Speakers argue AI stocks are undervalued relative to fundamentals, citing liquidity constraints and infinite demand.
Summary
Raoul Pal and Andreas Steno argue that the market is mispricing AI companies, trading them at significant discounts to consumer staples and the broader market, which they view as the inverse of the 2000 dot-com bubble. They contend that AI demand is effectively infinite, with intelligence output doubling rapidly, and that infrastructure build-outs are driven by existing capacity constraints rather than speculative excess. The speakers attribute recent sell-offs in AI and crypto assets to tight financial conditions and a strong dollar rather than fundamental deterioration. They highlight specific advantages held by players like Elon Musk and CoreWeave, while predicting that AI assistants will replace traditional operating systems and financial terminals.
Key takeaways
- 1 AI forward P/E multiples are at half the market average and a quarter of stable names like Costco and Walmart, mirroring the 2000 bubble in reverse.
Unidentified speaker (speaker_0) · 2:03
- 2 AI intelligence output per unit of energy is doubling every 3.4 months, indicating infinite demand for compute.
Unidentified speaker (speaker_3) · 4:30
- 3 I think they paid five billion dollars just for SpaceX um services in the third quarter, which is why they had a slightly weaker third quarter compared to the second quarter when you look at margins. They simply had to pay up for spot compute. compared to uh to the forward deals that you can price at much better terms.
Unidentified speaker (speaker_0) · 6:14
Episode sections
- 0:00 – 1:01
Why the Market Is Mispricing the AI Boom
Speaker_0 argues that AI companies are priced like dying businesses despite building critical technology. Speaker_1 agrees, noting the market's extreme nervousness and comparison to the dot-com era, where stable consumer staples were cheap and tech stocks were expensive, which is the current inverse.
- 1:01 – 4:20
Why This Is the Opposite of the Dot-Com Bubble
Speaker_0 highlights that AI stocks trade at half the P/E multiples of the average market and a quarter of stable names like Costco and Walmart, mirroring the 2000 bubble in reverse. He suggests portfolio managers are fighting the last war, being cautious due to past crashes.
- 4:20 – 8:32
Could the AI Buildout Eventually Overshoot?
Speaker_1 counters the oversupply fear by stating AI has infinite demand, with intelligence output doubling every 3.4 months. Speaker_0 adds that Anthropic faces capacity constraints, not demand issues, and notes Elon Musk's Colossus One compute allocation.
- 8:32 – 11:40
Elon Musk’s Compute Advantage
Speaker_3 details how Musk pre-secured gas turbines and compute capacity, allowing him to benefit from spot pricing while others are sold out. He claims Musk could achieve a one-year payback on his compute investments due to high margins and engineering efficiency.
- 11:40 – 14:29
The Semiconductor Crash the Market Is Pricing In
Speaker_0 disputes the market's expectation of a 40-50% semiconductor crash by mid-next year, arguing that even modest 25% annual CapEx growth would double investment by 2029. He cites CoreWeave's ability to fund CapEx from operations as evidence against a funding crisis.
- 14:29 – 16:10
AI Productivity Is Already Showing Up
Speaker_3 points to productivity rising from 1.1% to 3.6% in recent quarters, driven by new efficiencies. Speaker_0 provides a personal anecdote of producing more research with fewer people, attributing the gain to AI tools.
- 16:10 – 18:49
Why Claude Could Become the New Windows
Speaker_3 argues that AI assistants like Claude are becoming the new operating system for human attention, replacing traditional terminals like Bloomberg. He suggests that software companies must integrate into the Claude ecosystem to survive, similar to how Bloomberg adapted to Windows.
- 18:49 – 29:31
Is the AI Selloff Really Just a Liquidity Problem?
Speaker_0 attributes recent AI stock sell-offs to tight financial conditions, strong dollar, and wide credit spreads rather than fundamental issues. He cites the NVIDIA rumor as an example where selling occurred only after liquidity tightened, not when the rumor broke.
- 29:31 – 32:27
Geopolitics, Dollar, and Fed Policy Risks
Speaker_0 discusses the need for geopolitical resolution to allow the Euro to appreciate against the Dollar, noting a short dollar bias. Speaker_0 warns that unresolved wars could force central banks to hike rates, killing the business cycle, and criticizes Fed Chair Powell for inconsistent forward guidance while allowing lieutenants like Waller to signal aggressive hikes.
- 32:27 – 33:49
K-Shaped Economy and Consumer Staples Valuation
Speaker_3 describes a K-shaped economy where Main Street is struggling while hyperscalers spend heavily. Speaker_0 questions the high valuation multiples of consumer staples like Walmart compared to the rest of the economy, suggesting staples might benefit from automation.
- 33:49 – 36:03
Amazon and Walmart as AI/Robotics Beneficiaries
Speaker_3 argues that Walmart and Amazon are positioned to survive and thrive by using robotics to cut costs, citing Amazon's hiring freeze and productivity gains since 2019. Speaker_0 agrees, noting Walmart's robotics partnerships could significantly increase margins in a low-margin business.
- 36:03 – 44:21
Bloomberg vs. Claude and Attention Scarcity
Speaker_0 discusses canceling his Bloomberg subscription in favor of Claude, citing Claude's utility and lower cost. He identifies 'human attention scarcity' as a key theme, where AI agents filter information and direct communication shifts away from email to platforms like WhatsApp.
- 44:21 – 45:56
Crypto Outperformance and Liquidity Constraints
Speaker_0 notes Bitcoin's surprising outperformance against the Nasdaq despite a strong dollar, suggesting a rotation. Speaker_3 attributes the lack of broader market traction to household liquidity constraints caused by high costs, predicting a catch-up rally once liquidity returns.
- 45:56 – 47:31
AI as the New Operating System
The speakers conclude that AI models like Claude or Gemini may become the new desktop operating systems, replacing traditional platforms. They agree that the market is mispricing AI and crypto assets due to a lack of liquidity, not a lack of fundamental value.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| NVIDIANVDA | equity | mixed | With when you look at the forward price earnings of everything from Nvidia to Sandesk to Micron to even |
| Crude oil | commodity | neutral | that Venezuelan oil will come on market and refineries will be built for it and all of the things. And if you can resolve both of these two situations, the price of oil collapses and stays low for an extended period of time. |
| AmazonAMZN | equity | mixed | So we can test this because Amazon did it. So by 2019, they were piling money into robotics and CapEx. |
| BitcoinBTC | crypto | bullish | Now I looked at the the ratio of the Nasdaq versus Bitcoin. It was two standard deviations oversold in its log trend, so it was really oversold. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
AI stocks trade at half the multiples of the average market and a quarter of the multiples of Costco and Walmart.
Unidentified speaker (speaker_0) ·2:03
How to verify: Calculate the forward P/E ratios of Nvidia, TSMC, Micron, and Anthropic (via IPO filings or financial data providers) and compare them to the S&P 500 average and the P/E of Costco (COST) and Walmart (WMT). For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceWalmart and Costco traded at 10 P/E multiples in the year 2000, while Cisco traded at 100 P/E.
Unidentified speaker (speaker_0) ·2:23
How to verify: Check historical stock price data and earnings reports for COST, WMT, and CSCO for the year 2000 from SEC EDGAR filings or financial history databases. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceAnthropic's third quarter results were dominated by a lack of capacity to fulfill market demand.
Unidentified speaker (UNKNOWN) ·5:37
How to verify: Review Anthropic's Q3 financial statements or earnings call transcripts (available via their investor relations page or SEC filings upon IPO) for revenue vs. capacity utilization metrics. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceAnthropic paid $5 billion for SpaceX services in the third quarter.
Unidentified speaker (speaker_3) ·6:11
How to verify: Verify this figure in Anthropic's Q3 financial statements or 10-K/10-Q filings (when available) under operating expenses or related party transactions with SpaceX. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceElon Musk bought a mobile gas turbine business 18 months prior to the recording.
