Bitcoin está a punto de dispararse: la Fed se ha quedado sin herramientas

[0:00] I am convinced that all it
[0:01] takes to
[0:02] ignite the
[0:03] Bitcoin community is a
[0:04] technical tipping
[0:05] point. I think
[0:07] this is just the beginning, and we'll
[0:09] look back on the last
[0:10] two weeks as an indication of
[0:11] that.
[0:12] So, ladies and gentlemen,
[0:13] this week we have a
[0:14] great conversation with
[0:15] Jordi Visser.
[0:16] Jordi explains
[0:17] why he believes the
[0:17] last two weeks have been
[0:18] so significant for
[0:19] Bitcoin. He thinks
[0:20] we'll look
[0:21] back and realize
[0:21] that these were the most
[0:22] important two weeks in
[0:23] Bitcoin's history. We will
[0:23] also discuss what is
[0:24] happening with
[0:25] AI-based trading.
[0:26] Why did AI
[0:26] company stocks rebound so sharply
[0:27] ? Was
[0:28] this the bottom? Or does he think
[0:29] they will fall even
[0:30] lower? Then we
[0:31] will move on to
[0:32] physical AI. What's
[0:33] happening with
[0:33] the construction of
[0:35] Tesla and SpaceX factories? And how exactly will
[0:36] this affect the
[0:37] valuation of these
[0:38] companies? Next, I
[0:39] will talk in detail
[0:40] with Jordi. I'll ask
[0:41] him some pretty tough
[0:42] questions, like
[0:42] how does he evaluate Google
[0:43] now that
[0:44] so much
[0:44] talent is leaving the company? Is
[0:45] he going to
[0:46] buy SpaceX? What does he
[0:47] think about the current
[0:48] rating? How does he
[0:49] feel about how
[0:50] stocks are traded? This
[0:51] and much more in
[0:52] this conversation with Jordi
[0:53] Visser. So,
[0:54] Jordi, you have a
[0:55] bold idea that the
[0:56] last two weeks
[0:57] have been perhaps the
[0:57] most important for
[0:58] Bitcoin in a long
[0:59] time. Tell me what
[1:01] you think is
[1:02] happening here.
[1:02] Yes, I'm just finishing up the
[1:04] article.
[1:05] Last week
[1:07] reminded me a lot of my time
[1:09] in Brazil in 1998.
[1:12] This will take some time.
[1:15] You can ask
[1:16] follow-up
[1:17] questions because you don't
[1:18] know where I'm going,
[1:19] but when you
[1:20] find yourself in a
[1:20] situation where, first of all,
[1:22] you have situational
[1:23] awareness, which was the
[1:24] main topic. And
[1:29] you have a month during which
[1:30] we have
[1:30] seen a
[1:31] historic increase in
[1:32] volatility in the
[1:33] stock market. Where
[1:37] AI company stocks
[1:38] trade with
[1:39] volatility that
[1:40] far exceeds
[1:41] even the highest
[1:42] levels Bitcoin
[1:43] has had in the last
[1:44] decade.
[1:47] Most of these
[1:48] large companies
[1:49] have volatility
[1:51] in the 80 to
[1:52] 120 range. And the losses that
[1:57] occurred showed that
[1:58] even in a
[1:59] fundamentally
[2:00] sound trade, where
[2:01] profits are
[2:02] growing rapidly and everything
[2:03] seems simple, these
[2:04] assets have fallen sharply
[2:05] because of one fund. So,
[2:08] that's one side.
[2:09] Think of LTCM in 1998.
[2:12] The second moment—Varsh
[2:13] passed his first
[2:14] test. We went
[2:17] into this meeting with a
[2:18] probability of about
[2:20] 50%, well, about 40%, that
[2:22] he would raise
[2:22] interest rates. And
[2:27] everyone who
[2:27] came out of there began
[2:28] to express criticism
[2:30] . He talked about
[2:31] long-term
[2:32] returns doing the
[2:34] work for him. And we
[2:35] were discussing this
[2:36] last week, where
[2:37] he literally said,
[2:38] “Well, just
[2:38] see what
[2:40] the market does. I follow the market
[2:41] .” Okay,
[2:43] long-term
[2:44] yields were rising,
[2:45] but something more important happened on Friday after the
[2:46] market closed,
[2:48] which I
[2:49] think
[2:50] got a little too much
[2:51] attention. The US coordinated with the Bank of Japan to intervene in the yen for the first time since
[2:55] 1998, the same
[2:56] year as LTCM.
[3:06] This has all the signs of
[3:07] desperation. And when you
[3:09] read about this, sometimes it's like,
[3:10] "
[3:11] Okay, we're going to
[3:13] bail someone out. We're
[3:14] going to lower
[3:15] rates. We're going to create a
[3:17] bailout
[3:18] for Silicon Valley Bank." They
[3:19] express themselves very
[3:20] clearly. You don't
[3:21] need to read between the
[3:22] lines. This is intended to
[3:23] stop panic. In
[3:27] that case, why would
[3:28] the US intervene
[3:30] with Japan in the
[3:31] dollar-yen exchange rate, which has
[3:32] been weak for some time
[3:33] , the Bank of Japan
[3:34] intervened, and everyone
[3:35] went back to the fact,
[3:37] "Oh, what else happened
[3:38] last week? We
[3:42] hit 20-year
[3:43] highs in the US 30-
[3:45] year bond
[3:46] . The bond market has
[3:51] been a red line
[3:52] for the administration for the last two years. It's
[3:55] happened in
[3:56] each of the three
[3:57] shocks that we've
[3:58] seen, including "
[3:59] liberation day"
[4:00] when they said we don't
[4:01] care about the
[4:02] stock market. Everyone just
[4:03] panics a little bit, but
[4:04] as soon as rates
[4:05] start going up,
[4:06] we have a problem.
[4:09] And remember, all of this is
[4:10] because
[4:12] we still have a huge
[4:13] deficit, and rates can't
[4:14] go up. So
[4:16] , on the one hand, he
[4:17] was advocating for
[4:18] short-term
[4:19] rates to go down. Well, now
[4:20] he's got a problem
[4:21] with long-term rates.
[4:22] We've got 30-year
[4:24] rates going up, and who is the
[4:25] biggest owner of
[4:26] 30-years? paper?
[4:27] Japan. So put
[4:29] all of that together, and then
[4:30] I'll see how the
[4:31] market reacts. Here it is
[4:33] , Monday morning,
[4:34] the market opens.
[4:36] I'm buying silver again.
[4:39] And I want to see how
[4:40] this all
[4:40] trades
[4:41] over the course of the week. I
[4:42] want to see if
[4:44] my
[4:45] intuition is correct: if you don't
[4:46] let bonds
[4:47] fall, you're actually
[4:48] going to be
[4:50] printing money. Then
[4:52] we get the
[4:53] refinancing announcement, where
[4:55] nothing changes,
[4:56] which to me means
[4:57] the game is on, meaning there's
[4:58] nothing we can
[4:59] do. And then
[5:00] today comes the
[5:01] labor market report
[5:03] . I just want to
[5:03] tell people, because I'm
[5:04] tired of hearing
[5:05] economists
[5:06] talk about a strong
[5:07] labor market while we
[5:08] just sit and
[5:09] listen. Wages are
[5:11] falling all the time.
[5:12] Today's report
[5:13] was weak across the board
[5:14] . It wasn't
[5:17] just weak:
[5:19] if you take the total
[5:20] payroll, which is the number of
[5:22] hours worked
[5:23] , the salaries, and the
[5:24] number of
[5:26] people hired or laid off,
[5:27] and you look at it over the
[5:28] last 6 months—not
[5:30] 3, not 4, but 6—we're
[5:31] at the
[5:32] lowest level since
[5:34] 2012 on a 6-month
[5:35] rolling basis,
[5:36] excluding the COVID period.
[5:40] So that suggests that
[5:41] we're
[5:41] running out of
[5:42] policy options
[5:43] to deal with this
[5:44] problem. But I'm
[5:46] sure of one thing:
[5:47] if you're not going to
[5:49] cut rates and you
[5:50] 're trying to tell the Fed from the
[5:51] Treasury position
[5:52] that you
[5:53] want to use the
[5:54] repo facility for
[5:56] Japan, this FEMA, and
[5:57] you want to get rid of
[5:58] the restrictions. While I'm
[6:02] sure they're
[6:03] trying to give
[6:04] Japan the ability to
[6:05] buy
[6:07] Treasuries, I think
[6:08] everyone needs to realize that
[6:09] we're at a critical point
[6:11] where the labor market is clearly
[6:12] weakening. I've talked
[6:14] about inflation
[6:15] swaps being well below
[6:17] 10-year rates. So
[6:18] wages are still
[6:19] falling, and
[6:20] today's
[6:20] hourly wage rate
[6:22] is now 3.2%
[6:23] annualized—it's
[6:24] certainly the
[6:25] weakest
[6:26] in this trend. I
[6:28] think we have a
[6:29] problem because AI
[6:30] agents are coming. So So far it's
[6:33] been a hiring issue.
