Howard Marks argues AI investing requires calibrated humility due to low predictability and high uncertainty.
Summary
Howard Marks discusses the challenges of investing in artificial intelligence, noting that the technology is 'grown' rather than designed and its internal workings remain largely uninterpretable. He highlights the difficulty in forecasting profitability, competitive durability, and intrinsic value due to rapid technological shifts and unknown usage patterns. Marks warns against the dangers of certainty, suggesting that investors face a low combined probability of being right across multiple variables. He distinguishes between equity investors assessing tech progress and lenders evaluating tenant creditworthiness. The episode concludes that accepting ignorance and maintaining calibrated humility is essential given the high stakes and unpredictability of the AI sector.
Key takeaways
- 1 Investing in AI requires being right about technology evolution, compute demand, competitive structure, beneficiaries, financing availability, and timing, making the combined probability of any thesis being correct unusually low.
Unidentified speaker (speaker_0) · 2:23
- 2 Nvidia and TSMC are currently racking up big profits, demonstrating the value of strong competitive positions in the AI infrastructure chain.
Unidentified speaker (speaker_0) · 9:56
- 3 until we know something about how AI will be used and who will profit, we can't assess the wisdom of the company's spending. And until knowledge of these things enables us to reasonably estimate the company's earnings and the profitability of the infrastructure. we can't judge the appropriateness of the prices being paid for stocks or the yields being offered on debt.
Unidentified speaker (speaker_0) · 35:46
Episode sections
- 0:05 – 3:42
Introduction and the Challenge of Forecasting AI
Howard Marks introduces his memo on AI, acknowledging his lack of technical expertise but emphasizing the technology's massive impact. He argues that investing in AI requires assessing the probability of being right across multiple complex variables, noting that the combined probability of success is low.
- 3:42 – 8:51
Understanding AI Mechanics and Evolution
Marks uses Claude to explain that AI is 'grown' rather than designed, with internal workings that are largely uninterpretable. He questions whether AI can reason, align with human values, and perform second-level thinking, noting that few people have significant insight into these matters.
- 8:51 – 12:04
Profitability and Competitive Landscape
The speaker discusses whether AI will become a commoditized service or create monopolies, highlighting Nvidia and TSMC as current beneficiaries. He questions the durability of today's leaders like OpenAI and Anthropic compared to hyperscalers, noting the rapid pace of change in the industry.
- 12:04 – 18:27
Profitability Projections and Data Center Infrastructure
Marks argues that projecting earnings for AI companies is exceptionally difficult due to uncertainties in product capability, pricing, and depreciation. He suggests investing in data centers as a 'picks and shovels' strategy, though he warns of risks like obsolescence, regulatory hurdles, and financing constraints.
- 18:27 – 27:43
Market Concentration and Credit Market Implications
The transcript highlights the disproportionate contribution of AI to S&P 500 earnings and stock gains, creating a concentrated market fault line. Marks notes that AI-related debt issuance is massive, potentially crowding out government borrowing and raising interest rates, while also raising concerns about leverage and circular transactions.
- 27:43 – 32:51
Predictability Scores and Geopolitical Risks
Using Claude, Marks assigns predictability scores to AI subsectors, with frontier labs being the least predictable (score 2) and regulated utilities the most (score 6-7). He also raises geopolitical risks, specifically regarding US-China chip tensions and the reliance on TSMC.
- 32:51 – 35:46
Is AI a Bubble?
Marks defines a bubble as irrational exuberance leading to excessive investment and high prices. He assesses whether AI fits this definition by evaluating the technology's hype, the sustainability of capital expenditures, and investor risk-taking, concluding that the fundamental view hasn't changed significantly since his previous memo.
- 35:46 – 36:11
Assessing AI Investment Wisdom and Value
The speaker argues that without knowing how AI will be used or who will profit, it is impossible to assess the wisdom of corporate spending or estimate earnings. Consequently, determining appropriate stock prices or debt yields is currently unfeasible due to this lack of clarity.
- 36:11 – 37:07
The Core Question of AI Profitability
The central investment question is whether AI will provide sufficient value to users to justify high compute prices that generate adequate returns on trillions in investment. The speaker concludes that answering this is currently a matter of faith rather than fact.
- 37:07 – 38:48
Investing in Uncertainty and Different Roles
AI is described as the most consequential yet least predictable development for investors. The appropriate response is not paralysis but calibrated humility, recognizing that different investors (equity vs. debt) face different critical questions and risks.
- 38:48 – 39:48
Price, Intrinsic Value, and General Investment Questions
While low prices can compensate for high uncertainty, calculating intrinsic value for AI is complicated. The speaker lists universal investment questions regarding required assumptions, downside risk, and portfolio impact, rather than AI-specific answers.
- 39:48 – 43:06
The Danger of Certainty and the Value of Uncertainty
Investors lose money whether they know nothing or claim to know everything. The speaker emphasizes accepting ignorance in the face of AI's surprises and notes that uncertainty benefits long-term value buyers, whereas paying for consensus is expensive.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| AmazonAMZN | equity | neutral | If it's possible, consider the vast resources and potential of the so-called hyperscalers. Amazon, Microsoft, Alphabet, Meta, and Oracle. |
| MicrosoftMSFT | equity | neutral | If it's possible, consider the vast resources and potential of the so-called hyperscalers. Amazon, Microsoft, Alphabet, Meta, and Oracle. |
| AlphabetGOOGL | equity | neutral | If it's possible, consider the vast resources and potential of the so-called hyperscalers. Amazon, Microsoft, Alphabet, Meta, and Oracle. |
| Meta PlatformsMETA | equity | neutral | If it's possible, consider the vast resources and potential of the so-called hyperscalers. Amazon, Microsoft, Alphabet, Meta, and Oracle. |
| AppleAAPL | equity | neutral | Microsoft, Amazon, Apple, and Google came through that era as winners, but Napster, Yahoo, and Netscape fell by the wayside. |
| S&P 500 | index | neutral | For fourteen consecutive quarters through June thirtieth 30th of this year, earnings grew faster for the Magnificent Seven. 7, mostly the Hyperscalers, than for the rest of the S P 500. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
The probability of a top-down investor being right about the economy, industry, company, and market, assuming a 60% chance for each, is 13% (0.6 to the fourth power).
Unidentified speaker (speaker_0) ·2:15
How to verify: Mathematical calculation: 0.6^4 = 0.1296. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
AI systems are 'grown rather than designed,' with rules emerging from nudging billions of internal settings.
Unidentified speaker (speaker_0) ·4:27
How to verify: Review AI interpretability research papers or Anthropic/OpenAI technical reports on model training and emergence. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Alphabet and Amazon reported $151 billion in gains on equity investments in Q2.
Unidentified speaker (speaker_0) ·19:28
How to verify: Check Alphabet (GOOGL) and Amazon (AMZN) Q2 2026 earnings reports on SEC EDGAR for 'gains on equity investments' or 'investment income'. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Primary sourceA group of 42 AI-related companies generated around two-thirds of the S&P 500's price gains since January 2024, according to J.P. Morgan.