Unidentified speaker (speaker_3) ·9:27
How to verify: Search for press releases or SEC filings related to SpaceX or The Boring Company regarding acquisitions of gas turbine manufacturing or leasing assets. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceCoreWeave stated it can fund its CapEx via existing data center operations.
Unidentified speaker (UNKNOWN) ·10:24
How to verify: Locate CoreWeave's quarterly report or investor presentation referenced by the speaker (likely Q3 2026) and check the cash flow statement and management discussion for CapEx funding sources. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Michael Burry's thesis is that GPU depreciation is max 3-4 years.
Unidentified speaker (speaker_0) ·11:43
How to verify: Review Michael Burry's public statements, podcast appearances, or filings (e.g., 13F or public posts) discussing GPU depreciation schedules and semiconductor asset lifecycles. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceProductivity rose from 1.1% to 3.6% in the last three quarters.
Unidentified speaker (speaker_3) ·14:51
How to verify: Check US Bureau of Labor Statistics (BLS) data for nonfarm business sector productivity or specific sector productivity indices for the relevant quarters in 2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
More than 100 macro hedge fund PMs had to leave their portfolios in October.
Unidentified speaker (speaker_0) ·21:15
How to verify: This is a subjective claim about internal fund operations; verify via industry reports from hedge fund indices (e.g., HFR, CIS) or news articles from October 2026 detailing hedge fund closures or redemptions. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Oil prices could collapse to the 30s or 40s if Ukraine and Iran conflicts are resolved.
Unidentified speaker (speaker_3) ·28:48
How to verify: Analyze oil price forecasts from the IEA, EIA, or major banks under scenarios of lifted sanctions on Russia/Iran and increased Venezuelan output. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Speaker_0 has a short dollar bias in his portfolio.
Unidentified speaker (UNKNOWN) ·29:56
How to verify: Check speaker_0's public portfolio disclosures or Real Vision membership notes if available; otherwise, treat as personal statement. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Unresolved wars could cause central banks to hike interest rates to a level that kills the business cycle.
Unidentified speaker (speaker_0) ·30:26
How to verify: Monitor Federal Reserve and ECB meeting minutes for rate hike signals correlated with geopolitical events; track business cycle indicators like ISM Manufacturing. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Kevin Waller suggested 50 basis point rate hikes.
Unidentified speaker (speaker_0) ·30:56
How to verify: Search Federal Reserve press releases or transcripts for remarks by Governor Waller in the verification window (2026-08-19 to 2026-11-19). For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Amazon stopped hiring people in about 2019 and has not really hired again.
Unidentified speaker (speaker_3) ·35:06
How to verify: Check Amazon's SEC 10-K filings for employee headcount changes between 2019 and 2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Primary sourceAmazon's number of robots will exceed the number of humans by next year.
Unidentified speaker (speaker_3) ·35:19
How to verify: Verify via Amazon's annual reports or credible tech industry reports on warehouse automation deployment. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Speaker_0 canceled his Bloomberg subscription.
Unidentified speaker (UNKNOWN) ·36:43
How to verify: Self-reported claim; verify via public statements or Real Vision content where this was disclosed. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Bloomberg charges $30,000 a year for its subscription.
Unidentified speaker (speaker_0) ·37:00
How to verify: Check Bloomberg Terminal pricing on Bloomberg's official website. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Speaker_0 creates 90% of his charts within the Claude ecosystem.
Unidentified speaker (speaker_0) ·37:00
How to verify: Self-reported claim; verify via Real Vision content or public demonstrations of his workflow. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
The Nasdaq vs. Bitcoin ratio was two standard deviations oversold in its log trend.
Unidentified speaker (speaker_0) ·44:36
How to verify: Calculate the log trend of the Nasdaq/BTC ratio using historical price data from Yahoo Finance or CoinGecko for the period up to Nov 2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Household earnings have not been going up, and mortgage/rent/food costs are high.
Unidentified speaker (speaker_0) ·45:35
How to verify: Check US Bureau of Labor Statistics (BLS) data for Personal Income and Consumer Price Index (CPI) components for the verification window. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.
Do your own research
- Verify AI company P/E ratios against S&P 500 and consumer staples (Costco, Walmart) using Yahoo Finance or Bloomberg Terminal.
- Check Anthropic's Q3 financial disclosures and SpaceX service contracts for capacity constraint evidence.
- Review CoreWeave's latest earnings call transcript for CapEx funding sources and depreciation assumptions.
- Confirm Elon Musk's Colossus One data center power infrastructure details via SpaceX or Tesla filings.
- Analyze Federal Reserve meeting minutes for consistency between Chair Powell and lieutenants like Neel Kashkari or Christopher Waller.
- Assess Bitcoin's correlation with the Nasdaq 100 using CoinGecko or TradingView data for oversold signals.
- Compare Amazon and Walmart's robotics adoption rates and profit-per-employee metrics via their 10-K filings.
- Monitor global liquidity indicators (M2 money supply) and credit spreads to assess liquidity-driven market moves.
Full transcriptexpand
0:00Unidentified speaker (speaker_0): It's like who the fuck's gonna stay in business and who's gonna go out of business?
0:03Unidentified speaker (speaker_1): It's really incredible.
0:04Unidentified speaker (UNKNOWN): So
0:04Unidentified speaker (speaker_0): quickly, before we get to today's show, this episode is brought to you by Pith. Pith Network is the fastest growing financial data company today, trusted by Fidelity Investments, the US Department of Commerce, Revolut, Kelchi, Jane Street, Coinbase, and many more. They provide real-time financial market data across over 3,000 equities, commodities, crypto, FX, rates, and again. a lot more. They're the first and largest provider of 24-7 financial indices and offer the widest array of market data for the lowest cost. So go to pith.network to take advantage of their free trial.
0:39Unidentified speaker (speaker_2): Join me, Raul Pell, as I go on a journey of discovery through the macro, crypto, and exponential age landscapes. In The Journey Man, I talk to the smartest people in the world. so we can all become smarter together.
0:57Unidentified speaker (speaker_3): Andreas, always good to have you on.
1:00Unidentified speaker (UNKNOWN): Good
1:01Unidentified speaker (speaker_1): to see you, my friend.
1:02Unidentified speaker (speaker_3): There's so much to dig into. Um, there's a lot going on, and I think people are going to be really interested in your thoughts and my thoughts and how we are thinking through this. So, what's on your mind?
1:13Unidentified speaker (speaker_0): It strikes me that the combination companies that are building the most important technology in our lifetime are priced like dying businesses at the moment.
1:24Unidentified speaker (speaker_3): It's extraordinary. I don't know. You you wrote that yesterday on X, which is like, I think this is one of the most ridiculous things I've seen. I'm kind kinda of with you on this. It's insane.
1:35Unidentified speaker (speaker_0): Yeah. It's it's obviously always. something that could end up humbling you when you say that the market is incredibly wrong on something. Uh so let me try and be humble about it to begin with at least. With when you look at the forward price earnings of everything from Nvidia to Sandesk to Micron to even Anthropic, uh based on the numbers we know ahead of of their IPO this month.
2:01Unidentified speaker (UNKNOWN): I'd
2:01Unidentified speaker (speaker_0): argue that we're pricing the whole AI spectrum at the very least at half the multiples of the average market and at a quarter of the multiples of Costco, Walmart, and all of those stable names, right? That is the exact mirror image of what happened in the year 2000. The exact mirror image. Because Walmart costs Costco, etc. were priced at 10 price earnings multiples back in 2000, with Cisco and all of the names at 100 price earnings. So unless there is a huge bubble in earnings, which I struggle to see right now, there's something odd going on here. And I've been tempted to say, and I've said so for a while, that um most PMs tend tend to fight the last war. And the last time that they've seen something like this, everything went wrong. Which is why they're extremely cautious. And I get why if if you're if you've lived through the dot com
3:03and maybe the GFC as well, I I get why you're a little bit cautious and I get why you're nervous. The market is incredibly nervous. I mean, while we're recording the market is up 15% fifteen percent in all of the names that I just mentioned, right? Um so it one day down 20%, the next day up 15%. Never seen anything like it, but it just feels like the exact mirror image of the dot-com to
3:24Unidentified speaker (UNKNOWN): me.
