructure trade. Once the
[36:29] agentic side starts to accelerate, once
[36:32] AGI gets here,
[36:33] I believe the ROI C is headed towards
[36:36] private companies. I believe small
[36:38] private businesses are going to be the
[36:40] winners of this. The enterprises will
[36:41] find a way to reduce their expenses over
[36:43] the next 2 years. We are seeing that
[36:45] happen inside the public companies. If
[36:47] you haven't seen it so far, about 40 of
[36:49] the S&P 500 companies have reported so
[36:52] far, and just like happened in Q1, the
[36:54] surprise ratio right now, we're we're
[36:56] outperforming earnings. It's been 16% in
[37:00] terms of the beat.
[37:01] Um so, we're again headed towards
[37:04] another big, big, big month, uh big, big
[37:07] quarter for earnings, and a lot of these
[37:10] have happened in the banks and in
[37:12] insurance companies.
[37:14] >> Today's episode is brought to you by
[37:15] Blofin. If you're actively trading
[37:16] [music] crypto, then you already know
[37:18] the platform matters. Execution speed
[37:20] matters, liquidity matters, and
[37:21] reliability during volatile markets,
[37:23] that definitely matters. That's why more
[37:25] traders are starting to use Blofin. They
[37:26] built [music] a trading platform focused
[37:28] on fast execution, deep liquidity,
[37:30] advanced futures products, and a really,
[37:31] [music] really clean user experience.
[37:33] They don't overcomplicate things. You
[37:35] can trade hundreds of spot and perpetual
[37:37] pairs. They integrate directly with
[37:38] TradingView, and they built the platform
[37:40] for both active traders and people who
[37:41] are just getting [music] started.
[37:43] Security is a major focus, obviously.
[37:45] Assets are custody with Fireblocks. They
[37:46] provide proof of reserves, and they've
[37:48] built a strong reputation for platform
[37:50] stability during those high volatility
[37:51] [music] environments. And to celebrate
[37:53] the partnership with us, Blofin is now
[37:55] giving away $100,000.
[37:57] That's right. Blofin is going to give
[37:58] away $100,000 in deposit and trading
[38:00] rewards.
[38:01] >> [music]
[38:01] >> So, if you're looking for a better place
[38:02] to trade crypto, go check them out today
[38:04] using our link in the description.
[38:06] Again, that's Blofin, the trading
[38:08] platform built for serious [music]
[38:09] crypto traders. Go sign up. Pay
[38:11] attention, cuz they're giving away
[38:12] $100,000 in deposit and trading [music]
[38:14] rewards.
[38:16] Today's episode is brought to you by
[38:17] TikTok for Business. If you run a
[38:19] company, you are probably already
[38:20] spending a bunch of money on ads,
[38:22] search, X, maybe meta. But there's a
[38:24] platform your competitors are sleeping
[38:26] on and the window to move first is
[38:28] closing. TikTok has over 200 million
[38:31] monthly active users in the US alone and
[38:33] 51% of that reach is unique, meaning
[38:36] that those users aren't reachable on any
[38:38] other major platform where you are
[38:40] already buying. Unlike follower-based
[38:42] platforms, TikTok's algorithm rewards
[38:44] strong creative, not audience size. That
[38:46] means a well-made video from a brand no
[38:48] one's heard of can reach millions of
[38:49] relevant users for companies trying to
[38:51] acquire users cost-effectively, that's a
[38:53] fundamentally different model. TikTok
[38:56] for business gives you full-funnel tools
[38:58] built for performance. Whether you're
[39:00] looking to drive awareness, introduce
[39:01] your products, or convert new audiences,
[39:03] TikTok is becoming an important
[39:05] discovery engine for businesses and
[39:07] cryptocurrency alike. If you're a
[39:09] founder, a CMO, or growth lead and you
[39:11] haven't run a TikTok campaign yet, now's
[39:14] the time to change that. Click on the
[39:16] link in the description or go to
[39:18] anthonypompliano.slashthat.com.
[39:21] Again, click on the link in the
[39:22] description and go check out TikTok for
[39:24] business today.
[39:26] >> What are the areas that you're actively
[39:28] avoiding? We talk a lot about where
[39:30] you're putting capital. Is there any
[39:31] areas that you see maybe people talking
[39:33] about that you're more nervous about?