[6:35] I don't think we're
[6:36] going to see mass
[6:37] layoffs, but we've
[6:38] reached a point where the
[6:40] market is pricing in
[6:41] more than one more
[6:42] hike before the end of
[6:43] this year. And
[6:45] strangely enough, right now as we're
[6:46] talking about this,
[6:47] the probability of a
[6:48] rate hike in
[6:49] September
[6:50] is
[6:51] 44%. A lot of that has to
[6:52] do with the
[6:53] inflation data coming out
[6:55] next week. But
[6:57] my reaction to just
[6:58] one thing is: gold
[6:59] is breaking out,
[7:00] platinum is breaking out
[7:01] , silver is breaking out
[7:03] . Bitcoin hasn't
[7:05] broken out yet, but that's
[7:06] because the
[7:07] Clarity Act is still
[7:08] hanging over it like a
[7:09] weight. But one thing
[7:11] that Bitcoin has done
[7:12] in the face of
[7:14] bad news is that the
[7:16] Clarity Act
[7:17] has gone from 50%, well,
[7:19] let's say 40% to 17%
[7:21] this year. And at that very
[7:24] moment, the strategy
[7:25] was selling Bitcoin. And it's
[7:27] not only held there
[7:29] , it's
[7:30] coming close to
[7:30] setting new
[7:32] short-term
[7:32] highs. So, I
[7:35] believe now that
[7:36] all it takes to
[7:37] get the Bitcoin crowd going
[7:39] is a technical breakout
[7:41] . I've said before that
[7:46] for me, when Dogecoin
[7:48] breaks the 20-day or 50-
[7:49] day moving average,
[7:50] when we see strong
[7:52] breakouts from
[7:53] retail
[7:53] investors, that's
[7:54] the beginning, and we'll
[7:55] look back at the last two
[7:57] weeks as a sign that
[7:58] we're overheating
[7:59] this AI game. We're
[8:04] spending a lot of
[8:05] money, but we're not
[8:06] creating the
[8:07] jobs we need, and the
[8:08] yen intervention
[8:10] shows that
[8:11] we're running out of
[8:12] tools. So
[8:13] we're heading very quickly
[8:14] towards a
[8:16] Federal
[8:16] Reserve shortage of tools. What do
[8:17] you think
[8:19] the Fed should do?
[8:19] I don't think they have a
[8:21] choice in the
[8:22] matter. As far as I'm concerned, you
[8:23] know, I
[8:25] think this train
[8:26] should keep going.
[8:29] There's no stopping this train.
[8:30] Lynn Alden's whole take on
[8:31] this, you know, well, Well, that's the
[8:33] question. When you
[8:34] still have a fiscal
[8:36] deficit of 5% to 6%.
[8:37] You've done as much
[8:39] as you can. You've
[8:40] had all this
[8:41] stock market growth and all that stuff that's
[8:42] happened. We're not
[8:43] creating
[8:44] jobs. I don't even
[8:46] know how to say it.
[8:47] We've even talked
[8:48] about this before, but the
[8:50] labor force participation rate is falling. And
[8:51] it's falling pretty
[8:52] fast. People need to
[8:54] stop listening to
[8:55] economists who
[8:56] say we have to
[8:57] raise rates,
[8:58] we have to do this or
[8:59] that, we have to
[9:00] get through this. We
[9:01] 're at the
[9:02] beginning of the era of AI
[9:03] agents. And if there's one
[9:04] thing that every person,
[9:06] including me—well,
[9:07] maybe not you, because you're
[9:09] not an optimist. You're kind of a
[9:10] super-optimist.
[9:12] Everyone underestimated
[9:13] how quickly AI is going to have an
[9:15] impact.
[9:17] When it starts to have an
[9:17] impact, it's going to
[9:18] happen very
[9:19] quickly. And I think the
[9:20] jobs situation
[9:22] is starting to escalate
[9:23] into something more serious now, where
[9:24] it's not just
[9:25] about replacing
[9:26] jobs. But I
[9:27] just don't think that
[9:28] when the
[9:29] labor force participation rate is falling,
[9:30] labor supply
[9:31] is shrinking, and
[9:32] wages are not
[9:33] rising, they have
[9:34] to—they have to
[9:35] wait, but they have to
[9:36] be able
[9:37] to control
[9:38] long-term rates
[9:39] , and that's very hard to
[9:40] do when
[9:40] nominal GDP is
[9:41] strong.
[9:42] One of the things that
[9:44] I keep
[9:45] thinking about in my
[9:46] head is the
[9:47] rise of these
[9:48] socialist
[9:48] policies, the increased
[9:50] role of governments at the
[9:50] local and
[9:51] state levels. If
[9:53] they continue to
[9:54] blow their
[9:55] balance sheets and
[9:55] run
[9:56] huge deficits,
[9:57] their strategy
[9:58] historically has been to
[9:59] go
[10:00] to the federal
[10:01] government to
[10:02] cover those costs.
[10:03] You know, if
[10:04] you look at it in
[10:05] microcosm, you can
[10:05] see New York
[10:06] going to the state, in
[10:08] Albany, and saying, "Hey
[10:09] , we need,
[10:09] I think, about
[10:10] $4 billion or
[10:11] something." Now
[10:12] they claim to have
[10:13] balanced the budget
[10:15] or some other
[10:15] nonsense, but
[10:16] in fact the smaller government
[10:17] went to the
[10:18] middle level and
[10:19] said, "Give me some
[10:20] money to help
[10:21] close this "Well
[10:23] , then the state
[10:23] runs a deficit, and
[10:24] they turn around
[10:25] and go to the
[10:26] federal government.
[10:27] So, oddly enough, it's not
[10:28] just federal
[10:29] spending, it's also the
[10:30] federal
[10:31] government's role as a "
[10:31] safety net"
[10:33] for all the miscalculations
[10:34] at the state and local levels
[10:35] that only
[10:36] accelerates the growth of the
[10:37] national debt
[10:38] and exacerbates this
[10:39] problem.
[10:41] Yes, I see what
[10:42] you're saying, and let me
[10:43] go back to the
[10:44] Japan-
[10:45] U.S. comparison, because there's a certain connection between the two
[10:47] .
[10:48] When I was in Japan in
[10:50] 1998, Japan was
[10:51] the only player in the
[10:52] debt market. Everyone
[10:53] was trying to play the
[10:55] JGB down. They
[10:55] called it the "
[10:57] widow killer." That's the
[10:57] whole point.
[10:59] Well, now we're
[10:59] 30-
[11:01] plus years ahead.
[11:02] The U.S. is in
[11:03] exactly the same
[11:04] situation as Japan. The
[11:05] debt-
[11:06] to-GDP ratio may not be at the
[11:07] same level,
[11:08] but now we have a
[11:09] situation where Japan
[11:10] has inflation and the U.S. has
[11:11] inflation. Is it
[11:13] unmanageable?
[11:13] Absolutely not. Is it
[11:15] inflation? Yes, it's
[11:16] above the
[11:17] 2% target or
[11:18] whatever number
[11:20] people want
[11:20] to operate with. So, for
[11:22] your consideration and for
[11:23] anyone listening, the
[11:24] reason why Bitcoin is
[11:26] important now is that
[11:27] the catalyst was
[11:28] when the US government
[11:30] stepped in and sold
[11:31] euros to buy yen.
[11:34] And you have to look at all the
[11:36] reasons why they
[11:37] did that. The
[11:38] stabilization fund
[11:39] —they can't
[11:40] use it anymore.
[11:42] We're running out of
[11:42] tools, and when the
[11:45] refinancing announcement comes out,
[11:46] we're really at a
[11:46] point
[11:47] where we have to
[11:49] reflate our way out of this. So,
[11:50] think of gold as a
[11:51] kind of trigger, and
[11:52] the way I like to
[11:53] look at the markets is to
[11:54] analyze what
[11:55] happened last
[11:56] week? Well, the
[11:57] macro market
[11:57] situation is telling
[11:59] us to buy gold,
[12:00] silver, and Bitcoin
[12:01] fits in there too.
