Latest / Deep Values: Value Investing Deep Dives / [Deep Value] Howard Marks: Is It A Bubble?
Transcript
- 0:00Welcome back to The Deep Dive, the only show that's really
- 0:03engineered to deliver that comprehensive, synthesized
- 0:05insight directly to you, the learner.
- 0:08And today we are wading into probably the most volatile
- 0:12paradigm shifting market phenomenon of the decade.
- 0:15We have to be talking about the artificial intelligence
- 0:17revolution specifically, you know, generative AI, large
- 0:21language models. It is a phenomenal moment.
- 0:23I mean, it's one where technology is genuinely changing
- 0:26the way we all work and live. But.
- 0:28For investors, that tension is It's almost unbearable.
- 0:32It is. You have trillions of dollars
- 0:34valuation resting on a single, really a binary question.
- 0:37Right. Is this rise, this thing being
- 0:39led by the so-called Magnificent 7 stocks?
- 0:42Is it a true secular economic shift that actually justifies
- 0:46these prices? Or is it simply a classic
- 0:48speculative mania? Just, you know, wearing a new
- 0:51digital disguise. And that kind of ambiguity, it
- 0:53demands an analysis that looks way beyond just a quarterly
- 0:56earnings reports. We need a guy who specializes in
- 0:59investor behavior. Someone who's survived and
- 1:01documented multiple cycles of excess and panic.
- 1:05Exactly. We need a philosopher of the
- 1:07market, not just a spreadsheet jockey.
- 1:10And that is precisely why we're dedicating this entire deep dive
- 1:13to the work and wisdom of Howard Marks.
- 1:15The Co chairman and cofounder of Oak Tree Capital Management.
- 1:18Right Marks is recognized not just as a successful investor,
- 1:21but really as an intellectual giant in finance.
- 1:24His expertise is in dissecting investor psychology, mapping
- 1:28those market cycles. And translating these complex
- 1:31historical risks into something you can actually use.
- 1:33Exactly. His famous memos, especially on
- 1:37Bubble Watch and the more recent one, Is it a bubble?
- 1:40They're going to be the source material for today.
- 1:42So our mission is to really dissect his incredibly nuanced
- 1:46diagnosis of where we are right now.
- 1:48And he's not calling for a crash tomorrow, but he is sounding a
- 1:51very clear alarm. He labels the current
- 1:53environment as lofty, but not nutty.
- 1:56A very specific choice of. Words it is, and our goal is to
- 2:00take you through his entire intellectual framework, you
- 2:02know, integrating historical ghosts like the TMT bubble and
- 2:05the Nifty 50 with all the unique structural risks of the AI
- 2:09ecosystem. So that you can understand
- 2:10exactly what lofty but not nutty means for your portfolio today.
- 2:15OK, before we embark on this analysis, we do have to share a
- 2:18necessary note for you, the learner.
- 2:21All the analysis we're sharing here is strictly based on
- 2:23publicly available sources, and it's intended for educational
- 2:26purposes only. This content does not, under any
- 2:29circumstances, constitute investment advice.
- 2:32We really urge you to conduct your own due diligence before
- 2:35making any investment decisions. And if you find these deeper
- 2:38insights compelling, we'd encourage you to join the Deep
- 2:40Value Club membership for more advanced, in depth content.
- 2:44You can also subscribe to us on Apple and Spotify podcast.
- 2:47Just search for deep value investing.
- 2:50And for those interested in tracking the total market
- 2:53temperature, that essential gauge Marx tells us to watch,
- 2:55we'd recommend searching for and downloading the iOS app, a
- 2:58Buffett Indicator from the Apple App Store.
- 3:00Yeah, that's a fantastic tool for monitoring current market
- 3:03valuations against, you know, historical norms.
- 3:06A really useful piece of kit. Definitely.
- 3:08OK, thank you for that. With the context set, let's now
- 3:11begin our deep dive into Marx's intellectual architecture.
- 3:14Right. Understanding his definition of
- 3:16a bubble is really the first step to understanding the AI
- 3:20market today. So Marx argues that if you rely
- 3:22solely on, say, quantitative metrics to define a bubble,
- 3:27you're going to miss it every single time.
- 3:29So to really appreciate his lofty but not nutty verdict, you
- 3:34have to accept his foundational premise.
- 3:36Which is that bubbles are fundamentally psychological
- 3:38events, they're driven by emotion, and only secondarily do
- 3:42they manifest in the, you know, the mathematics evaluation.
- 3:45So we've got to look beyond the spreadsheets.
- 3:46What you're saying is that traditional metrics like PE
- 3:49ratio of 30 or 40X, they're just symptoms of a bubble.
- 3:52Exactly right. They're not the cause itself.
- 3:54So what's the cause? The cause is a breakdown of
- 3:57human reason. Marx defines a bubble as a state
- 3:59where that intense emotional desire to get spectacularly rich
- 4:04just completely overwhelms the rational, calculated fear of
- 4:07loss. The collective mind in the
- 4:08market just it shifts. It transitions from prudent
- 4:12calculation to a mass suspension of disbelief.
- 4:15The numbers only become impossible because the
- 4:17psychology became impossible first.
