The Noble Update Podcast

Speakers debate AI investment viability, citing commodity risks and bankruptcy fears versus long-term growth potential

1h 27m Listen to the original Episode page
Artificial IntelligenceNVIDIAVenture CapitalMarket BubblesSemiconductorsEnterprise SoftwareBanking RiskTesla

Summary

Speaker 0 argues that the AI sector is a bubble driven by vendor financing, accounting gimmicks, and unsustainable debt, predicting bankruptcies among neo-clouds like CoreWeave. Speaker 1 contends that AI is in the 'third inning' of wealth creation, viewing current valuations as a buy opportunity despite near-term volatility. The discussion contrasts the transformative potential of AI with historical parallels to the 1999 internet bubble and 2006 housing crisis. Both speakers acknowledge AI's technological importance but disagree sharply on the financial sustainability of current infrastructure spending and enterprise adoption rates.

Key takeaways

  1. 1
    If I keep lending you the money to buy chips from me, I can make up whatever number I want. All this vendor financing, Ponzi financing, the Bank of AI, also known as NVIDIA, would not have to be going through all these machinations. The market would freely give the money to these companies, but they're not.

    Unidentified speaker (speaker_0) · 0:00

  2. 2
    yeah, the internet changed the world. It's certain the bulls were correct, but if you own Global Crossing or Newcent or Nortel, you lost everything.

    Unidentified speaker (speaker_0) · 3:50

Episode sections

  1. 0:00 – 0:28

    Vendor Financing and the 'Bank of AI'

    Speaker 0 argues that NVIDIA acts as a 'Bank of AI' providing vendor financing, which allows it to manipulate revenue numbers. Speaker 1 counters that this is smart business, noting NVIDIA puts in $250 billion and takes out $420 billion.

  2. 0:28 – 2:45

    Opening Statements: Third Inning vs. Lousy Investment

    Speaker 1 (Gene Munster) claims the AI trade is in the 'third inning' of wealth creation and is a buy opportunity. Speaker 0 (George Noble) agrees AI is a wonderful product but a lousy investment, citing the disconnect between product value and stock performance.

  3. 2:45 – 4:06

    Anthropic Economics and Commodity Status

    Speaker 0 argues AI is a commodity, noting he spends ~$200/month on Anthropic for ~$10,000 in value. He compares the current AI boom to the 1999 internet bubble, where bulls were right about the technology but wrong about the stocks (e.g., Nortel, Global Crossing).

  4. 4:06 – 7:39

    Energy to Intelligence and Investment Case

    Speaker 1 defines AI as converting energy to intelligence, arguing it solves major global problems. He acknowledges Anthropic may be losing money on specific trades but believes the market will eventually create value. He notes Anthropic retains 70-80% of raised cash.

  5. 7:39 – 11:05

    CapEx, Depreciation, and Cash Flow Negativity

    Speaker 0 highlights the disconnect between trillion-dollar CapEx and current revenues, noting hyperscaler depreciation could exceed $500 billion/year. He compares the AI sector to the 2006 housing bubble, where slowing price appreciation signaled trouble for cash-flow-negative enterprises.

  6. 11:05 – 19:53

    TAM, Adoption, and Enterprise Throttling

    Speaker 0 claims enterprise adoption is throttling due to cybersecurity concerns, citing Boeing and Caterpillar. Speaker 1 argues the Total Addressable Market (TAM) is tens of trillions, citing average AI spend per employee ($25) vs. traditional software ($700) as evidence of massive growth potential.

  7. 19:53 – 24:45

    Market Sentiment and Small Improvements

    Speaker 1 argues that small improvements in autonomous driving or robotics can fuel market optimism, even if ROI is distant. He compares investing to religion, requiring faith in future breakthroughs. Speaker 0 counters that market sentiment is shifting, citing canceled IPOs and debt issues.

  8. 24:45 – 30:04

    Reflexivity and Capital Access

    Speaker 0 uses the 'Euro Disney' and 'Euro Tunnel' examples to illustrate how concept stocks fail when reality hits. He argues AI is 'reflexive on steroids,' where share prices dictate access to capital, and if the market stops funding, growth rates and valuations collapse.

  9. 30:04 – 32:25

    Internet Traffic Growth vs AI Inference

    Speaker_1 argues that internet traffic growth was underestimated in the past, citing a drop from expected 44% compound growth to low 20%, and uses this to predict AI inference will similarly exceed expectations due to its paradigm-shifting nature.

  10. 32:25 – 36:22

    2027 Revenue Forecasts for AI Companies

    Speaker_2 asks for a 2027 revenue threshold for Anthropic, OpenAI, and Grok. Speaker_1 estimates a $300 billion run rate for Anthropic and OpenAI by 2027, with potential 100% growth in 2028, totaling $600 billion for two companies.

  11. 36:22 – 44:29

    Tesla as a Cautionary Tale for AI Predictions

    Speaker_0 uses Tesla's history of missed revenue and margin targets (e.g., 6 million cars vs 1.7 million) and stock performance to argue that long-term AI revenue forecasts are unreliable and that market sentiment can detach from fundamentals.

  12. 44:29 – 47:02

    Regulatory Hurdles for Tesla FSD and Robo-Taxis

    Speaker_0 contends that Tesla lacks Level 4 autonomous driving capability, which regulators will not allow for robo-taxis, contrasting this with Speaker_1's view that the market is pricing in future FSD success despite current misses.

  13. 47:02 – 52:14

    Market Signals and Financing Concerns

    Speaker_0 points to slowing growth signs, such as OpenAI raising less capital at lower valuations and canceled IPOs, arguing that rising cost of capital and tightening financial conditions pose a risk to the AI build-out.

  14. 52:14 – 1:27:06

    Vendor Financing and NVIDIA's Role

    Speaker_0 criticizes NVIDIA's vendor financing as a sign of weak demand, while Speaker_1 defends it as logical business, noting NVIDIA earns significant margins on the financed chips and that the financing pool is backed by major institutional investors.

Assets discussed

AssetKindStanceContext
NVIDIANVDAequitymixedAll this vendor financing, Ponzi financing, the Bank of AI, also known as NVIDIA, would not have to be going through all these machinations.
OpenAIprivate companyneutralAs we all know, most roads lead through either anthropic or uh OpenAI.
Anthropicprivate companyneutralAs we all know, most roads lead through either anthropic or uh OpenAI.
SpaceXprivate companyneutralthis comes back to the SpaceX, Anthropic, the S1s about these massive $20, $30, $40 trillion TAMs.
AppleAAPLequityneutralThis is not 3D TV. This is not Vision Pro from Apple.
TeslaTSLAequitymixedI freely admit that I lost money in the past shorting Tesla, but I also happily report that this year has worked out pretty well so far.
BlackRock, Inc.BLKequityneutralNVIDIA has there's different levels of pledges that they've made, and some of them have gone up and down, but just for like rough math here, has is committed like 250 billion or so in this. these types of financing this vendor financing that is amongst a package so they make a commitment and a lot of times other people blackrock
Blackstone Inc.BXequityneutralblackstone goldman um Apollo will invest alongside that that pool is about a trillion dollars
CoppercommodityneutralYeah, but that's not quite right because in the early two thousands, copper was going up in price because pr supplies is in markets, we as we always know, is relatively price analysis and uh uh in elastic in the short run.
AlphabetGOOGLequitymixedUh Google, Microsoft.
MicrosoftMSFTequitymixedUh Google, Microsoft.
GoldcommoditybullishI'm I'm buying. things I like, energy, gold, etc. And I'm short tech and consumer discretionary, just to kind of embro brush in broad brush strokes.

“Stance” is how the speakers talked about the asset, not a recommendation.

Claims worth verifying

NVIDIA puts in $250 billion and takes out $420 billion, which Speaker 1 calls 'smart business' rather than a Ponzi scheme.

Unidentified speaker (speaker_1) ·0:16

How to verify: Check NVIDIA's Form 10-K or 10-Q filings on SEC EDGAR for cash flow from operations and investing activities, or analyze vendor financing disclosures in notes to financial statements. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

AI is in the 'third inning' of wealth creation and is a buy opportunity.

Unidentified speaker (UNKNOWN) ·0:28

How to verify: This is a subjective market view/forecast. Verify by checking Gene Munster's public statements or Deepwater Asset Management reports from the release period (2026-08-05 to 2026-11-05). For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Speaker 0 spends $200 a month on Anthropic tokens, receiving ~$10,000 worth of value.

Unidentified speaker (UNKNOWN) ·3:16

How to verify: This is a personal anecdote. Verify by checking Anthropic's pricing page or API documentation for token costs versus estimated compute value, though the 'value' is subjective. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Chinese competitors offer AI services at 95% less cost.

Unidentified speaker (UNKNOWN) ·3:39

How to verify: Compare pricing data for AI inference/training services from Chinese providers (e.g., Baidu, Alibaba) versus US providers (e.g., OpenAI, Anthropic) using public API pricing sheets. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Investors in Nortel, Lucent, and Global Crossing lost everything during the 1999 internet bubble.

Unidentified speaker (speaker_0) ·3:50

How to verify: Check historical stock price data for Nortel (NTDO), Lucent (LU), and Global Crossing (GLBCK) from 2026-2026 via Yahoo Finance or SEC filings to confirm bankruptcy or >90% decline. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

Anthropic retains 70-80% of the cash it has raised.

Unidentified speaker (speaker_1) ·6:21

How to verify: Check Anthropic's S-1 filing or latest private market valuation reports and balance sheets to compare cash on hand against total equity raised. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

Hyperscaler depreciation will run upwards of $500 billion a year, exceeding all their earnings.

Unidentified speaker (speaker_0) ·13:16

How to verify: Sum the depreciation and amortization expenses for major hyperscalers (Microsoft, Amazon, Google, Meta) from their latest 10-K filings and compare to net income. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

Average AI spend per employee is $25/month, compared to $700 for traditional software.

Unidentified speaker (UNKNOWN) ·18:50

How to verify: Verify the source of this data, cited as 'RAMP' (a payment company), by checking RAMP's published reports or press releases from the relevant period. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

1% of total ride-sharing is autonomous this year.

Unidentified speaker (speaker_1) ·22:48

How to verify: Check ride-sharing volume reports from Waymo, Uber, or Lyft for the current year to calculate the percentage of autonomous miles/trips. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Oracle CDS and CoreWeave debt issues indicate market stress and canceled IPOs.

Unidentified speaker (speaker_0) ·29:21

How to verify: Check CDS spreads for Oracle on Bloomberg/Reuters and search for news regarding CoreWeave's IPO status and debt financing terms from financial news outlets. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Internet traffic growth decelerated from a projected 44% compound annual growth to the low 20% range over the last five years.

Unidentified speaker (speaker_1) ·31:15

How to verify: Check historical internet traffic data from sources like Cisco Visual Networking Index or ITU reports for compound annual growth rates between 2026-2026. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Anthropic and OpenAI will have a combined 2027 run-rate revenue of $300 billion.

Unidentified speaker (speaker_1) ·33:36

How to verify: Monitor quarterly earnings reports or official guidance from OpenAI and Anthropic upon their IPOs or public disclosures, then annualize for 2026. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

AI companies will grow 100% in 2028.

Unidentified speaker (speaker_1) ·33:56

How to verify: Compare actual revenue growth of major AI firms (OpenAI, Anthropic, etc.) in 2026 against 2026 figures. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Tesla will sell 1.7 million vehicles in 2026.

Unidentified speaker (UNKNOWN) ·43:23

How to verify: Check Tesla's quarterly delivery reports for 2026. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Tesla made approximately $1.60 per share last year, missing an estimate of $8.00 from three years prior.

Unidentified speaker (speaker_0) ·43:43

How to verify: Check Tesla's GAAP earnings per share (EPS) for the most recent fiscal year and compare with historical analyst consensus estimates from three years prior. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

OpenAI raised money recently at a lower price and amount than expected.

Unidentified speaker (speaker_0) ·47:23

How to verify: Review press releases or financial news reports regarding OpenAI's most recent funding round, including valuation and amount raised. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

NVIDIA has committed approximately $250 billion in vendor financing.

Unidentified speaker (speaker_1) ·53:51

How to verify: Check NVIDIA's 10-K or 10-Q filings with the SEC for commitments to customer financing programs. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

76% of voters opposed data centers.

Unidentified speaker (speaker_0) ·50:34

How to verify: Identify the specific poll referenced by Speaker_0 (likely a local or state-level survey) and verify the results. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

NVIDIA captures 60% of revenue and 70% gross margin on financed chips.

Unidentified speaker (UNKNOWN) ·54:36

How to verify: Check NVIDIA's financial statements for gross margin percentages and review the terms of its vendor financing agreements to determine revenue recognition and margin impact. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

Speaker_0 bets that video margins will be a fraction of present levels in five years, and the bubble is in profit margins, not revenue.

Unidentified speaker (speaker_0) ·56:33

How to verify: Check historical margin trends for video streaming companies (e.g., Netflix, Disney) and current analyst forecasts for 2026. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Speaker_0 claims CoreWeave has $5 billion in equity and $70 billion in debt.

Unidentified speaker (UNKNOWN) ·57:43

How to verify: Review CoreWeave's private financial statements or SEC filings if publicly traded; verify via financial news reports on their capital structure. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

Speaker_0 states CoreWeave accounts for approximately 8% of NVIDIA's revenues.

Unidentified speaker (speaker_0) ·1:01:03

How to verify: Check NVIDIA's quarterly earnings reports (10-Q/10-K) for customer concentration disclosures or third-party analyst estimates. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Primary source

Speaker_1 claims Meta's entire business next year will be based on 'Muse' (AI agents), with millions of users.

Unidentified speaker (speaker_1) ·1:03:33

How to verify: Review Meta's recent earnings call transcripts or press releases regarding 'Meta AI' or 'Muse' user metrics and business strategy. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Speaker_0 claims Chinese AI models cost 95% less than Western equivalents.

Unidentified speaker (speaker_0) ·1:09:32

How to verify: Compare token pricing data from major AI providers (OpenAI, Anthropic, Google) with Chinese providers (Alibaba, Baidu, Moonshot) via public API pricing pages. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

The speaker claims the market is a 'stock picker's paradise' with significant volatility and dispersion, rewarding analysis over momentum trading.