[12:03] It's going to take some time,
[12:04] but I keep
[12:05] coming back to the same
[12:07] topic. AI
[12:08] agents—that's the main
[12:10] theme of this year.
[12:11] Next year's theme is
[12:12] consumer agents, which
[12:14] means transactions
[12:15] and volumes using
[12:16] stablecoins. This
[12:18] will happen regardless
[12:19] of the
[12:20] clarity act, and I
[12:21] think the
[12:22] clarity act is officially over for this
[12:23] year
[12:24] . We've heard that
[12:27] the SEC and CFTC are
[12:28] going to come out with
[12:29] measures right away to
[12:30] actually supplement
[12:32] this. Maybe that
[12:33] will be the
[12:34] catalyst that
[12:35] starts the process, and
[12:36] then everything will follow on its own
[12:37] .
[12:38] Today's episode
[12:39] is produced with the
[12:40] support of Galaxy One.
[12:41] Many people make
[12:42] their money work,
[12:43] but the question is
[12:44] how effectively
[12:45] it works? Galaxy
[12:47] One takes a slightly different approach.
[12:48] Everyone gets 3.5%
[12:50] per annum with Galaxy One Cash. And
[12:52] if you're making
[12:53] over 200k a year,
[12:54] that means you're
[12:55] rich. If you're
[12:56] making over
[12:57] 300k a year, you're
[12:58] likely a
[12:59] qualified
[13:00] investor, which
[13:00] gives you access to the 8%
[13:01] return
[13:02] guaranteed by Galaxy Digital
[13:03] . But the best part is,
[13:04] you can
[13:05] set up
[13:06] automatic
[13:06] reinvestment of
[13:07] the interest in your
[13:08] favorite stocks, ETFs, or
[13:10] cryptocurrency. You don't
[13:11] have to think
[13:12] about it every month.
[13:13] Right now, this
[13:14] second, Galaxy One
[13:15] is running a crazy
[13:16] promotion. You can
[13:17] get a
[13:18] $3,000 bonus
[13:19] if you deposit
[13:20] $10,000 or more in
[13:21] cash or
[13:22] cryptocurrency. If
[13:23] your money is
[13:24] n't working as hard as you are, it's
[13:25] time to change things up.
[13:27] Sign up and
[13:28] start earning
[13:29] today at galaxy.app/pomp
[13:31] to be eligible for a
[13:32] $3,000 bonus.
[13:33] I repeat, galaxy.app/pomp.
[13:36] Go and
[13:37] check them out
[13:38] today. We've seen
[13:39] that AI stocks have obviously taken a
[13:40] pretty big hit
[13:42] in the 2 weeks since
[13:43] we last
[13:44] talked about
[13:45] situational
[13:45] awareness. They're,
[13:47] you know, waiting for
[13:48] Ken Griffin
[13:48] to step in and take over
[13:50] this pretty
[13:51] bad situation, things
[13:52] happen. We
[13:53] got word that
[13:54] situational
[13:54] awareness is still
[13:55] relevant, they've
[13:56] got money left.
[13:57] Rumor has it that they
[13:58] just invested
[13:59] $400 million in
[14:00] another private
[14:00] company. Whether that's true
[14:02] or not, we'll see how
[14:03] that plays out
[14:04] . However, as soon as the
[14:06] Citadel news
[14:07] hit the news feed,
[14:08] AI stocks started to
[14:09] go up really fast.
[14:12] I have a friend who
[14:13] has invested heavily in
[14:14] energy companies
[14:16] , data centers, you
[14:17] know,
[14:18] hyperscalers and
[14:19] HPC-type companies. And
[14:21] he literally
[14:22] sent me a
[14:23] screenshot of his
[14:24] stock tracker. Everything
[14:25] went up 20-30% in one
[14:27] day. Right? So
[14:28] the market
[14:29] bounced back really fast. Is this a
[14:30] misstep or is this the
[14:31] next phase of
[14:32] growth that
[14:33] we should expect in
[14:34] these AI companies?
[14:35] No, I mean, we
[14:36] discussed
[14:38] this briefly last week. In
[14:40] my video, I
[14:42] basically asked, "
[14:43] Was this the bottom for AI?" "and
[14:46] I think it is.
[14:49] Again, you
[14:50] never know if there will be
[14:51] further
[14:52] asset dumps, but the way
[14:53] they have reacted
[14:54] this week, and more
[14:55] importantly, I
[14:56] use a lot of
[14:57] different metrics
[14:59] to gauge whether
[15:00] they were in a
[15:01] panic. We have
[15:03] clearly seen a
[15:04] significant dump
[15:05] . Because
[15:06] the volatility of these
[15:07] stocks is very high, I
[15:09] don't expect them
[15:10] to go back to their all-time
[15:12] highs overnight. But
[15:14] at the same time, I have a
[15:15] lot of these assets, and
[15:16] yes, not only are they
[15:17] up 10-20% in a day
[15:19] in some cases,
[15:20] but the positive
[15:21] momentum is
[15:22] continuing
[15:23] this morning. You're
[15:25] starting to see that
[15:26] good stocks
[15:27] respond well to good
[15:29] news. The area that
[15:31] I'm most focused on right now
[15:32] is
[15:33] optical technology. And
[15:34] the reason for all the attention
[15:36] to optical companies
[15:37] is because
[15:38] Vera Rubin is coming out. I have a
[15:41] solid stake in Nvidia
[15:42] because I think
[15:44] Nvidia is not only
[15:45] trading well and
[15:46] cheap, but
[15:47] Vera Rubin is coming out, and I
[15:49] think people looking for
[15:50] something new. The
[15:52] memory companies are going to be
[15:54] fine. I
[15:55] still, you know, I told
[15:56] you I bought Micron.
[15:58] It's a
[15:59] niche there. I'm not really
[16:00] building that position
[16:01] because I have stocks in the
[16:01] optical companies
[16:02] that I like
[16:03] better. I
[16:04] still like Marvell and
[16:05] those companies
[16:06] because I think they have a
[16:07] boost coming in from their
[16:09] earnings reports. But I
[16:09] think for those
[16:11] who have been able to get through
[16:12] this and have some
[16:13] cash, I would buy
[16:14] on any dip right
[16:15] now because I believe the earnings
[16:17] fundamentals are
[16:19] going to push them
[16:20] up. And I think
[16:22] overall, despite
[16:24] Michael Hartnett's report that the
[16:26] bull-bear indicator is
[16:27] signaling a
[16:28] strong sell-off, I
[16:29] want you to
[16:30] look at this
[16:31] chart from February of this
[16:32] year. We
[16:33] did see a slight
[16:34] pullback and then we went
[16:35] up. I will say right now
[16:36] that the lows for
[16:38] AI stocks are not only behind us, but
[16:39] by the end of the year, many
[16:40] of them will be reaching new
[16:41] all-time
[16:42] highs.
[16:43] So, look, I
[16:44] think there's a
[16:45] pretty interesting
[16:46] investment
[16:46] strategy: you could
[16:48] buy early and just
[16:49] hold. Let's take
[16:50] Micron for example,
[16:51] right? You invested
[16:53] in this asset early, you bought
[16:53] it, it went up a lot,
[16:54] and then you
[16:55] sold it. I think
[16:56] a lot of people are asking, "
[16:57] Why
[16:58] sell it?" Well,
[16:58] now you're back
[16:59] in the market. So
[17:00] you're essentially thinking,
[17:01] "It could get
[17:02] overvalued and
[17:03] pull back."
[17:04] You're re-opening
[17:05] a position, and I would classify
[17:06] you as a
[17:07] sophisticated
[17:08] investor when it
[17:09] comes to these things
[17:10] . Another strategy that I
[17:12] think
[17:13] people in the crypto world have mastered is that
[17:16] you
[17:17] start to build
[17:18] a position, and as
[17:19] the price goes up, your
[17:20] confidence increases,
[17:22] so you allocate more and more
[17:23] money. But you
[17:25] have to be careful
[17:26] because if it's a "bubble
[17:27] ," a lot of
[17:28] your capital
[17:29] is invested right at the
[17:30] peak. But with AI
[17:32] trading—I
[17:33] keep looking at
[17:34] the data, and the
[17:35] fundamentals are getting
[17:36] stronger and stronger.
[17:38] So it seems like
[17:39] the strategy
[17:40] of averaging the value of your
[17:41] investments in these
[17:42] assets over time,
[17:43] whether
[17:44] you buy on the
[17:45] dip or on the
[17:46] uptrend, has
[17:47] been fantastic so far
[17:48] over the last
[17:49] two to three years. I don't
[17:52] see any reason for that to
[17:53] change in the next two to
[17:54] three years unless
[17:55] something happens.