- 4:19That mass psychology of envy is just such a powerful force in
- 4:23markets isn't. It oh, it's everything.
- 4:24It's what breaks the discipline of even the most seasoned
- 4:27investor. Which reminds me, Marx
- 4:30frequently points back to the the canonical source on this
- 4:34psychological breakdown. He does.
- 4:36He often cites that powerful insight from economists Charles
- 4:39Kingelberger and Robert Alliber. From their book Manias, Panics
- 4:42and Crashes. Right.
- 4:43The central line is there is nothing so disturbing to 1's
- 4:47well-being and judgment as to see a friend get rich.
- 4:50That is the idiot neighbor paradox in its purest form.
- 4:53Yes, you know, you, the disciplined investor.
- 4:56You're sitting there on your steady maybe 8% returns and your
- 5:00neighbor who bought some random highly leveraged tech stock
- 5:03based on a social media rumor is bragging about a 300% game.
- 5:07That envy it just starts to eat away at your own certainty.
- 5:10You start questioning your proven rules.
- 5:12That's the seed of irrationality.
- 5:13And that seed, it blossoms very quickly into what Marx calls the
- 5:18No Price Too High syndrome. OK, break that down.
- 5:21Well, in any efficient, healthy market, there's a natural
- 5:23governor, right? It's risk aversion.
- 5:26As an asset's price rises, its potential for future returns
- 5:29goes down and its risk goes up, so demand should eventually
- 5:33fall. But in the middle of a
- 5:35speculative mania, market physics just seem to invert.
- 5:39Precisely. Rising prices start to generate
- 5:41increased demand, and it's driven entirely by the fear of
- 5:44missing out. FOMO.
- 5:45The higher it goes, the more people want in.
- 5:47Exactly. The higher the asset goes, the
- 5:49more validated the early buyers feel, and the more anxious the
- 5:53leak comers become about being left behind.
- 5:55And Mark's pinpoint points that crucial transition point.
- 5:58It's when investors stop asking the necessary discipline
- 6:01question. Is this asset worth the price
- 6:03based on its projected future cash flows?
- 6:05And instead, they capitulate. They start asking.
- 6:07I must own this at any price or I will miss the future.
- 6:10And once you start hearing those verbal markers, you know you're
- 6:13deep in the woods. You hear fund managers and
- 6:15commentators dismissing basic principles.
- 6:18Oh, you hear it all the time. Faiza's like valuation doesn't
- 6:21matter for a company this good or.
- 6:23PE ratios are an artifact of the Industrial age.
- 6:28Those statements are the loudest alarm bells a disciplined
- 6:31investor can possibly hear. They signal that the narrative
- 6:34is trying to provide cover for bad math and these manias they
- 6:38require specific breeding ground.
- 6:40The new thing? The necessity of the new thing.
- 6:43Bubbles rarely form around, you know, established, mature
- 6:46industries like electric utilities or consumer staples.
- 6:50They need a technological or financial innovation like.
- 6:53Railroads. The transistor, The Internet.
- 6:55Or today, generative AI, something that promises to
- 6:59fundamentally change the world. It's that seductive power of the
- 7:02this time is different argument. Which John Templeton famously
- 7:05called the foremost dangerous words in investing and that
- 7:08narrative it it provides the intellectual justification for
- 7:11suspending your valuation discipline.
- 7:14Marx emphasizes the crucial fallacy here.
- 7:16It is investors conflate the validity of the technology with
- 7:19the value of of the equity. OK, that's a key distinction.
- 7:22We can all agree that AI is transformative.
- 7:24That's the validity. But the fallacy is in the next
- 7:27step, the belief that because the technology is
- 7:30transformative, the profits must be infinite and therefore the
- 7:33equity must be worth any price. And that's what allows the
- 7:36multiples to expand to levels that are just mathematically
- 7:39impossible to justify under any reasonable assumption.
- 7:42That's the trap. That framing is essential.
- 7:44OK, OK, let's unpack this with Marx's specific diagnostic tool,
- 7:49The Three Stages of a Bull Mark Market.
- 7:51This gives us a structural way to figure out where we are right
- 7:55now. Yeah, this heuristic is
- 7:56fundamental to how oak tree approaches cyclical positioning.
- 8:00It defines the psychological shift in three clear stages.
- 8:03Stage 1 is the recovery phase. This.
- 8:05Is right after a deep recession or crash.
- 8:06Exactly. Only a few unusually insightful
- 8:09people are buying assets. They're recognizing deep value
- 8:12when everyone else is just paralyzed by fear.
- 8:14This is the period of maximum discomfort and maximum potential
- 8:18returns. That's where you make the
- 8:19easiest money, but it feels the hardest to do.
- 8:22Then comes stage 2. The realization phase.
- 8:25Right. The economy improves, earnings
- 8:27growth accelerates, and most people accept that improvement
- 8:30is actually taking place. This is the healthy, rational
- 8:33bull market. So prices are rising because
- 8:35fundamentals are genuinely getting better.
- 8:37Correct earnings are increasing and the risk reward trade off
- 8:40still remains reasonable. And Stage 3?
- 8:43That's the danger zone. The transition from Stage 2 to
- 8:45Stage 3 is the single most critical judgement an investor
- 8:49has to make, isn't it? It is.