Unidentified speaker (speaker_0) ·1:23:20

How to verify: Review market volatility indices (VIX) and sector dispersion metrics (e.g., S&P 500 sector performance variance) for the period 2026-08-05 through 2026-11-05. For recent events, verify 2026-08-05 through 2026-11-05. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify NVIDIA's vendor financing terms and total outstanding loans by reviewing NVIDIA's quarterly 10-Q filings and investor presentations for 'financing receivables' disclosures.
  • Check CoreWeave's debt maturity schedule and interest coverage ratios in their latest S-1 or 10-K filings to assess bankruptcy risk mentioned by Speaker 0.
  • Compare Anthropic's and OpenAI's actual revenue growth against the $300 billion run rate forecast cited by Speaker 1 using public funding rounds and earnings reports.
  • Research enterprise AI adoption rates and cybersecurity rejection rates for AI tools by reviewing reports from firms like Gartner or Forrester to validate Speaker 0's claims about throttling.
  • Analyze the depreciation schedules of hyperscalers (Microsoft, Meta, Google) to understand the lag between CapEx spending and expense recognition cited by Speaker 0.
  • Review Tesla's Level 4 autonomous driving regulatory status and court rulings to assess the feasibility of robo-taxi deployments discussed by Speaker 0.
  • Compare Chinese AI inference costs against US providers using public pricing data from cloud providers like Alibaba Cloud or Tencent Cloud to verify the 95% cost difference claim.
  • Examine canceled IPOs and recent secondary financing rounds for AI startups to evaluate the 'reflexivity' and capital access constraints mentioned by Speaker 0.
Full transcriptexpand
0:00
Unidentified speaker (speaker_0): If I keep lending you the money to buy chips from me, I can make up whatever number I want. All this vendor financing, Ponzi financing, the Bank of AI, also known as NVIDIA, would not have to be going through all these machinations. The market would freely give the money to these companies, but they're not.
0:16
Unidentified speaker (speaker_1): They will put in 250 billion, they will take out 420 20 billion. What is a Ponzi scheme about that? That is just smart
0:26
Unidentified speaker (UNKNOWN): business.
0:28
Unidentified speaker (UNKNOWN): I
0:28
Unidentified speaker (speaker_1): think this is a buy opportunity in the market for AI.
0:36
Unidentified speaker (speaker_2): We've got an incredibly important conversation today. Uh it's going to be a debate about AI, the future of this investment cycle. Is it a bubble? Is it a tremendous opportunity? Has have wonderful assets been sold at farmers. far too cheap a price with the the uh drawdown in semiconductors from June
0:57
Unidentified speaker (speaker_2): getting quite significant. We've got a bull and we've got a bear. We've got uh Gene Munster, managing director at Deepwater Asset Management, and George Noble of the Noble Update Substack. Gentlemen, great to have you here. Um, I think that not only is this an incredibly important conversation. there's two people who are
1:17
Unidentified speaker (speaker_2): got very opposing views. So even though, of course, this will be respectful, you know, don't hold back. Let's let's uh let's have this be uh full of lots of lots of sparks. Um, and of course, this is a very important time with uh the this the semiconductor stocks and the AI infrastructure stocks more broadly having had
1:37
Unidentified speaker (speaker_2): a very rough day after the uh uh the blue tsunami of the of election day yesterday. Gene, let's start with you. So let's go opening statements. Gene, what is your secular view of AI and the associated stocks with it? Where are we in the cycle? And and your
1:57
Unidentified speaker (speaker_2): views on the current drawdown. And then George, you you'll get your opening statement.
2:01
Unidentified speaker (speaker_1): Uh thank you, Jack. The highest level, we are nowhere near the top of a bubble. I think we are in the third inning of wealth creation. You probably want to own the trade from the third to the seventh inning.
2:15
Unidentified speaker (UNKNOWN): You
2:15
Unidentified speaker (speaker_1): don't want to stay to the end. I think uh in terms of the actual adoption, we are at the first pitch of AI. So uh yes, we've had some head headwinds here more recently, but I think the ultimately, this is the most important part. This everyone would agree that AI is going to be
2:35
Unidentified speaker (speaker_1): impactful. The question is, will the substance exceed the hype? And I think that will in fact happen.
2:41
Unidentified speaker (speaker_2): George, what's your opening statement? Where are we right
2:44
Unidentified speaker (UNKNOWN): here?
2:45
Unidentified speaker (speaker_0): Um, and agree with much of what Gene said. Um,
2:49
Unidentified speaker (UNKNOWN): I'm
2:49
Unidentified speaker (speaker_0): actually quite bullish on it. AI. AI the technology. Um
2:55
Unidentified speaker (speaker_0): but uh is uh my former boss Peter Lynch used to say, the product's not the stock and stock is not the product. You can have a wonderful product and have it be a lousy stock, vice versa.
3:07
Unidentified speaker (UNKNOWN): Um,
3:08
Unidentified speaker (speaker_0): you know, I was discussing this table with my producer,
3:11
Unidentified speaker (UNKNOWN): and
3:11
Unidentified speaker (speaker_0): I think I I think this is a graphic example of why I think it's a bad business.
3:15
Unidentified speaker (UNKNOWN): Um,
3:16
Unidentified speaker (UNKNOWN): you know,
3:16
Unidentified speaker (speaker_0): we're using anthropic. I think we're spending 200 bucks a month on tokens or something like that. And if I'm not mistaken, we're getting in the order of like something like $10,000 worth of stuff for it.
3:25
Unidentified speaker (UNKNOWN): So I
3:26
Unidentified speaker (speaker_0): don't know if I'm single-handedly bankrupting anthropic or not. So the the problem really is there i it it's the product versus the business model. And it's a commod, commodity in my view, in my opinion, it's a commodity.
3:39
Unidentified speaker (UNKNOWN): Um,
3:39
Unidentified speaker (speaker_0): the Chinese are doing it for 95% less. I think the open models are going to dominate the industry. And just like was in the case, just as an example, you know, in dot com,
3:49
Unidentified speaker (UNKNOWN): uh
3:50
Unidentified speaker (speaker_0): yeah, the internet changed the world. It's certain the bulls were correct, but if you own Global Crossing or Newcent or Nortel, you lost everything.
4:00
Unidentified speaker (speaker_2): Thank you. Gene, what do you make of that? And in particular, the
4:06
Unidentified speaker (speaker_2): criticism not of AI as a technology or that it's going to change the world, which you know it very likely is, but that of criticism of the AI business model, the business model of serving tokens.
4:18
Unidentified speaker (speaker_1): So at the the core, the most basic, the fundamental principle around AI is taking energy and converting it into intelligence. And I understand that we're drawing an important distinction here between the use of AI and how impactful it'll be, and separately the investment case. But when it comes
4:38
Unidentified speaker (speaker_1): to the investment case, uh, my view is that taking that energy and moving into intelligence, we have an opportunity to solve some very big problems. It could be for a more successful grid. it could be for curing disease, it could be for potentially solving for world peace. And at the
4:57
Unidentified speaker (speaker_1): most basic level, if you have a belief like I do, that it will help solve some of the biggest challenges, how do you put economic an economic value on that? That's the question. Uh George is accurate in terms of whenever he is doing do whatever he whatever
5:14
Unidentified speaker (UNKNOWN): is.
5:14
Unidentified speaker (speaker_1): he's doing with Anthropic, Anthropic is losing money on that trade right now. But ultimately, I believe that if in fact these models can do, and we can talk about the intelligence and where they're at and where they're going, but if they do help solve the world's biggest problems and a lot of other problems along the way, I
5:34
Unidentified speaker (speaker_1): think that the natural hand of the market will create value here.
5:39
Unidentified speaker (speaker_2): Okay. I also want to introduce Gene. I thought you were going to say that anthropic is profitable. So there are tons of people and I was one of them and George it sounds like he's is one of them who are uh probably anthropic has a negative gross margin on, but overall the the gross margins are quite high and the operating margins are quite high as well. U2 and Q3.
6:01
Unidentified speaker (UNKNOWN): I
6:01
Unidentified speaker (speaker_1): was giving uh George the benefit of the doubt here uh that there is uh a question about just in terms of how fast they're depreciating some of those, all that infrastructure and the impact and profitability. So you can, I think, look at this and objectively say that maybe it's not as profitable as they're saying it is. But at again, at the end of the day, I think their cash position, I would agree with that, Jack,
6:21
Unidentified speaker (speaker_1): jacket's it's much higher than what most people have. I think they've retained almost 70%, 80% of the cash that they've raised, they still have.
6:27
Unidentified speaker (speaker_1): which is remarkable given the kind of growth. And so, um, you know, from I I just want to maybe throw another kind of point, I don't want to mix concepts here too. But when we think about transformation, think about what has happened. Typically, when you see big things happening, uh, there are
6:47
Unidentified speaker (speaker_1): crazy things that happen in anticipation of that. And we can go and and think about how fast we've gone for the last few years, which is awesome. And George would agree with all that. But ultimately the question is, is that momentum sustainable? And uh I think I've got a case that in fact it is in fact that
7:07
Unidentified speaker (speaker_1): despite everything we've seen today with these moratoriums on data centers, I still believe that the brain of AI over the next few years is going to be meaningfully two, three times bigger than most investors expect.
7:19
Unidentified speaker (speaker_2): So, George, of course, you can respond to you know anything that that was said before, but I I also um you've called AI the greatest capital misallocation in history, and Gene says it's the third inning. So clearly you have different expectations about what 2027, 2028, 2029, and 2030 are gonna going to look like. So George. could you kind
7:39
Unidentified speaker (speaker_2): of paint a picture? Um, you know, obviously perhaps a bullish one about like adoption and stuff, but in terms of the economics, the pricing, how you kind of see this this movie playing out, you know, have you seen this movie before? Um, and then I want to get Gene's reaction to
7:52
Unidentified speaker (UNKNOWN): it.
7:52
Unidentified speaker (speaker_0): Thanks. It's very good. Well, it's a lot to unpack there. So um again,
7:59
Unidentified speaker (UNKNOWN): I
7:59
Unidentified speaker (speaker_0): think A is a wonderful thing. I just don't think it's a good investment. I'd like to uh by way of uh example,
8:07
Unidentified speaker (UNKNOWN): uh
8:09
Unidentified speaker (UNKNOWN): if
8:09
Unidentified speaker (speaker_0): you look to go back to nineteen ninety nine, the internet, and listen, this is not dot com. History rhymes, it doesn't repeat itself. There are ways in which it's very different, there are ways in which it's similar, and we can go through all that. But if one was bullish on the internet in nineteen 1999, ninety nine,
8:25
Unidentified speaker (UNKNOWN): the
8:25
Unidentified speaker (speaker_0): bulls were actually right. The bulls were actually right. Peter Barrison, the bank credit analyst, points out that
8:32
Unidentified speaker (UNKNOWN): I
8:32
Unidentified speaker (speaker_0): think internet usage compounded at 43% a year for the next 25 years. Went up by like, I don't know, 500,000% or 500 million, some crazy, crazy number.
8:43
Unidentified speaker (UNKNOWN): Um
8:43
Unidentified speaker (speaker_0): so hold that thought. But at the same time, you know, most m many, if not most of the companies that were involved. in the internet in 99, Global Crossing, Lucent, Nortel, Cisco, AOL,
8:59
Unidentified speaker (speaker_0): go down the list. Mainly went bankrupt, mainly saw saw their share prices uh fall 90%.
9:06
Unidentified speaker (UNKNOWN): And
9:06
Unidentified speaker (speaker_0): this is the way that investments usually work. Go back to the railroad boom. It's it History is replete with examples. And so, you know, railroad. railroads wind up to be a very important technology. And it's wound up to be a very important technology. I'm sure AI will probably w wind up being a very important technology.
9:22
Unidentified speaker (UNKNOWN): It
9:23
Unidentified speaker (speaker_0): doesn't mean the stocks are attractive. Uh and you know if you put some numbers to it and you I Gene knows the numbers better than I do, but you talk about CapEx running on I don't know a trillion dollars a year or a trillion and a half next year, whatever is the anal you sum it all up.
9:38
Unidentified speaker (UNKNOWN): How much
9:39
Unidentified speaker (speaker_0): money is being spent on AI over a five-year period? And I think people come up with numbers like approaching $10 trillion, whatever the numbers are. The point of it is, I'm not I'm interested in in accuracy, not precision here. You work backwards, um, how much gross profits have to be generated in order to justify this There's trillions of dollars of
9:58
Unidentified speaker (speaker_0): expense. And I can't get from here to there. Uh and so there's a disconnect between the product, the technology, and um in and the economics of it. And so as an example, you know, if you look at um individuals, it's either free for them or they're paying twenty bucks a month. You pay up, you can pay two hundred bucks a month.
10:18
Unidentified speaker (speaker_0): I'm told we're paying two hundred bucks a month, we're still gonna get zillion dollars worth more value out of it. Um, and so you know, the TAM. and Gene knows this better than I, he's been very correct on a lot of the uh you know Netflix and all the other guys. The TAM at 20 bucks a month is much different than the TAM at 200 bucks a month. And then you they say, well,
10:38
Unidentified speaker (speaker_0): you know, how about all the corporate enterprise market? And we all read the same stuff. There's been information out recently, uh, they I think came it in the S1, you know, 1% of the customers' account for 80% of the business, whatever the numbers
10:49
Unidentified speaker (UNKNOWN): are. And then
10:51
Unidentified speaker (speaker_0): it is a well, how many of those actually are making money? How many are loss making? All this other stuff. Here's a very important point. I'm on a ramble here, but you you really uh triggered me, Jack, with the question. If you look in the corporate arena, the enterprise level,
11:05
Unidentified speaker (speaker_0): almost universally, um the big corporate clients, the who's supposed to be the savior for AI, they're not taking this up. Uh the adoption is not. occurring the way the bulls think. In fact, you're seeing a lot of throttling back. But more importantly, what gets lost on a lot of folks, talk to companies themselves. Cannot or will not give their hand over their
11:25
Unidentified speaker (speaker_0): data to these large language models because of concerns over cybersecurity and compliance. And so you look at Boeing, Caterpillar, That go down the list. So the way I look at it, yeah, there's a huge, there's a huge market for it if it's free. But then if it's 20 bucks, 200 bucks, I think get some ice water.
11:44
Unidentified speaker (UNKNOWN): So
11:44
Unidentified speaker (speaker_1): let me ask you this question. You uh you talked about I have a hard time getting from here to there in relation to the amount that's being spent on CapEx and then some sort of a multiple on what the earnings is going to be a few years from now. out. Basically, there's just not going to be enough revenue generated from that.
12:01
Unidentified speaker (speaker_2): Correct.
12:02
Unidentified speaker (speaker_1): And uh
12:03
Unidentified speaker (speaker_1): what are you basing
12:06
Unidentified speaker (speaker_1): the I I can't get from here to there on? Is it that the $200 a month that you just mentioned, uh it's too expensive to really have wide scale adoption? Like what uh help me understand what that gap is, w what where where that that belief is that we're not going to get from here
12:26
Unidentified speaker (speaker_1): to there.
12:27
Unidentified speaker (speaker_0): Yes. Thanks for the question. And I just give you a framework on how to answer the question. It's obviously garbage in, garbage out. You know,
12:33
Unidentified speaker (UNKNOWN): you
12:33
Unidentified speaker (speaker_0): know, you could put your own assumptions a couple where you want. But if you start making calculations like how many customers does Netflix have, and, you know, 20 bucks and how big would the market be at 200 bucks, and start putting pencil to paper, and then plug some gross margin on that, and then look at the enterprise. level and go through all that. You can't,
12:54
Unidentified speaker (speaker_0): and then and then you you think about
12:56
Unidentified speaker (UNKNOWN): um,
12:56
Unidentified speaker (speaker_0): you know, in a similar fashion, just to put this in perspective, um, the hyperscalers, you you look, you the the you look out a few years, the depreciation is gonna be running upwards of $500 billion a year, which it which exceeds all their earnings. And so I just no matter how I try, I can't. cannot make the numbers
13:16
Unidentified speaker (speaker_0): add up. And this is all the system is inherently unprofitable. It gets to a very important point I want to make. Um, Jay Martin put out a really good piece a couple months ago on X,
13:29
Unidentified speaker (UNKNOWN): uh,
13:29
Unidentified speaker (speaker_0): drawing the analogy to the housing bubble. And the point is, you know, the the system itself, the whole AI uh edifice is very cash flow negative. It needs external financing. As we all know, most roads lead through either anthropic or uh OpenAI.
13:47
Unidentified speaker (UNKNOWN): The
13:47
Unidentified speaker (speaker_0): and if you let me ramble here for a second, I will land this plane. Um the key thing that Jay Martin likes to point to, the key, the key moment in the 2008 housing debacle, was not 2008, it was actually 2006, because that's when the rate of rate of house price depreciation slowed. So the Rube Bilber petrol money machine where
14:07
Unidentified speaker (speaker_0): guys were buying things that they couldn't afford, but don't worry, you know, go up 30% and I'll sell it. That started to freeze up. It didn't work.
14:14
Unidentified speaker (UNKNOWN): And
14:15
Unidentified speaker (speaker_0): the read across here to um open a to to AI is that it's a very cash flow negative um