3:25Unidentified speaker (speaker_3): Yes. And I think people are scarred by that, and they're assuming cyclicality. And what they think is you can't price the forward because of the cyclicality. And the question you've got to ask is: are revenues gonna go down? going to or maybe stabilize for a bit? And I just don't see revenues going down. Every single component part of this trade, you and I were chatting about this recently, every single component part of this trade, from TSMC to NVIDIA
3:55Unidentified speaker (UNKNOWN): to
3:55Unidentified speaker (speaker_3): the hyperscalers to anthropic, are producing supernormal returns. They have so much demand that they cannot physically build enough compute. There's the inability to be comp to be able to build compute may slow down chips for a little bit. But then what we've also found, and the labs are telling you this, and nobody's listening, is that more compute brings RSI, which is AI training. AI, which creates needs for more compute because it keeps going exponential. And I've got the rate of compute output of intelligence per unit of energy, my exponential age index, doubling every 3.4 months. And the other way I frame this to people is this is the only good ever with infinite demand. Because the more compute you build, the the better the air. AI becomes, the more agents it uses, the more it does. And we barely started.
4:55Unidentified speaker (speaker_1): The thing I hear the most from some of the
4:59Unidentified speaker (speaker_0): veteran PMs involved in this trade is that a build-out always ends in tears. Because a market is by design always overshooting when there's a scarcity, so we'll end. up building too much compute capacity. That is the pushback I get every single time.
5:20Unidentified speaker (UNKNOWN): Uh,
5:20Unidentified speaker (speaker_0): and another pushback I get is that well, in two, three, four years, most of our simple tasks will just run on our iPhone or whatever uh mobile um device that we have. Let me start with the with the first one.
5:37Unidentified speaker (UNKNOWN): If
5:37Unidentified speaker (speaker_0): if we look at the computer. compute capacity, uh if you listen to all of the labs, right, that they still say that they have a a capacity issue. If you look at Anthropic's third quarter, it was heavily dominated by lack of capacity to fulfill the market demand. It wasn't a demand issue, it was a capacity issue. They're now getting closer to solving the capacity issue. We've seen great momentum in in the anthropic business over the past four to six weeks But I've said continuously that the biggest risk to the anthropic case is that they have not secured enough compute on a forward basis. I
6:11Unidentified speaker (speaker_3): mean they took all of Elon's Colossus One, all of it. Yeah,
6:14Unidentified speaker (speaker_0): I think they paid five billion dollars just for SpaceX um services in the third quarter, which is why they had a slightly weaker third quarter compared to the second quarter when you look at margins. They simply had to pay up for spot compute. compared to uh to the forward deals that you can price at much better terms. Having said that, right, uh now that we're on the other side of the midterms, I don't really buy the notion that the Democrats will end the AI build-out, but we do see issues related to everything from regulation uh to pushbacks from uh from the population in terms of the pace of the build build-out, out right? So it is actually difficult to build out.
6:51Unidentified speaker (UNKNOWN): in
6:52Unidentified speaker (speaker_0): the pace needed to match the demand for compute uh i think that's a very very different environment to to the year 2000 build out of uh of fiber lit right where you could just you just built right it wasn't an issue
7:04Unidentified speaker (speaker_3): yeah a you could just build but you built before there was any demand this is the opposite you're trying to build as fast as you can because the demand is completely outstripping your ability to supply
7:15Unidentified speaker (speaker_0): And then this notion that you'll end up with a lot of
7:19Unidentified speaker (UNKNOWN): uh
7:19Unidentified speaker (speaker_0): oversupply in the data center space because we end up using local models on our phones and what what have you. Uh I think it is similar to stating that, oh, as soon as you got I don't know, sixteen 16 gigabyte of memory space on your phone, no one would use cloud after that because you have so much. space on your phone locally that you would never end up uh consuming
7:42Unidentified speaker (UNKNOWN): uh
7:43Unidentified speaker (speaker_0): uh
7:43Unidentified speaker (UNKNOWN): space in in
7:44Unidentified speaker (speaker_3): yeah. And people misunderstand that AI is going to be in every single object. Every single electronic object will have AI. The it's not it's a G Von's paradox. It gets cheaper and you sell more of it. Right now I've got demand double doubling for. every unit it goes down in price. So we're not seeing demand fall either, which was the other argument. They're like, oh well, it's just going to get so cheap, all the margins are going to get crushed. No, margins have gone up at the frontier models as the actual AI has got cheaper. It's extraordinary.
8:19Unidentified speaker (speaker_0): Exactly. And you know, think of it this way yes, for example, Enthropic, Anthropic, they've spent a lot of dollars trading training their models. But if you look at the ratio between training and inference, we're moving so fast in the direction of you know paid services inference relative to training that it won't be an issue in two, three, four quarters from now. It's not an issue already now.
8:45Unidentified speaker (UNKNOWN): Um,
8:46Unidentified speaker (speaker_0): they are cash flow positive by now, which is extraordinary given how much they spent, right. Um and and and therefore I really struggle with this notion that AI is going to slow down incredibly fast into 2028, 2029, which is what you get if you look at forward pricing, basically.
9:07Unidentified speaker (speaker_3): So the other one that's interesting in this equation is SpaceX. So Elon is always early in this stuff, and he's like it was eighteen. like 18 months ago talking about transformers. Everyone's like, what the fuck's a transformer? And he's like, there's issues, there's gonna be. And so what basically he did, he pre-secured all the stuff. He then went and bought a gas turbine, a mobile gas turbine business before anybody had figured out what was going to go on. So now he's gonna put, I don't know how many gigawatts of compute, like 13 thirteen gigawatts of compute, out next year when everybody. else can't get stand up a data center. Amazon, Google, Microsoft have all pre sold all of theirs, and he's taking advantage of the spot pricing. It looks like if he gets this right, he'll have had a one year payback.
9:54Unidentified speaker (UNKNOWN): I
9:54Unidentified speaker (speaker_3): mean, the return on capital is fucking extraordinary, and yet people don't want to believe it. They just don't believe it can happen.
10:01Unidentified speaker (speaker_0): Also this week, right? Well, we got the quarterly report from Callweave, you know, one of the neo clouds that has faced the wrath of the market with higher credit spreads, and they haven't really been able to access the bond market for a little while. And they basically told us that, you know, within the foreseeable future, they can just fund
10:24Unidentified speaker (UNKNOWN): their
10:24Unidentified speaker (speaker_0): capex via the existing. data center operations. What if we relatively soon get to the point where the existing data center operations will fund the next data center operations? But we don't need the funding markets at all. It's it's not a bizarre scenario.
10:43Unidentified speaker (speaker_3): Well because people misunderstand how much cash flows out of this and Elon's about to prove it. Now yeah look he he does everything at a marginal uh yeah at higher margins than everybody else because he's just a brilliant engineer. But still, if he can do that, then he can just build infinite data sensors because he'll just fund them out of cash flows.
11:02Unidentified speaker (speaker_0): Yeah. And it's something that could happen say late 27, early 2028. It's it's not five, 10 ten years down the road. It's already happening. Um, And even more interesting is that this whole notion that old GPUs will not be able to provide the return needed for them to be funded going forward. Uh, it is, you know, the actual evidence is wiping the floor with Michael Burry on this thesis, right? Because his thesis is that the depreciation of a GPU is at max three, maybe four years. And again, Coley, for example, told us that they're renting out their H100s for the next three, four, five years on forward deals, and they're already paid back.