Unidentified speaker (speaker_0) ·20:08
How to verify: Search J.P. Morgan research reports or press releases from mid-2026 referencing '42 AI-related companies' and S&P 500 price gains. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
AI-related bonds represent around 6% of the US high yield bond market.
Unidentified speaker (speaker_0) ·22:34
How to verify: Check Bloomberg or ICE BofA data for high yield bond sector allocation. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Frontier Labs (OpenAI, Anthropic, XAI) have a predictability score of 2 (practically unknowable) according to Claude.
Unidentified speaker (UNKNOWN) ·30:13
How to verify: Review the transcript or Claude's output as cited by Marks; verify via public statements from these companies regarding their business models and moats. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Regulated power utilities serving AI load have a predictability score of 6, comparable to regulated utilities (6-7) and large industrials (5-6).
Unidentified speaker (speaker_0) ·29:27
How to verify: Compare Claude's scoring with standard industry volatility metrics (e.g., beta, earnings stability) for utilities vs. industrials. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Nearly all leading-edge chips come from Taiwan's TSMC.
Unidentified speaker (speaker_0) ·31:39
How to verify: Check semiconductor industry market share reports (e.g., from Gartner, IDC, or TSMC earnings) for advanced node manufacturing share. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Assessing the wisdom of company spending on AI is impossible until usage and profit distribution are known.
Unidentified speaker (UNKNOWN) ·35:46
How to verify: Review Oaktree Capital Management LP memos or Howard Marks' published writings on AI investment criteria for stated prerequisites for capital allocation. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Answering whether AI will provide adequate returns on trillions in investment is currently a matter of faith.
Unidentified speaker (UNKNOWN) ·36:36
How to verify: Check Howard Marks' published memo 'A Layman Thinks About AI' for the specific phrasing regarding faith versus proof in AI valuation. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
AI is the most consequential and least predictable development for investors in at least a generation.
Unidentified speaker (UNKNOWN) ·37:12
How to verify: Verify against the text of Howard Marks' memo released on 2026-11-11 for this specific characterization of AI. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Lenders financing data centers focus on tenant creditworthiness, contract terms, power availability, and residual asset value.
Unidentified speaker (speaker_0) ·38:03
How to verify: Review Howard Marks' memo for the list of factors relevant to debt investors in AI infrastructure. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
High uncertainty is acceptable if the price is low enough relative to intrinsic value.
Unidentified speaker (speaker_0) ·38:23
How to verify: Check standard value investing principles and Howard Marks' writings on risk/return trade-offs for confirmation of this general principle. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Two kinds of investors lose a lot of money: those who know nothing and those who know everything.
Unidentified speaker (speaker_0) ·39:48
How to verify: Search for this specific quote in Howard Marks' published works or speeches regarding investor psychology and overconfidence. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Investors pay a very high price in the stock market for a cheery consensus.
Unidentified speaker (speaker_0) ·40:35
How to verify: Verify against the Warren Buffett quote from Forbes magazine, August 6, 2026, cited by Howard Marks. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Uncertainty is the friend of the buyer of long-term values.
Unidentified speaker (speaker_0) ·40:55
How to verify: Check the Warren Buffett quote from Forbes, August 6, 2026, for this specific assertion regarding uncertainty and value investing. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
The memo concludes with a quote from Warren Buffett appearing in Forbes magazine on August 6, 1979.
Unidentified speaker (speaker_0) ·41:03
How to verify: Locate the Forbes article from August 6, 2026, featuring Warren Buffett to verify the existence and content of the cited quote. For recent events, verify 2026-08-11 through 2026-11-11. Use another period only when the speaker explicitly names it.
Do your own research
- Verify the current market share and revenue growth rates of Nvidia and TSMC in the AI infrastructure supply chain.
- Review the capital expenditure plans and debt issuance volumes of major hyperscalers and AI startups to assess leverage risks.
- Analyze S&P 500 earnings concentration metrics to quantify the disproportionate contribution of AI-related stocks to index performance.
- Check regulatory filings for potential government interventions or taxes on AI companies or displaced workers.
- Assess the geopolitical risk exposure of US tech firms relying on TSMC for advanced chip manufacturing.
- Compare current valuations of AI leaders against historical precedents of disruptive technology adoption cycles.
- Evaluate the credit quality and tenant concentration of data center REITs and infrastructure funds.
- Research the technical capabilities and reasoning limits of current frontier AI models as reported by independent researchers.
Full transcriptexpand
- 0:05
- Unidentified speaker (UNKNOWN): This
- 0:06
- Unidentified speaker (speaker_0): is the memo by Howard Marks.
- 0:10
- Unidentified speaker (UNKNOWN): A
- 0:10
- Unidentified speaker (speaker_0): layman thinks about AI.
- 0:17
- Unidentified speaker (speaker_0): Should we invest in artificial intelligence and can we do so intelligently? That's what I'm going to address in this memo.
- 0:26
- Unidentified speaker (UNKNOWN): On one
- 0:26
- Unidentified speaker (speaker_0): hand, I'm not a tech person, and I don't claim to have above average knowledge of artificial intelligence, or technology investing for that matter.
- 0:35
- Unidentified speaker (UNKNOWN): On
- 0:36
- Unidentified speaker (speaker_0): the other, AI is the biggest development of our time, and could be one of the most impactful things the world has ever seen. So investors don't have the option to abstain from thinking about
- 0:47
- Unidentified speaker (UNKNOWN): it. With
- 0:48
- Unidentified speaker (speaker_0): that disclaimer, I'm going to share my deliberation deliberations, with put forth the questions I'm wrestling with, not all of which are directly relevant to investing in AI, and provide my thinking about
- 0:59
- Unidentified speaker (UNKNOWN): them.
- 0:59
- Unidentified speaker (speaker_0): While my limited knowledge prevents me from definitively answering these questions, I hope to provide a service by assembling them in one place.
- 1:08
- Unidentified speaker (UNKNOWN): And
- 1:08
- Unidentified speaker (speaker_0): while news on AI comes hot and heavy these days, with security prices for AI and AI adjacent. companies gyrating up and down, I'm not going to try to keep up, since what I have to say relates to the long-term outlook for AI, not what security prices will be tomorrow.
- 1:26
- Unidentified speaker (speaker_0): Investing consists of just one thing: positioning capital to benefit from future events, and most investors cope by relying on forecasts. But not everything is equally predictable. and not all forecasters are equally skilled. Thus, to invest in AI based on expectations about the future, and I
- 1:46
- Unidentified speaker (speaker_0): think that's all we have, we must consider both what we think the future may hold and the chance we can get it right.