14:22
Unidentified speaker (UNKNOWN): uh
14:22
Unidentified speaker (speaker_0): enterprise. The rate of spending is accelerating, not diminishing. Um, and therefore it depends on ever larger amounts of external capital. And that thus,
14:33
Unidentified speaker (UNKNOWN): if
14:33
Unidentified speaker (speaker_0): there and you you look at the financing that have been done, you know, the OpenAI and Anthropocene doing deals up 50%, up 75%. Not because people are expecting a cash return, but it's because of prices ever going up. If the rate of price appreciation slows, as is now happening,
14:49
Unidentified speaker (UNKNOWN): you
14:49
Unidentified speaker (speaker_0): got a problem.
14:51
Unidentified speaker (UNKNOWN): So
14:51
Unidentified speaker (speaker_1): I I'm still don't understand.
14:58
Unidentified speaker (speaker_1): Because that's a fair question of like ultimately where's the revenue, where's uh it where's it going to come from? And your argument is valid from what I would consider an academic basis, a historical basis. And uh I've recently received a t-shirt. I should have worn it
15:18
Unidentified speaker (speaker_1): tonight, that says it's going to be different this time. Somebody was um was nice enough to send me that that that t-shirt. Of course, um I'm careful when I say that this is different, but this is different.
15:33
Unidentified speaker (UNKNOWN): What
15:34
Unidentified speaker (speaker_1): we're what's going on right now in AI, the significance of it, is hard. I think
15:38
Unidentified speaker (UNKNOWN): that's a good idea.
15:39
Unidentified speaker (speaker_1): for humans to wrap our minds around. Now, this conversation, this my comments may seem kind of fluffy, but the substance is what does it mean right now? The average intelligence, average intelligence, pop quiz, Jack. Can I put you on the spot?
15:53
Unidentified speaker (UNKNOWN): Pop
15:54
Unidentified speaker (speaker_1): quiz. Average intelligence of a human is what IQ? 100. 100. Thank you. That is impressive. I've asked a lot of people. Yeah.
16:02
Unidentified speaker (UNKNOWN): most
16:02
Unidentified speaker (speaker_3): people. say 70. I don't know why.
16:05
Unidentified speaker (speaker_1): I digress.
16:07
Unidentified speaker (speaker_1): The average intelligence of a frontier model right now is 130. We're at general intelligence. There is a hundred and fifty billion or three hundred billion between OpenAI from the model labs, OpenAI and Anthropic, 300 billion is gonna going to be spent in the next 12 months from enterprise customers. That's real
16:27
Unidentified speaker (speaker_1): spending. But that still doesn't answer the question how do you get from 300 billion to trillions of spending?
16:32
Unidentified speaker (UNKNOWN): And
16:33
Unidentified speaker (speaker_1): the answer is that if we look at the adoption, talk about like the Netflix and start going through anything that's knowledge work related and think about the disruption around knowledge work. And we can talk about pricing of tokens and premium tokens, but
16:50
Unidentified speaker (UNKNOWN): I
16:50
Unidentified speaker (speaker_1): think that the essentially the addressable mark, and this comes back to the SpaceX, Anthropic, the S1s about these massive $20, $30, $40 trillion TAMs. I think, I think it's hard to argue that these aren't tens of trillions of dollars of TAM. Any knowledge worker today. And then once we
17:10
Unidentified speaker (speaker_1): start to get to physical AI,
17:12
Unidentified speaker (UNKNOWN): and
17:12
Unidentified speaker (speaker_1): this is going to take longer.
17:15
Unidentified speaker (speaker_1): then we talk about the whole global economy. And I'd say one other piece. Part of my enthusiasm about what these models can do, not only do I kind of research and study the markets, but at Deepwater, we also have an affiliate company that's been around for three years called Intelligent Alpha. They have a a ticker, it's an ETF called
17:35
Unidentified speaker (speaker_1): GPT. It's a global comp, it's a global growth. company. It's not a comp against the SP. GPT, great ticker.
17:43
Unidentified speaker (speaker_0): Yeah.
17:44
Unidentified speaker (speaker_1): But we've learned uh a lot about models, a lot about infrastructure by being practitioners. And we are blown away at how powerful this is getting. And this sounds again like well-traveled, well traveled, but let's look at the numbers. Gets back to the question about how do we get from here to there. If we look
18:04
Unidentified speaker (speaker_1): at the numbers of
18:06
Unidentified speaker (UNKNOWN): the
18:06
Unidentified speaker (speaker_1): big buzz right now is around the personalized agents, Muse,
18:12
Unidentified speaker (speaker_1): uh, dots, Grokbot, all that's kind of stormy. That was the big message around this last. So if we look at the numbers, the usage, it is 20 million dailies, and there's five and a half billion dailies on the internet. Like the how we get there. is two billion people spend a hundred bucks a month. Like it
18:32
Unidentified speaker (speaker_1): sounds crazy. Like how is that how and in what circumstance can that happen? Another way we can get there is right now the average spend, this is there's data from this company called RAMP, which basically is like a payment company, the average spend on AI per employee is 25 bucks a month right now.
18:50
Unidentified speaker (UNKNOWN): The
18:50
Unidentified speaker (speaker_1): average spend on an uh, if you look at software, is 700. So
18:56
Unidentified speaker (speaker_1): this has so much room to grow. As again, the value cases, yes, there's going to be competition. Yes, there's going to be some pricing that's going to go on. But ultimately, if you start to look at every industry that's knowledge today and physical in the future can be impacted by this, I think. it's hard to build a case that
19:16
Unidentified speaker (speaker_1): we're we're we're we're uh we're at the top. One last piece. I should have mentioned this at the top. I don't think that the AI trade is gonna go up into perpetuity, but I do think the next three to five years are gonna surprise people, not just surprise, shock people at how resilient the AI trade is.
19:33
Unidentified speaker (speaker_2): Gene, that thank you for that. And I I'm gonna put a little pressure on you because I cause I know you can handle it, because Cause pr you know, perhaps uh you know, George's argument about yes, AI is gonna be big, that doesn't mean that it's a it's a buy now. It's it's you know, you gotta look at the stocks, not the product. Um, you know, you you totally could be right about robotics and all this stuff. But if it happens in twenty thirty
19:53
Unidentified speaker (speaker_2): two and not twenty twenty seven and twenty twenty eight, we might have a very serious problem. Um tell us, you know, just to use George's language of of I'm getting from here to there. Um the the spectacular growth of AI revenues over the past year, uh, that that has been very high. And I just want to you know acknowledge that that fact.
20:12
Unidentified speaker (speaker_2): But in terms of AI revenues from the model companies, uh, the clouds, where do we have to go to justify that ROI in 2027 and 2028? Because again, you know, in 2033. 2036 there could be robots that do incredible things. You know, robots could move mountains and whatever, um,
20:32
Unidentified speaker (speaker_2): powered by AI. But that doesn't matter if you know there isn't a sh a very large growth in 2027 as there was in 2026.
20:41
Unidentified speaker (speaker_1): So I think that that you mentioned the ROI question, just to be clear, you're talking about the ROI on like the neo clouds or hyperscalers or just the broader AI ROI.
20:51
Unidentified speaker (UNKNOWN): Uh
20:51
Unidentified speaker (speaker_1): I think I think both. Um
20:59
Unidentified speaker (speaker_1): so the the way I think about it is there is the
21:03
Unidentified speaker (UNKNOWN): part
21:03
Unidentified speaker (speaker_1): of the market is about earnings, obviously, and part of it is about expectations. And uh when you invest, ultimately investing is a form of a religion. You have to believe something is gonna going happen to that hasn't happened. yet.
21:16
Unidentified speaker (UNKNOWN): And
21:17
Unidentified speaker (speaker_1): whether it's just continuation going on or whether it's a breakout or a breakdown, like there is some like faith component that you have to have on this. And that's where the market multiple comes from, is a sense of what is the the future opportunity of this company. I don't want to get too abstract here, but I used to cover as an analyst, I covered this little company called Sonic Solutions,
21:36
Unidentified speaker (speaker_1): SNCC. It was like constantly. comically small, like $100 million market cap. They missed, it was something like 32 consecutive quarters.
21:47
Unidentified speaker (speaker_1): Um
21:49
Unidentified speaker (speaker_1): they the stock never went down, not because it was trading below cash, but because for four or five years, five years, five years, five, six years, they
22:02
Unidentified speaker (UNKNOWN): had.
22:03
Unidentified speaker (speaker_1): this vision about high-definition DVD, as silly as that sounds. But the market always just kept saying that golden goose is out there. And someday it's gonna happen.
22:17
Unidentified speaker (speaker_1): When it comes to the ROI question, it's academic. The question is: is the we need to see movement towards that improvement.
22:28
Unidentified speaker (UNKNOWN): And
22:28
Unidentified speaker (speaker_1): we're seeing that with some of the anthropic news. We're hearing it from SpaceX in terms of what they're seeing, in terms of the return on capital. Uh, they say it's about a one-year return on capital, which is phenomenal, assuming that they're being honest, which I think they are. The ultimately, if the excitement is out there about what robotics potentially could
22:48
Unidentified speaker (speaker_1): do, all we need to see is small improvements. improvement. So, small improvement example number one would be today, this year, about 1% of total ride sharing is going to be autonomous. And for anybody who's been in a Waymo or CyberCab, uh robotaxi, it's awesome. It's so much better.
23:08
Unidentified speaker (speaker_1): And
23:10
Unidentified speaker (speaker_1): all that needs to do is go from 1% to 4%. And then people start filling in the blanks. All you need to do is see going from a very small percentage of surgeries that are done with physical AI robotics, and Jensen talked about this in Berlin last week, to a little bit more. And the market,
23:30
Unidentified speaker (speaker_1): I think, will reward it. So my view is that the use cases of AI is going gonna to go through the roof over the next two, three years.
23:39
Unidentified speaker (speaker_1): And that is gonna basically fuel investors' optimism that this can go on for 10 years. And whether it goes on 10 years or not, it's beside the point. But if it's as uh if we are as close to a breakout, I think like FSD, for example, and Tesla, as as silly as that sounds, like going from um,
24:00
Unidentified speaker (speaker_1): you know, what it whatever it is, five billion miles a quarter to
24:03
Unidentified speaker (speaker_1): um, you know, hundreds of billions. Uh the hope of that I think is going to inject enthusiasm across the market that all that infrastructure needs to happen. So
24:17
Unidentified speaker (speaker_1): my view is it's less about I think there is a path. We could talk about subscriptions, we can talk about in enterprise spending. I think there's a path there. But what's more important
24:29
Unidentified speaker (speaker_1): is the
24:31
Unidentified speaker (speaker_1): is the optimism in the market. Right now it's dark, understandable. I think we're gonna see some blowout things coming from AI that it's gonna be really exciting and really scary too.
24:44
Unidentified speaker (speaker_2): George.
24:45
Unidentified speaker (speaker_0): Gene raises a lot of really good points for which I think I have a lot of really good answers. Um
24:51
Unidentified speaker (speaker_0): first thing I'd like to go back uh to read about Warren Buffett and um
24:57
Unidentified speaker (UNKNOWN): one
24:58
Unidentified speaker (speaker_0): of his first big investments was actually in the textile industry in uh New England back in the 50s and 60s. And I think they I forget the name of the contraption, but there was this whiz-bang new technology, some type of fancy loom that was gonna going to go a million times faster than all the other stuff.
25:12
Unidentified speaker (UNKNOWN): And the
25:13
Unidentified speaker (speaker_0): idea was it was gonna going modernize to it's the textile industry, and efficiency would go through the roof, and productivity improves, and everyone's going to make a fortune, and so it would revive the American uh textile industry.
25:24
Unidentified speaker (UNKNOWN): Well,
25:24
Unidentified speaker (speaker_0): what happened was that got commoditized really fast.
25:28
Unidentified speaker (speaker_0): And yes, it brought cost down immeasurably, but then pricing followed it down. So it was a commodity. It quickly became a commodity. And I believe that AI. is a commodity. Um, and it remains to be seen what the pricing power is going to be like.
25:46
Unidentified speaker (speaker_1): Uh George, you've been really kind at letting me give me a lot of room here, but in the interest of kind of uh really emphasizing the points here, I'd be curious in your is it Jevins paradox that uh the guy from uh the British economists uh coal. Like what's your take on that?
26:06
Unidentified speaker (speaker_0): Yeah, i demand increases enormously, but but costs go down equally. That's the problem. So
26:12
Unidentified speaker (speaker_1): but in Jevins' paradox is that the net yes, costs went down for coal, and they thought the coal producers were in big trouble, but the reduction in costs with the corresponding, of course, increase in demand. was more than beneficial to the producers.
26:30
Unidentified speaker (speaker_0): And and and and and listen, Gene, you know, at the end of the day, it doesn't really matter what I think or what you think,
26:36
Unidentified speaker (UNKNOWN): but
26:37
Unidentified speaker (speaker_0): matters is what the market thinks. And agreeing me let me get on to a very crucial point here. Whether you're right or I'm right, it's what the market thinks. Thanks. Okay. And
26:47
Unidentified speaker (speaker_0): you know, when I was younger, I was more optimistic. Faith is a wonderful thing. Um I'm not sure when I invest, I do things on faith so much anymore. But in all seriousness, keep in mind that yes, I agree when you see change of the margin, which brings some of these things closer to reality. That provides
27:07
Unidentified speaker (speaker_0): grounds for hope.
27:08
Unidentified speaker (UNKNOWN): But
27:08
Unidentified speaker (speaker_0): also at the same time, keep in mind we're going through huge upheaval in capital markets. We've come through I started my career in 1981 working in Fidelity. And from 81 to 21, we basically had you know interest rates, inflation, one-way trade all the way down. And everything went up. Stocks, bonds, real estate,
27:28
Unidentified speaker (speaker_0): mines, everything went up. All right. And whether you're financing this stuff at, That you know, bond yields or are 60 basis points or 5% or whatever, and whatever P ratios are. that's a huge difference. So, yes, there'll be there'll be more innovation. I have no argument. I'm I'm sure you're right about that. But um, the market, there's this race
27:48
Unidentified speaker (speaker_0): between um, you know, expectations and what the market demands. And as, you know, it's funny.
27:55
Unidentified speaker (speaker_0): This is a great story, Gene. Two two great stories. So I remember, I'm old enough to remember when um Disney
28:03
Unidentified speaker (speaker_0): uh made plans. They announced they're gonna open up a new uh theme park outside Paris, Euro Disney.
28:09
Unidentified speaker (UNKNOWN): So
28:09
Unidentified speaker (speaker_0): everyone had these great things, you know, oh, they're gonna have X number of customers and blah, blah, blah, and how many pe how many tourists go to Paris and blah, blah, blah. And you can make up all the blue sky calculations you wanted.
28:21
Unidentified speaker (UNKNOWN): It
28:21
Unidentified speaker (speaker_0): was a concept stock with the emphasis on the syllable con. They made one cruise. crucial error. They opened the park.
28:29
Unidentified speaker (speaker_1): They did what?
28:31
Unidentified speaker (speaker_0): They opened the park.
28:32
Unidentified speaker (speaker_2): Disney.
28:33
Unidentified speaker (speaker_0): Disney.
28:34
Unidentified speaker (UNKNOWN): Oh, because
28:34
Unidentified speaker (speaker_1): hope the hope was no longer
28:37
Unidentified speaker (speaker_1): substance behind.
28:37
Unidentified speaker (speaker_0): Okay, okay. And the same thing, by the way, same thing. Euro tunnel. I was running a hedge fund back in the day. Euro tunnel, same thing. You know how many cars are going to the channel, and blah, blah, blah, blah, blah. You can make up anything you want, any scenarios you want, but then once the rubber hits the road,
28:51
Unidentified speaker (UNKNOWN): no
28:52
Unidentified speaker (speaker_0): pun intended, you got a problem. And so But
28:55
Unidentified speaker (speaker_1): by the way, we're with AI, we're gonna have many years of people being able to dream up where this goes.
29:01
Unidentified speaker (speaker_0): Well, yes, but yes, but the the party can only continue as long as the market keeps serving up the capital because there's cash flow negative. Agree. I And the market's telling you right now, witness, you know, Oracle, uh, CDS. Core Weave Debt, et cetera, et cetera, canceled IPOs. Market's not willing to take so much on faith these days.
29:21
Unidentified speaker (speaker_0): And the crucial part, if people can take away one thing from what I'm saying tonight, because you make me make a lot of really good points, Gene. And that is a reflexive nature of all this. So, you know, it's not that, oh, you have the stock, the stocks, uh the companies over here with certain revenues, and earnings, and then the stock prices over there, and somehow they're not. related. No. This is a this is
29:41
Unidentified speaker (speaker_0): reflexivity on steroid. In other words, the share prices are the fundamentals because the share prices, the access to capital is and if you in it, and if if if the companies are asking the market for more money and the market says no, or the market says we're going to give it to you at a only at a much higher price, that impacts the growth rate, and in turn, that
30:01
Unidentified speaker (speaker_0): impacts the valuations.
30:03
Unidentified speaker (speaker_1): Agree.
30:04
Unidentified speaker (speaker_0): So that's why that's why I come back. I totally agree with you. I mean, I'd really be keen to learn more about some of the work that you've done because I I will confess I'm not an AI nerd. I'm saying it affectionately. Okay, I haven't gone over the week to do