- 8:50Stage 3 is pure, unadulterated euphoria.
- 8:54After years of strong gains, investors just lose all
- 8:57perspective. The consensus concludes that
- 9:00things can only get better forever.
- 9:02Risk aversion evaporates. Speculation becomes the norm and
- 9:05skepticism is actively ridiculed.
- 9:07So Marx's current challenge, and really the focus of his latest
- 9:11memos, is pinpointing if the AI boom is still justifiable Stage
- 9:152 growth rally supported by exploding earnings, or if the
- 9:18psychology has already crossed into Stage 3 mania.
- 9:21He's got a diagnostic marker table to help with that.
- 9:24Let's use that to summarize the difference between those two
- 9:26psychological states. This helps synthesize the
- 9:29behavioral cues you need to be watching for.
- 9:31The contrast is pretty stark in a healthy bull market, Stage 2
- 9:35investor sentiment is optimistic, but it's grounded in
- 9:38economic data. But in a speculative bubble,
- 9:40Stage 3, it's outright euphoric. It's driven by your consensus
- 9:45and that overwhelming belief that yes, this time is
- 9:48different. And the practical difference in
- 9:50how assets are valued is maybe the most glaring sign.
- 9:53In Stage 2, valuation methodology is focused on
- 9:56tangible reality, earnings, cash flow, replacement cost.
- 10:01But in stage 3. It shifts entirely to
- 10:03abstraction. It becomes what Marx calls price
- 10:05to dream or new nonsensical metrics like user growth or
- 10:10eyeballs. We remember eyeballs.
- 10:12Right. People are no longer grounding
- 10:14their valuations in discounted future cash flows.
- 10:16They're pricing based on the maximum possible future
- 10:19scenario. And we also see that fundamental
- 10:21shift in risk attitude. Stage 2 is fundamentally loss
- 10:24averse. You're asking how much can I
- 10:26lose if this doesn't work out? But stage 3 is entirely
- 10:30dominated by FOMO. It's only asking how much can I
- 10:32make if I'm right. That shift from mitigating loss
- 10:36to maximizing gain is a hallmark of euphoria.
- 10:40And market breadth is critical to healthy Stage 2 markets.
- 10:44They see broad participation across sectors.
- 10:46A rising tide lifts all boats but.
- 10:48Bubbles are defined by extreme narrow concentration in the new
- 10:52thing, a phenomenon we are certainly witnessing today with
- 10:55the Magnificent 7. And finally, that treatment of
- 10:57skepticism in a healthy market, if you voiced out, you get a
- 11:01constructive debate. But in a stage 3 environment, if
- 11:04you voiced out, you're just dismissed as a dinosaur who's
- 11:07incapable of grasping the future.
- 11:08That's a huge tell. If you start to feel
- 11:11intellectually attacked for simply asking about valuation,
- 11:13you're very likely in Stage 3 territory.
- 11:16OK, that framework gives us the surgical tools for analysis now,
- 11:19Marks insisted. To understand the risk of AI
- 11:21today, we have to look backward. The market has been here before
- 11:24many times. So let's start with the
- 11:25precedent that really solidified his reputation as a cycle expert
- 11:29the TMT bubble. Technology, Media Telecom
- 11:3219992000. Marx's memo bubble.com was just
- 11:36a masterpiece of timing and an insight.
- 11:38It came out in January 2000, just weeks before the NASDAQ
- 11:42peak. And it was a direct, rational,
- 11:44intellectual assault on the prevailing mania.
- 11:47The specifics of that madness? They were astonishing.
- 11:49Even looking back, you had companies with 0 profits, often
- 11:53minimal revenue, commanding billion dollar market caps based
- 11:56on metrics that were pure fantasy.
- 11:59He identified the fundamental economic error of that era.
- 12:02The volume over margin fallacy. We joke about it now, but people
- 12:05really believed it. The actual stated operating
- 12:08belief for many dot coms was that they were losing money on
- 12:11every single transaction, but somehow they'd scale up, achieve
- 12:15massive volume, and then magically become profitable
- 12:18later. It just defies basic economics.
- 12:20And here's the key lesson for AI.
- 12:22Marx made that crucial distinction that the Internet
- 12:24did change the world. The validity of the technology
- 12:27was absolute. But that validity was not enough
- 12:29to save investors. Right, the enduring lesson is
- 12:32the validity valuation gap. Marx frequently reflects that
- 12:35the underlying premise that the Internet would transform society
- 12:38and commerce came absolutely true.
- 12:41Yeah, the vast majorityofthose.com and
- 12:43e-commerce stocks that benefited from that thinking.
- 12:45They're worthless now, or just gone entirely.
- 12:48Exactly. Being right about the profound
- 12:50impact of the technology is necessary, but it is not
- 12:54sufficient for making money. Valuation is the ultimate
- 12:57arbiter. So let's apply that gap directly
- 13:00to today's AI infrastructure. Today, NVIDIA and other chip
- 13:04makers are universally hailed as the essential pick and shovel
- 13:08suppliers of the AI gold rush. Right.
- 13:11So who are their closest equivalentsduringthe.com era?
- 13:14The infrastructure builders. We're talking about companies
- 13:16like Cisco Systems, Lucent Technologies, JDS, Uniphase.