- 1:55
- Unidentified speaker (UNKNOWN): I've
- 1:55
- Unidentified speaker (speaker_0): written in the past about the fact that The people who engage in top-down top down investing have to be right about the outlook for the economy, the industry, the company, and the market. If they have a sixty percent probability of getting each one right, which is probably generous, what's the probability of getting all four right? The answer
- 2:15
- Unidentified speaker (speaker_0): is thirteen percent. Zero point six to the fourth power, assuming the outcomes are independent of each
- 2:22
- Unidentified speaker (UNKNOWN): other. In the
- 2:23
- Unidentified speaker (speaker_0): case of AI, there are an unusually large number of important variables, each of which complicates the decision of how to invest. An investor in AI has to be at least somewhat right about the evolution of the technology, the demand for compute, the competitive structure, the beneficiaries of the
- 2:43
- Unidentified speaker (speaker_0): technology, the availability of financing, and the timing of all the above. Assessing these key elements. seems to me to be unusually challenging. In that case, what's the combined probability that any thesis is correct? That's just something to think about as you listen.
- 3:02
- Unidentified speaker (speaker_0): Lastly, bear in mind that A, we can never know everything about an investment. B, the biggest mistakes are made when people think they do, and C the incompleteness of our knowledge. regarding the future can't be permitted to completely prevent us from taking action. It's the investor's job to
- 3:22
- Unidentified speaker (speaker_0): bear risk in the pursuit of profit. How much and in what form are decisions we each must make? But given its vast potential, we can't decline to participate in AI just because the future is unclear. Question one. How does AI do what it does? Before getting more
- 3:42
- Unidentified speaker (speaker_0): granular, I'm going to start with a big picture, non-financial question that I find particularly interesting. Since I have absolutely no knowledge on this subject,
- 3:52
- Unidentified speaker (UNKNOWN): I
- 3:52
- Unidentified speaker (speaker_0): went to Anthropics Claude for some answers. Here's some of what it said.
- 3:58
- Unidentified speaker (speaker_0): We know what AI is made of and how it was built almost perfectly.
- 4:03
- Unidentified speaker (speaker_0): but we know surprisingly little about how it actually does what it does. We also know the training recipe. Show it enormous amounts of text, have it guess the next word, and nudge billions of internal settings slightly each time it guesses wrong. Repeat that trillions of times.
- 4:23
- Unidentified speaker (speaker_0): Nobody wrote the rules for how it answers the question.
- 4:26
- Unidentified speaker (UNKNOWN): Those
- 4:27
- Unidentified speaker (speaker_0): rules emerged from the nudging, spread across billions of numbers that mean nothing individually.
- 4:33
- Unidentified speaker (UNKNOWN): The
- 4:33
- Unidentified speaker (speaker_0): best phrase I know for this is that these systems are grown rather than designed. Researchers in interpretability are making real progress. They have found, for example, that a model writing a rhyming poem picks the rhyme word before it writes the line, which means it plans ahead.
- 4:53
- Unidentified speaker (UNKNOWN): But
- 4:53
- Unidentified speaker (speaker_0): they can explain only small slices of what goes on. It's like having a complete record of every trade in the market and still not knowing why prices moved.
- 5:04
- Unidentified speaker (UNKNOWN): When
- 5:04
- Unidentified speaker (speaker_0): I explain how I reached an answer, research suggests my explanation doesn't always match what actually happened inside.
- 5:13
- Unidentified speaker (UNKNOWN): A
- 5:13
- Unidentified speaker (speaker_0): model asked how it needed. a sum may describe the school book method while internally doing something very different. So the honest summary is that we've built a mind like thing we can inspect completely and still can't fully explain. And when it explains itself, it may be no more reliable than we are when we
- 5:33
- Unidentified speaker (speaker_0): explain ourselves. Interestingly, here Claude talks like it's a person.
- 5:40
- Unidentified speaker (UNKNOWN): As
- 5:40
- Unidentified speaker (speaker_0): you think about answering the questions that follow, consider these. Does anyone truly understand how AI works? How will AI evolve, and what will it be able to do and when? How will users put it to work? What impact will it have on users' profitability?
- 5:59
- Unidentified speaker (speaker_0): Investors may not have to know all these things, but it's my understanding that few people have significant insight into these matters.
- 6:07
- Unidentified speaker (UNKNOWN): If
- 6:07
- Unidentified speaker (speaker_0): that's the case, can anyone reliably answer the questions that follow? Question two. What will AI be able to
- 6:16
- Unidentified speaker (UNKNOWN): do?
- 6:17
- Unidentified speaker (speaker_0): What AI is great at is evolving. Just a few years ago, what AI could do was generally summed up as knowing everything that has occurred before.
- 6:27
- Unidentified speaker (speaker_0): identifying historical patterns, extrapolating them, and using this ability to perform tasks. This has already made huge changes in some areas, like coding, and promises to alter many aspects of life as we presently know
- 6:42
- Unidentified speaker (UNKNOWN): it.
- 6:43
- Unidentified speaker (speaker_0): But now AI has gone well beyond that, to something that probably can fairly be described as reasoning. Just this one area poses several questions that are both stimulating and daunting. AI's full potential will be subject to the answers to some of these.
- 7:02
- Unidentified speaker (speaker_0): Can AI's process be made to reflect our preferences, values, and goals? Or will it apply its own? Can AI be depended on to be honest, humble, ethical, and loyal?
- 7:15
- Unidentified speaker (speaker_0): Will it be accurate and objective, or subject to influence and manipulation?
- 7:21
- Unidentified speaker (UNKNOWN): Will
- 7:21
- Unidentified speaker (speaker_0): it admit when it's wrong or doesn't know an answer? Can it give us skepticism rather than sycophancy?
- 7:29
- Unidentified speaker (UNKNOWN): Will
- 7:29
- Unidentified speaker (speaker_0): it become a threat to humans and seek to advance its own interests instead of ours? How will it do in areas such as creativity? originality, aesthetics, subjectivity, and intuition? Will it be able to perform what I call second level thinking? Thinking in a way that's
- 7:49
- Unidentified speaker (speaker_0): different from that of most people and at a higher level, leading to superior solutions.
- 7:55
- Unidentified speaker (UNKNOWN): Will
- 7:56
- Unidentified speaker (speaker_0): it be able to think about the future and blue sky to the same degree as humans?
- 8:01
- Unidentified speaker (UNKNOWN): Can it
- 8:02
- Unidentified speaker (speaker_0): be instructed? to furnish primary sources, as opposed to published citations and provide counter arguments? Will one model be able to do it all? Or will different models be best at different things? If so, how will we know which one to use? Will artificial general
- 8:22
- Unidentified speaker (speaker_0): intelligence, the ability to do everything a human can do,
- 8:26
- Unidentified speaker (UNKNOWN): be
- 8:26
- Unidentified speaker (speaker_0): the standard?
- 8:27
- Unidentified speaker (speaker_0): Will AGI mean the ability to do the same things as A the average person, B, a person in the ninetieth percentile, or C, the smartest person in the world? In other words, will there be varying levels of knowledge and things AI doesn't know?
- 8:46
- Unidentified speaker (speaker_0): All these things bear on what work will be entrusted to AI.
- 8:51
- Unidentified speaker (speaker_0): Question three. Who will profit?