11:53Unidentified speaker (UNKNOWN): And
11:54Unidentified speaker (speaker_0): you know, we've been told that the H100s were useless as soon as we got the Vero Vera Rubins Rubens and uh what have you uh online. That has not proven to be the case because of this gas. the scarcity, right? So I
12:04Unidentified speaker (UNKNOWN): And
12:05Unidentified speaker (speaker_3): there's different use cases for different things. You know, inference doesn't require Vera Rubens. Training does, for example. You know, there's just they can use all the compute. Yeah.
12:17Unidentified speaker (UNKNOWN): Um
12:17Unidentified speaker (speaker_0): so
12:19Unidentified speaker (UNKNOWN): we're
12:20Unidentified speaker (speaker_0): kind of we're kind of circle jerking now, Ralph, that you know, if everyone's not getting this, but I I sincerely think that this is one of the most interesting crossroads I've ever been at from a market standpoint. Where you have the market basically prices the semiconductor cycle to be down, I don't know, forty-fifty percent by mid next year. I mean the crash needs to happen very, very soon for the market to be right on this. And
12:49Unidentified speaker (UNKNOWN): when
12:50Unidentified speaker (speaker_3): you say a crash, it's are you looking for rates of change, which could slow down, should slow down at some point, or are you looking for absolute dollars to reduce?
12:59Unidentified speaker (UNKNOWN): Absolute
12:59Unidentified speaker (speaker_0): dollars.
13:00Unidentified speaker (speaker_1): Yeah.
13:00Unidentified speaker (speaker_0): That's the bizarre thing. You know, sure Bloomberg had an article this week where they say the first stated that the first derivative of CapEx from the hyposcalers Typoscalers have now rolled over.
13:11Unidentified speaker (UNKNOWN): And
13:11Unidentified speaker (speaker_0): I was like, Yeah, that is pretty simple math. I
13:13Unidentified speaker (UNKNOWN): mean,
13:14Unidentified speaker (speaker_0): you cannot grow a at a one hundred and fifty percent forever. You cannot at least not increase the growth rate. No.
13:19Unidentified speaker (UNKNOWN): Uh
13:20Unidentified speaker (speaker_0): but but if you let's say that they grow fifty percent a year from here in CapEx terms, that is humongous in nominal dollars. Um I just made this simple exercise um
13:32Unidentified speaker (UNKNOWN): in
13:32Unidentified speaker (speaker_0): in in a research paper this week where I just said, Okay, what if they grow twenty 25% five percent capex from now on until twenty 20 uh 35?
13:41Unidentified speaker (UNKNOWN): It
13:41Unidentified speaker (speaker_0): it means that we've already doubled capex from the current um
13:45Unidentified speaker (UNKNOWN): uh
13:46Unidentified speaker (speaker_0): state in two and a half years from now, something like that, given that 2007 27 is almost carved in stone. So if 28 grows at 25, 29, 25, then we're already
13:58Unidentified speaker (UNKNOWN): more
13:58Unidentified speaker (speaker_0): than double up of the levels that we're currently at.
14:01Unidentified speaker (UNKNOWN): Um
14:02Unidentified speaker (speaker_0): I think. we're underestimating this, really underestimating this. Essentially as because the one thing that could kill this route
14:10Unidentified speaker (UNKNOWN): would
14:10Unidentified speaker (speaker_0): be the Bond vigilantes. But if the cash flow is strong enough to give the Bond vigilantes the middle finger, which is probably going to be the case in a year from now at the latest, then we shouldn't care about the bond bunt market.
14:22Unidentified speaker (speaker_3): Also, if you've look looked, if you look at how this has played out so far, and
14:26Unidentified speaker (UNKNOWN): you
14:26Unidentified speaker (speaker_3): see it, Amazon's you can see this a lot. They do this big capex cycle.
14:30Unidentified speaker (UNKNOWN): And then
14:31Unidentified speaker (speaker_3): by the time they finish the capex cycle, profits go up again. Because you produce new efficiencies or new products. So any investment that they're all making now in this stuff just turns into output later. So people assume that they're going to be cash burning forever. What we could find is that they cash burn for periods of time as of the heavy intensive stuff, because this is a big new CapEx build-out. Then as soon as it starts coming online, their profits go up again. Because you see productivity rising, and nobody wants to believe that yet either. Productivity's gone up from the last three quarters from 1.1% to 3.6%. It's accelerating. Um, and that's been the view to prove this out. If if productivity goes up, then everybody's
15:14Unidentified speaker (UNKNOWN): going to be able to do it
15:15Unidentified speaker (speaker_3): gonna win here.
15:16Unidentified speaker (speaker_0): Yeah. You know, I I sometimes le uh like to uh keep it simple, stupid around productivity. Productivity is very, very difficult to measure life. But I I had a research team of 15 people producing maybe 15, 20 research reports a week two two years ago.
15:35Unidentified speaker (UNKNOWN): Now
15:36Unidentified speaker (speaker_0): me and my partner Mickel produce produced more. We do produce more. than that timber 15. When I keep it as simple as that, I cannot come to the conclusion that it's not a productivity gain. I mean, it is a productivity gain of many X, right?
15:51Unidentified speaker (speaker_3): So one of the ways I'm looking at this is I think the scarce thing that we're now hitting is human attention.
16:01Unidentified speaker (UNKNOWN): Right?
16:01Unidentified speaker (speaker_3): Because everybody can do what. you're doing and everybody's doing it, right? So it's like, how do you read all this stuff? How do you absorb all this stuff? So everybody's doing it on their Claude or their Chat GPT. So ChatGPT or Claude is now reading all of your emails, reading all of your ex posts, you know, the the information, distilling it down for you, reading all the research reports, and it's now becoming the platform of our lives. That everything is being absorbed into these things. You either sort of plug in and access it through it and use the intelligence to overlay, or they just build it themselves. And you would just mention this just before we started, is you're like, this is sort of Microsoft Windows or the operating system all over again.
16:44Unidentified speaker (speaker_1): Yep. So
16:46Unidentified speaker (UNKNOWN): I've
16:46Unidentified speaker (speaker_0): taken my you know the last few quarters, every time. I've sort of faced the obstacle that I had to create workarounds to get some of my existing software licenses into Cloud, I've just emailed my account manager saying, Why don't you build it directly into Cloud? Because that is where my workflow is. And three years ago, I would have had the same view if they didn't provide an exit file for a Windows application, right? So think of it this way. You know, to begin with, a Bloomberg, for example, was a terminal, right? You sold hardware alongside that Bloomberg license. At some point, they took a big decision and said, okay, we have to build this for Windows because that's where the users
17:33Unidentified speaker (UNKNOWN): are. I
17:35Unidentified speaker (speaker_0): kind of struggled to wrap my head fully around it, but of course Bloomberg needs to be incorporated in Claude pretty soon.
17:42Unidentified speaker (UNKNOWN): I haven't come to mine.
17:43Unidentified speaker (speaker_0): Yeah. They haven't come to that conclusion yet. But if they want to exist in five, ten years, they probably need to function within the Claude ecosystem, in my opinion.
17:51Unidentified speaker (UNKNOWN): I mean
17:51Unidentified speaker (speaker_3): I just got rid of mine.
17:53Unidentified speaker (UNKNOWN): I
17:53Unidentified speaker (speaker_3): got rid of it because I can get all of the information I need or construct it Tip very simply. Because you nobody uses the 58,000 credit models and all that shit. that's for some specialist. You know, the basic stuff, it can do all of that very easily. You can access most of the information you'd ever need unless you're highly specialist in looking for off-the-run, you know, Lithuanian bonds from 1962, you know, shit like that. Yeah, fine. But other than that, you don't need it because Claude Cloud can do all of it, and there's plenty of MC. MCP connectors. I mean, we're looking at this for Real Vision as well. It's just saying, you know, we should just bring Real Vision inside of Claude, give you access to all of this stuff. And I'm looking at what I do with my GMI data as well. It's like, you might as well go where people are and let them use the intelligence layer that they're using.