- 13:19They weren't pets.com. No.
- 13:21They were building the physical backbone, the routers, the
- 13:23switches, the fiber optic lines that enabled the Internet to
- 13:26function at scale. They had real products exploding
- 13:29sales and they were profoundly profitable at the time.
- 13:31They were superior established businesses.
- 13:33And yet even Cisco, a fundamentally great company that
- 13:37was essential to the technology's long term success,
- 13:41suffered an absolutely catastrophic fate.
- 13:43It just shows how dangerous the entry price.
- 13:45Is the losses were breathtaking. Cisco lost nearly 90% of its
- 13:49market value from its peak in March 2000, and what should
- 13:53haunt today's investors is the time horizon.
- 13:56How long did it take to recover? It took Cisco, arguably one of
- 13:59the most successful technology companies of the past 30 years,
- 14:03roughly 20 years to simply regain its March 2000 high.
- 14:07Two decades of 0 return for buying a great asset at a
- 14:10catastrophically high price. And this illustrates the primary
- 14:13risk for AI today. You can be perfectly right about
- 14:16the technology that AI will dominate, but if your entry
- 14:19price discounts a future that is too distant, too optimistic, and
- 14:22ignores competition. Your investment outcome can
- 14:25still be devastating. Exactly.
- 14:26That analogy suggests the AI risk isn't just about the
- 14:29outright speculative garbage, but about buying high quality
- 14:32essential assets at an unsustainable multiple.
- 14:35But the TMT comparison focuses on companies with no earnings.
- 14:40Marx often references another historical parallel that feels
- 14:44closer to the current valuation structure of the Magnificent 7,
- 14:47the Nifty 50, the Nifty 50 from the late 60s and early 70s.
- 14:50These were the original one decision stocks.
- 14:52Right companies perceived as being so high quality, so
- 14:55reliable, so dominant that the decision was simply to buy them
- 14:58and never sell regardless of the price we're.
- 15:00Talking about the Titans of that era.
- 15:02Xerox, Polaroid. Coca-Cola.
- 15:04Disney. Procter and Gamble, IBM,
- 15:06McDonald's, these were not small cap speculative startups.
- 15:09They were the blue chip growth stocks of their day, right?
- 15:13But because of this universal, unquestioning consensus on their
- 15:16quality, they traded at PE ratios that are shocking by
- 15:20historical standards. How high are we talking?
- 15:22Often hitting 70X80X even 90X earnings.
- 15:26The prevailing theory was that quality was worth any price and
- 15:29they were too high quality to fail.
- 15:31Then the deep recession in bear market hit between 73 and 74.
- 15:35And the simultaneous collapse of multiples and earnings meant
- 15:37their valuations were just crushed.
- 15:39They were. But the critical warning for The
- 15:42Magnificent 7 isn't just the price collapse, it's the lack of
- 15:45persistence of that dominance, isn't it?
- 15:47That is the most salient point for today.
- 15:49The current assumption embedded in the market cap of the
- 15:52Magnificent 7, Google, Microsoft, NVIDIA, Apple,
- 15:54etcetera is that they will remain globally dominant and
- 15:57compound earnings at 20% to 30% annually forever.
- 16:02And the Nifty 50 history just violently counters that.
- 16:05Where are Polaroid, Kodak and Xerox today?
- 16:08You were disrupted. They didn't persist.
- 16:11Technology and dominance are transient.
- 16:13The assumption of infinite compounding is a bet against the
- 16:16very nature of aggressive capitalism and technological
- 16:19change, which you know it thrives on disruption.
- 16:22So whether we look at the pure speculation of TMT or the
- 16:25excessive quality pricing of the Nifty 50, the lesson is clear.
- 16:29Being right about the future doesn't make up for bad
- 16:31valuation discipline, right? Let's shift gears now and apply
- 16:34these historical lessons to the specific structure of the AI
- 16:37landscape today. Marks readily concedes that AI
- 16:40is the definitive new thing of the cycle.
- 16:42I mean, it's capacity for rapid learning, generating
- 16:45sophisticated code, passing the Turing test feels genuinely
- 16:50unprecedented. Almost like magic.
- 16:52And that sense of magic is what allows the investor to throw out
- 16:55the rulebook. If it's magic, how can you
- 16:57possibly value it? That's the intellectual trap,
- 17:00and Marx identifies 3 specific elements that make AIA
- 17:03particularly fertile ground for speculative excess, maybe even
- 17:07more so than the Internet bubble.
- 17:08OK, what are they? 1st is the promise of
- 17:10omnipotence. The narrative surrounding AI
- 17:13suggests it's going to solve everything from climate change
- 17:16to curing disease, creating an infinite total addressable
- 17:19market mindset. And if the market is infinite,
- 17:22then any current price can be justified.
- 17:24That's the thinking. The second is the complexity,
- 17:27the black box nature of the underlying technology.
- 17:30Right, the lack of understanding.
- 17:31This is a major differentiator from the Internet.