- 8:56
- Unidentified speaker (UNKNOWN): As
- 8:56
- Unidentified speaker (speaker_0): Warren Buffett pointed out about the internet during the tech bubble, transformative technologies don't necessarily reward investors. We know AI will increase efficiency, but will it increase profitability? And whose? If several AI models can become equally smart, on a given subject,
- 9:16
- Unidentified speaker (speaker_0): will their service become an undifferentiated commodity that sells on the basis of price? If it does, will profitability and ROI be ground down? In which case the economic benefits may go more to the users of AI than to the vendors? What does all this imply for the return
- 9:36
- Unidentified speaker (speaker_0): on the trillions of dollars being invested in the AI infrastructure?
- 9:41
- Unidentified speaker (UNKNOWN): Or
- 9:41
- Unidentified speaker (speaker_0): on the other hand, will some AI companies acquire capabilities that make them monopolies whose products are essential for company success and even survival? If so, think about the prices they'll be able to command.
- 9:55
- Unidentified speaker (UNKNOWN): The
- 9:56
- Unidentified speaker (speaker_0): big chip producers, Nvidia and TSMC, are racking up the big profits these days, demonstrating the value of the a strong competitive position. Lastly, on this subject, I haven't heard much about whether and how AI will be taxed. Given that one of AI's most likely effects will be replacing
- 10:15
- Unidentified speaker (speaker_0): workers, and given governments need to replace the tax revenues that previously came from taxing the people AI puts out of work, what's to say AI won't be taxed like people? Many of today's top earners are taxed at combined rates of 50% or more. What tax bracket will AI be in?
- 10:36
- Unidentified speaker (speaker_0): Question four. What will be the competitive landscape? Can we know which companies will succeed? With what degree of certainty? How strong are the motes surrounding today's leaders?
- 10:50
- Unidentified speaker (UNKNOWN): Perhaps
- 10:50
- Unidentified speaker (speaker_0): the most startling characteristic of AI is the speed at which it makes progress.
- 10:56
- Unidentified speaker (UNKNOWN): In
- 10:56
- Unidentified speaker (speaker_0): that connection, I'm struck by the fact that today's leading model labs, OpenAI and Anthropic, were unknown to most of us just a few years ago. If companies can forge ahead that fast in the world of AI, can they likewise be supplanted by others, or at least challenged? If it's possible,
- 11:16
- Unidentified speaker (speaker_0): consider the vast resources and potential of the so-called hyperscalers. Amazon, Microsoft, Alphabet, Meta, and Oracle. Who will win among these incumbents? Or will it be companies we've never heard of? How about companies that don't exist yet? Think about the probability twenty 20 to
- 11:36
- Unidentified speaker (speaker_0): twenty-five 25 years ago that the leaders this then would still be the leaders now. Microsoft, Amazon, Apple, and Google came through that era as winners, but Napster, Yahoo, and Netscape fell by the wayside. It takes a lot to maintain a lead in a rapidly changing technology. Add in AI's novelty and
- 11:56
- Unidentified speaker (speaker_0): speed of To progress and ask, what's the probability today's AI leaders will still be the leaders in five? or ten years?
- 12:04
- Unidentified speaker (UNKNOWN): If
- 12:05
- Unidentified speaker (speaker_0): the competitive picture can shift, as I suggest, how much is a stake in today's leaders worth? Question five. How profitable will the companies be?
- 12:17
- Unidentified speaker (speaker_0): The value investor estimates a company's earnings or cash flows in the future, discounts them back to the present, and compares the result with today's stock price. To predict earnings in the future, we need answers to questions like these.
- 12:32
- Unidentified speaker (UNKNOWN): How
- 12:33
- Unidentified speaker (speaker_0): many units of its product will it sell? What will be the selling price per unit? What will each unit cost to produce? How much of the profit will need to be reinvested in developing the next product? What will be the company's resulting return on investment?
- 12:51
- Unidentified speaker (speaker_0): While answering these questions for any company requires dealing with uncertainty, doing so in connection with AI introduces particularly novel considerations, like these.
- 13:03
- Unidentified speaker (speaker_0): What will its product be capable of doing and how will customers put it to work? Will the product be priced as a commodity or a unique essential good?
- 13:14
- Unidentified speaker (UNKNOWN): How will
- 13:15
- Unidentified speaker (speaker_0): long will chips last? And how fast will data center depreciation have to be charged off? For how long will today's breakneck spending race continue? Will enough money be made to produce a decent return on today's capital expenditures? I could certainly I would be wrong, but I don't
- 13:34
- Unidentified speaker (speaker_0): remember puzzling over uncertainties of this magnitude in my day's analytical. analyzing the conglomerates in the nineteen seventies, or in connection with the Internet and e-commerce in nineteen ninety nine to two thousand. Those businesses, as novel and complex as they were, seem simple and straightforward compared to AI.
- 13:54
- Unidentified speaker (UNKNOWN): So
- 13:55
- Unidentified speaker (speaker_0): I wonder, can we project twenty 2030 thirty earnings for today's hyperscalers and leading frontier labs? Let alone customer. companies that mostly consist of conjecture? Do we have to be able to predict earnings reasonably accurately to invest?
- 14:11
- Unidentified speaker (UNKNOWN): If
- 14:11
- Unidentified speaker (speaker_0): we come up with estimates, how much confidence should we have in them?
- 14:15
- Unidentified speaker (UNKNOWN): And
- 14:15
- Unidentified speaker (speaker_0): if we can't, how can we think about investing in AI companies? Here's the test I propose nowadays.
- 14:23
- Unidentified speaker (UNKNOWN): Tell
- 14:23
- Unidentified speaker (speaker_0): me what you think the EPS for a given AI company would be. will in five years. Will you bet that the eventual answer will be within plus or minus 50% of your figure? For example, if you guess $10 per share, how confident are you that the actual figure will be between $5 and $15?
- 14:45
- Unidentified speaker (speaker_0): Question six. What's the outlook for data centers? Historically, data centers were digital warehouses that stored data and hosted computer applications and websites. The term data center or AI factory has evolved to refer to the workplaces of AI. Analogous to
- 15:04
- Unidentified speaker (speaker_0): auto plants, there where AI is produced, ranging from storing information and training models to providing time on the system, called compute. Data centers are a critical element in the development of the AI industry, a construction of data centers, and thus the performing of compute is subject
- 15:24
- Unidentified speaker (speaker_0): to constraints, physical and otherwise. While the future of AI is exciting, some people might prefer to take a cautious approach to participating in
- 15:36
- Unidentified speaker (UNKNOWN): it.
- 15:36
- Unidentified speaker (speaker_0): This suggests investment via tangible infrastructure rather than developers and vendors of the emerging technology.
- 15:44
- Unidentified speaker (UNKNOWN): A
- 15:44
- Unidentified speaker (speaker_0): strategy akin to selling picks and shovels to gold miners rather than trying to figure out which miners will find gold. In theory, this makes determining the ultimate winners less critical since whoever ends up winning would still need to use the infrastructure. Compute is in short supply to. today, adding to the
- 16:04
- Unidentified speaker (speaker_0): attractiveness of data centers for investment. However, a number of factors must be considered.