[17:56] existential
[17:57] situation.
[17:58] Okay. There's a big... I
[17:59] don't want to
[18:00] get back into
[18:01] crypto again, but
[18:02] we're at a major
[18:03] turning point right now. I
[18:04] really want
[18:06] people to realize:
[18:07] earlier this year,
[18:08] one of the things we
[18:09] talked about was
[18:11] that trading
[18:12] AI stocks was pretty easy
[18:13] . That means
[18:14] companies
[18:15] are beating
[18:16] expectations, earnings
[18:17] are up, all the news is
[18:18] positive, these stocks are
[18:19] skyrocketing
[18:21] and sucking all the
[18:22] capital out of Bitcoin.
[18:23] So, we talked about that
[18:24] . So,
[18:25] Bitcoin is not becoming
[18:26] something that
[18:27] we talk about at
[18:28] the beginning of the show. We're
[18:28] talking about things
[18:29] that
[18:30] most traders care about
[18:31] and that are actually
[18:32] moving. How do you
[18:33] talk about something that's
[18:34] essentially staying
[18:35] in the 10% range while
[18:36] everything else is going up? Well
[18:37] , that's changed now.
[18:39] And I've said it
[18:40] before, and I'll say it
[18:41] again: this
[18:42] AI part is over. And
[18:43] now for the fun part.
[18:45] If everyone was paying attention, the
[18:46] biggest
[18:47] negative for
[18:48] crypto that What
[18:49] keeps hedge funds
[18:50] from getting into them is
[18:51] that
[18:52] they're just going
[18:53] on momentum, not
[18:54] just price, but
[18:56] fundamentals. And
[18:58] the problem with cryptocurrencies
[18:59] is that they didn't
[19:00] have, and don't have, any
[19:01] fundamentals to
[19:03] fall back on. So what's
[19:05] going to change now is
[19:06] that we're going to have a
[19:07] transition. AI trading,
[19:10] fundamentals
[19:11] behind it 100
[19:13] % of the time, is going to last for the
[19:14] next 10 years.
[19:16] It's not going to stop. I
[19:17] don't believe in the
[19:18] bubble theory. I don't believe in the
[19:20] debt problem. I
[19:21] don't care about CDS. I
[19:23] don't care about any of that
[19:23] . These companies,
[19:25] the hyperscalers,
[19:26] are going to make
[19:27] money. It's probably
[19:28] going to happen faster than
[19:29] people think. I
[19:30] wrote an article about it
[19:31] this week. And I think
[19:32] we both listened to
[19:33] Gavin
[19:34] Baker's interview on "Invest Like the
[19:36] Best." He said it
[19:37] perfectly. He said he
[19:38] went to Silicon
[19:40] Valley. He kind of
[19:40] said, let
[19:41] me see if he'll give
[19:42] me at least one
[19:43] company has some
[19:44] negative news about
[19:45] something. And if
[19:46] you listen to that hour-long
[19:47] interview, he
[19:48] actually said that the
[19:49] fundamentals are
[19:50] even better than he
[19:52] expected going into
[19:53] it. So that's what
[19:55] cryptocurrency is going to have. When
[19:57] AI agents come along
[19:58] , and once we
[19:59] start valuing these
[20:01] things based on the
[20:02] revenue
[20:02] generated by
[20:03] transactions in the
[20:04] network effects of
[20:05] cryptocurrencies, I'm just
[20:07] telling you, the argument
[20:08] is going to be strong for what
[20:10] people are saying. I remember
[20:13] when Circle went
[20:14] public and people
[20:15] started telling me
[20:16] that it was going to replace
[20:17] Bitcoin, and why is
[20:19] n't it worth a trillion
[20:20] dollars? I started hearing
[20:22] that kind of stuff and I realized
[20:23] that we were in trouble in this
[20:24] space when I heard
[20:26] the fund guys
[20:26] trying to do
[20:28] the math. But the
[20:29] reason it's
[20:30] important is because
[20:31] once
[20:32] the math is
[20:33] tied to the
[20:34] cryptocurrency in the
[20:34] ecosystem, hedge
[20:35] funds are going to start
[20:36] investing there and
[20:37] everyone is going to be looking for
[20:38] something based on the
[20:39] Sharpe ratio.
[20:41] There's a high risk in trading AI
[20:43] . So when prices They
[20:44] go up 20%,
[20:45] guess what? They
[20:47] can go down 20%.
[20:48] When you run a
[20:49] hedge
[20:50] fund portfolio and you have an asset
[20:52] that moves 20% a
[20:53] day, you can
[20:54] only hold a
[20:55] limited amount of it,
[20:56] otherwise you're going to run
[20:57] into risk limits.
[20:59] So I think that's
[21:01] very, very good for
[21:02] cryptocurrency, and I think the
[21:03] tipping
[21:04] point will forever
[21:05] be July of
[21:06] this year, when we
[21:08] started to put pressure on the Fed
[21:09] . AI agents are putting pressure on the
[21:12] labor market. We
[21:13] needed to
[21:13] see how we were
[21:14] going to solve
[21:15] the deficit problem. We have a new
[21:16] Fed chairman.
[21:18] The question before him was, are you
[21:18] going to be
[21:19] hawkish or
[21:20] dovish?
[21:21] And everything we're
[21:21] learning is that
[21:22] even if
[21:23] they wanted to be
[21:24] hawkish, they
[21:25] couldn't. And now
[21:26] that the labor market is giving
[21:27] them a break, and
[21:28] we've had negative inflation for the time being
[21:30] . Let's see
[21:31] if that gives them the
[21:32] argument they need,
[21:33] and if not, then I think
[21:34] people are going to look
[21:35] back and be really
[21:36] mad at themselves, like
[21:37] Micron when
[21:38] it was below 100.
[21:39] Let's talk about
[21:40] physical AI and the
[21:41] reindustrialization of
[21:42] America. There
[21:43] were two big
[21:44] announcements last week that I
[21:45] think were really
[21:46] eye-opening.
[21:47] The first one was Hadrian. It's
[21:49] kind of a "factory
[21:50] as a service" or a
[21:51] factory of the future.
[21:52] They're doing
[21:54] manufacturing automation, and they
[21:54] seem to have three
[21:55] different locations,
[21:56] about 3
[21:57] million square
[21:58] feet in total. And their whole
[21:59] focus is how
[21:59] to use
[22:00] robotics and
[22:01] physical AI to
[22:02] reindustrialize
[22:03] the country. And they're
[22:04] starting with what you'd
[22:05] expect.
[22:06] They're making,
[22:07] you know, munitions and
[22:09] other kinds of very
[22:09] complex
[22:10] industrial
[22:11] equipment. But they
[22:13] raised $1.3 billion
[22:15] at a valuation of
[22:16] almost $8 billion
[22:18] . And the
[22:19] most interesting thing to me was
[22:20] who was involved.
[22:22] There's JP Morgan's Strategic
[22:23] Defense Initiative
[22:25] . There's Apollo,
[22:27] which did its
[22:28] first-ever
[22:29] venture capital investment.
[22:31] Right? I mean,
[22:32] these are really big
[22:33] organizations that are very,
[22:35] very experienced.
[22:36] They're not playing a
[22:38] lottery-like venture game
[22:39] where you just place a
[22:40] bet and hope
[22:41] something goes up. They
[22:43] almost view
[22:44] this type of
[22:45] industrial venture
[22:46] as a private
[22:47] equity vehicle. You know,
[22:49] it's like new private
[22:50] equity. So, I
[22:51] think that was one
[22:52] of the indicators. The second
[22:54] indicator is that
[22:55] Elon actually
[22:56] lifted the curtain on
[22:57] TerraFab and said, this is what
[22:58] we're building. And he
[23:00] said it's 100
[23:01] million square
[23:02] feet or more. He
[23:03] thinks it's going to be the
[23:04] most valuable building
[23:06] ever built in
[23:07] the world. And these images,
[23:09] the scale of this
[23:09] building—you just
[23:10] look at it and
[23:11] you realize, if it even
[23:13] comes close to being
[23:14] realized, it's
[23:15] obviously going to be
[23:16] extremely valuable.
[23:18] So how do you look
[23:19] at physical AI for the
[23:20] manufacturing sector
[23:22] and
[23:23] perhaps the
[23:24] impact on jobs
[23:25] ? Because they create
[23:26] jobs. I think
[23:28] TerraFab will create 3,000
[23:29] jobs. But there's going to be a lot more robots than people in
[23:31] 100 million
[23:32] square feet,
[23:35] right?