30:15
Unidentified speaker (UNKNOWN): it.
30:18
Unidentified speaker (speaker_0): AI geek. Okay. But you know, as I as I uh I'm not saying you're a liar, please don't taste it this the a long way. But you've heard the phrase before, you know, liars figure to figures don't. lie. Look at the numbers.
30:28
Unidentified speaker (UNKNOWN): Yeah,
30:28
Unidentified speaker (speaker_1): but
30:29
Unidentified speaker (speaker_0): I just and so going back to your question you threw at
30:31
Unidentified speaker (UNKNOWN): me
30:31
Unidentified speaker (speaker_0): when I when I used the phrase can't get from here to there, I I start there and I end there. You know, internet consumption has gone up by umpti million percent in the last 25 years. It didn't mean it meant to still almost. Well,
30:41
Unidentified speaker (speaker_3): didn't we
30:42
Unidentified speaker (UNKNOWN): uh
30:42
Unidentified speaker (speaker_1): when here's a question?
30:43
Unidentified speaker (speaker_0): Yeah.
30:43
Unidentified speaker (UNKNOWN): Because
30:44
Unidentified speaker (speaker_3): we we've we use historical
30:45
Unidentified speaker (UNKNOWN): uh
30:45
Unidentified speaker (speaker_1): examples too.
30:47
Unidentified speaker (UNKNOWN): You
30:47
Unidentified speaker (speaker_1): mentioned
30:47
Unidentified speaker (speaker_3): earlier about the 40%,
30:49
Unidentified speaker (speaker_1): that's a correct number too of all all I'll all agree with that number about that compounding growth of the internet. Today now more than half of it is pot traffic.
30:59
Unidentified speaker (speaker_0): Right.
30:59
Unidentified speaker (speaker_1): But
31:01
Unidentified speaker (speaker_1): let me uh ask the question: Is like, what what do you can can we learn anything about what's happening with the internet traffic? I think we can, as potentially what could happen with inference.
31:14
Unidentified speaker (UNKNOWN): Meaning
31:14
Unidentified speaker (speaker_2): why? what?
31:15
Unidentified speaker (speaker_1): Meaning what? Basically, if we thought like internet traffic was good, I thought in 2005 it was going to cap out or 2002 and grow at 8% a year. And based on my numbers, that 44%, but the last the look at the last five years, it's more like low 20% Percent compounding annually.
31:35
Unidentified speaker (speaker_1): And my point is like these transformations. all the underlying uh driver or the use case is just much bigger and much more sustainable. When you have these paradigm shifts, this is a paradigm shift. This is not 3D TV. This is not Vision Pro from Apple. Uh, when you
31:55
Unidentified speaker (speaker_1): have a paradigm shift, the size and duration of it is much bigger than you think. And I think that's what we learn from internet traffic.
32:05
Unidentified speaker (UNKNOWN): The
32:05
Unidentified speaker (speaker_1): the analogy, of course, is how much inference is going on today.
32:09
Unidentified speaker (UNKNOWN): And
32:09
Unidentified speaker (speaker_1): if we roll forward, let's I let's let's talk I I I think I fully interrupted your your thought there, George. I want to be uh poignant but yet respectable. Do you want to finish your thought?
32:20
Unidentified speaker (speaker_0): It's okay. There's so many things we can talk about. I don't even remember what I was saying.
32:25
Unidentified speaker (speaker_2): Okay. I'll I'll be I'll be the real rude one now and I'll um uh interrupt you, Gene, and I'll I'll say um peep I'll ask you the the question, which is you know, I people don't like the number ARR because you take you know a favorable period of twenty-eight days, multiply it by twelve. I understand, but like let's just say a quarter quarterly revenue times four, so annualized revenue of Q4
32:45
Unidentified speaker (speaker_2): For 2027 anthropic plus OpenAI plus grok. At what level? of monetization of in terms of annualized revenue of those three, which you your estimate, of course, now will be better than mine, but I'll say is you know, probably maybe like uh 130 to 150 billion. At what
33:03
Unidentified speaker (UNKNOWN): uh
33:03
Unidentified speaker (speaker_2): at what number is it in 2027 that you can give you know your new friend George Noble a call and say, actually, George, you know, look, you had a legendary bear call from you know June to to November. We had quite a panic in this in the semiconductor and AI trade. But actually I think the ROI has kind of so far yet improved. What's kind of your threshold? And then I want to hear if George agrees with
33:23
Unidentified speaker (UNKNOWN): that.
33:24
Unidentified speaker (speaker_1): So um the the numbers that I look at from anthropic and OpenAI is the 2027 run rate of 300 billion. So we're looking slightly different. Is that okay if we kind of start there? Sure sure, sure, yeah.
33:35
Unidentified speaker (UNKNOWN): Yeah.
33:36
Unidentified speaker (speaker_1): And so the question is like what's the number need to be to feel like we're moving in the right direction? And my sense is that they're gonna grow in 2028, probably a hundred percent. I think they'll probably do their guidance and an analyst will come out probably something like seventy, eighty percent kind of growth, a much lower number. But
33:56
Unidentified speaker (speaker_1): I think it's gonna be much higher than people believe. So the question that I have is six $600 hundred. billion
34:03
Unidentified speaker (UNKNOWN): for
34:04
Unidentified speaker (speaker_1): two companies. That's two companies. That's pretty impressive. But when you look at all the trillions that are being spent on infrastructure, does that justify it? And I think the answer is yes. But what's more important, it gets back to this internet usage kind of analogy that I was making relative to inference, is that if the view is at the end of
34:23
Unidentified speaker (speaker_1): 2027 that 2029 is
34:26
Unidentified speaker (UNKNOWN): going
34:26
Unidentified speaker (speaker_1): to grow at 80%,
34:29
Unidentified speaker (speaker_1): then I think this holds together. And ultimately I think that that is going to happen because I think these tools, like, just try them. I mean, I beg you, go Tesla, go and do uh you can take a two day test drive, uh, see what FSD is like. If you haven't, I'm sure you've heard
34:49
Unidentified speaker (speaker_1): you've been annoyed by people who have tried and said it it's a amazing. Try it, you will be amazed. Try Muse, try Grockbot. You'll be amazed. It's changed like our entire world at how we do things at deep water is totally changed based on using these bots, these personalized agents, over since
35:08
Unidentified speaker (speaker_1): mid-September. I It's mean, crazy. So my point is if we're if like that group that's saying, oh my, this is crazy, like crazy good, if that group is really small today, which it is,
35:23
Unidentified speaker (speaker_1): like fractions of a fraction of a percent, and it's gonna be bigger in two years, I think these companies can grow much faster. So back to that, that is the those are two great ones to watch. What's What's gonna going happen to
35:34
Unidentified speaker (UNKNOWN): with
35:35
Unidentified speaker (speaker_1): um
35:35
Unidentified speaker (UNKNOWN): uh
35:36
Unidentified speaker (speaker_1): OpenAI and and then anthropic, we're investors in open. AI and and feel very good about holding this in through the IPO.
35:44
Unidentified speaker (speaker_3): And so I gotta, George, um,
35:46
Unidentified speaker (speaker_2): Jean's
35:47
Unidentified speaker (UNKNOWN): uh
35:47
Unidentified speaker (speaker_2): a
35:47
Unidentified speaker (speaker_3): vision
35:48
Unidentified speaker (UNKNOWN): of
35:48
Unidentified speaker (speaker_2): $300 billion in revenues, annualized revenues by the end of 2027, and then a double of that from 2028. So 600 billion, is that good enough for you? Or do you think even with that, we're gonna have some problems?
36:00
Unidentified speaker (speaker_0): I think it's very difficult to have much faith in those numbers looking out. Um
36:08
Unidentified speaker (UNKNOWN): you know,
36:08
Unidentified speaker (speaker_0): in bull markets, time horizons get extended.
36:11
Unidentified speaker (speaker_1): But why, but George, why? Because history has said that. That's why it's hard to it's I mean, the law of large numbers say it's impossible. Is that I just want to understand understanding.
36:22
Unidentified speaker (speaker_0): why you're
36:22
Unidentified speaker (UNKNOWN): not gonna be able to do that?
36:22
Unidentified speaker (speaker_0): No, no, no, no, but but again, you know, we're talking about the top line results. but there's also the bottom line. So for instance, now we'll go from the flying pan into the fire. Um this is supposed to be about AI. Well, SpaceX is an AI company, really, when you get down to it, three quarters AI. Okay, yeah. But I but I'm gonna do the read across to the other Elon
36:42
Unidentified speaker (speaker_0): Musk company. Um I started my career as an automobile analyst in 1981. nineteen eighty one. I was a the guy carrying Peter Lynch's bags, visiting LeI Lee Iacoka.
36:51
Unidentified speaker (speaker_3): Lee Iacosa.
36:52
Unidentified speaker (speaker_0): I I I was in the room, okay? So autos I know something about. And uh I that is I freely admit that I lost money in the past shorting Tesla, but I also happily report that this year has worked out pretty well so far. And the reason I'm mentioning this is, you know, one can make um
37:12
Unidentified speaker (speaker_0): big predictions about revenues It's hard to know what's going to happen six months from now, about you know, three years from now. If some if you look at what t what hey, not only have the Tesla revenues been way off, but even worse, the the margins associated with those revenues, it's been a complete and utter disaster. All right. And so um, you know, I uh
37:33
Unidentified speaker (speaker_0): history is replete with examples of where, you know, as the Yo OG Gi Bear once famously said, you know. predictions are difficult, especially about the future. Okay. You want to try the big dig on for size in Boston, which is supposed to be three billion, it wound up coming at 19 billion. I mean, I go on and on and on. And so I get I get in the bull market where people are extending
37:53
Unidentified speaker (speaker_0): time horizons and cost of capital is the minimum. Um, that might have, you know, dreams have more appeal. But in a world we have a rising cost of capital, I think before we start. debating whether or not 28 or 29 numbers, what that might look like, if the markets are not going to give you enough, they're not going to give you the capital just to get through the next year because
38:12
Unidentified speaker (speaker_0): they're starting to say, Whoa, wait a second. And keep in mind, Gene, we we know they always say the bond guys are smarter than the equity guys. All right. Bond guys are like, they got they're like, you've got some explaining to do. All right. The equity guys are the dreamers. And so, you know, let's first get past this. hurdle. Then we can worry about whether the 29 estimates are correct. You may well be right. But I'm just saying right now, where I
38:32
Unidentified speaker (speaker_0): sit, I don't think you're right. And more importantly, doesn't matter what I think. The market is not buying what you're selling.
38:38
Unidentified speaker (speaker_2): So so George, I I uh, you know, when Gene said, you know, he he articulated laid out a vision of robots and stuff, and that could I said, okay, that could be 2032, but it doesn't matter. I thought, George, on your behalf, I got George uh I I got Gene to make a commitment or a prediction about the next 24 months, which is the here and now, because of course you're right that in bull markets tight
38:58
Unidentified speaker (speaker_2): times get extended. So I think that is, you know, relatively near term. And so you you're basically saying that you don't think that it's good that's gonna happen.
39:06
Unidentified speaker (speaker_0): I know I I listen, listen, I I will put this way.
39:10
Unidentified speaker (UNKNOWN): I
39:10
Unidentified speaker (speaker_0): think there's an article of faith that one has to believe in to to to go with that. Um, there are signs. I mean, I go back a few months ago, um, kind of funny. I had I had Gary Marcus and Ed Zitron on a podcast, and it was a very week that the Tokomax thing was peaking. And Gary made the comment, just
39:30
Unidentified speaker (speaker_0): watch, and he was 100% right. And then I find it curious that, you know, somehow um, you know, folks like Gavin Baker and um uh uh Brad Gerstner, you know, forgot to mention what the third quarter numbers look like, or Anthropic forgot to mention what the third quarter revenues looked like when they came out with the
39:50
Unidentified speaker (speaker_0): S1. Of course they knew. And so there's there's tangible signs that things are already slowing. And therefore, I may be wrong, but I'm just saying there's a lot of doubt here. And therefore, to to take Gene's forecast, forecasts, which may be right in the end, but to just take them as an article of faith, uh, that's a bridge too far for me.
40:06
Unidentified speaker (UNKNOWN): Well,
40:06
Unidentified speaker (speaker_2): as well, it was say, uh, George, on the anthropic S1, you're right, the Q3 revenue didn't come out, but I b I'm actually pretty sure before it came out, revenue was released. And and then and then the updated one, we have the the revenue.
40:19
Unidentified speaker (speaker_0): Yeah, they but there was there was a lot they withheld. No, there was a lot they there, there was a lot they withheld. Which which would you and I want wouldn't to say anything illegal was done. It's just it was how should I say? Information was released. The truth was they're rather economical with the disclosures, and it was done in a way which kind of cast the company in the best light. But that's not really the point. The point is
40:39
Unidentified speaker (speaker_0): there are signs that things are slowing already.
40:44
Unidentified speaker (speaker_1): Um
40:46
Unidentified speaker (speaker_1): signs of things slowing already. Yeah,
40:48
Unidentified speaker (speaker_2): give us your thoughts on that. That this the revenue slowed from a reduction. ridiculous quarter over quarter number, well in excess of 100%, to a number in the 40%, quarter over quarter, which is still very high. But again, it's the deceleration.
41:02
Unidentified speaker (speaker_1): We're still so early in this adoption curve that
41:07
Unidentified speaker (speaker_1): we can easily have
41:09
Unidentified speaker (UNKNOWN): a
41:10
Unidentified speaker (speaker_1): massive reacceleration.
41:12
Unidentified speaker (UNKNOWN): Um,
41:13
Unidentified speaker (speaker_1): I would put it this way: two years ago. ago, companies were talking about doing AI. Over the past year, they've started to dabble. And the dabble, and I would say, like, give me some hard use cases of like where AI is having a big impact. I don't think we could come out with a ton of them today.
41:31
Unidentified speaker (speaker_1): But yet you have
41:33
Unidentified speaker (speaker_1): OpenAI and anthropic at their what's the the run rate uh probably egg. exiting this year, like you said, 150 billion, going from basically zero to 150 billion over three years. Something is happening. When you see that kind of growth and that low of adoption, that is a sign that something bigger is going on.
41:55
Unidentified speaker (UNKNOWN): I
41:55
Unidentified speaker (speaker_1): don't know. My uh so this idea, I think if as I'm kind of uh from a high level.
42:03
Unidentified speaker (speaker_1): it seems like we're kind of talking about two different things. Like
42:07
Unidentified speaker (UNKNOWN): you're
42:07
Unidentified speaker (speaker_1): talking about costs of capital and history of railroads and
42:12
Unidentified speaker (UNKNOWN): a
42:12
Unidentified speaker (speaker_1): very academic approach to this, back to that theme. But I think if you look at what's actually going on with this technology and use it, I think you'll be blown away and start to believe things are going gonna to change. We're going to get gonna some high. unemployment stuff. There's gonna be some very hard things
42:29
Unidentified speaker (UNKNOWN): that are
42:29
Unidentified speaker (speaker_1): gonna happen.
42:30
Unidentified speaker (speaker_0): So, Gene, much of what you say may come, may in fact come to pass. However, I I want to go back. It's not so academic. I want to go back on Tesla as an example. Okay. You look at the forecasts that were made. But
42:40
Unidentified speaker (speaker_3): Tesla stock has gone up. No, hold on, Gene.
42:43
Unidentified speaker (UNKNOWN): So Gene Gene,
42:44
Unidentified speaker (speaker_0): Gene, this is where I beg to differ with you, please. Up until 2021, Tesla stock was a monster.
42:52
Unidentified speaker (UNKNOWN): I
42:52
Unidentified speaker (speaker_0): never owned it.
42:53
Unidentified speaker (UNKNOWN): I
42:53
Unidentified speaker (speaker_0): was short from time to time. I wish I'd owned it. Shame on me. We all make mistakes.
42:58
Unidentified speaker (UNKNOWN): Uh
42:59
Unidentified speaker (speaker_0): by the
42:59
Unidentified speaker (speaker_3): way, I want to put it, it went from 70 billion to a trillion.
43:04
Unidentified speaker (speaker_0): Right, right. So, J Gene, please, please let me make my point. Um, Tesla stock over the last, With, but the bloom is off the road right now. Tesla's Tesla stock is flat over the last five years. against a market which is up 50 or 60 percent, if not more. All right. More importantly, there's a stock price and the fundamentals. Those two aren't always the same.
43:23
Unidentified speaker (UNKNOWN): Uh
43:23
Unidentified speaker (speaker_0): Tesla is going to sell about 1.7 million, 1.7 million vehicles this year. Okay. The estimate uh five years ago for Tesla was six million cars to be sold this year. Okay. Tesla, I'm out of date now, but already last year they were supposed to have made I think over eight dollars a share. You look at the estimate going three years prior, it was it
43:43
Unidentified speaker (speaker_0): was it was sorry, they made about a dollar sixty a share from memory. They were ex the the estimate three years prior was for eight dollars a share. So I'm not saying it that AI is necessarily Tesla. I'm just trying to illustrate how again predictions are difficult, especially about the future.
43:58
Unidentified speaker (speaker_1): But but but if we agree with those numbers, I agree with those numbers six million cars supposed to be 2027, they'll do two million, something like that.
44:07
Unidentified speaker (speaker_0): Right.
44:07
Unidentified speaker (speaker_1): But the stock is at the same place. That to me says the market is looking beyond all the
44:16
Unidentified speaker (UNKNOWN): the
44:16
Unidentified speaker (speaker_1): downward revisions and saying, what else is to come? Back to this idea about behind door number two. And the behind door number two is
44:26
Unidentified speaker (UNKNOWN): obviously.
44:27
Unidentified speaker (speaker_1): FSD, robo taxi.
44:29