- 17:34Few investors, even the most sophisticated professionals, can
- 17:37independently verify the fundamental claims made by
- 17:40management teams about their proprietary models or capability
- 17:44loops. So this complexity forces you to
- 17:46rely on hype on the narrative and the credibility of the
- 17:50speaker, rather than hard, independently verifiable
- 17:53fundamentals. You're buying on faith, which is
- 17:56the definition of stage 3 psychology.
- 17:59And the third element is that sheer corporate terror that's
- 18:01driving immediate spending. The fear of obsolescence.
- 18:05This is an existential threat driving institutional capital.
- 18:08Every CEO, every board member is terrified that if they don't
- 18:12invest in AI infrastructure immediately, their company will
- 18:15become irrelevant within five years.
- 18:17So this drives what Marx calls forced buying capital deployment
- 18:21regardless of price or proven utility, right?
- 18:24And this pressure is internal and almost impossible to resist,
- 18:27which just artificially inflates demand.
- 18:29OK, this analysis leads us to a crucial and very sophisticated
- 18:33point that Marx and Oak Tree have raised regarding the
- 18:36structural risks of the current revenue model.
- 18:38The risk of the circular economy.
- 18:40This is a deep analytical cut that needs far more exploration.
- 18:43Yeah, let's break it down. We are currently deep in what
- 18:46the economist Carlota Perez calls the installation phase,
- 18:50the colossal build out of physical and digital
- 18:52infrastructure. And the obvious beneficiaries
- 18:54are the chip makers, followed by the cloud hyperscalers.
- 18:57Right. And the market is pricing these
- 18:59two groups for perpetual exponential growth.
- 19:03Marx draws an uneasy historical parallel here to the late 1990s
- 19:07telecom bubble, which was riddled with capacity swaps.
- 19:11Let's walk through that parallel.
- 19:12Yeah, in the 90s, telecom companies sold fiber optic
- 19:15capacity to each other to book revenue and meet Wall Street
- 19:18growth expectations. So the demand was internal to
- 19:21the industry. It wasn't based on proven end
- 19:23user subscription growth. Exactly.
- 19:25It created an illusion of robust demand and profitability that
- 19:29just vanished when the internal trade dried up, and Mark
- 19:32suggests we may be seeing a modern AI version of this
- 19:35financial loop. Let's follow the money trail in
- 19:38the AI circular economy because this is where the risk
- 19:41crystallizes for the learner. OK, it starts with the AI chip
- 19:43makers who are reporting massive revenue growth, trillions in
- 19:47value added in just two years. That revenue is largely funded
- 19:51by the vast capital expenditures, the CapEx of the
- 19:53cloud hyperscaler. Amazon Web Services, Microsoft
- 19:57Azure, Google Cloud, they're essentially buying the new
- 20:00digital oil rigs. Right, The hyperscalers are
- 20:02spending hundreds of billions on hardware now.
- 20:04Where does their money come from?
- 20:06Well, they attempt to recoup that spending by selling
- 20:08computational time and services to the third layer, the
- 20:12thousands of AI startups, smaller enterprises and
- 20:15corporate development teams. And crucially, the hyperscalers
- 20:18often fund this layer of demand by giving out huge amounts of
- 20:22cloud credits. Or by taking equity stakes in
- 20:25the startups that are then forced to spend their operating
- 20:27capital on their cloud platform. So the startups spend their cash
- 20:31or credits on the cloud services of the hyperscalers, who in turn
- 20:35use that massive revenue to justify buying the next
- 20:38generation of expensive chips. It's a closed loop.
- 20:40A self referential circle of cash.
- 20:42And Mark's central question, the one every investor has to ask,
- 20:45is where is the actual proven external user paying cash for
- 20:50productivity gain that justifies this entire massive cycle?
- 20:54We're still waiting for the deployment phase.
- 20:56We need actual businesses paying cash to AI services that
- 20:59genuinely improve margins or productivity outside of this
- 21:03tech and startup bubble. So the potential reversal, the
- 21:07Wiley Coyote moment you mentioned earlier, comes if
- 21:11those hyperscalers can't monetize their data center
- 21:13capacity fast. Enough, or if the startups run
- 21:16out of runway, if that external real world monetization doesn't
- 21:19materialize soon, the hyperscalers will see their
- 21:22enormous CapEx as stranded assets.
- 21:25And they will immediately slash their orders for chips.
- 21:27Leading to a sudden dramatic contraction across the entire
- 21:30valuation pyramid, starting with the chip makers.
- 21:33Marks explicitly warns that the fierce competition among the
- 21:36leading tech firms means they may be content to experience
- 21:40losses in their AI businesses for years in order to gain
- 21:42share. Which suggests A margin
- 21:44destroying battle is ahead and that's not currently factored
- 21:47into the stock prices. Not at all, and this structural
- 21:49fragility is made exponentially worse by the sheer concentration
- 21:53of the market. The Magnificent 7 concentration
- 21:56risk. The degree of market
- 21:57concentration is historically extreme.
- 21:59A handful of names now account for the overwhelming majority of
- 22:02the S&P 500's performance over the last two years.
- 22:06And while these are fundamentally superior
- 22:08businesses to the.com era, they have massive of cash flows and
- 22:11established moats. The concentration itself creates
- 22:14systemic fragility. Meaning that if the AI trade on
- 22:18wines, passive investors are inherently making a massive,
- 22:22unacknowledged active bet that will hit the entire market.