[23:38] Yeah, I'm... well, here's how I'll
[23:40] put it. When I say that
[23:45] there's an insatiable
[23:46] demand for
[23:46] computing
[23:47] power and that we're
[23:48] never going to
[23:49] have enough of it for
[23:51] what these people
[23:51] are investing in these
[23:52] seemingly relatively
[23:54] small amounts of money.
[23:57] Let's leave Lana
[23:58] aside. We
[23:59] need more AI, and
[24:00] for more
[24:01] AI, we
[24:01] need more
[24:02] computing
[24:03] power. So if you
[24:04] combine what
[24:05] Gavin Baker said, which is that
[24:07] whether it's open
[24:08] source or
[24:08] cutting-edge, you
[24:09] still need the
[24:10] same amount of
[24:11] computing. It doesn't
[24:12] matter. So
[24:13] computing is
[24:14] computing. If we
[24:16] thought of
[24:16] computing the same way we thought
[24:17] of copper, we'd be
[24:19] very interested in
[24:20] buying
[24:21] as much of it as possible. Well,
[24:24] you need computing for TerraFab. By Essentially,
[24:26] Elon Musk is saying that there's
[24:28] not
[24:29] enough
[24:29] computing
[24:30] power for my humanoids. So if
[24:32] I want to have billions of
[24:33] humanoids, if I
[24:34] want to get to Mars,
[24:35] we don't have enough
[24:36] intelligence to do that.
[24:37] I need
[24:38] intelligence in every
[24:39] machine, in every
[24:40] car, in every
[24:41] single thing. So that
[24:42] brings us back to
[24:43] the question of what
[24:44] comes first. Well,
[24:45] first we have chat
[24:47] agents, we have agents for
[24:48] writing code, we
[24:49] have chatbots, code
[24:50] agents. After code
[24:52] agents, well, then we
[24:53] move on to the
[24:56] allure of consumer
[24:57] agents and
[24:58] each of us having
[24:59] our own Jarvis in our
[25:00] pocket. Once we
[25:01] get to that,
[25:02] we have
[25:03] embodied agents. We'll
[25:04] actually have
[25:04] humanoids
[25:05] running around,
[25:06] but they all need
[25:07] intelligence.
[25:07] Cars need
[25:08] intelligence, everything
[25:09] needs intelligence.
[25:10] So in all these
[25:11] conversations, on the
[25:12] investment side,
[25:13] when people start
[25:15] getting ahead of themselves, like
[25:16] with SpaceX, I've
[25:17] noticed, I don't even
[25:19] know what means the
[25:20] TAM space. Like,
[25:21] what is it? This
[25:23] TAM space. I don't know,
[25:25] I... I don't even know
[25:26] if we'll have anything
[25:27] then. From an investment
[25:28] perspective, I want
[25:29] to focus on the
[25:30] next two to three
[25:31] years. The next two to
[25:32] three years are
[25:33] compute,
[25:34] compute. We
[25:35] need more
[25:36] intelligence, we
[25:37] need more
[25:38] compute, and that's why
[25:39] semiconductors
[25:40] are a
[25:41] public infrastructure industry
[25:43] ; they're
[25:44] early winners
[25:45] compared to
[25:45] private businesses,
[25:46] but these private
[25:47] companies are a signal
[25:48] of what's coming
[25:49] in 3, 4, 10 years.
[25:50] Everyone should pay
[25:51] attention to this, because this train is
[25:52] unstoppable because
[25:53] we need more
[25:54] compute
[25:55] for
[25:56] intelligence.
[25:56] What conclusions can you
[25:58] draw from this
[25:59] logic? So I'm
[26:00] an investor, I'm listening to
[26:01] Geordie. I think
[26:03] Geordie is the Wizard of
[26:03] Oz, and I hear
[26:04] him say that we need there's a
[26:05] lack of
[26:05] compute
[26:06] power. I hear
[26:07] about physical AI. I
[26:08] see what's happening
[26:10] with advanced
[26:10] models. How do I
[26:11] relate this trend
[26:13] and this belief to
[26:14] my portfolio?
[26:16] Should I just
[26:17] buy
[26:17] hyperscalers? Should I
[26:19] try to
[26:20] detect things like
[26:21] memory shortages and
[26:22] so on, or just
[26:23] tell me how you
[26:24] think the average
[26:25] individual
[26:26] investor might
[26:27] approach this? Not
[26:28] necessarily
[26:29] recommending
[26:30] specific stocks, but
[26:31] just how I might
[26:32] think about how I might allocate
[26:33] my portfolio given what
[26:35] you're talking about?
[26:36] Well, I know that there are people in
[26:37] our audience and
[26:38] listeners who
[26:39] can be divided into
[26:41] two camps on this
[26:42] issue. One is
[26:44] investors and the other is
[26:45] traders. In my
[26:49] video this weekend, I'm going
[26:50] to prepare a kind of
[26:51] lesson on what
[26:52] happened this year
[26:53] for everyone who heard me
[26:54] talk about the
[26:55] impending
[26:57] mid-cycle AI slowdown, and
[26:59] I'm going to explain that it's because
[27:01] people are rushing too much right now
[27:02] . We
[27:05] need the point where
[27:07] people start to
[27:08] question the growth. And that's what
[27:10] we've seen.
[27:11] At that point, stocks
[27:12] lose momentum and
[27:14] fall, which is what we've
[27:15] seen. If
[27:19] people are investors,
[27:20] that's why I
[27:21] created a 100-
[27:23] stock index that includes not
[27:24] only what we've
[27:25] talked about, but also bitcoin,
[27:27] silver, and Palantir. The things
[27:30] that didn't work. Those are the
[27:32] things that I'm
[27:34] thinking about, and I've talked
[27:35] about getting out of
[27:36] Micron and moving
[27:38] into silver and bitcoin.
[27:41] It may
[27:42] not have worked
[27:43] perfectly, but I'll tell
[27:44] you what I did in July
[27:45] , because
[27:46] they all went up in July
[27:47] , and AI trading went down. I
[27:49] cut it. So from a
[27:52] trader's perspective,
[27:53] when I say
[27:54] AI slowing down, I
[27:56] mean, we made a good
[27:57] profit on this trade
[27:58] , so we cut
[27:59] positions, we increase
[28:01] cash,
[28:02] we pause, and when
[28:03] we see signs of panic, and
[28:04] that's a situational issue
[28:06] awareness—we're
[28:07] ready for new
[28:08] investments. This is advice
[28:11] for traders, for
[28:12] those who want to get
[28:13] through this.
[28:14] The easiest way to
[28:15] gauge when
[28:16] things are getting
[28:17] crazy, and this is what I'll
[28:18] show you in the video, is the 50-
[28:19] day rate of change,
[28:20] which is how
[28:21] much
[28:23] the index has changed. At the peak, or the
[28:25] point where I was
[28:26] cutting
[28:28] Micron, my thematic
[28:29] portfolio of 100 names was
[28:31] up 50% in a 50-day
[28:32] period. So I
[28:35] cut it. Now I'm
[28:36] showing that over the
[28:37] last five years
[28:38] it's been peaking at
[28:39] about 20% and then
[28:40] coming back down to
[28:41] just below
[28:42] zero or about
[28:43] minus 10%. Guess where
[28:45] we are after
[28:46] that 50%? We're
[28:47] at minus 10%
[28:49] on a 50-day moving
[28:50] average. So this is the first thing
[28:51] traders need to do: you need to
[28:52] start
[28:53] using the
[28:54] rate of change indicator.
[28:55] The beauty of the rate of
[28:56] change is that it's a
[28:57] mechanism that
[28:58] combines time and price. What that
[28:59] means is that when you
[29:00] working in AI,
[29:01] where things are moving
[29:02] faster, you get
[29:03] parabolic
[29:04] growth and
[29:05] bubbles on one
[29:06] side, and then you get a
[29:06] crash to
[29:07] reset
[29:08] sentiment, and now you
[29:09] can get in
[29:10] at a much
[29:11] better price.
[29:12] As far as investors, I would
[29:13] n't worry about that.
[29:15] Here's what Elon Musk
[29:16] said on
[29:17] SpaceX's earnings call. And
[29:19] all of that goes back to the
[29:20] reason for Terafab
[29:21] . ​​He said that
[29:23] memory demand is growing
[29:25] 200% a year.