18:42Unidentified speaker (speaker_0): Yep. I I think it's it's spot on, and it's That's it's also something that um
18:47Unidentified speaker (UNKNOWN): you know
18:47Unidentified speaker (speaker_0): I've been actively trying to promote within the companies that I I own or or I've co-founded. And
18:55Unidentified speaker (UNKNOWN): in
18:55Unidentified speaker (speaker_0): that sense, okay, if if we try to, you know, finish this historical analogy, uh Windows kind of made the rounds in ninety-four and ninety five, something like that.
19:04Unidentified speaker (speaker_3): Yeah.
19:05Unidentified speaker (speaker_0): If we were to make that comparison to the dot com, then you know, we are We're in we are incredibly early, right? Because I don't even think that people have come to the conclusion that Claude is the new Windows. I think it will happen soon, but but um um so again I I really struggle with this notion that we're in ninety-nine or two thousand very late in this cycle. But um yeah.
19:24Unidentified speaker (UNKNOWN): So
19:25Unidentified speaker (speaker_3): let me put it another way. And we had this conversation on the RV pro chat. Is what happens if this is just all narrative and the fact is that financial conditions tightened because of of Iran and liquidity got tight. So prices have gone down because of liquidity, and everyone's trying to pin a narrative on it.
19:44Unidentified speaker (speaker_0): Okay, the best example of why I think that is true is the story about Jensen Huang. The story that he had an affair with an alleged Chinese spy was more than a week old when he emailed his employees and told him that he had an affair. Nobody sold NVIDIA when Walking. Wall Street Journal wrote about it eight days prior to that email. When that email made the rounds, everyone started selling NVIDIA, probably because they found credit spreads to be tied. You know, it it's it's difficult to
20:15Unidentified speaker (UNKNOWN): uh
20:16Unidentified speaker (speaker_0): to measure it outside of doing it via financial conditions, but I think you're right that
20:23Unidentified speaker (UNKNOWN): the
20:23Unidentified speaker (speaker_0): dollar has been too strong, credit spreads spritz have been too wide, the oil price. has been too high, you know, all of that.
20:29Unidentified speaker (speaker_3): It just sucks liquidity out of the system, and therefore any sort of bad news or or or something that's kind of misunderstood or the market just can't take it because there's not enough liquidity around.
20:43Unidentified speaker (speaker_0): Yeah, and if if you look at the degrossing that we've seen from especially hedge funds, uh both low long short hedge funds fund, but also multiple. multi-strat hedge funds through the month of October and into the early innings of November, they've all been probably been running very, very close to their max capacity uh set by the risk manager, right? And when something like this shows up, um, okay, there is a headline. I know that I'm running close to my risk capacity, I'll bring it down. I think that that is what happens in practice. Uh at as I've also um posted on on social media outlets, I think more than a hundred of of um macro hedge fund PMs had to leave their pod through the month of October. I'd never seen anything like it in my life. I mean uh you've been on the hedge fund coverage for forever as well, Rao, Val, and you talk to many of them still. I I think it's the worst month ever. for hedge funds, October, because of
21:42Unidentified speaker (UNKNOWN): uh you
21:42Unidentified speaker (speaker_0): know financial conditions forcing them to simply exit positions.
21:47Unidentified speaker (speaker_3): And also,
21:48Unidentified speaker (UNKNOWN): you know,
21:48Unidentified speaker (speaker_3): I I've talked about this for a long time. There is no diversification, everything is risk-on, risk off. And risk on, risk off is just a function of liquidity. Liquidity is risk on when liquidity withdraws would draws is this risk off. That's why everybody observes this trend. So if you're a hedge fund and you're directional. you have no choice but to be in the momentum trade because you don't make any money either. Um, and so what happens is they all get washed out at the same time.
22:12Unidentified speaker (speaker_0): And the leverage has increased in the system since the GFC and also since year 2000. I I know that many like to look at this uh margin debt from the FINRA organization and and you know take that as a ratio to GDP or M2 to to get a grip of of how much leverage there is. I I think it vastly underestimates the amount of leverage in the system because it's a very old school way of using leverage, right? You have leverage via ETFs now. Uh options have become much more widely available for retail investors. Uh I I have lost count of the amount of subscribers and um
22:52Unidentified speaker (UNKNOWN): uh
22:53Unidentified speaker (speaker_0): ex members, etc., that have messaged me in my inbox. Do you think I should go 3x long the semiconductor index via an ETF? And I'm always like, no, it's a crap product, but it's available to everyone, right? Um and it's just so visible when you look at, for example, the pr price action in semiconductors that there are so many levels. levered players that it goes in and out of these ETFs and other structures on a daily basis. It it exact it exaggerates the narratives,
23:24Unidentified speaker (UNKNOWN): uh
23:24Unidentified speaker (speaker_0): the way that the market structure is there.
23:26Unidentified speaker (speaker_3): So a quick break in your regular programming. If you're serious about your future, grab my free report called Prepare for 2030. I think you've got five years to make as much money as possible, and this guide will help you. navigate what's coming. The link is in the description. Download it now. And you know, that that's the way that bull markets continue. My old boss used to call them the pause that refreshes. And it's like if you're in a serious bull market, you have a bunch of these. Um, and they wipe everybody out because everyone gets over their ski tips, and then everyone looks back and go, Well, why am I out of the position? Used to happen to me. I and I'll never forget because I was running a book via the um during the uh um the 2000s bubble, and I was yeah, very short and had a good few years based on that. But half the time you close out your positions, you know, when the market went against you, and then the next
24:19Unidentified speaker (UNKNOWN): day
24:19Unidentified speaker (speaker_3): you're like, fuck, I just want to put the positions back on again. Because you know where the trend is, you know what's going on. Um, and so you know what other what what else are you going to trade?
24:29Unidentified speaker (UNKNOWN): And
24:30Unidentified speaker (speaker_3): you said that there are prices where everything is cheap again. We've got off the frothy levels. Yeah, the Nasdaq was quite overbought at the top, and it's now unwound itself pretty decently. It's back to its kind of logarithmic trend, fair value. It's like, yeah, okay, what's there not to like here? Unless the liquidity picture is gonna get worse. And I didn't like what happened yesterday. and today in bonds and the dollar. It's that that's like fuck, I could do without this, because if the dollar keeps going up from here, then we do actually have bigger problems.
25:03Unidentified speaker (UNKNOWN): I
25:04Unidentified speaker (speaker_0): think the simple answer to to this question, Raoul, is that financial conditions are currently too tight to get a very strong rally in everything AI related and crypto related, for example.
25:15Unidentified speaker (UNKNOWN): Um,
25:16Unidentified speaker (speaker_0): even though the fundamentals are incredibly strong and improving by as we speak.
25:21Unidentified speaker (speaker_3): In all of it, in crypto, everything, everything's doing this, but prices aren't. Uh
25:26Unidentified speaker (speaker_0): but it doesn't matter as long as as the liquidity is not there. Um and financial conditions,
25:33Unidentified speaker (UNKNOWN): you
25:33Unidentified speaker (speaker_0): know, the dollar is obviously the most important thing for financial conditions, in my opinion, but the second most important thing is probably. energy. And
25:44Unidentified speaker (UNKNOWN): my
25:44Unidentified speaker (speaker_0): famous last words of this podcast, because I've said those exact same words three or four times already this year, is that it looks like the worst is behind us in relation to energy prices and the Iran war and all of that. And
25:58Unidentified speaker (UNKNOWN): uh
25:59Unidentified speaker (speaker_0): you and I had a chat earlier this week around
26:02Unidentified speaker (UNKNOWN): you know.
26:03Unidentified speaker (speaker_0): the potential for
26:05Unidentified speaker (UNKNOWN): you know
26:05Unidentified speaker (speaker_0): Trump spending the last two years of his presidency, you know, basically trying to solve these things um out of legacy reasons. Yeah.