- 22:25Exactly. Anyone invested in a broad index
- 22:27fund like the S&P 500 is now implicitly making a huge
- 22:31leveraged bet on the perpetual dominance and valuation of these
- 22:35seven companies. And Marks cautions that we must
- 22:37not forget the law of large numbers.
- 22:39It's encapsulated in his famous maxim Trees don't grow to the
- 22:43sky. Right.
- 22:44The underlying assumption that these companies will continue to
- 22:47compound earnings at 2030% indefinitely is mathematically
- 22:51highly skeptical. And it's fundamentally a bet
- 22:53against competition, economic gravity, and the entire history
- 22:56of corporate dominance. So we've built the necessary
- 22:59intellectual structure, we've reviewed history and we've
- 23:02assessed the current structural risks, especially the fragility
- 23:05of that circular economy. And this brings us back to
- 23:08Marx's central conclusion. But not nutty.
- 23:12This profoundly important distinction Why is he so
- 23:15unwilling to use the word bubble today?
- 23:17The distinction lies entirely in separating an expensive market
- 23:20from an irrational, manic one. If the market is merely
- 23:24expensive, say it's trading AT25X or 30X PE, the risk
- 23:28profile is stagnation, poor future returns, or maybe a
- 23:31cyclical correction of 2030%. But if it were a full blown
- 23:35nutty bubble where PE is infinite or just irrelevant like
- 23:38in 2000. Then the risk is a catastrophic
- 23:4050% to 80% market crash. The firewall preventing the
- 23:44nutty status then is real profitability.
- 23:46It is. Marx classifieds the AI market
- 23:49as lofty because the key drivers, the chip makers and the
- 23:52hyper scalers have massive, undeniable profitability.
- 23:55NVIDIA, Microsoft, Google. These are not pets.com.
- 23:58No, they have huge existing businesses, vast cash reserves.
- 24:01And crucially, their earnings growth has for the most part,
- 24:05been phenomenal enough to keep pace with the price
- 24:07appreciation, although just barely.
- 24:09So the multiples are elevated and stretched, making them
- 24:12lofty, but they're not based on non existent profits.
- 24:15Which was the definition of nutty in 2000?
- 24:18This suggests the air in the balloon is just very thin, not
- 24:20entirely absent. But how does Marx gauge the
- 24:24psychological temperature? We established that bubbles are
- 24:27defined by psychology. So what's the final indicator of
- 24:29mania? He relies on his most reliable
- 24:32real world indicator, a man in the street indicator, right?
- 24:35Marx believes unsustainable manias only happen when
- 24:39speculation permeates the entire society.
- 24:42It has to move beyond professional investors and
- 24:44dedicated tech enthusiasts. In 1999 it was the taxi drivers,
- 24:48the bartenders, the hairdressers giving you stock tips.
- 24:50Telling you to quit your job to day trade and his current
- 24:54observation, he states. Clearly I don't detect that
- 24:57level of mania at this time. So the euphoria is largely
- 25:00contained. It's contained within the
- 25:01professional investment community, the Silicon Valley
- 25:04ecosystem, and those already actively trading.
- 25:07The average household isn't yet pulling out their home equity to
- 25:10invest in AI micro caps. In the absence of that's society
- 25:15wide, universal mania is the key psychological factor preventing
- 25:19him from upgrading the verdict from lofty to nutty.
- 25:22That's the firewall. But if the man in the street
- 25:25indicator is the only thing preventing nutty status, doesn't
- 25:28that make the current situation incredibly fragile?
- 25:31It's absolutely fragile, and Marx acknowledges that
- 25:34psychological stability can break very quickly.
- 25:36So he's really relying on the absence of evidence, the absence
- 25:39of universal euphoria, rather than positive evidence of
- 25:43valuation support. That's seems inherently risky.
- 25:46It is, which is why he categorizes the current
- 25:48environment as late Stage 2 or early Stage 3.
- 25:52The difference is critical for positioning.
- 25:54Let's look at his comparative analysis of the three major
- 25:56bubbles he's documented. The TMT bubble of 2000 was
- 25:59nutty. It was characterized by
- 26:01nonexistent profitability, speculative business models, low
- 26:04corporate debt. The housing bubble of 2008 was
- 26:07toxic. Driven by extreme systemic
- 26:10credit risk and bank leverage, the underlying assets were
- 26:13priced reasonably, but the debt structure was fatally flawed.
- 26:16And the AI boom of 2024 is lofty.
- 26:20Characterized by high PE ratios applied to real massive earnings
- 26:24and crucially, high corporate cash reserves rather than
- 26:27extreme systemic debt leverage. That context shows the risk
- 26:30isn't toxic credit, but simply stretched pricing.
- 26:34But this entire lofty market is now colliding with a massive
- 26:38macroeconomic sea change. This is perhaps the most
- 26:40important context for long term investors.
- 26:42The gravity of rates. Yes.
- 26:44Marx argues that the 40 year era of secularly falling interest
- 26:47rates from 1980 to 2021 is fundamentally over.
- 26:51We've entered a new epoch defined by persistent structural
- 26:54inflation and a real persistent cost of capital.