[29:28] Memory capacity is growing 20%
[29:29] a year. Here's what I'm
[29:31] saying: the reason
[29:32] memory demand is
[29:34] growing so much is
[29:35] because
[29:36] the demand is coming from agents
[29:37] . The more we
[29:38] let digital
[29:40] agents do their work,
[29:41] the more memory
[29:42] we need, because we
[29:43] need more
[29:44] processing
[29:45] power. This is a
[29:46] digital thing. My son is
[29:49] in an internship right now,
[29:51] he's a junior in
[29:52] college, and he
[29:53] sent me
[29:54] a message that I
[29:55] forwarded to you this
[29:56] week, which says, "
[29:57] My AI chief of
[29:58] staff every day creates
[29:59] new subagents."
[30:03] For its operation. So
[30:06] it uses
[30:07] more
[30:07] computing
[30:08] power every day, and
[30:10] so the demand for memory
[30:11] increases. For the 20% of
[30:12] capacity that
[30:13] Elon Musk is talking about, it's going to take
[30:16] some
[30:17] time to build Terafab. ​​Any
[30:19] capacity build takes
[30:20] time, because it's physics,
[30:21] people, regulations,
[30:22] all that stuff
[30:23] . AI agents are
[30:25] not constrained
[30:25] by regulations. So I
[30:26] just want to
[30:27] make sure people
[30:28] understand the essence of what's
[30:29] going on here,
[30:30] why supply and
[30:31] demand are completely out
[30:32] of balance, and that's how
[30:32] I would act in
[30:33] this situation. As
[30:34] an investor, don't
[30:35] worry about it:
[30:36] stay in the deal,
[30:37] sit back and be
[30:38] happy that you're getting an
[30:39] extra 15% above the S&P.
[30:41] Traders have to be
[30:42] very, very careful
[30:43] not to overestimate
[30:43] their strength when the pace of
[30:44] change gets too
[30:45] fast.
[30:46] Today's episode
[30:47] is presented
[30:48] by Arch Public. Arch
[30:49] Public has just expanded
[30:50] its
[30:51] agent trading platform
[30:51] beyond
[30:52] crypto assets. So
[30:53] listen up watch out, this is
[30:54] important. Now they are
[30:55] automating
[30:56] strategies for stocks,
[30:57] commodities
[30:58] and ETFs, and I think it will
[30:59] be awesome. You
[31:00] can automatically
[31:01] lock in profits
[31:02] when one market
[31:02] hits new
[31:03] all-time
[31:04] highs and
[31:04] redirect that
[31:05] capital to other markets
[31:06] where there is more
[31:06] opportunity.
[31:08] Whether you are moving
[31:09] capital into AI stocks,
[31:10] gold, investing in the S
[31:11] &P 500 or accumulating
[31:12] Bitcoin, Arch Public adds
[31:13] real
[31:13] discipline and
[31:14] automation to your
[31:15] investment
[31:16] strategy. In addition,
[31:17] they have launched a
[31:17] powerful new
[31:18] tool for
[31:19] tax
[31:19] optimization of losses.
[31:20] Since the crypto market is
[31:21] very volatile and does not
[31:22] fall under the
[31:23] “wash sale” rule, Arch Public can
[31:24] offset profits with
[31:25] losses without
[31:26] jeopardizing your
[31:27] long-term
[31:27] positions. This is exactly what
[31:29] every
[31:29] serious investor does.
[31:31] Institutional-grade automation
[31:32] that works across all
[31:33] major
[31:34] asset classes. No
[31:35] emotional trading,
[31:36] no missed
[31:37] tax
[31:37] opportunities, just
[31:38] smarter
[31:39] automatic
[31:39] execution of your
[31:40] strategies.
[31:41] Head over to
[31:42] archpublic.com right now.
[31:43] Contact their
[31:44] team, set up
[31:45] a meeting,
[31:46] connect
[31:47] your account if you need to—and
[31:47] see what
[31:48] automated
[31:49] trading can do for
[31:50] you. archpublic.com
[31:52] So, I have a few
[31:53] thoughts. Number one: You know
[31:54] how to become a
[31:55] long-term
[31:56] investor? Start
[31:57] as a trader, and then
[31:58] let the positions
[31:59] go against you. Then
[32:00] you just become a "
[32:00] sad owner" of the
[32:01] assets, right?
[32:02] So we have two groups.
[32:04] Sometimes some of them
[32:04] move from one
[32:06] group to the other. Number two
[32:07] : I had to do
[32:08] some research
[32:10] for another project this week. I
[32:12] remembered that Elon
[32:13] built the Colossus data center
[32:14] pretty quickly.
[32:16] But I couldn't remember
[32:18] how long, so when I
[32:19] checked, it turned out
[32:21] that Elon built it
[32:22] in 19 days. Normally it
[32:25] takes 4 years.
[32:27] So again, how do
[32:28] you calculate that?
[32:29] It probably took
[32:30] a couple of months, right? Not
[32:31] 19 days, but he
[32:32] compressed 4 years into a
[32:34] few months.
[32:35] Enough impressive. I
[32:36] mention this because
[32:37] Terafab would take, I
[32:38] don't know. If someone
[32:39] else built it,
[32:40] it would take 20 years in
[32:41] America. Right?
[32:42] Who knows how
[32:44] long it would take. If
[32:46] they built it in two
[32:47] years, it would probably be
[32:48] one of the greatest
[32:49] engineering
[32:51] feats of our
[32:52] lifetime. So that would be
[32:53] exciting. I don't know.
[32:55] Maybe it would take five years,
[32:56] right? Maybe it would take one
[32:58] year. I have no idea
[32:59] how long it would take
[33:00] , but I think we're
[33:01] putting
[33:02] our Michael
[33:03] Jordan out there and saying, "
[33:04] Hey, come on, get to work."
[33:06] Let's see how fast
[33:07] you can do it,
[33:08] which is pretty cool. I
[33:09] saw someone on
[33:10] the Internet saying, "
[33:11] Let's get this straight."
[33:13] Basically, our entire
[33:14] national
[33:15] strategy is one
[33:16] guy against the
[33:17] capabilities of the entire
[33:18] country of China. That's our
[33:21] bet. Like, this one
[33:23] guy can do it.
[33:24] But maybe
[33:25] we'll see. Another thing
[33:27] I want to talk about
[33:28] talk: every now and
[33:30] then, in what
[33:31] I do every day,
[33:32] I get
[33:33] the feeling that I'm
[33:34] looking into the future. And
[33:36] when I look into the
[33:37] future, I can
[33:38] go back and
[33:39] say to people, "Hey, I
[33:40] saw something, and this is something
[33:41] worth paying attention to
[33:43] ." And what I've seen
[33:47] in the last 2 weeks,
[33:48] that I can't
[33:49] forget, is that people
[33:50] overestimate the
[33:51] possibilities and the
[33:52] prevalence of
[33:53] advanced models in
[33:54] their use
[33:55] , and underestimate
[33:56] how much
[33:57] proprietary
[33:58] technologies, specialization
[33:58] ,
[33:59] specialized
[34:00] workflows, and all
[34:01] that stuff will eventually be
[34:03] used in
[34:04] manufacturing in different
[34:05] industries. I know
[34:11] you have some thoughts
[34:12] about the role of the advanced
[34:13] model and where
[34:14] it will be in all of
[34:16] this. So tell
[34:18] us how you think about
[34:19] advanced models, which
[34:21] historically have been the
[34:22] only option. Everyone
[34:23] has used them
[34:24] for everything. Now, I
[34:25] think you've
[34:26] evolved to
[34:27] the idea of, "Hey, advanced
[34:28] models are going to have a very
[34:29] specific but very
[34:30] important role in the
[34:31] future."
[34:33] Okay, I'll start by saying
[34:34] that people can
[34:35] read my I wrote an article
[34:36] on Substack
[34:37] this
[34:38] week about
[34:43] the abundance of intelligence
[34:44] and its impact on,
[34:45] say, what we
[34:47] invest in as
[34:48] investors. So I
[34:51] bring this up because of
[34:52] your question; I
[34:53] want people to
[34:54] look at it from two
[34:56] different perspectives.
[34:58] One is an AI startup, a
[35:00] company that does
[35:01] n't have bureaucracy. It's an
[35:03] entrepreneur. We're
[35:06] seeing a massive
[35:06] increase in
[35:08] companies that are in
[35:08] the millions of dollars in revenue that are
[35:10] run by just
[35:11] one person. My
[35:13] company fits
[35:15] that model. You can
[35:18] grow a
[35:19] business quickly, scale it up,
[35:20] and do
[35:21] things for
[35:22] people. It's an
[35:25] unprecedented
[35:25] lack of
[35:27] capital and
[35:28] people. So any
[35:32] business out there will go down the
[35:33] path that you're
[35:34] describing. They don't
[35:37] need Quad. They don't
[35:38] need ChatGPT. They don't
[35:40] need the big
[35:41] picture. The problem
[35:42] comes when you
[35:43] get to the
[35:44] larger companies, and that's what
[35:46] everyone
[35:47] listening to us is investing in. Public
[35:50] companies are just
[35:51] giant
[35:52] bureaucratic machines.