[39:36] >> I I still have no interest in in SaaS
[39:38] seat-based software. So, you know, a lot
[39:41] of people are trying to uh
[39:43] pick bottoms in in that. Um aside from
[39:46] that, I
[39:49] I think you're going to get a lot of
[39:51] sectors that are S&P 500 related. The
[39:54] things that I mentioned, um
[39:57] like I think between now and the end of
[39:58] the year at these levels, regardless of
[40:01] whether Micron sells off another 20%.
[40:05] I I don't see how the memory trade
[40:08] the earnings don't grow and how the S&P
[40:10] doesn't go higher. So, that's one of the
[40:12] things when people get worried about
[40:14] things, they extrapolate a fall in a
[40:16] certain stock market to something else.
[40:19] So, you've heard a lot of people talk
[40:20] about the debt side. Oh, this is
[40:22] unbelievable. The debt's going to go.
[40:23] And I want to make people feel very,
[40:25] very comfortable about this. Um one of
[40:27] the reasons I care so much about crypto
[40:30] is because of the budget deficit. We
[40:33] still have a deficit of 5 to 6%. So
[40:36] again, the government is spending 5 to
[40:38] 6% a year more than they're getting in
[40:41] receipts.
[40:42] The printing press is still happening.
[40:44] Meaning, this is good for nominal GDP.
[40:46] At the same time, the government is in a
[40:49] race against China. So, they're going to
[40:51] make sure this AI trade works. They're
[40:53] going to do whatever is necessary. If
[40:55] the funding markets
[40:57] If we see long-term rates go higher,
[40:59] they got to do something because they
[41:01] can't lose to China on this. And China
[41:03] just gives money and opens up things in
[41:05] a very different way. So, I'm very
[41:07] fixated on the CapEx that's going to
[41:09] happen.
[41:10] AI is going to continue to accelerate.
[41:13] The companies that people are worried
[41:14] about the hyperscalers, their debt to
[41:16] equity is It's nothing. Um so, unless
[41:19] their stock price falls violently, they
[41:21] have tons of ability to raise capital.
[41:24] They have $2 trillion of RPOs still
[41:27] sitting there. And again, so people
[41:29] understand, I've never understood over
[41:31] the next 2 years, if they took debt for
[41:33] all of the CapEx, we're talking around
[41:35] $2 trillion.
[41:37] Well, there's $2 trillion between
[41:39] Amazon, Microsoft, Google, and Oracle
[41:42] contracted for the compute.
[41:45] So again, we already have demand for
[41:47] this stuff. And these guys are going to
[41:49] have money to go through. So, I'm not
[41:50] worried about the debt. I'm not worried
[41:52] about that. So, I think people should be
[41:54] focused on the areas that are going to
[41:56] have a beta associated with them. And
[41:58] I'm going to stick with AI
[41:59] infrastructure.
[42:00] All the names that I've mentioned, plus
[42:02] the crypto side. So, I'm more focused on
[42:04] the beta. But if you want to just go by
[42:05] the S&P, I think you're going to get
[42:08] great returns in the S&P for at least
[42:09] the next couple of years as well.
[42:11] >> What about the Nasdaq?
[42:15] >> The Nasdaq has um
[42:17] a beta side associated with I think is
[42:19] is is good. You can you know, if you're
[42:22] going to buy S&P, I think you're
[42:23] generally going to get more on there.
[42:25] The area on on uh
[42:28] on the Nasdaq that I'm most interested
[42:30] in is is probably biotech at this point.
[42:33] Um
[42:33] the good thing about the Nasdaq
[42:36] it is
[42:37] obviously heavily weighted towards the
[42:38] Mag 7, which is a negative, but the
[42:41] reason that it still has outperformed is
[42:43] cuz it's not a pure cap weighted side,
[42:45] so you're not overly weighted towards it
[42:48] in there. And I think there are a lot of
[42:50] areas within side the Nasdaq which will
[42:52] do well, but I I I'm not sure they're
[42:54] going to be able to outperform the S&P
[42:56] in in a in a big way. I got to ask the
[42:59] question which
[43:00] I think was a good a good question.
[43:03] There's a lot of people have talked
[43:04] about the fact that this won't end until
[43:06] we have a true bubble.