- 26:57The end of free money Essentially, we need to explain
- 26:59clearly to the learner why near 0 Interest rates were the
- 27:02lifeblood of high growth tech valuation.
- 27:04It all comes down to the mechanics of discounted cash
- 27:06flow. How you value any asset that
- 27:08promises future profits today. Right.
- 27:10When interest rates are 0 or near 0, the discount rate, the
- 27:14required rate of return you use to calculate present value, is
- 27:18negligible. If money is free, the value of
- 27:21those massive theoretical AI profits promised 10 years from
- 27:24now is exponentially higher. Potentially infinite.
- 27:27It creates the perfect gravity free environment for long
- 27:30duration growth assets like tech stocks to soar.
- 27:33But now we live live in a world where the Federal Reserve is
- 27:36holding rates at 5% or higher and the risk free return on cash
- 27:40is 4% to 5%. That changes the gravity
- 27:43entirely. When the discount rate moves
- 27:44from 1% to 5%, the present value of those far off future profits
- 27:49just collapses dramatically. Give us an example.
- 27:51OK. A $1 billion profit projected 10
- 27:54years in the future, discounted back at 1%.
- 27:56That profit is worth about $905 million today.
- 27:59OK. Discounted back at 5%, the same
- 28:02$1 billion profit is only worth about $614 million today.
- 28:06That's a 32% valuation haircut just from the change in the
- 28:10interest rate environment. It illustrates the point
- 28:12perfectly. Higher interest rates
- 28:13structurally raise the hurdle rate for equities.
- 28:16Making current profits more valuable than future profits.
- 28:19Absolutely. Mark's warns that investors who
- 28:21are today applying the 2021 multiples that zero rate
- 28:24valuation standards to 2024 interest rates are engaged in a
- 28:28massive structural mismatch bed. Higher rates naturally pull down
- 28:33valuations. Even if AI succeeds, the
- 28:36increased cost of capital will test the validity of every high
- 28:39multiple in the current market. So we have a market that is
- 28:42lofty, high risk, high prices, but not completely divorced from
- 28:46reality. And that market is colliding
- 28:49with the gravity of high interest rates, right?
- 28:51Mark seamlessly avoids market timing, but he does advise
- 28:54investors on how to define where we are in the cycle.
- 28:57So what is the prescriptive advice for the learner?
- 28:59His core message is that knowing where we are late Stage 2, early
- 29:03stage 3, where comfort is high and skepticism is low, dictates
- 29:07a necessary defensive shift in your risk posture.
- 29:10So he advises moving to what he calls an investment readiness
- 29:13condition. It's characterized by increasing
- 29:15defensiveness, trimming back risk exposure, and optimizing
- 29:18your portfolio to survive the inevitable collision between
- 29:22infinite expectations and economic reality.
- 29:24The goal is survival and preservation, not speculation.
- 29:28Exactly. What does that defensive posture
- 29:30look like in practice? Let's detail the defense
- 29:33tactics. First and foremost is the
- 29:35discipline of trimming winners. Marx argues that if you've been
- 29:39heavily invested in the AI trade and have seen violent
- 29:42appreciation, you have to take some chips off the table.
- 29:45This isn't about selling everything.
- 29:46No, it's about rebalancing your portfolio.
- 29:49When a single assets appreciation dictates your
- 29:51entire risk exposure. You are speculating, not
- 29:54investing. Trimming removes that exposure
- 29:57and frees up capital. You're rotating away from the
- 29:59assets where the consensus is highest and the comfort is
- 30:02strongest. Exactly.
- 30:03Secondly, emphasizing value, you need to rotate the capital you
- 30:06just freed up into neglected sectors or assets that have
- 30:09demonstrable current cash flows and reasonable grounded
- 30:13valuations. So these are often found in the
- 30:15old economy sectors like traditional energy, utilities,
- 30:18specific industrials. Things that haven't been subject
- 30:21to the same speculative fever. You're moving from a price to
- 30:24dream model to a price to cash flow model.
- 30:26And the most important attitude shift is the commitment to
- 30:29skepticism. You have to apply a much higher
- 30:32burden of proof to every new investment idea, especially
- 30:35those linked to the AI narrative.
- 30:37Stop asking the stage 3 question.
- 30:39How much can I make? And force yourself to focus on
- 30:42the discipline. Stage two question What?
- 30:44Could go wrong? What is the downside scenario?
- 30:48The current market narrative is universally positive.
- 30:50The professional second level investor must adopt A
- 30:54universally skeptical position. Now, as the Co chairman of
- 30:57Owetree, Marx is fundamentally a credit and distressed debt
- 31:00investor. He highlights a crucial
- 31:03structural advantage of debt in a lofty market like this.
- 31:06Yes. Why does credit become
- 31:07relatively more attractive when equity is priced for perfection?
- 31:11The difference lies in the payoff structure.
- 31:13In equity investing, you need the price to rise.
- 31:16You need the company to out form expectations.
- 31:19It's a highly variable and unpredictable outcome.
- 31:22But in credit investing in debt, your outcome is fixed.
- 31:25You just need the company not to go bust to receive your fixed
- 31:29contractual return principle plus interest.
- 31:32When equity is priced for absolute perfection, that
- 31:35variable outcome is highly dangerous.