26:13Unidentified speaker (speaker_3): Because he still wants the Nobel Peace Prize.
26:15Unidentified speaker (speaker_0): Yeah.
26:16Unidentified speaker (UNKNOWN): He
26:16Unidentified speaker (speaker_0): he really wants it.
26:17Unidentified speaker (speaker_3): He really wants it.
26:20Unidentified speaker (speaker_0): I can tell you from speaking to my Norwegian friends up there, they're pretty pissed with his this phone calls after I find out But have. having said that, well, um,
26:29Unidentified speaker (UNKNOWN): there
26:30Unidentified speaker (speaker_0): there is some momentum, in my opinion, um, in relation to the Ukraine-Russia concept.
26:35Unidentified speaker (UNKNOWN): Yeah, we
26:35Unidentified speaker (speaker_3): saw that today, right? The Kushner thing that they uh prisoner exchange and stuff like that's positive.
26:40Unidentified speaker (speaker_0): Yeah, and if Putin wants to cease this moment, he needs to cease it before there's a new president in the White House, right?
26:50Unidentified speaker (UNKNOWN): Um
26:50Unidentified speaker (speaker_0): what I struggle with is how the Europeans will respond to this because they're kind of left outside of the room um in many ways. It seems like Russia and the US are trying to sort of redefine how how's a view Europe between them, um, which which is obviously a a a weird thing to experience as a European not being present. But um just earlier this this morning, uh um I saw the Ukrainian president uh Volidija Selinski here in Copenhagen. He visited uh our prime minister and gave a speech in the parliament. And Denmark is by far, and I repeat, by far, the country that provides the most support to Ukraine per capita and as a percent of GDP. But many countries in Europe basically do not support Ukraine. any longer economically. So I think several of these countries, without them having explicitly stated it so far, they have sympathy for what's going on between Trump and Putin now. Um, and that that is my cynical view of what's going
27:53on. I think there is a fatigue uh that is getting more and more uh present also within the European Union uh on this topic. And if they can solve the Ukraine war, they can also solve the Iran war. I think the Ukraine war is the more difficult one of the two. Uh and it it still seems likely to me that the Republicans have basically concluded that they lost the midterms because of those wars.
28:18Unidentified speaker (UNKNOWN): Uh
28:18Unidentified speaker (speaker_0): so they need to get rid of that if they want um a as seat he'd at the table after 28.
28:23Unidentified speaker (speaker_3): And people don't really understand because of what the US did with Venezuela, that Venezuelan oil will come on market and refineries will be built for it and all of the things. And if you can resolve both of these two situations, the price of oil collapses and stays low for an extended period of time. I mean, what do Russia and Iran do if they get open market access to sell oil? They'll sell as much of it as they. possibly can to make up the revenue streams that they've been missing. And the Saudis were aware of it, they're all aware of it. Um it'll just happen. And that's oil back down in the 30s or 40s, which is a huge difference. But we've got to get that to the other side, because I've been thinking like you, I've been like, this is gonna get resolved. It's in everybody's interest to resolve it, but it hasn't happened yet. Oil looks trading better, but fucking dollars not still.
29:19Unidentified speaker (speaker_0): Probably as a lot of the oil products aren't really trading better yet. I mean diesel is still incredibly elevated, right, because of the lack of of refiners capacity. And that isn't of course an issue for
29:31Unidentified speaker (UNKNOWN): the
29:31Unidentified speaker (speaker_0): Eurozone, Japan, some of the importers, right? So I th I think we need to get a proper resolution to to to the Iran war and to the strikes on the refiners' capacity in Russia and all of that to to get the energy market um to a place where the euro can appreciate versus the dollar again. But I think we're getting there. And you know I
29:56Unidentified speaker (UNKNOWN): I
29:56Unidentified speaker (speaker_0): admittedly have a short dollar bias in my portfolio right now. A slight lean, not not a big lean, but I I don't think the risk reward is super good in chasing the dollar higher here.
30:05Unidentified speaker (speaker_3): Yeah, but we've all got a short dollar bias by the nature of our bets, because if the dollar rate of change goes up a lot from here, then that's the unwinding of liquidity, and therefore that's the unwinding
30:19Unidentified speaker (UNKNOWN): of all
30:20Unidentified speaker (speaker_3): the beta to trades. trade.
30:21Unidentified speaker (speaker_0): I think if you look into 2027,
30:25Unidentified speaker (UNKNOWN): if
30:26Unidentified speaker (speaker_0): these wars are not resolved, then I think there is a clear risk that the central banks will hike interest rates
30:34Unidentified speaker (UNKNOWN): to
30:34Unidentified speaker (speaker_0): an extent that will kill
30:37Unidentified speaker (UNKNOWN): the
30:37Unidentified speaker (speaker_0): business cycle due to financial conditions. I think that risk is clear. Also given that this supply shock has now gotten some of the co co-members members of the Federal Reserve Committee talking about 50 basis points at a time. and all of that shit.
30:51Unidentified speaker (UNKNOWN): What was
30:52Unidentified speaker (speaker_3): that? What's Waller going on about?
30:56Unidentified speaker (speaker_0): Raoult is that you know, essentially by by not providing forward guidance, Kevin Warsh has now handed the keys to the Federal Reserve to some of his lieutenants. Because they're still providing forward guidance. And Wallow said 50 fifty basis points, and the market just chased that. Even though,
31:13Unidentified speaker (UNKNOWN): you know,
31:13Unidentified speaker (speaker_0): we don't have a clue whether. the other members support it or whether Kevin Walsh supports it. But I think it is such a bad idea for the chairman not to provide forward guidance when all of his lieutenants run around and provide forward guidance on a running basis. Either or, right?
31:28Unidentified speaker (speaker_3): Or say nothing and let the market decide, but you can't have one and not the other.
31:32Unidentified speaker (speaker_0): Because Wars wants the market to decide. W which I hold s I I've I sincerely hold sympathy for for for the director direction he's trying to take the Fed in. But then he needs to silence his lieutenants as well. Tell them to shut up. It's as simple as that. If Walla keeps stating that, okay, we're considering 50 basis points, then there's no point in Wars not providing forward guidance because he's providing it for him.
31:54Unidentified speaker (speaker_3): But how can they raise I I just don't understand. We're already talking about financial conditions of time, because rates have gone up. the dollar's gone up, and oil was high. So what good is it to raise rates? Does it stop the CapEx boom? No, because the hyperscalers don't really care. And they've already locked in so much funding at, you know, a 10-year funding plus. Um the only person that hurts at the front end is actually the US government, because they're the biggest issue at the short end.
32:27Unidentified speaker (UNKNOWN): And
32:27Unidentified speaker (speaker_3): if rates go up, their the whole interest payment curve goes up again, yet another.
32:32Unidentified speaker (speaker_0): I think if you look if you look at the construction cycle, for example, in the U as a good old business cycle indicator, outside of data centers and you know the grid expansion, the business cycle is doing very bad. Yeah.
32:47Unidentified speaker (speaker_3): We've got a K-shaped economy, and the Fed don't want to acknowledge it. because they see the shiny thing of the biggest revenue generating machines in history spending a lot of money on stuff to build more revenue.
33:00Unidentified speaker (UNKNOWN): And
33:00Unidentified speaker (speaker_3): they're not seeing the fact that Main Street's pretty fucked right now.
33:05Unidentified speaker (UNKNOWN): You
33:05Unidentified speaker (speaker_3): know, the whole commercial office space, the the construction industry, the mortgage markets, all of it is a mess.
33:11Unidentified speaker (speaker_0): And especially given that case shape,
33:14Unidentified speaker (UNKNOWN): I
33:14Unidentified speaker (speaker_0): don't understand why we trade Walmart at plus forty forward earnings and the part of the economy that I can actually cope with high interest rates at I don't know five four five X forward earnings. That's weird to me because you know
33:31Unidentified speaker (UNKNOWN): good
33:31Unidentified speaker (speaker_0): old school consumption will suffer.