- 16:11
- Unidentified speaker (UNKNOWN): When
- 16:11
- Unidentified speaker (speaker_0): we think about infrastructure, the phrase bricks and mortar often comes to mind,
- 16:16
- Unidentified speaker (UNKNOWN): and
- 16:16
- Unidentified speaker (speaker_0): it's that solidity that can make data centers appear the safer AI option. But a relatively small part of the cost of a data center goes for the building itself. the shell. Chips or GPUs make up most of the cost, and that, and the risk of obsolescence, is generally borne by the
- 16:36
- Unidentified speaker (speaker_0): center user. Large building sites will have to be located and purchased, and data center developers will have to overcome regulatory hurdles and resistance from citizens opposed to data centers in their backyards.
- 16:51
- Unidentified speaker (speaker_0): Connectivity to users of inference via available fiber will be a necessity. Data center builders will have to deal with potential bottlenecks in terms of the needed electric power, cooling equipment, and construction services. Limitations on the availability of financing for data centers could
- 17:11
- Unidentified speaker (speaker_0): constrain the pace at which they're constructed,
- 17:13
- Unidentified speaker (UNKNOWN): and the
- 17:13
- Unidentified speaker (speaker_0): cost of financing for future projects could surprise.
- 17:18
- Unidentified speaker (speaker_0): Interestingly, the potential bottlenecks just mentioned could complicate life for would-be builders of data centers and limit their growth.
- 17:28
- Unidentified speaker (UNKNOWN): On
- 17:28
- Unidentified speaker (speaker_0): the flip side, any constriction of the growth in the supply of compute, as implied by the success last week, of a number of moratoriums on data center construction would should make the ones that presently exist, or that can be built in the future, more profitable and thus more valuable.
- 17:47
- Unidentified speaker (UNKNOWN): The
- 17:47
- Unidentified speaker (speaker_0): key here is that data centers aren't a layup. The intelligent investor has to find opportunities that are insulated from the uncertainties just listed, and won't be subject to overbuilding and thus oversupply of compute. As in all forms of investing, there will be successful data center investors and
- 18:07
- Unidentified speaker (speaker_0): less successful ones. Given the dynamics of AI, however, I think the successes of the winners and failures of the losers may be even more pronounced than is usual in infrastructure. Question seven. Should we be worried about our unusual reliance on this?
- 18:27
- Unidentified speaker (speaker_0): sector?
- 18:29
- Unidentified speaker (speaker_0): Today, AI is responsible for a disproportionate share of what's going on in the investment environment, introducing considerable uncertainty for all investors and all participants in the economy, not just for investors in AI. AI represents a small part of the U.S. economy, but is
- 18:49
- Unidentified speaker (speaker_0): responsible for an outsized part of economic growth. AI's contribution to GDP growth currently comes overwhelmingly through capital expenditures, primarily for construction of data centers. These are rising rapidly, while the rest of construction is in decline. For fourteen
- 19:08
- Unidentified speaker (speaker_0): consecutive quarters through June thirtieth 30th of this year, earnings grew faster for the Magnificent Seven. 7, mostly the Hyperscalers, than for the rest of the S P 500. Some AI companies have been generating large gains from equity investments in fellow AI-related companies, making them top contributors
- 19:28
- Unidentified speaker (speaker_0): to SP 500 earnings. Fact Set reported that in the second quarter, Alphabet and Amazon notched a total of one $151 hundred fifty one billion of dollars gains on equity investments, mostly on Anthropic and SpaceX, which now includes XAI. AI's impact on aggregate
- 19:48
- Unidentified speaker (speaker_0): earnings doesn't stop with the sector. Sellers of associated components are reporting strong earnings as well, and there are knock on effects as AI workers and AI adjacent workers spend their pay on rent, food, cars, and other things far removed from AI itself. AI has
- 20:08
- Unidentified speaker (speaker_0): disproportionately contributed to stock market gains, with a group of 42 AI-related companies generating around two-thirds of the S P 500's price gains since January 2024, according to J.P. Morgan. The sum of these trends has left the SP 500 five hundred highly concentrated. in a
- 20:27
- Unidentified speaker (speaker_0): few AI and tech companies.
- 20:31
- Unidentified speaker (UNKNOWN): The
- 20:31
- Unidentified speaker (speaker_0): key observation is that the factors our economy and markets depend on today are relatively undiversified. Exposure to AI constitutes an unusually powerful fault line that runs through portfolios and has the potential to dominate all aspects of the investment environment.
- 20:50
- Unidentified speaker (UNKNOWN): Now
- 20:50
- Unidentified speaker (speaker_0): Negative developments with regard to chip life, corporate strategy, or demand for compute could have far reaching consequences. It's incontrovertible that disproportionate reliance on a single sector makes the economy and stock market more vulnerable than they would be if the factors underlying them were broad based.
- 21:10
- Unidentified speaker (UNKNOWN): When
- 21:11
- Unidentified speaker (speaker_0): Michael Milken pioneered the issuance of sub investment grade debt, in nineteen seventy seven to seventy eight, he ushered in an investment industry wide willingness to accept risk as long as it's viewed as being accompanied by enough potential return to compensate for the risk.
- 21:31
- Unidentified speaker (speaker_0): A. Incremental yield that potentially compensates for the losses on the bonds that default, and B.
- 21:39
- Unidentified speaker (speaker_0): enough diversification so that a few defaults won't spoil the overall portfolio outcome. But today, our stock and debt markets are heavily reliant on a single factor. The success of AI.
- 21:54
- Unidentified speaker (speaker_0): Question eight. What are the implications Applications for the credit market? AI-related AI related debt issuance has totaled nearly $600 six billion. hundred dollars globally so far this year, according to Goldman Sachs, comprising debt issued by the hyperscalers in data centerslash project financings and by other AI linked
- 22:14
- Unidentified speaker (speaker_0): borrowers. Around four hundred billion dollars of this went to the U.S. investment grade debt market, representing over one one-fifth fifth of total IG supply and taking the share of AI AI-related related bonds in the Bloomberg API. USD IG corporate index to around 12%. For the U.S. high yield bond market, into
- 22:34
- Unidentified speaker (speaker_0): which the hyperscalers don't directly issue, the weighting is more like 6%.
- 22:39
- Unidentified speaker (speaker_0): Clearly, AI-related debt issuance is one of this year's major capital market developments. Elements. Going forward, AI seems likely to require an average of $1 one trillion of dollars new capital per year in the period ahead, and hundreds of billions of that are likely to continue to be raised each year through the issuance
- 22:59
- Unidentified speaker (speaker_0): of debt. What will be the impact on interest rates of AI's demand for debt capital?
- 23:06
- Unidentified speaker (speaker_0): Since expanding capacity and developing new products are imperatives for the hyperscalers and frontier labs, and since they don't want to see competition crop up, these companies likely can't afford to slow their investment in AI just because interest rates may be higher than they expected. That means they have to be
- 23:26
- Unidentified speaker (speaker_0): takers of financing, even if it is in short supply and increases in cost.