- 31:38So instead of betting on which AI giant will win the long term
- 31:41competition, you might just lend money to the essential
- 31:43infrastructure that supports all of them.
- 31:45That's the strategic pivot in a winner take all environment
- 31:48where equity prices are disconnected from reality.
- 31:51Lending to AI, adjacent infrastructure, data center
- 31:54operators, power generation facilities, networking
- 31:57providers. It offers high yields with debt
- 31:59like protections. You get paid well as long as the
- 32:02underlying company survives. Regardless of whether Open AI,
- 32:05Google, or some startup eventually dominates the AI
- 32:08model space. This all ties back to the
- 32:09fundamental intellectual discipline that marks preaches,
- 32:12the discipline of second level thinking.
- 32:14This is the intellectual insurance policy against being
- 32:17swept up in the crowd. It's the difference between
- 32:19making comfortable decisions and making profitable ones.
- 32:22First level thinking is simplistic, consensus driven and
- 32:26comfortable. For example, the first level
- 32:28thinker observes AI is going to be huge, NVIDIA makes the best
- 32:32chips. I should buy NVIDIA.
- 32:34And everyone thinks that. And that consensus is already
- 32:37priced into the stock. So that comfort guarantees
- 32:39inferior risk adjusted returns going forward.
- 32:42Right? Second level thinking is
- 32:43complex, non consensus and it requires genuine discomfort.
- 32:48It necessitates questioning the obvious trade.
- 32:51So the second level thinker acknowledges the premise.
- 32:53Everyone knows AI is going to be huge and NVIDIA is priced to own
- 32:57the entire market forever with no execution risk.
- 33:00But then the second level thinker asks, what if
- 33:02competition from Amds new chips accelerates?
- 33:05What if Google, Microsoft and Amazon successfully build and
- 33:08deploy their own custom AI chips, eroding Nvidia's 90%
- 33:12market share and crushing their margins?
- 33:14The consensus is too optimistic. The price is too high.
- 33:18So I will wait for a deep pullback or I'll look for
- 33:21secondary beneficiaries like specific copper mining stocks or
- 33:24power grid providers that the consensus is completely
- 33:28ignoring. That counter argument is the
- 33:30true discipline. Marx warns that any investment
- 33:32that feels comfortable and is the consensus trade is highly
- 33:36unlikely to yield superior risk adjusted returns going forward.
- 33:39The best investments are inherently uncomfortable because
- 33:42they are non consensus. Hashtag tag outro.
- 33:46So to summarize our deep dive into Howard Marks's analysis,
- 33:49the signal regarding the current AI boom is a flashing yellow.
- 33:53The technology of AI is genuinely revolutionary, but the
- 33:56prices associated with the leaders are elevated, stretched
- 34:00and lofty. And the combination of extreme
- 34:02concentration, the risk of a circular economy and revenue,
- 34:05and the structural gravity of higher interest rates means the
- 34:07risk reward profile is now asymmetric to the downside.
- 34:10The easy money in this cycle has undoubtedly been made.
- 34:13It has the knowledge we've gained today understanding the
- 34:16psychological traps, analyzing that historical Cisco analog,
- 34:20and recognizing the structural collision between infinite
- 34:23expectations and hard economic reality.
- 34:26This allows you to construct a portfolio that is both defensive
- 34:30and ready for the inevitable correction.
- 34:33And Marx's ultimate philosophical warning, the one
- 34:36that should guide every decision, remains timeless.
- 34:38What's that? The most dangerous thing in the
- 34:40world is to watch your neighbor get rich.
- 34:43It always comes back to that. The investor who can resist that
- 34:46universal envy and stick firmly to their valuation discipline,
- 34:50who remembers that trees don't grow to the sky, will be the one
- 34:53who survives and thrives when the inevitable bust follows the
- 34:57boom. That is profound wisdom worth
- 34:59internalizing and as a final provocative thought for you, the
- 35:02learner. If the current market is lofty
- 35:05but not nutty, and the one thing keeping it from becoming nutty
- 35:09is the absence of the man in the street mania, what single
- 35:12concrete sign, a particular type of mass media article, a
- 35:16specific social media trend, or a new consumer product will you
- 35:20look for over the next six months that tells you the
- 35:23general public has finally capitulated to FOMO?
- 35:26That's a great question. That signal will tell you the
- 35:28psychological peak is truly here, and it will be the time to
- 35:30ensure your portfolio is battened down.
- 35:33An excellent point to ponder and if you want to access more
- 35:36advanced, in depth content about these topics, welcome to join
- 35:39the Deep Value Club membership or subscribe to us on the Apple
- 35:42and Spotify Podcasts channel. Deep value investing.
- 35:45And remember to track that essential total market
- 35:48temperature that objective measure of valuation Marks
- 35:51relies on. Welcome to search and download
- 35:54the iOS app A Buffett Indicator from the Apple App Store.
- 35:57Disclaimer All analysis is based on publicly available sources
- 36:00and may be subject to revisions or differing interpretations.
- 36:04The content is for educational purposes only and does not
- 36:06constitute investment advice. Conduct your own due diligence
- 36:09before making investment decisions.
- 36:11Thanks for joining the deep dive.
- 36:13We'll see you next time.