Unidentified speaker (speaker_0): Right.
44:29
Unidentified speaker (speaker_1): And uh
44:31
Unidentified speaker (speaker_0): and and and I would say a couple things. So A, it's it's always what's the next thing, but you know, at a certain point, and I would beg to strenuously differ with your characterization of uh of uh what's going on with the cyber cobs and fsd and that whole thing. The fact is they don't have level four performance. Um FSD, they're selling a bunch, but the regulators aren't gonna aren't gonna allow. the robo
44:51
Unidentified speaker (speaker_0): taxis because that they had not been able to prove level four.
44:54
Unidentified speaker (UNKNOWN): Well,
44:55
Unidentified speaker (speaker_1): we could Jack, we take a bet on that.
44:57
Unidentified speaker (UNKNOWN): Not
44:57
Unidentified speaker (speaker_2): a bet, but
44:58
Unidentified speaker (speaker_1): just make a prediction on that.
44:59
Unidentified speaker (UNKNOWN): Um
45:00
Unidentified speaker (speaker_2): sure, but then I want to move away
45:01
Unidentified speaker (UNKNOWN): from
45:01
Unidentified speaker (speaker_2): Tesla because I feel like you both are so strongly held in your views. No,
45:05
Unidentified speaker (speaker_0): no, I'll be br I'll be brief. I'll I'm gonna land a tough blow, but uh a brief one. Um
45:11
Unidentified speaker (UNKNOWN): you
45:11
Unidentified speaker (speaker_0): uh Elon Musk is the only person, only serious. person in the world. I mean, I take that back. He even gave up a few months ago saying they weren't at low. That's not the point. Tesla um is,
45:24
Unidentified speaker (UNKNOWN): you
45:24
Unidentified speaker (speaker_0): know, if if I come to you, Gene, if I come to you for 12 consecutive years and I say we're gonna have self-driving cars, level four, level four, level four,
45:31
Unidentified speaker (UNKNOWN): at
45:31
Unidentified speaker (speaker_0): about the third or fourth year, you might say George Georgia's is.
45:35
Unidentified speaker (speaker_0): Some might say George is committing securities fraud. right
45:40
Unidentified speaker (speaker_0): it's just not credible I mean you they they're not they're no they don't have level four it's gonna take level four to be able to get to where they need to to get all the all the cyber cabs it's just not happening and Tesla's admitted with their own technology they can't get there and they admitted in court that they don't have level four for. but then they say to the investing community something totally different.
46:00
Unidentified speaker (speaker_0): Let's get back to AI.
46:02
Unidentified speaker (speaker_3): Thank you
46:04
Unidentified speaker (UNKNOWN): So
46:04
Unidentified speaker (speaker_0): I guess what the bigger point really here is, Gene, you could be totally right. I'm just saying, you know, before we get to the promised land that you're suggesting, you know, the the the and you know the stories you get involved on stock. Peter Lynch always used to say, can you upgrade the story, downgrade the story? What do you learn incrementally? You know, is the TAN getting bigger? Is
46:24
Unidentified speaker (speaker_0): it smaller? This, that, whatever. And I guess what I'm saying is, you know, right here, right now, this clearly is evidence by it's, and it's not just the stock prices going down. There's tangible signs that things are slowing. But then on top of that, the reflexive nature of this will really, in a way, very fundamental. The stock prices are the fundamentals because of the reflexive nature of it, that these
46:44
Unidentified speaker (speaker_0): companies have to earn the credibility of the market, the trust of the market. Gene, I'd be happy to come back six months from now. with Jack and say, you know what? You were totally right. I was totally wrong. Okay. But this is not, it's a show me situation. It's not a trust me situation.
47:02
Unidentified speaker (UNKNOWN): And
47:03
Unidentified speaker (speaker_0): you know, in a bull market, people extend horizons, they're willing to believe anything. In a bear market, it's like, well, wait a second, hold on. You said, what? And that's kind of where I think we are. And then so the real problem, I right now is because it'd be one thing if they weren't so dependent on external capital, but they are. And you know, when when OpenAI last week, they got to
47:23
Unidentified speaker (speaker_0): raise some money and they they raised much less and a lower price than expected. Many you see IPOs canceled all over the place, and and who knows what's going to happen with the anthropic. I mean, the c caught their cost of cap, capital, their stock price, that's the cost of capital. Forget about the bond yields, is It's going up. going way up. And the question, the easiest way to make this problem go away, show me the
47:43
Unidentified speaker (speaker_0): money. Show me. I hope, listen, for humanity's sake, I hope you're right. I I just I need to be convinced. That's
47:50
Unidentified speaker (UNKNOWN): all.
47:51
Unidentified speaker (speaker_2): Yeah. So Gene, uh George's argument is in the long term, you could totally be right, but financial conditions are tightening sufficiently now that it may not happen. What do you make? What do you make of that? And do you think that like if uh you know we have a 2008 or 2020 moment where just the
48:11
Unidentified speaker (speaker_2): credit markets freeze and shut down and Microsoft has to borrow at 9%, like then what happens to the
48:17
Unidentified speaker (UNKNOWN): 600?
48:18
Unidentified speaker (speaker_2): And that is by definition kind of a very negative case, but I'm throwing at you and seeing how seeing how you react.
48:23
Unidentified speaker (UNKNOWN): My
48:24
Unidentified speaker (speaker_1): my answer is today is different than yesterday. I think that that is a real conversation about the near term. And I I feel that um, but when you talk to we do venture and public investing, when when people like me say something is 10 years away, they don't know when it's gonna be. Five years,
48:44
Unidentified speaker (speaker_1): that means they think it's gonna happen, but they're not sure when. If they say it's a year or two away, that's there's pretty high confidence in it.
48:49
Unidentified speaker (UNKNOWN): I think
48:50
Unidentified speaker (speaker_1): in the next one. or two years, we're gonna have breakout.
48:53
Unidentified speaker (UNKNOWN): The
48:53
Unidentified speaker (speaker_1): next six months, this isn't like me trying to give a get out of car jail in case the market flips upside down. But the reality is what happened with the blue wave is is negative for the AI trade. And we saw that getting priced in Than today. The question is is it more negative? And ultimately the question is
49:13
Unidentified speaker (UNKNOWN): uh
49:13
Unidentified speaker (speaker_1): and essentially dance. to play this through the cost of capital, is that if it gets difficult to get these data centers up and running, if it's that can have moving them from Texas to Ohio or to some other place can impact the cost of capital that can get investors nervous, that's a problem. But at the end of
49:33
Unidentified speaker (speaker_1): the day, the that basically intelligence at scale.
49:39
Unidentified speaker (speaker_1): I think it's it's an eventuality that it will impact almost everything. And um I think that means that means that eventually these data centers are going to get built or based Elon's tweet today about I mean I can't believe SpaceX wasn't up on this data center news. I understand that it it's a negative trade against
49:59
Unidentified speaker (speaker_1): Elon because of the change in the parties,
50:01
Unidentified speaker (UNKNOWN): but
50:02
Unidentified speaker (speaker_1): this is like they're gonna get the data center thing fixed, whether it's orbital or whether it's some other state. It will get fixed. And so I do believe this is a riskier part. We're entering a riskier part, especially going into thank goodness the the earnings just happened. Maybe, but when that earnings, that December earnings calls at the
50:22
Unidentified speaker (speaker_1): end of January, that's gonna going to be the first time when people are really gonna going to sharpen up on terms. of what those moratoriums mean to the pace of the build-out because that's important to keep this whole thing going.
50:34
Unidentified speaker (speaker_0): If I'm not mistaken, I Gene or Jack, did I see something 76 per I think some polls? 76% of those who voted were opposed to uh data centers.
50:44
Unidentified speaker (UNKNOWN): I
50:44
Unidentified speaker (speaker_1): believe that. I didn't see that. But yeah,
50:46
Unidentified speaker (UNKNOWN): I mean people are in
50:47
Unidentified speaker (speaker_1): total denial. Like if you're opposed to it, you cannot. stop this. I'm sorry. The free hand of the market is gonna push this into reality.
50:56
Unidentified speaker (speaker_0): Well, but Gene, but hang on, the free hand of the market, it's not a Mr. Market. The free hand of the market right now is you know Oracle, uh CDS to the moon, core weave, uh uh bonds that fixed.
51:07
Unidentified speaker (UNKNOWN): Sure, it's
51:08
Unidentified speaker (speaker_2): totally a financing issue. We we've addressed that, and I you
51:10
Unidentified speaker (UNKNOWN): know,
51:10
Unidentified speaker (speaker_2): we can we can I
51:11
Unidentified speaker (speaker_0): don't you know no you know no Jack, the the key point I want to make is yeah.
51:17
Unidentified speaker (speaker_0): I'm just saying what the free hand of the market. The free hand of the market is saying not so much right
51:21
Unidentified speaker (UNKNOWN): now.
51:22
Unidentified speaker (speaker_2): Sure, but Gene is making a point that the market wants it and that's what will get it. So it the data centers will be built somewhere. But Gene, you know, we don't live in China. Like if China wanted to build as many data centers as possible, they would do it. And if you know a million people lived in a rural village and they you know those people would leave, and that
51:42
Unidentified speaker (UNKNOWN): we
51:42
Unidentified speaker (speaker_2): have parallels of that. Um ironically, China's
51:45
Unidentified speaker (speaker_2): we
51:46
Unidentified speaker (speaker_1): don't live in China.
51:47
Unidentified speaker (speaker_2): Yeah.
51:48
Unidentified speaker (speaker_1): That's a statement of fact. We live in a country that is in fierce competition with China.
51:54
Unidentified speaker (speaker_2): Yes.
51:56
Unidentified speaker (UNKNOWN): In
51:56
Unidentified speaker (speaker_2): in the realm of AI.
51:58
Unidentified speaker (speaker_1): Yes.
51:59
Unidentified speaker (speaker_1): That's everything. Everything is AI.
52:01
Unidentified speaker (speaker_0): Right. And I would say talking about the market, if the value proposition, if the ROI, the prospect of ROI was so
52:11
Unidentified speaker (speaker_0): clear, so evident,
52:13
Unidentified speaker (UNKNOWN): then
52:14
Unidentified speaker (speaker_0): I humbly submit to your consideration, all this vendor financing, Ponzi financing, the bank of AI, also known as NVIDIA, would not have to be going through all these machinations. The market would freely give the money to these companies, but they're
52:28
Unidentified speaker (speaker_0): Maybe maybe they're maybe they're looking a little bit a scance at the value proposition. That that's my issue. It's it's like there should not be a need for any of this vendor financing coming from NVIDIA. It's it's can we talk vendor financing, Jack?
52:41
Unidentified speaker (speaker_2): Yeah, let's let's talk vendor financing. And I guess I'll put a point on it is you know, so Gene, you uh began your career in the nineties, 90s so you ha you know had it had us a as a a front row seat to um the vendor financing i read books about the dot com but obviously i was not a you know a a market participant there educate me educate your audience just what happened with the vendor financing and i presume
53:01
Unidentified speaker (speaker_2): you think what's going on right now is different why do you think
53:04
Unidentified speaker (UNKNOWN): that
53:05
Unidentified speaker (UNKNOWN): well
53:05
Unidentified speaker (speaker_1): the the big thing that it wasn't necessarily vendor financing it was the circular contracts that they had around advertising where companies would raise money uh they would spend money on with other companies. It was these are ad driven businesses, a lot of the initial uh internet businesses. And it was like uh basically one business raising money,
53:25
Unidentified speaker (speaker_1): propping up another business of a partnership. And they uh they did it until that capital w ran out, and then there was no buyers for the ads. hats. So
53:36
Unidentified speaker (UNKNOWN): uh
53:37
Unidentified speaker (speaker_1): that got us to two thousand. 2000. The um again, this different kind of thing. I think this is totally logical. I don't think this is a Ponzi screen. It's totally logical. Put some quick numbers on
53:50
Unidentified speaker (UNKNOWN): it. Is
53:51
Unidentified speaker (speaker_1): NVIDIA has there's different levels of pledges that they've made, and some of them have gone up and down, but just for like rough math here, has is committed like 250 billion or so in this. these types of financing this vendor financing that is amongst a package so they make a commitment and a lot of times other people blackrock
54:11
Unidentified speaker (speaker_1): blackstone goldman um Apollo will invest alongside that that pool is about a trillion dollars
54:23
Unidentified speaker (speaker_1): NVIDIA so if we put NVIDIA out as the the the the group that is causing this Ponzi screen, it's it's totally logical for NVIDIA to do this. NVIDIA is going to take 60% of that revenue. They have
54:36
Unidentified speaker (UNKNOWN): uh
54:36
Unidentified speaker (speaker_1): call it 70% gross margin. They will put in 250 billion, they will take out 420 billion. What is a Ponzi scheme above about that? That is just smart business.
54:50
Unidentified speaker (speaker_0): George. Um it's pretty simple. Uh if they're having to lend the money, yeah, it's logical they're doing it because they have it it it directly out of uh agency that owns financial incentive, of course. But if they're selling all these, if they're lending the money to s to the folks to buy all these chips, meanwhile, some ungodly percentage of the chips are
55:10
Unidentified speaker (speaker_0): sitting at in warehouses, uh, you know, not in use. And then on top of it, we now have a added complication where I don't
55:17
Unidentified speaker (UNKNOWN): know about that
55:18
Unidentified speaker (speaker_2): though.
55:18
Unidentified speaker (speaker_0): Well, I I think it's been widely documented that is the case. But then furthermore, whatever you agree or disagree on, we all agree that what's happening right here, right now, probably is certainly not going to accelerate the pace of uh of uh of of of uh uh build out if anything it's gonna slow it and so I would argue that they're wildly uh over earning
55:38
Unidentified speaker (speaker_0): water. wildly overinflating their
55:43
Unidentified speaker (speaker_0): revenues. That you know, here's a c eighty eighty percent, eighty-five percent gross margins, whatever the numbers are, you know, it just stands completely at odds at history. And you know, people say, well, there's no bubble because the the valuation, the PE, this, that, and everything else. The problem with and I want to make it not just about NVIDIA, but more generally, the problem is the the
56:03
Unidentified speaker (speaker_0): bubble is in the earnings and the margins, not in the share prices necessarily in some of these companies. And that um
56:10
Unidentified speaker (UNKNOWN): uh
56:10
Unidentified speaker (speaker_0): say
56:11
Unidentified speaker (UNKNOWN): that
56:11
Unidentified speaker (speaker_3): again, the bubbles in the earnings and the
56:13
Unidentified speaker (speaker_0): when you have when you have an investment boom, when you have an investment boom like we have, the profits accrue to those selling, you know, NVIDIA sales Sells chips, Bitcoin sales sells chips, that's great. So they're showing all the profits, but the expenses come later. the depreciation. And so of course earnings are accelerating as they are. But that's that happens in every
56:33
Unidentified speaker (speaker_0): investment, boom. And so um I I I would I would, you know, I bet you a beer gene five years from now, the video margins are a fraction of what they are presently. The bubble is in the profit margins, not in the revenue. And listen, the reason they keep beating, if I keep lending you the money to buy chips from me, I
56:50
Unidentified speaker (UNKNOWN): can.
56:51
Unidentified speaker (speaker_0): make up whatever number I want.
56:52
Unidentified speaker (UNKNOWN): Of
56:52
Unidentified speaker (speaker_0): course they're gonna beat. You just tell me what income what number I gotta make up. To me, I mean, have they broken any laws? No. But is gap accounting, while it's intended to try to portray the economic situation as best as possible, does it have its limitations? It certainly does. And I I think that the that that the all the all the vendor financing, the special
57:12
Unidentified speaker (speaker_0): purpose vehicles, we've seen this movie for time and time again. And uh the fact they now want to go they want to go um uh c uh securitize chips. Are you kidding me? I mean, it's this is this is you guys are gonna crack up. I'm sure you've heard this before. These these neo-cloud guys, Korweve, Nebbius, Iron, all those
57:32
Unidentified speaker (speaker_0): guys, in my opinion, in my opinion, they're all going bankrupt, okay, is is is nobody's nobody special Jack Gamblecoin. Gamble coined it. Courtweave is basically we works with GPUs.
57:43
Unidentified speaker (UNKNOWN): All
57:43
Unidentified speaker (speaker_0): right. It's a pilot. I mean, uh I was looking I I could not believe the Core Weave financial statements. Sit down and actually look at I know you you probably have Gene. I don't know, Jack, if you have. I have I was looking at the thing, I
57:53
Unidentified speaker (UNKNOWN): was
57:53
Unidentified speaker (speaker_0): like, Are you kidding me? Five billion in equity, 70 billion in debt, the losses are accelerating, you're spending more and more money, there's no operating leverage. I'm like, in what world is this possible? And as one of my good friends said to me, going back on anthropic, looking just the numbers, leave the dream aside for a second. Just looking at the numbers, it's
58:13
Unidentified speaker (speaker_0): more easy to see how that's a zero. It's a bankruptcy anthropic that's worth $2 trillion. The numbers are completely insane.
58:21
Unidentified speaker (UNKNOWN): Well
58:21
Unidentified speaker (speaker_3): Well, certainly
58:21
Unidentified speaker (UNKNOWN): the
58:21
Unidentified speaker (speaker_2): bankruptcy risk, because it's so indebted, the bankruptcy risk is much higher than every other company. we've we've talked about.
58:28
Unidentified speaker (UNKNOWN): I think
58:28