[43:08] Um and they talk about it like the
[43:10] dot-com bubble where Cisco's trading at
[43:11] a 100 PE and that we're still in the
[43:14] early stages. And I want to I want to
[43:17] the people who are bullish, I want to
[43:19] take them a different direction on this
[43:20] cuz when I was asked the question, what
[43:22] do I think on it? And I said number one,
[43:25] demographics are completely different
[43:26] than the dot-com bubble. Um the money
[43:28] being made, you know, with the baby
[43:31] boomers in this, they were still
[43:32] working, they were still taking risk.
[43:35] Now they're passive investments, they've
[43:36] got advisers, they're playing on the
[43:38] golf course, they're not sitting out
[43:39] there gambling. Plus they've made so
[43:41] much money, they're not trying to make
[43:43] money the same way. There's obviously
[43:45] active traders that are doing it, that's
[43:46] fine. But then the other element that to
[43:48] me is very different
[43:50] and I think that's something people have
[43:51] to get involved in, it gets back to this
[43:54] concept that if in three years we have
[43:55] AGI, at some point the stock market is
[43:57] going to start saying,
[43:59] "Hey uh
[44:01] all businesses are going to be
[44:02] disrupted."
[44:05] I I don't think people realize what that
[44:07] means. One of the reasons that oil
[44:09] stocks basically really never go higher
[44:13] in a major way and see multiple
[44:14] expansion
[44:16] is because it's dependent on the price
[44:18] of oil. So if oil stays at $60 forever,
[44:21] or if natural gas stays at $3 forever,
[44:25] how is how's a company that generates
[44:27] their revenues based on the underlying
[44:29] asset moving actually going to get
[44:31] there? Well, I think if tokens are
[44:33] commoditized to zero, they don't hurt
[44:35] the model companies necessarily, but I
[44:37] think what they do do is they make it
[44:39] easier for AI native people to compete
[44:41] with all public companies because the
[44:43] inability for public companies to
[44:45] embrace AI
[44:47] and and and make it work is going to be
[44:49] much more challenging. So I only bring
[44:51] that up. I think the market is going to
[44:53] go through multiple compression. So I
[44:54] think earnings are going to be great. I
[44:56] think the S&P is going to grow
[44:58] less than earnings like we've seen so
[45:00] far this year, and I think that's going
[45:02] to be a trend at some point in the next
[45:04] 3 years. I actually believe that
[45:06] multiple compression story will become a
[45:08] much bigger story.
[45:10] >> Yeah, it's um
[45:12] it's going to be fascinating, I think,
[45:13] to watch this play out because
[45:15] part of what you mentioned earlier, I
[45:17] think, is underrated. I see a lot of
[45:20] people trying to pick bottoms in SaaS
[45:22] software.
[45:23] And I think that you have held steady on
[45:26] a view that AI is going to have a
[45:28] significant impact on those businesses,
[45:30] and it's more of a structural change
[45:32] than a than kind of a sell-off that is
[45:34] somewhat temporary. Um
[45:38] and it does feel like AI now, like I
[45:41] don't hear anyone saying AI is not going
[45:43] to be valuable, or AI doesn't work,
[45:46] right? You know, 3 years ago I heard a
[45:47] lot of like, "Oh, I talked to the the
[45:49] chatbot, and it was stupid. It you know,
[45:51] it lied to me." type stuff, right? That
[45:53] that conversation's gone. Now what I
[45:55] hear a lot is like, "Who are the winners
[45:56] going to be? How big can can the winners
[46:00] be? How profitable can they be? What's
[46:02] the impact of Chinese open source, you
[46:04] know, like we're in the nuance now. And
[46:06] so I take that as everyone is convinced
[46:08] this is real.
[46:10] Um, but I think that the one area,
[46:11] [snorts] um, I still, but and I
[46:14] did some of the math, uh, with Sylvia
[46:16] this week, um, a very large portion of
[46:18] my portfolio is exposed to physical AI
[46:20] and robotics.
[46:22] And it just feels like, you know, the
[46:24] software side of AI has now we've gotten
[46:27] so into the weeds because everyone is
[46:29] looking there that the next big, you
[46:32] know, kind of move, um, of capital in
[46:35] within the AI trade will be that, uh,
[46:37] that robotics hopefully, you know, that
[46:38] at least that's what I'm betting on. Um,
[46:40] do you have any exposure there? Are you
[46:42] thinking at all about that the robotics
[46:44] stuff?