- 23:32
- Unidentified speaker (UNKNOWN): In
- 23:33
- Unidentified speaker (speaker_0): the past, economists used to talk about crowding out, in which aggressive government borrowing makes it harder. hard for the private sector to find the debt capital it needs. At present, it feels to me that the demand for debt from rate insensitive hyperscalers willing to pay whatever it takes has the potential to elevate the interest rates the
- 23:52
- Unidentified speaker (speaker_0): government must pay. Hasn't that contributed to the rate rise we've seen to date?
- 23:58
- Unidentified speaker (UNKNOWN): In
- 23:58
- Unidentified speaker (speaker_0): the same vein, will the equity markets be capable of providing the hundred? hundreds of billions of new equity investment required each year, supplementing the company's cash flow?
- 24:09
- Unidentified speaker (UNKNOWN): The
- 24:09
- Unidentified speaker (speaker_0): amount raised in traditional US IPOs was below $50 billion in eight of the last ten years, the exceptions being almost $80 billion in 2020 and $140 billion in the boom year of 2021. twenty twenty one. I suspect there's just so much new capital available. in a given year. Won't
- 24:29
- Unidentified speaker (speaker_0): AI's appetites for both debt and equity ultimately compete against each other for it?
- 24:35
- Unidentified speaker (UNKNOWN): In
- 24:36
- Unidentified speaker (speaker_0): the past, the highly profitable non-AI businesses of the hyperscalers produced massive free cash flow that rendered them debt-free, cash rich, and highly rated. Things are different now, and these companies are spending their cash. hoards and current cash flow, and they're borrowing more. Their strong
- 24:56
- Unidentified speaker (speaker_0): investment grade ratings enable them to easily finance their expansion in AI. But as this process wears on, might their creditworthiness and or ratings change? What are the possible implications for those who lend against the high ratings today and watch as the companies increase their leverage?
- 25:15
- Unidentified speaker (speaker_0): Some of the data center financing taking place is facilitated by the investment grade ratings of the hyperscaler tenants. Other debt is collateralized by chips. But investors have to be very conscious of out clauses in some data center leases that can enable tenants or guarantors to
- 25:35
- Unidentified speaker (speaker_0): terminate their obligation under certain conditions, such as the failure of the center to operate and ninety nine point nine percent reliability, and of the potential risk that collateral in the form of chips won't last as long as the debt. What are the implications of the industry's so called circular
- 25:54
- Unidentified speaker (speaker_0): transactions, in which companies finance purchases of their products?
- 25:59
- Unidentified speaker (UNKNOWN): We
- 25:59
- Unidentified speaker (speaker_0): see frequent instances of lending and related buying. between chipmakers, hyperscalers, frontier labs, and data center builders. Sometimes money goes from A to B and back to A. Sometimes it goes from A to B and then to financing for C, which spends it with A. Are these
- 26:19
- Unidentified speaker (speaker_0): transactions the parties would enter into on an arm's length basis? Or are vendors subsidizing buyers to support
- 26:26
- Unidentified speaker (UNKNOWN): the process?
- 26:27
- Unidentified speaker (speaker_0): demand and protect market share for their products? What are the risks associated with this interdependence?
- 26:34
- Unidentified speaker (UNKNOWN): And
- 26:34
- Unidentified speaker (speaker_0): what is the impact on the hyperscaler's creditworthiness of borrowing on the part of special purpose vehicles that keep debt off balance sheets? Is the parent behind the debt or isn't it? Finally, what are the effects of borrowing on the
- 26:50
- Unidentified speaker (UNKNOWN): debt?
- 26:50
- Unidentified speaker (speaker_0): in a time of exuberance surrounding an exciting technological advance?
- 26:55
- Unidentified speaker (UNKNOWN): In
- 26:56
- Unidentified speaker (speaker_0): the end we have the confluence of several factors a massive desire to borrow on the part of AI companies, investor excitement surrounding AI and thus a willingness to lend, heavy concentration on AI in the capital markets, as just described, and so so-called called circular target. transactions
- 27:15
- Unidentified speaker (speaker_0): between participants in the AI industry.
- 27:19
- Unidentified speaker (speaker_0): Lenders to AI face a new era, marked by debt concentrated in a single industry, rapid change in industry fundamentals, and a high degree of reliance on the level of future revenues from sales of compute. Lenders will have to be exceptionally vigilant regarding the creditworthiness of the borrowers they back and the
- 27:39
- Unidentified speaker (speaker_0): structureslash terms through which they do
- 27:42
- Unidentified speaker (UNKNOWN): it.
- 27:43
- Unidentified speaker (speaker_0): Question nine. How predictable is AI? How answerable are these questions?
- 27:51
- Unidentified speaker (UNKNOWN): In
- 27:51
- Unidentified speaker (speaker_0): memos and in my book, Mastering the Market Cycle, Getting the Odds on Your Side, I've written that for someone attempting to deal with the future, a prediction isn't enough. In addition to your prediction, you need a view on the probability your prediction is correct.
- 28:09
- Unidentified speaker (UNKNOWN): It's
- 28:09
- Unidentified speaker (speaker_0): my view that with regard to AI, given its technical complexity, unprecedented nature, and high rate of change, that probability is unusually low.
- 28:21
- Unidentified speaker (UNKNOWN): I
- 28:21
- Unidentified speaker (speaker_0): wanted to quantify AI's predictability, but I worried that doing so was well beyond my competence.
- 28:27
- Unidentified speaker (speaker_0): Thus, I asked Claude to help by ranking AI on a scale from one, unknowable, to ten, fully predictable. Its answers were highly informative. Claude summed up on the subject, reflecting some of the uncertainties just discussed.
- 28:45
- Unidentified speaker (speaker_0): My overall judgment? AI as a whole scores about a three.
- 28:49
- Unidentified speaker (UNKNOWN): That
- 28:50
- Unidentified speaker (speaker_0): number hides. a split. The direction is fairly predictable, maybe a six or seven. The technology will keep improving and spread through the economy. The things investors get paid for are much less predictable around a two, how fast it happens, how big the profits are, and who
- 29:09
- Unidentified speaker (speaker_0): keeps
- 29:10
- Unidentified speaker (UNKNOWN): them.
- 29:11
- Unidentified speaker (speaker_0): I also ask asked Claude to rate subsectors of the AI industry. on predictability and identify parts of the non AI economy it considers comparable in that regard. Here are what it describes as its judgment calls.
- 29:27
- Unidentified speaker (speaker_0): Regulated power utilities serving AI load, predictability score six. Regulated utilities six to seven. Data centers least leased long term. to big tech companies, predictability score of five to six, large industrials five to six.
- 29:47
- Unidentified speaker (speaker_0): Hyperscalers, predictability score five, comparable industries, not applicable.
- 29:55
- Unidentified speaker (speaker_0): Chips, predictability score four. Airlines shipping three to four. speculative or neo cloud data centers, predictability score of two to three, telecom slash fiber ninety-eight to zero one,
- 30:13
- Unidentified speaker (UNKNOWN): two
- 30:13
- Unidentified speaker (speaker_0): to three, Frontier Labs, OpenAI, Anthropic, XAI, etc. Predictability score of two, early stage VC,
- 30:24
- Unidentified speaker (UNKNOWN): one
- 30:24
- Unidentified speaker (speaker_0): to two.