33:37Unidentified speaker (speaker_3): Well Walmart do they get really hit? Because they're they're just like the Microsoft Windows of stuff, right? It's like everybody go everybody else will die and Walmart and Amazon will live.
33:49Unidentified speaker (speaker_1): Yeah. That's that's a fair point. I mean
33:52Unidentified speaker (UNKNOWN): Because they
33:52Unidentified speaker (speaker_3): kind of take everything because they can be the lowest cost producer of everything. Costco as well.
33:57Unidentified speaker (speaker_0): The the interesting counter-argument counter argument to that Walmart trait, right?
34:03Unidentified speaker (UNKNOWN): S
34:03Unidentified speaker (speaker_0): some people because consumer staples have traded incredibly well through this AI cycle, which is an anomaly.
34:10Unidentified speaker (speaker_3): Well consumer discretionary has been killed.
34:13Unidentified speaker (speaker_0): Yes. But
34:14Unidentified speaker (UNKNOWN): look
34:15Unidentified speaker (speaker_0): at it this way, Raoult.
34:17Unidentified speaker (UNKNOWN): I
34:17Unidentified speaker (speaker_0): haven't made up my mind on that thesis yet, but Walmart they've had a um fantastic cooperation uh with one of the um
34:29Unidentified speaker (UNKNOWN): you know
34:29Unidentified speaker (speaker_0): manufacturers of um
34:31Unidentified speaker (UNKNOWN): of
34:32Unidentified speaker (speaker_0): warehousing robotics and they've made incredible gains using robotics in their warehousing over the past two or three years.
34:42Unidentified speaker (UNKNOWN): And
34:42Unidentified speaker (speaker_0): given that Walmart is a low margin business to begin with, maybe that is the exact type of business that we'll see. will
34:51Unidentified speaker (UNKNOWN): from
34:51Unidentified speaker (speaker_0): a percentage standpoint the biggest increase in margins if robotics
34:56Unidentified speaker (UNKNOWN): you
34:56Unidentified speaker (speaker_0): know work.
34:57Unidentified speaker (speaker_3): So we can test this because Amazon did it. So by 2019, they were piling money into robotics and CapEx. And out the other side, they stopped hiring people. They stopped hiring people in about 2019. And they haven't really hired again. Yet they're not their number of robots will exceed the number of humans by next year. And their productivity's gone up. So their profits per employee have gone up massively. Their margins have gone up. So Amazon is like a giant experiment of the US economy as it implements AI and robotics. And they're proving it, and Walmart's proving it or will prove it. And everybody will follow suit in the end because they'll have to, because if not you're going out of business. So there's that case shape that's going to continue as well. It's like who the fuck's going to stay in business and who's going to go out of business?
35:46Unidentified speaker (UNKNOWN): It's
35:47Unidentified speaker (speaker_1): it's really incredible. And
35:49Unidentified speaker (UNKNOWN): that
35:49Unidentified speaker (speaker_0): is why, you know, I had the discussion with a portfolio manager earlier today on a call. Which sectors can we actually price with a decent visibility over the next five to ten years?
36:03Unidentified speaker (UNKNOWN): Uh
36:03Unidentified speaker (speaker_0): I'd argue that Amazon can be priced with a decent visibility.
36:08Unidentified speaker (UNKNOWN): Uh
36:09Unidentified speaker (speaker_0): Walmart may be the same, at least they're doing some of the right things. They're maybe doing it too slow, but they're doing some of the right things. Uh Bloomberg, can we price that with a decent visibility? I don't think so, because it it it it takes they need to take some incredibly bold decisions to stay alive, in my opinion. I've said the same to my all of my financial software account managers. I've said it three years running. If you do not take the right strategic decisions, I will cease being a client within five years. I'm certain about that.
36:40Unidentified speaker (UNKNOWN): Yeah,
36:40Unidentified speaker (speaker_3): you well, I I had this conversation with Bloomberg and they said, Fuck you, and I
36:43Unidentified speaker (UNKNOWN): said,
36:43Unidentified speaker (speaker_3): Thank you. You know, I've been a client of Bloomberg at since nineteen 1992. ninety two.
36:48Unidentified speaker (UNKNOWN): And
36:49Unidentified speaker (speaker_3): I've I just cancelled. canceled. I said, No, I'm just not doing this anymore. It's pointless. And you can't charge me claiming you have a monopoly, you know, $30,000 a year for this, when basically I don't need you.
37:00Unidentified speaker (speaker_0): And look at it this way, Rao. I spent substantially less on Claude than on Bloomberg. Even though my workflow is in Claude now and I probably create 90% of my charts within the Claude ecosystem. occasionally with Bloomberg data short.
37:18Unidentified speaker (speaker_3): When you're looking at 1962 off-the-run Lithuanian bonds, maybe.
37:21Unidentified speaker (speaker_0): Yeah. Enough about my portfolio now, but
37:25Unidentified speaker (UNKNOWN): but
37:26Unidentified speaker (speaker_0): maybe the only moat that's left, you know, speaking of this lack of human attention, uh, whatever you called it early in the in in the in this show,
37:35Unidentified speaker (UNKNOWN): the
37:36Unidentified speaker (speaker_0): only moat that's left is that many of my counterparts they do not reply to emails anymore if you try to, you know, write them, okay, your your subscription is up for the annual renewal. That is not being brought forward by their clawed scanner or whatever they have in place, right? So when you ping them on Bloomberg, you can actually get in contact with people. That's that's very valuable for me still. But it's the only reason why I have the subscription, more or less. It's the only reason. And then and I pay $30k. thirty K dollars for that a year. And I I happen to admit to it because you know it's it's still a people business to sell subscriptions to many people um within within the uh fund management industry, etc. But I think maybe in five, ten years from now, you know, who cares about the Bloomberg chat because everyone's already connected via the MCP that I've built within the cloud ecosystem and all of that. But
38:23Unidentified speaker (speaker_3): but I've even noticed a slowdown in email traffic. from you know like GMI subscribers. And it's not because they're less engaged, it's because people are overwhelmed and there's just they don't do that. People just stop doing it now. WhatsApp seems to be the most effective way of speaking to everybody now for me. Because nothing else people people don't use people stop using email.
38:47Unidentified speaker (speaker_0): But I I cannot
38:48Unidentified speaker (speaker_3): f
38:49Unidentified speaker (speaker_0): I I I'm still trying. to figure out whether this is a paradise or a nightmare for me because everyone's turning into the same kind of person that I've been for all of my life. You know, I've always lacked the attention span to read stuff from uh my counterparts via email and stuff. I never responded to emails and never been like that, right? I I I only respond to things that I get notified with on my phone, right? And um but now everyone's like that. And it it's it's something that I need to learn how to deal with because I've I've typically dealt with someone actually reading what I said to them. And they don't do that anymore.
39:26Unidentified speaker (speaker_3): No. And that's a weird thing because you know how much of your own research do you write do they actually sit down and read versus put it into their Claude plaud and say, you know, just take this into account. How Andreas is thinking about this. Does this change my positions? It's like, it's kind of weird.
39:45Unidentified speaker (speaker_0): One of our one of our members at Real Vision, Rao, uh, sent me a screenshot of an uh agent that she set up to only provide her with the summary of the stuff's To stop s stuff that I put in caps lock in in the research papers I write. And it was actually damn good because. I always put it in caps lock when I sincerely read something by this.
40:08Unidentified speaker (UNKNOWN): Yeah, I
40:08Unidentified speaker (speaker_3): always put mine in bold. Exactly. But I don't know how the world deal I don't I know there's something bigger in this and I'm still trying to think it through. This scarce scarcity t of human attention. We've saturated it. I can see it on X, I can see it everywhere, I can see it in businesses. We've saturated human attention. And I don't know how that changes things and who it changes for what. But I think we'll end up with less platforms and more super platforms. Yeah. Um, you know, it's hard enough to flip between you know, X and then your WhatsApp and then your email and then Realvision and then something uh the YouTube and then whatever, all of these things, like, it's too hard. There's too much context switching in different platforms and different things when it feels that a lot of this is all gonna be served in one place in a way that's digestible. But then you it's weird 'cause there's AI writing articles for
41:17AIs who are digesting articles.