Unidentified speaker (speaker_0): Core V is a bankruptcy. We've been short since 107. And I actually think it's a better short now in the low 50s, wherever it is right now than it was at 107 because the cost of capital is going up. You look at that thing, you look at the interest expense as a percentage of the revenues. It's like you can't grow your way out of it to the contrary. The whole the the hole is getting deeper and deeper and deeper. I think
58:48
Unidentified speaker (speaker_0): Corey was is a bankruptcy, and what's important about that. the whole ecosystem with all these neo-cloud guys, one of them goes down, there's a ripple if it they're all dependent on each other. And that's why I think it's a much more tenuous situation than some might believe.
59:02
Unidentified speaker (speaker_2): George, I I'm not saying you're wrong. I just want to add the context that part of the reason their financial statements that uh you know you have poured through and have found them wanting, part of the reason they look like they do, is because they locked into these long-term contracts.
59:15
Unidentified speaker (speaker_2): with a lower rate of compute than is current. And in order to borrow so much money, they had to do that to get off-tick agreements. But if they had done what Elon Musk had done had did and basically just build the compute and then sell it in the spot market, they're they would be doing a lot better. And and obviously, we don't live in that world, but you know, computer
59:33
Unidentified speaker (UNKNOWN): prices have
59:33
Unidentified speaker (speaker_2): gone up a lot.
59:34
Unidentified speaker (speaker_0): Right. And Jack, to your point, okay, and you have spoken about this, okay, the spa spot mark market is a tiny part of the overall market. Okay. It's just like shipping. Okay. The spot market's interesting, but it doesn't tell you really where the market is. All right. And so I take exception with your statement. Um, and by the way, um, lest you think that I'm completely off base there, I'd like to point out
59:54
Unidentified speaker (speaker_0): for the record that Magnetar, one of the big private investors, has sold every stinking share they have. And the management of Core Weeb is selling stock as fast as they can. All right. So the you know, you know, the incredible thing is, the fight despite the fact the stock's fallen fifteen percent in the last few months, it's still above the IPO price. I think it's a zero. And you
1:00:14
Unidentified speaker (speaker_0): know, there's a natural there's a natural tendency. People say, Well, the stock is down. Oh, you think about covering it. No, no. It it it's the opposite. It's just like when you gee, when you have a good stock on the upside, you know, and take Netflix or wherever you want, going back a number of years, you know, the the the the incremental margins are better than you thought, the TAM is bigger than you thought, you know, the Jevons paradox, the whole thing. Just because it's up doesn't mean you
1:00:34
Unidentified speaker (speaker_0): sell it. In fact, you might want to buy it could be a it could be a bigger buy at a higher price than it was previously. It's like Peter Lynch would say, you know, you don't want you don't want to pull the flowers and the
1:00:43
Unidentified speaker (speaker_0): in in and into water the weeds, right? I'm telling you, coral Coralweed weed is going down the drain. I think coral Coraled weed is a total bankruptcy. And I believe Iron and Cypher and Nebius and all these other guys aren't too far behind. And when the important thing here is not just about core weed, is what does it mean for the whole system, for the whole ecosystem? Because this whole thing's been cobbled
1:01:03
Unidentified speaker (speaker_0): together. If that goes down, I think core weave accounts for like eight percent of NVIDIA's revenues or some number like that from memory. This It's is gonna have real ripple effects. And so uh
1:01:11
Unidentified speaker (UNKNOWN): I
1:01:12
Unidentified speaker (speaker_0): don't know, stay tuned. I mean, look G, I hope you're right. I really hope you're right. And you know what's probably gonna wind up happening? I think we're both gonna be right. That 10 years from now, this is all gonna be great, just like the internet turned out to be great. But in the meantime, had you been one of the poor, unfortunate souls that fought Nortel or Lucent or Global Crossing 99, you got wiped
1:01:31
Unidentified speaker (speaker_0): out. And so I
1:01:32
Unidentified speaker (speaker_1): would, yeah, maybe on that note.
1:01:35
Unidentified speaker (speaker_1): Yeah. I I do think this is I said, you know, who knows what's going on in the mortar. moratoriums. I think this is a buy opportunity in the market for AI.
1:01:44
Unidentified speaker (UNKNOWN): We
1:01:45
Unidentified speaker (speaker_2): got it. We got it mark timestamped evening of November 4th. So, you know, uh right after November 4th. Gene, you know, I I of course you can respond to Decorey however you want, but just to ask about the uh, you know, what what uh George calls pot Ponzi Finance I about Nvidia. So so even though I don't believe NVIDIA is actually
1:02:04
Unidentified speaker (speaker_2): lending the money, they are, you know, adding support via three three ways, uh residual value guarantees, um at guaranteeing the the lease, as well as another way, which I actually can't find in my notes. But uh uh so there's there's three ways. But it you know it in fact is is financial support
1:02:19
Unidentified speaker (UNKNOWN): as
1:02:19
Unidentified speaker (speaker_2): well as buying the stocks by the way. Like they own Core Weave, they own other stuff. Um your point that it's Nvidia's best interest to do that, I agree with, and perhaps you know George might agree with too. But in what way is that not uh kind of inflating demand, A, and B increasing the systemic risk within the system in the same way, like,
1:02:39
Unidentified speaker (speaker_2): yeah, in 2005, it was in cities' interest and Bank of America and Goldman's interest to do all the things that led to the financial crisis, of yeah, of course. Um, but that doesn't mean that it it uh obviously made the system better. So do just I understand your point. Yeah, yeah.
1:02:53
Unidentified speaker (speaker_1): The the the the there's the revenue part of this equation, these infrastructure, these clouds, they're all being built to service. Some of them will go away, but the net need for the size of the AI brain, whether it's through a neo cloud or a hyperscaler or sovereign, is gonna Going go to up dramatically, thirty, 30, fifty, 50, sixty 60% percent growth
1:03:13
Unidentified speaker (speaker_1): annually for the next couple of years. give or take a moratorium here or there. But the to answer your question, Jack, is that I believe that the reason why Nvidia stands behind this, yes, they benefit, but they see what the demand is, what the curve is coming. The reason why Zuckerberg gets on the call last week and says we have
1:03:33
Unidentified speaker (speaker_1): millions, not billions, but millions of Muse users, basically. a drop in the bucket, but our entire business next year is going to be based on Muse. I mean, that is incredible.
1:03:46
Unidentified speaker (UNKNOWN): The
1:03:47
Unidentified speaker (speaker_1): reason why Elon says we go from uh you know uh one and a half gigawatts of of uh capacity to I
1:03:54
Unidentified speaker (UNKNOWN): mean
1:03:55
Unidentified speaker (speaker_1): he said 10 to 20, but let's say they get to seven, something like that. My point is is that if the demand is there for inference
1:04:04
Unidentified speaker (speaker_1): this works. If the and I talk about these the low penetration, the low use cases and how that we're gonna have these breakouts in the next few years. If that ends up happening, the demand will be there and these businesses will be thriving. And if that doesn't happen, if the use
1:04:22
Unidentified speaker (speaker_3): cases don't present itself, we don't have a breakout, then the whole thing's gonna fall down.
1:04:29
Unidentified speaker (speaker_0): George? I I completely agree. And by the way, I just got this this afternoon. I don't know if you saw this, Gene.
1:04:34
Unidentified speaker (UNKNOWN): Uh
1:04:35
Unidentified speaker (speaker_0): and I quote two interesting data points on AI just like Muse from Earnings just reported. Uber recorded 320 million rides and Uber Eats orders during Holland Week alone. A Of vact that total, 320 three hundred and twenty million,
1:04:49
Unidentified speaker (UNKNOWN): only
1:04:49
Unidentified speaker (speaker_0): four rides and five. hundred Uber Eats orders came from agents. I think this Muse thing, this agent thing, just a commodity thing. I guess it's exciting. Gee, everyone's you know jumping through oops initially. This too shall pass.
1:05:02
Unidentified speaker (speaker_1): No way. This is
1:05:03
Unidentified speaker (speaker_0): and you G Gene, my point my my my whole point is all this is a commodity. I mean that that that's the identity. Like, you know, one company comes out with one, another company comes out with one. Go Warren Buffett, you know New England, nineteen fifty-eight, New Bedford, please call your office. All the returns are going to get arbitraged away. That's the problem. It's a commodity.
1:05:22
Unidentified speaker (speaker_1): It's a whole different way that we're gonna use technology. It'll be our starting point, and there are gonna be a lot of ways that they can make money.
1:05:30
Unidentified speaker (speaker_2): How is it not a commodity? And how is it not like shirts in um
1:05:35
Unidentified speaker (speaker_2): in i in uh you know the nineteen 1950s fifties? Uh I'd say uh one similarity to tokens and shirts is that you know the in china they they make a a a lot of them that that that probably would be where the threat is um i mean china i don't think makes that you
1:05:48
Unidentified speaker (UNKNOWN): know
1:05:48
Unidentified speaker (speaker_2): it's mostly vietnam now i think but in other countries but um you know the the token prices are going down how is it not a commodity that eventually will be served at like a 10% five 5% percent gross margin or perhaps a negative gross margin as is the case for many commodity products, whether it's copper or shipping, sugar, whatever.
1:06:08
Unidentified speaker (speaker_1): Two two things. Number one is as the pricing goes down, usage is going to go up. But also, if you look at the pricing of the frontier models, and they're so close, like it's like a 128 IQ versus 132 IQ is you know, you're paying
1:06:27
Unidentified speaker (speaker_1): 2x for that five IQ improvement kind of a thing. I'm simplifying it here. But why is that? And the the answer is that the hardest problems are getting a lot thrown at it. And I think that will always be the case. We're always going to solve the hardest problems 10 years from now. We're not going to look around and
1:06:47
Unidentified speaker (speaker_1): say, we got this is a quote from Jeff Bezos when he was talking with the Blue Origin investors. 10 years from now, we're not going to look around and say, we got plenty of intelligence. Let's just get a little bit dumber. That frontier is always going to be there. It will. It's like human nature that's going to be there. And so, yes, there's going to be commoditization is going to happen, but there's
1:07:07
Unidentified speaker (speaker_1): also going to be this gap, and there's always going to be the premium market for the the the best, even if it's fractionally better. Like if you're putting your life on the line for something and your business on the line for something, you will do whatever it takes to get that fractional advantage and you'll pay up for
1:07:23
Unidentified speaker (UNKNOWN): it. Uh
1:07:24
Unidentified speaker (speaker_2): Gene, I'm curious, what about
1:07:27
Unidentified speaker (UNKNOWN): the
1:07:27
Unidentified speaker (speaker_2): the the claim that that like most fields are actually
1:07:31
Unidentified speaker (speaker_2): number one, good enough is good enough. And number two, you literally can't get better. Like you know, the best an AI could scroll score on a test, the SAT test, um, is you know, uh 1600, right? And if if the the if they the current model, that IQ of 130 scores 1600, the model that will be in five years with a
1:07:51
Unidentified speaker (speaker_2): 200 IQ will just score 1600. I understand that there are unbounded domains like AI research and cybersecurity, but for fields um that aren't hacking, if isn't right what we have right now uh quite similar to to like good enough?
1:08:10
Unidentified speaker (speaker_1): I think 90% of what we're doing is gonna be in the segment of the inference is gonna come from the segment that is gonna have a lot of pricing pressure, but the usage is gonna go up exponentially. So I think that's an important dynamic. I mean, the the use case, like again, just the percentages of the amount of use is just still so small. First
1:08:30
Unidentified speaker (speaker_1): pitch kind of a thing. So I would agree with that. And
1:08:35
Unidentified speaker (UNKNOWN): I'm
1:08:35
Unidentified speaker (speaker_1): I'm in agreement. I'm just also saying is that that high end, that that top 10%
1:08:40
Unidentified speaker (speaker_1): might be 50% of the spend. The top 1% might be 40% of the spend. I think it's going to be exponential. Those are going to be the hardest problems to solve. And one last piece is I don't think there's a cap. There is, there's some like mind-bending kind of things when you start to morph from general intelligence to superintelligence. And like what is
1:09:00
Unidentified speaker (speaker_1): super intelligence? Like how what's the IQ of superintelligence is infinite? What does that mean? But I think that
1:09:09
Unidentified speaker (UNKNOWN): I
1:09:10
Unidentified speaker (speaker_1): mean, what was the what was what was the mockumentary with the the rock? and roll band with uh uh eleven on the the ten, the volume.
1:09:18
Unidentified speaker (UNKNOWN): Yeah,
1:09:18
Unidentified speaker (speaker_2): yeah. Spinal tap.
1:09:20
Unidentified speaker (speaker_1): Thank you, spinal tap. I think everything's gonna always hunt the eleven.
1:09:24
Unidentified speaker (speaker_2): Yeah. George?
1:09:27
Unidentified speaker (speaker_0): I I don't know. I I um
1:09:32
Unidentified speaker (speaker_0): Yeah, there will always be a demand for the cutting edge stuff. Um but that's a very tiny part of the market. All right. Uh, these models are basically all the same. What is the difference? It's a commodity. The Chinese do it for 95% less. Okay. Um
1:09:51
Unidentified speaker (speaker_0): there's gonna be no pricing power. It's a freaking commodity. Yeah, there'll be 1% of the market where they mean state of the art stuff. And then on top of it, do keep in mind again the point I raised earlier, which we're talking much about, is the enterprise is not gonna going to be the answer. The corporates don't want it. There's there's cybersecurity and compliance issues there as well. So it's a commodity. I don't see where the revenue is going to come
1:10:11
Unidentified speaker (UNKNOWN): from.
1:10:12
Unidentified speaker (speaker_0): It's a commodity, prices will fall to marginal costs. That's how commodity businesses work.
1:10:17
Unidentified speaker (speaker_2): Gene, do you think that's true? I understand you don't think that's true for the frontier, but do you think that George's claim that prices will fall to the marginal cost? Do you think that is true for generic tokens?
1:10:27
Unidentified speaker (speaker_2): non-frontier tokens?
1:10:31
Unidentified speaker (speaker_1): I think uh prices will fall.
1:10:35
Unidentified speaker (UNKNOWN): Uh,
1:10:35
Unidentified speaker (speaker_1): but I I'm I'm repeating myself here too, but I think that the there's this usage part that we're not talking about. That's relevant to the conversation. It's not just about what the cost of a token is.
1:10:48
Unidentified speaker (speaker_1): If
1:10:50
Unidentified speaker (UNKNOWN): the use
1:10:50
Unidentified speaker (speaker_1): of it
1:10:52
Unidentified speaker (speaker_1): is goes up, I'll give you one like use case example. We talked about these bots, a typical of the uh personalized agents that were like Muse and GrockBot.
1:11:02
Unidentified speaker (UNKNOWN): If
1:11:02
Unidentified speaker (speaker_1): it's 10 to 15x, the tokens,
1:11:06
Unidentified speaker (UNKNOWN): and
1:11:06
Unidentified speaker (speaker_1): that ends up being the new way that you know billions of people use every day, there's just going to be a lot, even if some of that stuff is marginalized on the pricing side.
1:11:15
Unidentified speaker (speaker_1): So
1:11:17
Unidentified speaker (speaker_1): I'm of you, yes, we're going to get pricing pressure, but I think that uh the usage is going to more than offset it and keep these labs in a good position.
1:11:26
Unidentified speaker (speaker_0): And I guess I respond by saying the very definition of a commodity market, okay, um,
1:11:32
Unidentified speaker (speaker_0): is when we don't have pricing power where prices do fall to the marginal cost of production. Okay. You're talking about you know, a $40 trillion TAM
1:11:42
Unidentified speaker (speaker_0): you know and $100 to $120 trillion global GDP. I mean, how much does global GDP have to grow in order to make that work? It's just the numbers don't a commodity market by definition, by definition, prices fall to marginal costs. And I haven't heard you other than the one 1%, percent which is specialized, okay. I haven't heard you say anything which
1:12:01
Unidentified speaker (UNKNOWN): I think
1:12:01
Unidentified speaker (speaker_1): prices are going they they have. been going down. There's no question about it. No,
1:12:04
Unidentified speaker (speaker_0): but they're gonna go down a lot. No, but by definition, they're gonna keep going down. So it's that any excess return is arbitraged away. That's our that's the moral obligation of capitalism, that's how commodity markets work.
1:12:14
Unidentified speaker (speaker_2): George, I might say if if I Gene, and tell me if you I've you know speaking improperly for you, that Gene might say that like it's it you're right, AI is copper now and it it's 2000, two thousand and the China demand is just going to overwhelm it. So, yes, of course. over time that your claim is true, George, but that like we're just in a, you know, w it's early. Gene, did I is that fair?
1:12:35
Unidentified speaker (speaker_2): Yes. Okay. What's
1:12:36
Unidentified speaker (UNKNOWN): it, George?
1:12:37
Unidentified speaker (speaker_2): Okay.
1:12:39
Unidentified speaker (speaker_0): Yeah, but that's not quite right because in the early two thousands, copper was going up in price because pr supplies is in markets, we as we always know, is relatively price analysis and uh uh in elastic in the short run. But that's not the case right now. Prices are collapsing. I mean it it's so it's it's it's
1:12:57
Unidentified speaker (UNKNOWN): and
1:12:57
Unidentified speaker (speaker_1): demand prices are
1:12:59
Unidentified speaker (speaker_1): demanding faster.
1:13:01
Unidentified speaker (speaker_0): Excuse me?
1:13:02
Unidentified speaker (speaker_1): Prices are collapsing and demand is going up faster than prices are collapsing.