[35:56] And I was
[35:56] pushed out of that position.
[35:58] my peak careers. How did
[36:00] you guys and everyone
[36:01] watching manage to get
[36:02] to know me? I
[36:04] was in my early 30s. My
[36:06] best friend and the
[36:07] groomsman at my
[36:08] wedding was killed on
[36:09] September 11th. And I
[36:11] literally had to
[36:12] ask myself why I was
[36:13] working in a place where
[36:15] my job at
[36:16] 33 years old would be to
[36:17] fire people and
[36:18] sometimes hire them. This was
[36:20] supposed to be my job.
[36:22] I reached a management
[36:23] level at a very
[36:24] early age. Everything
[36:25] was happening so fast for me, and I did
[36:26] n't want to be
[36:27] caught up in it. When
[36:29] such companies
[36:29] make decisions,
[36:31] everything happens
[36:32] slowly.
[36:33] Hundreds of
[36:34] people have to check them. So will
[36:35] they allow
[36:37] some kind of
[36:38] open source model to be used there? No,
[36:41] they want to have
[36:42] protection from
[36:42] liability: "
[36:44] Hey, it happened because of a
[36:45] system error." We'll
[36:48] blame Anthropic. We'll
[36:50] blame ChatGPT. In
[36:51] the documents, they are the ones who
[36:52] take the risk
[36:53] if the model
[36:54] makes a mistake, if
[36:55] a hallucination occurs,
[36:56] or if something is
[36:57] sent incorrectly
[36:58] . There is no chance
[36:59] that they will quickly
[37:00] implement
[37:01] open source models in
[37:02] large corporations.
[37:03] So, on the one hand,
[37:05] you're right. I
[37:06] think small
[37:07] businesses will
[37:07] thrive because of
[37:09] this. On the other hand,
[37:10] now big business is
[37:12] gaining
[37:12] productivity,
[37:13] making profits, and I
[37:15] think this
[37:15] will continue for some
[37:17] time. But I think
[37:19] eventually there will be a
[37:20] problem when
[37:21] public companies
[37:22] start to suffer both
[37:23] from competition from
[37:24] AI-native
[37:25] businesses that are
[37:26] growing rapidly with better
[37:28] margins, and from the
[37:29] ability
[37:29] to compete with everyone
[37:31] else, including
[37:32] hyperscalers. I
[37:33] just think that in
[37:34] Anthropic and OpenAI, we're going to
[37:35] see
[37:36] revenue growth
[37:37] over the next
[37:38] year, let's say a year and
[37:39] a half. Most of
[37:40] this revenue will
[37:41] come from
[37:42] businesses and the
[37:43] research
[37:44] sector, because I
[37:45] believe they
[37:45] will be more
[37:46] aligned with the government on
[37:48] issues like cancer and
[37:49] things like that. And
[37:50] finally, you saw
[37:52] what happened at Google
[37:53] this week. Jeff
[37:54] Dean is gone. And at the same time,
[37:56] Demis Hassabis
[37:57] is moving on to another
[37:58] role. And I can't
[37:59] emphasize this
[38:00] strongly enough.
[38:01] Think of AGI as an
[38:03] AI-native world, which is something that
[38:04] Jeff
[38:06] Dean and Demis
[38:07] Hassabis are very excited about. And about Google
[38:10] as a company that
[38:12] needs to generate
[38:13] revenue from consumer
[38:14] products to
[38:15] cover capital
[38:16] expenditures. Even in a
[38:19] large company,
[38:20] talented people
[38:21] begin to strive to
[38:23] focus on
[38:23] things that create a
[38:25] personal legacy for them
[38:26] , instead of
[38:27] increasing the
[38:28] revenue of the business. And this
[38:29] means that talents
[38:31] will gradually leave these
[38:32] places. This all bodes poorly
[38:33] for public stocks
[38:35] over the next 5
[38:36] years in terms of
[38:37] growth rates. This is
[38:38] very, very good for
[38:39] entrepreneurs and small
[38:41] businesses. Are
[38:41] you worried about Google?
[38:43] Eventually, some
[38:44] key people left. I
[38:45] can give two
[38:46] arguments. One
[38:47] argument is that
[38:48] Google
[38:49] is changing. Jeff Dean
[38:50] , who was there for 27 years,
[38:52] you know, he left. Or
[38:54] Demis Hassabis. I
[38:55] mean, there are
[38:56] all these
[38:57] changes happening. This is a negative
[38:58] point. Why do your
[38:59] best people leave,
[39:00] change,
[39:01] step aside, etc.?
[39:02] I can make another
[39:04] argument: hey, I've
[39:05] been here for 27 years. We have
[39:07] succeeded.
[39:08] You know, I don't think
[39:09] anyone questions
[39:10] my skills. I
[39:11] want to try to
[39:12] do something on my
[39:13] own, and I'll take
[39:14] three of my
[39:15] friends with me, who are also
[39:16] extremely
[39:16] successful people. There's a
[39:17] real "
[39:18] gold rush" in
[39:19] AI right now, and if there was a time to
[39:20] go, if there was a
[39:21] technology to
[39:22] try to
[39:23] build something of our own,
[39:24] this is the time.
[39:25] So the question is: are
[39:26] they running away from
[39:27] something or running towards
[39:29] something? We'll probably
[39:30] never know
[39:31] the answer, but how do you
[39:32] look at it as an
[39:33] investor valuing
[39:34] Google?
[39:34] You just said it
[39:35] perfectly. This is
[39:36] uncertainty. So,
[39:38] as we've already talked
[39:39] about how it's impossible
[39:40] to value a company
[39:41] after 3 years, Google is
[39:42] no
[39:42] different. Um, that's the
[39:44] whole point. If
[39:45] you ask most
[39:46] people in Silicon
[39:47] Valley what
[39:48] the key to success is, it would
[39:49] be talent. Um, in a
[39:50] technology
[39:51] company you constantly
[39:52] have to
[39:54] innovate,
[39:54] come up with something new,
[39:55] and get through it.
[39:58] I think over the last 10,
[40:00] maybe 14 years,
[40:01] because of the protective barriers
[40:02] around Google, Apple, and
[40:04] Amazon, we've been lulled into a
[40:05] false sense of
[40:06] security, and
[40:08] their stock multiples have
[40:09] gone up significantly. So now we
[40:12] have accumulated a lot of debt.
[40:14] I mean, Google
[40:14] did
[40:15] another
[40:17] $25 billion bond issue this week.
[40:17] So I completely agree with everything you
[40:18] said
[40:19] . We don't know. So
[40:21] what if you don't
[40:21] know, shouldn't a
[40:23] company's multiples go down,
[40:24] especially if it's
[40:25] in a
[40:26] heavily indebted industry and you
[40:27] don't know anything? And I
[40:28] think that's exactly what's
[40:29] happening. That's
[40:31] why I said I want to
[40:32] hold stocks of
[40:33] companies that
[40:34] get cash
[40:35] from these companies that are not
[40:36] prone to excess,
[40:37] i.e. code. They
[40:40] actually benefit
[40:42] from the physical
[40:43] limitations that
[40:44] exist in the world. This
[40:45] game won't last
[40:46] forever, because all the
[40:47] investments in
[40:48] humanoids and
[40:48] robotics will
[40:49] eventually lead to the
[40:52] destruction of this business. But
[40:53] for now, for the
[40:54] next 18 months --
[40:55] and I'll always say "
[40:56] 18 months," because when I
[40:58] see signs
[40:59] that
[41:00] we're moving
[41:01] too fast in the
[41:02] physical realm in less than 18 months, I
[41:03] 'll sound the alarm
[41:04] or at least say,
[41:05] "Hey, this doesn't look
[41:07] good." Now, I'm
[41:08] saying the same thing as
[41:09] Gavin Baker: the
[41:10] fundamental data
[41:11] confirms that we're
[41:13] still in the early
[41:14] stages of
[41:15] consumer agents, and
[41:16] we have to get through this
[41:17] stage before we have
[41:18] humanoids
[41:19] capable of a real
[41:21] breakthrough.
[41:22] August, I think
[41:23] a lot of people in
[41:23] corporate
[41:24] America consider it to be a
[41:25] time of slowdown.