41:21Unidentified speaker (UNKNOWN): But
41:21Unidentified speaker (speaker_3): I And compressing it for humans 'cause we're too stupid to be able to process this amount of bandwidth.
41:27Unidentified speaker (speaker_0): But f from an entrepreneurial standpoint, Rao, I I'm involved in a lot of uh, you know, angel investments within software in Northern Europe. And many of the cases that I've invested in um, they have they already have the clients. And given the speed of their coding now, they can basically deliver on all promises to all clients at the same time within you know 45 minutes. They just ask their account managers to sit out there, record the conversation with a client, and then it is basically automatically coded within the software ecosystem within days at maximum. max, right? I think that will make their moat even stronger compared to the two guys in a basement trying to create a new software product. Because what those two guys in the basement cannot do now compared to earlier is to build something that will take a lot of time to replicate.
42:21Unidentified speaker (speaker_3): Yes, but those guys that you invest in would do better. to allow it to also plug into Claude. Sure. So then people have one operating system and it becomes one of their tools that Claude uses. Like, you know, my Trading View. Yes, I use the thing, but I also have the feed into my Claude. I've got the MCP connector and I get my stuff from Trading View there, and I can analyze analyze charts and go back and say, Hey, what do you think of the chart of the dollar? And Claude will give me a you know a look at it and say, Yeah, you like DMAX, and this is where I think the DMARCs are, all of that stuff, without ever opening my trading view. That kind of I think is where it's all going. Yeah. Because the intelligence layer is too important. But that yeah, that yeah, I'm I d I don't know, I haven't kind of figured it all out yet, but there's something about this human attention scarcity that everything has to now get compressed for us. Because the reality is, is intelligence ran on carbon for the entire existence of us understanding anything. You know, the planet, everything runs on carbon intelligence. You basically put an energy through carbon and get intelligence. And it's just switched substrate to silicon.
43:28Unidentified speaker (UNKNOWN): And
43:28Unidentified speaker (speaker_3): so we now put fucking electricity through sand and we generate super intelligence, and it's a million times faster.
43:36Unidentified speaker (UNKNOWN): We
43:36Unidentified speaker (speaker_3): can't deal with that. It's Why there's a million times. more output of articles, and nobody can deal with it.
43:43Unidentified speaker (UNKNOWN): And the
43:43Unidentified speaker (speaker_3): only thing they can deal with is AI itself,
43:45Unidentified speaker (UNKNOWN): because
43:46Unidentified speaker (speaker_3): it's a million times faster.
43:47Unidentified speaker (speaker_0): And that is why I've launched a live stand-up tour in Denmark because I don't know what to do about
43:52Unidentified speaker (UNKNOWN): it. This makes
43:53Unidentified speaker (speaker_3): total sense. Yeah. And people want us to be chatting on camera because it
43:59Unidentified speaker (UNKNOWN): it
43:59Unidentified speaker (speaker_3): compresses information.
44:01Unidentified speaker (UNKNOWN): All
44:01Unidentified speaker (speaker_3): of the writing, all of the stuff, it comes down to the conversation in a way that's digestible that people think, okay, I kind of feel like I I'm I'm in touch with all of that now. Yeah. So it doesn't
44:12Unidentified speaker (UNKNOWN): go
44:12Unidentified speaker (speaker_3): away.
44:13Unidentified speaker (UNKNOWN): So
44:14Unidentified speaker (speaker_3): anything else you're looking at, or is that pretty much the whole world? I mean these are the most important things. I can't think of anything else that really matters.
44:21Unidentified speaker (speaker_1): Well, the only thing
44:24Unidentified speaker (UNKNOWN): that
44:25Unidentified speaker (speaker_1): has
44:25Unidentified speaker (speaker_0): surprised me a little bit during these months of a stronger dollar is actually how crypto has outperformed a lot of peers through this
44:35Unidentified speaker (UNKNOWN): now.
44:36Unidentified speaker (speaker_0): Uh I probably wouldn't have had that on my bingo card.
44:40Unidentified speaker (speaker_3): Now I looked at the the ratio of the Nasdaq versus Bitcoin. It was two standard deviations oversold in its log trend, so it was really oversold. But yeah, it's done pretty well. Was
44:50Unidentified speaker (UNKNOWN): it
44:50Unidentified speaker (speaker_3): seemed that there was a rotation and it although it still can't
44:56Unidentified speaker (UNKNOWN): you know
44:56Unidentified speaker (speaker_3): it can't get traction because there's not enough liquidity yet
45:00Unidentified speaker (UNKNOWN): but
45:01Unidentified speaker (speaker_3): I think as soon as liquidity comes back as soon as it comes back then you know it's got like catch up.
45:06Unidentified speaker (speaker_0): I just struggled with the notion that it's either a crypto or AI.
45:10Unidentified speaker (speaker_3): Yeah. And
45:11Unidentified speaker (speaker_0): I mean those two themes go hand in hand for a range of obvious reasons, but it just seems like that narrative exists because people do not have enough liquidity to i invest in both themes at the same time at the moment. And you can see that because you can see Robin Hood's earnings, you can see Coinbase people even Robin Hood, people don't have enough money really to invest. So it's you know, there's a compass capacity constraint. from households that they don't have a lot of money because earnings haven't really been going up. Mortgage costs are high, rents are high, food costs are high, and that's
45:47Unidentified speaker (speaker_3): not getting resolved. So, you know, the only way around that is liquidity and lending and all of that stuff. And that's not really happening because the yield curve's too flat.
45:55Unidentified speaker (UNKNOWN): It
45:56Unidentified speaker (speaker_3): all comes
45:56Unidentified speaker (UNKNOWN): back
45:56Unidentified speaker (speaker_3): to the same thing. It's always liquidity in the end.
45:58Unidentified speaker (speaker_0): Yeah. And um Claudus, did the you new windows? I won't I wanted to say that as a one liner before rounding off here. And I don't know whether OpenAI is the Mac of the desktop space. That who knows? Or or is
46:13Unidentified speaker (UNKNOWN): it Gemini?
46:14Unidentified speaker (speaker_0): Yeah. Yeah,
46:15Unidentified speaker (speaker_3): I've thought it's kind of the Mac, but no. We'll see. It's a whole new world. Listen, mate, fantastic conversation. Hopefully we've added a bit of value and not confused everybody. But uh we both think that the market, unless something dramatically changes with liquidity, the market is mispricing all of this. Yeah,
46:33Unidentified speaker (UNKNOWN): I
46:33Unidentified speaker (speaker_1): agree.
46:34Unidentified speaker (UNKNOWN): All right,
46:34Unidentified speaker (speaker_3): see you next time, mate.
46:36Unidentified speaker (speaker_1): Take care.
46:37Unidentified speaker (speaker_0): This episode is brought to you by Pith. Pith Network is the fastest growing financial data company today, trusted by Fidelity Investments, the US Department of Commerce, Revolut, Calci, CalShi, Jane Street, Coinbase, and many more. They provide real real-time financial market data across over 3,000 equities, commodities, crypto, FX, rates, and again, a lot more. They're the first and largest provider of 24-7 financial indices and offer the widest array of market data for the lowest cost. So go to pith.network to take advantage of their free trial.
47:09Unidentified speaker (UNKNOWN): You
47:09Unidentified speaker (speaker_3): obviously enjoyed the episode because you're here with me at the end. But listen, don't forget to go to realvision.com forward slash join and grab a free membership. It's an incredible community packed with alpha, great investment ideas and the research that you need to help you unfuck your future. So get started now, go to realvision.com forward slash join.