- 30:26
- Unidentified speaker (speaker_0): For me, the surprise was the Frontier Labs score of two, which I take to mean practically unknowable. These seem to me to be impressive companies with big leads, but Claude notes as follows These businesses are close to binary. Capital needs are huge. Prices per unit
- 30:46
- Unidentified speaker (speaker_0): of capability keep falling fast, and it's unclear whether a lasting moat exists. Ironic, isn't it? Given that Claude belongs to Anthropic? Intriguingly, these most uncertain propositions, according to Claude, seem to be priced at a premium in the private market. It'll be interesting to see how the shares of
- 31:06
- Unidentified speaker (speaker_0): OpenAI and Anthropic, both of which have filed for IPOs, are priced publicly and how they trade afterwards.
- 31:14
- Unidentified speaker (UNKNOWN): And
- 31:14
- Unidentified speaker (speaker_0): note that much of the equity gains that goosed SP five hundred earnings in the second quarter, described earlier in the memo, were related to this sector, which is described as least predictable. How lasting will they prove to
- 31:28
- Unidentified speaker (UNKNOWN): be?
- 31:29
- Unidentified speaker (speaker_0): Finally, on the subject of predictability, I'll take this opportunity for a word about geopolitics, which with regard to AI generally means
- 31:39
- Unidentified speaker (speaker_0): China. Nearly all the leading edge chips come from Taiwan's TSMC.
- 31:45
- Unidentified speaker (UNKNOWN): The
- 31:46
- Unidentified speaker (speaker_0): US restricts sales to China of the most advanced chips. This encourages China to offer low-cost models that can undercut sales of more powerful US models. What happens to the US's competitive position if this changes?
- 32:01
- Unidentified speaker (UNKNOWN): Will
- 32:02
- Unidentified speaker (speaker_0): chips from China? catch up with or pass today's leaders? Will China ever be in position to discourage or even prevent sales of Taiwanese chips to the US? And will worries about China's intentions blunt efforts to impose safety guardrails on the AI industry?
- 32:21
- Unidentified speaker (speaker_0): Clearly, there's a lot investors don't know and can't predict about AI. The question is, how much risk will you take and in what form?
- 32:31
- Unidentified speaker (UNKNOWN): There
- 32:31
- Unidentified speaker (speaker_0): are many parts to the AI industry, and according to Claude, they range from the Frontier Labs, with regard to which we're largely in the dark, to securely leased data centers and their power providers, with predictability roughly comparable to that of large industrials and regulated utilities.
- 32:51
- Unidentified speaker (speaker_0): Question 10. Is AI a bubble? Worry over some of the questions just discussed has begun to surface in recent weeks.
- 33:01
- Unidentified speaker (UNKNOWN): A
- 33:02
- Unidentified speaker (speaker_0): confluence of negatives in the investment environment arose, centering around inflation, the federal deficits and debt, the resulting increase in interest rates, and the seeming intractability of the war in Iran. Iran. To these general issues were added some considerations specific to AI, such as
- 33:22
- Unidentified speaker (speaker_0): competition from new models, cheaper models with fewer bells and whistles, low priced Chinese models and open source models that are free to download. Declining prices per token, the unit in which model usage is priced. the possibility of slowing growth in the use of
- 33:41
- Unidentified speaker (speaker_0): compute, reported sales growth that falls short of expectations, and concern about some companies' high leverage levels and heavy dependence on future financing. Questions that are easily dismissed in good times can worry investors when their equanimity is shaken.
- 33:59
- Unidentified speaker (UNKNOWN): And
- 33:59
- Unidentified speaker (speaker_0): then we find, as I say,
- 34:01
- Unidentified speaker (UNKNOWN): that
- 34:01
- Unidentified speaker (speaker_0): whereas in real life, things fluctuate between pretty good and not so hot, in the minds of investors they go from flawless to hopeless.
- 34:10
- Unidentified speaker (UNKNOWN): The
- 34:11
- Unidentified speaker (speaker_0): things AI investors are worrying about these days are nothing new. But a changed investment climate has rendered them sufficient to darken the mood.
- 34:21
- Unidentified speaker (speaker_0): Even though the subject has been debated for well over a year, including by me, people
- 34:26
- Unidentified speaker (UNKNOWN): can
- 34:26
- Unidentified speaker (speaker_0): keep asking, is it a bubble? First, we need a definition of a bubble, and I want to supply mine. Irrational exuberance leading to imprudent behavior, excessive investment, and too high prices for participation.
- 34:43
- Unidentified speaker (UNKNOWN): I
- 34:44
- Unidentified speaker (speaker_0): think an AI bubble, like any possible bubble surrounding a technological advance, has to be considered in at least three.
- 34:51
- Unidentified speaker (speaker_0): ways. Is the technology itself overhyped and likely to disappoint? Are the amounts companies are spending excessive and fated to produce inadequate returns? And are investors taking too much risk and paying too much for stakes in this field? Reaching conclusions on these
- 35:11
- Unidentified speaker (speaker_0): subjects relies heavily on the answers to the questions just beginning. mentioned.
- 35:16
- Unidentified speaker (UNKNOWN): I
- 35:16
- Unidentified speaker (speaker_0): don't feel much has changed in these regards since I last wrote a memo on AI in February.
- 35:23
- Unidentified speaker (UNKNOWN): I
- 35:23
- Unidentified speaker (speaker_0): still think AI itself will prove powerful and consequential and change corporations, the economy, and life speaking broadly in ways we can't even imagine. So no, the technology isn't an illusion.
- 35:37
- Unidentified speaker (UNKNOWN): A
- 35:37
- Unidentified speaker (speaker_0): great deal of discussion is taking place regarding the revenues that will have to be generated in the future to justify the capital expenditures.
- 35:46
- Unidentified speaker (UNKNOWN): But
- 35:46
- Unidentified speaker (speaker_0): until we know something about how AI will be used and who will profit, we can't assess the wisdom of the company's spending. And until knowledge of these things enables us to reasonably estimate the company's earnings and the profitability of the infrastructure.
- 36:03
- Unidentified speaker (speaker_0): we can't judge the appropriateness of the prices being paid for stocks or the yields being offered on debt.
- 36:11
- Unidentified speaker (speaker_0): While there are a lot of questions swirling around AI, some of which I've just described, the key boils down to this rather complex one. Will AI provide sufficient value to users to warrant sales of compute and
- 36:26
- Unidentified speaker (UNKNOWN): the process?
- 36:27
- Unidentified speaker (speaker_0): prices for compute high enough to result in adequate returns on the trillions of dollars being invested in it? And for which companies?
- 36:36
- Unidentified speaker (UNKNOWN): The
- 36:36
- Unidentified speaker (speaker_0): conclusion I draw from my conversations with people I consider knowledgeable is that at this point, answering this question is largely a matter of faith.