1:13:08
Unidentified speaker (speaker_0): They're not going up fast enough to make to justify you know a two $2 trillion dollar valuation for OpenAI or anthropic. I mean, we're debating here is you know how to invest. Okay. And you know, we can all go on about, you know, your estimate of 2029 market and mine and this and that and everything else. A, it's fanciful, nobody really knows. I
1:13:28
Unidentified speaker (speaker_0): highlighted that with past predictions. Um, you know, in and then B, even if you're right, uh, you know, Mr. Market, depending on what type of convulsions he has right now.
1:13:39
Unidentified speaker (speaker_0): he turns off the tap, um, and it's game over because they can't, they don't have the capital that they'll be necessary to grow this thing the way that the bulls think will be the case. All right. Um, you know, it's like Cisco. Yeah, the market was right on Cisco, it just was off by 25 years.
1:13:57
Unidentified speaker (speaker_2): Yeah. Uh Gene, what do you make of the optical?
1:14:03
Unidentified speaker (speaker_2): revenues for I guess from ramp showing that open an anthropic revenue is not growing that much now? I know there's a not a lot of nuances with you know distillatum being canceled and then the increasing token things sold in the cloud, which is not on ramp, to my knowledge. Um, but just you know, people looking at that chart and saying, okay, yeah, that
1:14:23
Unidentified speaker (speaker_2): validates the huge bull move we had earlier this year when revenue just exploded. But from Q3 and Q3, and as we're in Q4, it's not looking so hot. What what's your reaction?
1:14:33
Unidentified speaker (speaker_1): We're still having these step functions that if we go back to almost like the day a year ago, this is just before Claude uh Claw came out, if I get that one right.
1:14:46
Unidentified speaker (UNKNOWN): Like
1:14:46
Unidentified speaker (speaker_1): we would have never like enthropic. I was negative on enthropic. Like the they're losing the consumer business take longer for enterprise to come through. And they pulled clodcoat out of the out of out of out of thin air. And that's what happens when these markets are establishing themselves is you have these breakthroughs. And uh a lot
1:15:06
Unidentified speaker (speaker_1): of it, I mean there's analogies to autonomy too here around March of Nines and the March of Nines around hallucinations and enterprises willing to embrace doing this tech. But when you have these improvements, and some of them can be marginal improvements that create a step function increase in demand. And so this is not a linear
1:15:26
Unidentified speaker (speaker_1): equation that's going on. These these growth rates are dynamic right now.
1:15:31
Unidentified speaker (speaker_0): They they they are dynamic indeed in both ways. And I'll be the first one. Gene check back in six months and you know, yeah.
1:15:39
Unidentified speaker (speaker_1): I'll do the same too. Yeah,
1:15:40
Unidentified speaker (speaker_0): yeah.
1:15:41
Unidentified speaker (speaker_0): It's very dynamic. It just goes to underscore,
1:15:43
Unidentified speaker (UNKNOWN): you
1:15:44
Unidentified speaker (speaker_0): know, yogi bear again. Predictions are difficult, especially about the future. You know, when we're a roaring bull market and yields are crashing and you know the the bartenders are up the drinks, you know, pigs fly. And and then when then when they when they have last call, but I think we're we at last call, but when they say we're gonna be uh be closing the bar down that at some point, uh, not so much.
1:16:03
Unidentified speaker (speaker_0): So, um
1:16:05
Unidentified speaker (UNKNOWN): I
1:16:05
Unidentified speaker (speaker_0): I don't know. I mean with Willis, listen, Gene, that's what makes the market. That's what makes a market.
1:16:09
Unidentified speaker (UNKNOWN): I
1:16:09
Unidentified speaker (speaker_1): appreciate
1:16:09
Unidentified speaker (UNKNOWN): that,
1:16:09
Unidentified speaker (speaker_1): George. It's been a lot of fun. Yeah.
1:16:11
Unidentified speaker (speaker_0): This is me right.
1:16:12
Unidentified speaker (UNKNOWN): Hey
1:16:12
Unidentified speaker (speaker_2): George, George, let me ask you a question before before we go. You've obviously been seeing the ball quite clearly. It's no it's no secret that, you know, the market's been moving in your direction. Um, what do you think the odds are that, you know, yeah, we're in month three. of the the crash and and June was the absolute top?
1:16:33
Unidentified speaker (UNKNOWN): We've
1:16:34
Unidentified speaker (speaker_0): got to be careful here because I don't like to engage in market predictions and out of that business. It's a market of stocks. Um, as Peter would always say, there, you know, we've we've been doing really well with energy stocks. Uh our gold stocks haven't been doing so well lately, but I'm optimistic that the turn is at hand and our tech shorts have been killing it.
1:16:54
Unidentified speaker (speaker_0): So, what the market itself does, I mean, you know, Jack, I think you pointed out to me, others have said it. You got 30% SP trading with a neck of the beta right now. Okay. So you put that in the blender with the tech stocks crashing and some of the defenses going up. And what does that mean for the index? I don't know. Now it's not lost upon me that I think upwards of
1:17:14
Unidentified speaker (speaker_0): 60% of the index is trading in sympathy with the tech trade. It's 40% in tech, and you know, Cata probably we're now told the tech stock, it trades like what, right? So what it means for the market, I don't know. So my real call has been less about what the market's gonna do and more about the dispersion. It's stock pickers nirvana. Even for Gene, you know,
1:17:34
Unidentified speaker (speaker_0): they're probably even within his, even if he begrudgingly begrudgedly says, okay, right here right now. George is right, but this is going to change. I'm sure he's got some things he really likes or doing well. He's got some things doesn't like. Okay.
1:17:45
Unidentified speaker (speaker_2): Uh Google, Microsoft.
1:17:47
Unidentified speaker (speaker_0): I got a sneaky suspicion, Jack. If I
1:17:50
Unidentified speaker (speaker_0): if I say to you, hey Gene, you know what? Uh, I want to go long Google and short uh Corweeb. How
1:17:56
Unidentified speaker (UNKNOWN): about
1:17:56
Unidentified speaker (speaker_0): you take the other side of that trade? I got a sneaky suspicion, you're not gonna take that trade, okay? So you have your own ideas of what you like and what you. don't like. Just going back to what's the stock pickers market. I is I think it's nirvana. Any fool can make money in a rip-roaring bull market just by momentum and narratives. But now we separate the men from the boys. Valuation counts, fundamentals
1:18:16
Unidentified speaker (speaker_0): count. And so, but I would say, I would say the proponents of the, if you ask me with a risk, like the most likely 10% on the index, probably down, but that's not why I'm investing. I'm I'm buying. things I like, energy, gold, etc. And I'm short tech and consumer discretionary, just to kind of embro brush in broad brush strokes. Gene,
1:18:35
Unidentified speaker (speaker_2): I'll bring bring you in in a moment and I and I will uh land this play and tease George's just phrase. But George, uh yeah, sorry, I understand you thought I meant SP. I I meant like the the AI tech trade. So whether it's the NASDAQ or the semis or a basket of semis and G E Vernova, like what do
1:18:50
Unidentified speaker (UNKNOWN): you
1:18:50
Unidentified speaker (speaker_2): think the odds are? that June was the hot in in that battle?
1:18:53
Unidentified speaker (speaker_0): Oh, I I'm very convicted on that. Very convicted on that. I mean, I didn't even get a chance to look at it. I see that's G, you
1:18:59
Unidentified speaker (UNKNOWN): know,
1:18:59
Unidentified speaker (speaker_0): G Vernova's a good example. Okay. You know, good company, right place, right time, you know, but at a certain price, at a certain time it's all on the price. I mean, Jack, something we can agree on, like shipping stocks where, you know, but it can't get any better. You're take refining stocks right now.
1:19:13
Unidentified speaker (UNKNOWN): Yeah.
1:19:14
Unidentified speaker (speaker_0): You
1:19:14
Unidentified speaker (UNKNOWN): know.
1:19:15
Unidentified speaker (speaker_0): Yeah, I agree the outlook for refining, there's not much compared. I get all that, okay? But if the trade in in Gene, maybe the read across is easier for you to see.
1:19:24
Unidentified speaker (UNKNOWN): You know,
1:19:25
Unidentified speaker (speaker_0): as we had a wonderful conference last week, and a couple of real sharp-eyed uh analysts, Jonathan Krinsky, BTI G and Christopher Adron Stategis, they both called out the refining stocks as are shorts. And they said, wait a second, if their firing trade is so great, why are crack spreads making a three-month low? Uh uh uh uh okay, and the answer, I think I found the
1:19:45
Unidentified speaker (speaker_0): answer yesterday, speaking of sharp energy analysis, he pointed out to me that I think there are a few uh new refineries coming online, stuff that was commissioned at the time of COVID, some Indian refineries. It's not a lot, but it's just enough that and in a tight market, it can upset the supply-demand balance. Anyway, the point of it is crack spreads are. of three-month lows. Even though the secular outlet
1:20:05
Unidentified speaker (speaker_0): for refiners are wonderful, right? And the stock selling went straight up and it don't be sandwich.
1:20:11
Unidentified speaker (speaker_0): In giving the elevator pitch, why you should short refiners? My way, I don't have a trade-on-I got other things that rather be short. Uh they said the refiners today are exactly where semiconductors were three months ago. I thought that was perfect, absolutely perfect. I look
1:20:22
Unidentified speaker (UNKNOWN): at
1:20:22
Unidentified speaker (speaker_0): sand Sandis. disc, I don't know if you cover it or not. Okay, micron, sandic. Sandis, get out of here. Okay. That is a complete joke. As Louis Gobb says, you know, when the Chinese enter the room, profits walk out the door. Okay. And you you you look at you look at the capacity ads that are coming, you know, and whether they don't come until
1:20:42
Unidentified speaker (speaker_0): 28 or 29, it doesn't matter. The markets are forward looking. And I found historically, I'm doing this 45 years. Whenever you get in a situation, I know you
1:20:50
Unidentified speaker (UNKNOWN): know
1:20:50
Unidentified speaker (speaker_0): this, Jane. Gene, and maybe you disagree, where it's not a question of my estimates are higher than your estimates, but instead we're arguing about the duration of the of the earnings, the surplus earnings. Like you own a tanker stock and keep it third person. It's easier. And Jack's in and I say, Hey Jack, I got this tanker thing. It's making, you know, excessively obscene windfall profits.
1:21:08
Unidentified speaker (UNKNOWN): And
1:21:09
Unidentified speaker (speaker_0): it's gonna go in Jack, you want wanna to know something? The shortage is gonna go on for three years, not two years. So I'm actually bullish because. consensus doesn't understand that. Meanwhile, my estimates are no higher than Jack's. That trade never works. Never. Okay. And that's where the refiners are right now. And that's in my humble opinion, come back to your question, Jack, that's where the semiconductor complex is
1:21:29
Unidentified speaker (speaker_0): and all the other related AI trades.
1:21:32
Unidentified speaker (speaker_2): Gene, why do you you said you uh did you say you you don't like Google and Microsoft now? Is that true?
1:21:38
Unidentified speaker (UNKNOWN): Uh
1:21:38
Unidentified speaker (speaker_1): that's true, yeah.
1:21:39
Unidentified speaker (speaker_2): Why, why?
1:21:41
Unidentified speaker (speaker_1): Uh Google has uh challenge around getting into the bot game, the uh personalized agent game. They did a great job of competing in on the generative side by basically taking their search traffic and moving over to AI overviews, then Than AI mode. Modestation is a little bit lighter right now, but that eventually is gonna going to wrap. That
1:22:01
Unidentified speaker (speaker_1): always happens. It's gonna going wrap to so the they'll
1:22:03
Unidentified speaker (UNKNOWN): be
1:22:03
Unidentified speaker (speaker_3): fine there. But ultimately, if
1:22:05
Unidentified speaker (UNKNOWN): if
1:22:05
Unidentified speaker (speaker_3): I'm right that these bots are the bots, these agents are the starting point of our lives every day and throughout the day, then it it just basically uh displaces Google.
1:22:16
Unidentified speaker (speaker_1): Now they may find a way to
1:22:18
Unidentified speaker (UNKNOWN): have
1:22:18
Unidentified speaker (speaker_1): another moment
1:22:19
Unidentified speaker (speaker_3): like they had
1:22:20
Unidentified speaker (UNKNOWN): a
1:22:20
Unidentified speaker (speaker_1): uh in spring of last year where they're able
1:22:23
Unidentified speaker (UNKNOWN): to take that traffic
1:22:23
Unidentified speaker (speaker_1): and somehow move it into
1:22:25
Unidentified speaker (UNKNOWN): the
1:22:25
Unidentified speaker (speaker_1): personalized agent. world.
1:22:27
Unidentified speaker (UNKNOWN): Uh
1:22:27
Unidentified speaker (speaker_1): but they've got to do that. And I think that's a hard, that's a hard turn. And Microsoft, um, I've been dead wrong on this. I thought a year ago we'd start to see some cracks now in terms of the growth. I think they guided the December quarter like up 19% or 18%. And like Microsoft grows at 12% every quarter for
1:22:47
Unidentified speaker (speaker_1): as long as I can remember.
1:22:50
Unidentified speaker (UNKNOWN): Uh
1:22:50
Unidentified speaker (speaker_1): So so something good's happening there, but ultimately I think that knowledge work is going to be more impacted than people realize. And that impacts their seat growth.
1:22:59
Unidentified speaker (speaker_2): So are you a are you a SaaS bear as well?
1:23:03
Unidentified speaker (speaker_1): Not uh yeah, like Salesforce, Adobe, bearish on those, uh, but uh positive on like CrowdStrike and we own some of them in our ETF.
1:23:13
Unidentified speaker (speaker_2): Got it. Um could we get a closing statement? Uh George starting with you and then Gene.
1:23:20
Unidentified speaker (speaker_0): Jack and Gene, you know, like there will be pain, there will be blood, there will be volatility, right? And I think I don't mean to be mealy mouthed about this any on one hand, on the other hand, seriously, there's gonna be a lot of volatility. There's a lot of dispersion. There's a lot to do here. Okay. This is a stock picker's paradise. I think Gene, even though you
1:23:40
Unidentified speaker (speaker_0): know some of the things aren't acting so great, this is a chance to really get an opportunity to show your metal that you can differentiate, you can add value. You know, just betting on a momentum trade and narrative and Fed liquidity, that's that's not the best thing, right? So I'm relishing this market because it really rewards analysis. And as a time, tie,
1:24:00
Unidentified speaker (speaker_0): you know, we've been in this make make-believe, believe. you know, weird, dystopian type of market. And whether you want a data from the GFC in 2008 or COVID, where it's always like, you know, it's liquidity, liquidity, liquidity. As they say on X, more and more and more M O A R, M O A R, M O A R.
1:24:17
Unidentified speaker (UNKNOWN): You
1:24:18
Unidentified speaker (speaker_0): know, the Fed's not coming to rescue
1:24:20
Unidentified speaker (UNKNOWN): you. All
1:24:21
Unidentified speaker (speaker_0): right. And so you gotta got to do your work. And um I think, you know, there's a lot to do here. I think, you know, Gene's had a lot of great ideas. I'm sure he'll continue to have them. We just disagree on the AI trade. Um, and I again I think it's a situation where we both can be right, you know, he could be totally right that, you know, four years from now the the revenue number is some ungodly,
1:24:40
Unidentified speaker (speaker_0): whatever. But, you know, it it still didn't mean that um Sandis doesn't Don't go down eighty 80% percent from here.
1:24:47
Unidentified speaker (UNKNOWN): It
1:24:47
Unidentified speaker (speaker_0): still doesn't mean that Korea Corviv doesn't go bankrupt. Both can be true. That's what I would say.
1:24:52
Unidentified speaker (speaker_2): Gene.
1:24:55
Unidentified speaker (speaker_1): I think that
1:24:57
Unidentified speaker (speaker_1): the usage is so small today
1:25:01
Unidentified speaker (speaker_1): that it will create enough demand that will restart the AI trade.
1:25:07
Unidentified speaker (UNKNOWN): I
1:25:07
Unidentified speaker (speaker_1): think we're gonna have a breakout moment, breakout moments, multiple ones,
1:25:13
Unidentified speaker (UNKNOWN): uh,
1:25:14
Unidentified speaker (speaker_1): whether it's orbital data center testing, whether it's autonomy, whether it's enterprises changing their workforce meaningfully in the next year that's going to increase investor optimism that this is going to go on for way longer than people think.
1:25:30
Unidentified speaker (UNKNOWN): And
1:25:31
Unidentified speaker (speaker_1): I would just leave people with an exercise, which is
1:25:35
Unidentified speaker (speaker_1): um
1:25:38
Unidentified speaker (UNKNOWN): go
1:25:38
Unidentified speaker (speaker_1): into and test drive FSD and and download Muse or Grockbot
1:25:45
Unidentified speaker (speaker_1): and ask yourself,
1:25:48
Unidentified speaker (UNKNOWN): is
1:25:48
Unidentified speaker (speaker_1): this 10x better than anything that I've seen? And if the answer is it is, then I think you're probably in my camp. The answer is no, I'm this is something I just don't really care about,
1:26:00
Unidentified speaker (UNKNOWN): then
1:26:00
Unidentified speaker (speaker_1): you're probably in George's camp.
1:26:02
Unidentified speaker (speaker_0): Gene, when I turn bullish, I'm calling you first to get to get your picks.
1:26:06
Unidentified speaker (speaker_1): All right. Well, I really enjoy it, George. You're a legend and a class act. And Jack, you are too. Thank you for allowing me. I feel like I interrupted you there, Jack. My apologies. You're a great host and it's been a lot of fun.
1:26:17
Unidentified speaker (UNKNOWN): Thank
1:26:17
Unidentified speaker (speaker_2): you. Two great guests. Hey, George, tell us where people can find you and your sub stack. And Gene, tell people where can people can find you as a management as well as research. On
1:26:26
Unidentified speaker (speaker_0): Noble. Uh Noble update uh on Substack uh GNOBLE79 on X. Also on YouTube.
1:26:33
Unidentified speaker (UNKNOWN): Uh
1:26:34
Unidentified speaker (speaker_1): just find me on Twitter. I just like calling it Twitter. Monster underscore Gene.
1:26:39
Unidentified speaker (UNKNOWN): Thanks,
1:26:39
Unidentified speaker (speaker_2): guys. Thank you everyone for watching.
This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.