[41:27] Everything is slowing down.
[41:30] Um, August also
[41:31] tends to bring
[41:32] smaller returns in the
[41:33] stock market. Do
[41:35] you do anything differently
[41:36] due to
[41:37] seasonality?
[41:38] You know what? I, uh,
[41:40] because I believe that
[41:42] artificial intelligence
[41:43] speeds up everything
[41:44] tenfold compared
[41:45] to how it was before.
[41:50] Well, the worst year for
[41:51] momentum, in general,
[41:56] wasn't as bad as what
[41:57] happened in
[41:58] July. So, we see
[42:00] price fluctuations. So
[42:03] I think if
[42:04] you look at July as a
[42:05] whole year for AI
[42:06] trading, it was a
[42:07] terrible year. This was the
[42:09] worst year for
[42:10] momentum. So what
[42:11] I'm looking at is a run through of the
[42:15] ruins, as I wrote
[42:16] last week in
[42:17] a tip for people.
[42:19] Um, I just think there's a
[42:20] good opportunity, and
[42:21] I don't pay attention to the
[42:22] seasonality of the calendar
[42:24] the way, let's say, we
[42:25] did before. I
[42:26] discarded all the old
[42:27] models. I just
[42:28] think everything is
[42:29] moving too
[42:30] fast. And so, whether it's a
[42:32] bullish or
[42:33] bearish indicator or something
[42:34] else, I think if you
[42:36] try to
[42:37] use
[42:38] old models that
[42:39] have 40
[42:40] years of backtesting, I think the
[42:41] AI ​​world is destroying them all. The
[42:43] last question
[42:44] I have for you is about
[42:46] SpaceX. Apparently the stock
[42:47] went from $135
[42:49] to $200 something. And then
[42:51] they dropped all the way down to almost
[42:52] 100. I think I
[42:54] saw 108, 106—that was the
[42:56] lowest I've ever seen.
[42:57] Um, now they're back at
[43:00] 127-128. So,
[43:01] obviously, it's very
[43:02] volatile. They
[43:03] had a big
[43:04] stock unlock.
[43:05] Um, there's a lot of news
[43:07] around TerraFab and so on.
[43:08] You have a briefcase. Are you
[43:10] adding SpaceX to the
[43:10] portfolio, or how do you
[43:11] value, you know, a
[43:12] business that
[43:13] has very high
[43:14] ambitions, operates in
[43:15] different markets, has, you
[43:16] know, a great
[43:17] entrepreneur behind it,
[43:19] but also, I think, there's a
[43:20] question about whether it's
[43:21] overvalued?
[43:23] I honestly don't...
[43:24] Their computing
[43:25] business is incredibly
[43:26] valuable. Let's
[43:29] just go back
[43:30] to basics, like, uh,
[43:32] Elon built Colossus,
[43:36] and not only in
[43:37] record time, but
[43:38] he also did something
[43:40] , which is build the
[43:41] largest GPU clusters
[43:42] in history. Now he's
[43:45] going to do
[43:46] things with Bear Reuben. Um, I
[43:48] think what
[43:49] Dwarkesh was talking about two
[43:50] weeks ago, and in his
[43:52] post, what Gavin
[43:53] Baker reiterated
[43:54] last week or this
[43:55] week in his
[43:56] interview, is starting to
[43:57] become something that people
[43:58] are realizing, which is that the
[44:00] companies that
[44:01] have spent money on building
[44:03] computing
[44:04] power are going to have an advantage, because
[44:05] it costs a lot
[44:06] more than it used to,
[44:08] because they're all...
[44:09] nobody expected the
[44:10] pace of adoption
[44:11] to accelerate as much
[44:12] as it has.
[44:15] And I think the problem is that
[44:16] we're going
[44:17] too fast. So,
[44:19] instead of
[44:20] focusing on SpaceX,
[44:21] I'll say this: value your
[44:22] business by
[44:23] computing
[44:24] power, and you'll
[44:25] get space
[44:26] for free. And that
[44:28] means to me that
[44:29] every time the price
[44:30] goes back to
[44:31] $100
[44:32] because of the calculations
[44:33] and his ability
[44:34] to build it all up,
[44:35] there's a value that
[44:37] makes the business more
[44:38] predictable, and
[44:39] you get a call option on space
[44:41] for free.
[44:41] Great perspective on this
[44:43] situation. What
[44:44] will you talk about in
[44:44] your video this
[44:45] week?
[44:46] Well, I mean, I'll
[44:47] go into detail about
[44:47] a lot of the things
[44:48] we talked about. Um,
[44:49] an interview with Gavin
[44:51] Baker. Lots of
[44:52] details on AI
[44:53] trading, including
[44:55] why it definitely
[44:56] looks like a
[44:57] market minimum. So, last
[44:59] week it looked like a
[45:00] panic low.
[45:01] It looks like a
[45:02] rebound now. Last
[45:03] week I talked
[45:05] about a
[45:05] confirmation day, which is
[45:06] after a big
[45:07] bounce you assess
[45:08] its strength to
[45:09] see if it's
[45:10] the end, based on
[45:11] data from the 4th to the 10th
[45:12] day. Well, we're about
[45:13] halfway through
[45:14] this process now, and so far,
[45:15] things are looking very,
[45:17] very good. And I
[45:18] like the fact
[45:19] that right now, memory companies
[45:21] are lagging behind
[45:21] optical companies and
[45:22] things like that, they're
[45:23] bouncing back, but not by that
[45:24] much. And,
[45:25] most importantly, again
[45:26] , I will talk
[45:28] about Bitcoin. Well, you
[45:29] know, I've talked
[45:31] about this before, and you get the idea,
[45:32] but
[45:33] I'm adding a crypto
[45:34] component to what I'm
[45:35] doing in AI, and it's
[45:37] starting in
[45:38] late September,
[45:39] early October. And just
[45:41] so people understand, the
[45:42] reason I'm adding
[45:44] this for a fee is because
[45:45] this
[45:46] coincides with
[45:47] a point where
[45:49] traditional finance people
[45:50] -- those people you would have
[45:51] come to ten
[45:52] years ago and they would have
[45:53] laughed at you --
[45:54] are going to be, first, forced
[45:56] to enter this space,
[45:57] and second, they're
[45:58] going to be interested in
[45:59] it because of the development of AI
[46:00] agents. So, for
[46:03] anyone who has been following
[46:04] the crypto world and the
[46:05] traditional
[46:06] finance world,
[46:07] I hope you
[46:08] learned one thing from
[46:09] me: I constantly write
[46:11] on my Substack and
[46:12] reflect on the
[46:13] connection between the two. I
[46:15] believe that the last 10
[46:16] trading days, that is,
[46:18] from Monday last
[46:19] week to Friday
[46:20] this week, and all the
[46:21] events that we
[46:22] talked about at the very
[46:23] beginning, will be a
[46:25] turning point for us
[46:26] when we realize
[46:27] that the world debt has become
[46:28] too large.
[46:31] Central banks
[46:32] have no choice; they
[46:33] have to literally
[46:34] find the tools
[46:35] to make sure that
[46:37] this can continue, and
[46:38] AI trading has shown
[46:39] that there will be no easy money
[46:40] , which means that
[46:42] money and capital
[46:43] will be distributed
[46:44] to other places, and I
[46:45] think that's where
[46:46] crypto is starting its
[46:47] rise. So, I'll
[46:49] cover all of this in a
[46:50] video over the weekend.
[46:51] Surprisingly. For all of
[46:53] you wonderful,
[46:54] beautiful men
[46:55] and women who really
[46:56] appreciated this video,
[46:57] we have two
[46:58] requests today. First—you
[47:00] have to subscribe
[47:01] to Jordi. Go to
[47:02] YouTube, type in Jordi Visser,
[47:03] click the
[47:04] subscribe button, "digital
[47:05] handshake,"
[47:06] thank him, or
[47:07] go to 22V Research,
[47:08] Jordi Visser, you can
[47:09] just Google it
[47:10] , ask your
[47:11] local AI model, and
[47:12] it will direct you to the
[47:13] right place.
[47:14] Sign up there. And
[47:15] then, if you want to
[47:16] use
[47:17] the latest, greatest
[47:18] AI products
[47:19] to
[47:20] better manage
[47:21] your finances,
[47:22] check out Silvia.com's CFO.
[47:23] You can
[47:24] register there for
[47:25] free. My
[47:26] friend Geordie, we'll
[47:27] talk again
[47:27] next weekend
[47:28] .
[47:28] See you
[47:29] next week,
[47:30] friend.