- 36:46
- Unidentified speaker (UNKNOWN): That
- 36:46
- Unidentified speaker (speaker_0): reminds me of a quote attributed to American economist and writer Stuart Chase.
- 36:51
- Unidentified speaker (speaker_0): For those who believe, no proof is necessary. For those who don't believe, no proof is possible. Try as I may, I can't come up with a more appropriate resolution to the questions surrounding AI than this.
- 37:07
- Unidentified speaker (speaker_0): Question eleven. How should one think about investing in AI?
- 37:12
- Unidentified speaker (UNKNOWN): I
- 37:12
- Unidentified speaker (speaker_0): consider artificial intelligence the most consequential and least predictable development for investors in at least a generation. But investors needn't dismiss something just because it's uncertain.
- 37:25
- Unidentified speaker (UNKNOWN): The
- 37:25
- Unidentified speaker (speaker_0): right response isn't paralysis, but a calibrated humility regarding strategy selection, position sizing and risk taking, and b
- 37:36
- Unidentified speaker (UNKNOWN): a
- 37:36
- Unidentified speaker (speaker_0): willingness to act in the presence of another. uncertainty. Different aspects of investing in AI require answering different questions. An equity investor in a frontier lab may need to be right about technological progress, competitive positioning, pricing power, future profitability, and
- 37:56
- Unidentified speaker (speaker_0): valuation. A lender financing a data center under a long term contract may need to get fewer things.
- 38:03
- Unidentified speaker (speaker_0): right, but the things that matter, the tenant's creditworthiness, contract terms, power availability, and residual asset value are quite different. Thus, a key question is which of the questions am I or my organization well suited to answer? And which uncertainties am I willing to
- 38:23
- Unidentified speaker (speaker_0): bear in the absence of that expertise? Everything else being equal, a high level of uncertainty can be acceptable to investors if the price of the investment is low enough relative to its intrinsic value.
- 38:37
- Unidentified speaker (UNKNOWN): At
- 38:37
- Unidentified speaker (speaker_0): least in theory, the sectors rated lower on clause predictability scale can be the better investment prospects if they're sufficiently cheaper than the more predictable ones.
- 38:48
- Unidentified speaker (UNKNOWN): At
- 38:48
- Unidentified speaker (speaker_0): the end of the day, intelligent investments. investing is largely about price. Note, however, that in the case of AI, thinking about intrinsic value and thus cheapness, is greatly complicated by the questions just raised.
- 39:03
- Unidentified speaker (UNKNOWN): In
- 39:04
- Unidentified speaker (speaker_0): an effort to tie this all together, I'll leave you with a list of questions that might help you reach your decisions. Note, however, that these are questions that apply to all investments.
- 39:15
- Unidentified speaker (speaker_0): not just AI.
- 39:17
- Unidentified speaker (UNKNOWN): How
- 39:18
- Unidentified speaker (speaker_0): many things have to go right for me to earn an acceptable return? If my central assumptions prove wrong, how much could I lose? Which risks am I being compensated for and which am I simply assuming?
- 39:33
- Unidentified speaker (UNKNOWN): How
- 39:33
- Unidentified speaker (speaker_0): much optimism is already incorporated in the price?
- 39:37
- Unidentified speaker (UNKNOWN): How
- 39:37
- Unidentified speaker (speaker_0): does this investment change? the overall risk profile of my portfolio? Does it amplify the risk I'm already carrying, or does it diversify it?
- 39:48
- Unidentified speaker (speaker_0): There's an old saying, and it seems highly applicable to AI. Two kinds of investors lose a lot of money. Those who know nothing and those who know everything.
- 39:59
- Unidentified speaker (UNKNOWN): I
- 39:59
- Unidentified speaker (speaker_0): think in the case of AI, it's unusually important to accept that you don't know everything. AI is sure to produce a lot of surprises, but they needn't be unbearable.
- 40:11
- Unidentified speaker (UNKNOWN): The
- 40:11
- Unidentified speaker (speaker_0): thing to ponder is what's the probability an AI investment will disappoint and what's the probability it won't? I'll conclude with a great quote I happened upon the other
- 40:23
- Unidentified speaker (UNKNOWN): day.
- 40:24
- Unidentified speaker (speaker_0): The argument is made that there are no just too many question marks about the near future. Wouldn't it be better to wait until things clear up a
- 40:33
- Unidentified speaker (UNKNOWN): bit?
- 40:33
- Unidentified speaker (speaker_0): You know the pros.
- 40:35
- Unidentified speaker (speaker_0): Maintain buying reserves until current uncertainties are resolved, etc. Before reaching for that crutch, face up to two unpleasant And facts. The future is never clear. You pay a very high price in the stock. market for a cheery consensus. Uncertainty actually is
- 40:55
- Unidentified speaker (speaker_0): the friend of the buyer of long term values.
- 40:59
- Unidentified speaker (UNKNOWN): The
- 40:59
- Unidentified speaker (speaker_0): perfect ending for a memo on AI, isn't
- 41:02
- Unidentified speaker (UNKNOWN): it?
- 41:03
- Unidentified speaker (speaker_0): Note, however, that it came from one of my oldest files, Timeless Wisdom from Warren Buffett that appeared in Forbes magazine on august sixth, nineteen seventy nine.
- 41:14
- Unidentified speaker (UNKNOWN): I'm
- 41:15
- Unidentified speaker (speaker_0): always happy to give Warren the last word.
- 41:19
- Unidentified speaker (speaker_0): November 11th, 2026.
- 41:25
- Unidentified speaker (speaker_0): Thank you for listening to the memo by Howard Marks.
- 41:28
- Unidentified speaker (UNKNOWN): To
- 41:29
- Unidentified speaker (speaker_0): hear more episodes, be sure to subscribe wherever you listen to podcasts.
- 41:50
- Unidentified speaker (speaker_1): Legal information and disclosures. This memorandum expresses the views of the author as of the date indicated, and such views are subject to change without notice. Oak tree has no duty or obligation to update the information contained herein. Further, Oak Tree makes no representation, and it should not be assumed that past investment performance is an indication of future results. Moreover, wherever there is the potential for
- 42:10
- Unidentified speaker (speaker_1): profit, there is also the possibility of loss. This memorandum is being made available for educational purposes only and should not be used for any other purpose. The information contained herein does not constitute and should not be construed as an offering of advisory services or an offer to sell or solicitation to buy any securities or related financial instruments. in any jurisdiction. Certain information contained herein
- 42:29
- Unidentified speaker (speaker_1): concerning economic trends and performance is based on or derived from information provided by independent third-party sources. Oak Tree Capital Management LP, Oak Tree, believes that the sources from which such information has been obtained are reliable. However, it cannot guarantee the accuracy of such information and has not independently verified the accuracy or completeness of such information or the
- 42:49
- Unidentified speaker (speaker_1): assumptions on which such information is based. This memorandum, including the information contained herein, may not be copied, reproduced, republished, or posted in whole or in part in any form without the prior written consent of Oak Tree.
- 43:03
- Unidentified speaker (speaker_2): Audition