More from The Noble Update Podcast

The Noble Update Podcast

NVIDIA's biggest bulls just showed you how scared they are

Episode discussion: Speakers predict NVIDIA revenue collapse and criticize investor letter on China exports

Speakers’ claims: Speaker 1 opens the episode by predicting a significant collapse in NVIDIA's stock price, forecasting a 30-40% revenue drop attributed to lost Chinese business and AI sector bankruptcies. The speaker then critiques an open letter from Gavin Baker and Brad Gerstner, arguing that their request to keep options open, including Jensen Huang's return, prioritizes their financial positions over national security concerns regarding chip exports to China. The episode focuses on bearish views regarding NVIDIA's future performance and the ethical implications of investor advocacy on export policies.

NVIDIAAI SectorChina Exports
3 min1 assets7 claims to verify
The Noble Update Podcast

Emergency break-in episode: Nvidia Is Enron 2.0

Episode discussion: Speakers allege Nvidia fraud, predict AI sector collapse, and critique AI IPOs as exit liquidity for VCs

Speakers’ claims: Speakers claim Nvidia engaged in massive fraud involving 'dark GPUs' and illegal chip smuggling to China, predicting a collapse of the AI infrastructure sector and an 80% drop in semiconductor stocks. They criticize non-GAAP metrics like adjusted EBITDA used by companies such as CoreWeave and Anthropic, arguing these mask true cash burn. The episode also discusses the timing of the Anthropic IPO as a potential market turning point for VC exit liquidity, while advising listeners to avoid tech index funds and consider shorting overvalued consumer stocks.

NvidiaAI InfrastructureIPOs
1h 4m2 assets22 claims to verify
All insights