[46:45] >> So, first of all, um,
[46:48] in this past
[46:49] weekend's video,
[46:51] uh, I I put up a chart and I got I got a
[46:54] lot of good responses from people on it,
[46:56] which was
[46:57] and I don't think you and I have talked
[46:58] about this.
[47:01] When people, um, want to compare the AI
[47:04] situation,
[47:06] yes, the dot com bubble would come up a
[47:08] lot, but the other thing that comes up
[47:10] that I think it's both more recent, but
[47:13] I think people are thinking about it the
[47:14] right way is they bring up fracking and
[47:16] what happened to all of those companies.
[47:18] Now, the reason I bring it up here,
[47:22] oil demand
[47:24] at the end of the day is driven by
[47:26] people.
[47:27] And that's why it grows with nominal
[47:29] GDP. So, it's a function of how many
[47:31] more people are on the planet, how much
[47:32] money do those people have to spend? Are
[47:34] they using more energy? Like, that's
[47:36] pretty much what drives the majority of
[47:39] energy price and that's why if you ask
[47:40] any oil person, so how much does oil
[47:42] demand go up a year?
[47:44] Well, what's nominal GDP? Nominal GDP is
[47:46] the answer.
[47:48] The problem with that is eventually with
[47:51] fracking, technology made the supply
[47:55] side go exponential. It actually
[47:56] outpaced what happened with the demand
[47:58] side. So, if the demand side stays
[48:01] linear and all of a sudden we get a
[48:02] technology that changes the supply side,
[48:05] well, then we end up in a situation
[48:06] we're in, which is no matter what seems
[48:08] to happen with oil, it ends up going
[48:09] back down.
[48:11] And we probably don't know,
[48:13] like with Hormuz, how much was shifted
[48:16] to natural gas. We have endless natural
[48:18] gas. What what All of these different
[48:20] things have clearly changed. The problem
[48:22] in what you're bringing up and the
[48:23] reason I'm so focused on the
[48:25] infrastructure side,
[48:27] we haven't reached a point yet where we
[48:28] can change the supply side of what's
[48:31] necessary for this build out, because we
[48:34] haven't been focused on it. That's why I
[48:36] say Micron was the same price in 2017.
[48:39] You know what companies don't do when
[48:41] their stock price doesn't rise for 7
[48:43] years? They don't build new capacity.
[48:45] They don't have it. So, that's why I'm
[48:47] convinced that on the memory side, when
[48:49] you haven't built the capacity for any
[48:51] of this stuff, it's not there. For
[48:53] robotics, it's the same thing. So, you
[48:55] have exponential demand for tokens.
[48:59] Tokens are driven by computers. They're
[49:02] driven by digital employees. They have
[49:03] nothing to do with human beings. The
[49:05] reason they're going parabolic is
[49:07] because you just put your computer on
[49:08] overnight and you say, "Okay, do this
[49:09] job for the rest of the day." What
[49:11] you're doing at Silvia, if you go
[49:12] through your token usage, it's not just
[49:14] the cost that's going higher, your usage
[49:16] is going higher, too. As you grow more
[49:17] people, as you guys build more systems,
[49:20] as you have more agents, it'll just use
[49:21] more and more tokens. It never goes
[49:24] down. And so, until we get the supply
[49:26] side to move up on an exponential to
[49:29] match off the exponential demand, this
[49:31] is something we haven't seen before and
[49:33] this is why whenever I say it, I'm like,
[49:35] "Guys, you're missing the point." Like
[49:37] when consumer agents come, we're going
[49:38] to need a lot more tokens. When
[49:40] humanoids come, we need a lot more
[49:42] tokens. When FSD comes, we need a lot
[49:44] more tokens. So, the way I'm playing the
[49:45] robotic side right now is the AI
[49:47] structure trade. It's a semiconductor
[49:49] trade. What is Elon Musk doing for the
[49:51] robotics trade? He's building Terafab
[49:54] because we don't have enough chips. So
[49:56] eventually, and I'm sure I will start
[49:57] buying some of these robotics things
[49:59] because I do think the next trade is
[50:00] embodied AI because, like you said with
[50:02] software, I don't want to fight this
[50:04] battle with software names. It's going
[50:06] to continually be an issue.
[50:09] >> I I completely agree.
[50:11] I appreciate your time today. The
[50:12] audience loves hearing from you. We'll
[50:14] do it again next week.
[50:15] >> All right, bud. Have a good one.