WSJ's Take On the Week

Anthropic IPO and AI Market Dynamics: Supply, Demand, and Valuation

35 min Listen to the original Episode page
AI InfrastructureAnthropic IPONVIDIACoreWeaveMarket ConcentrationEnterprise AIJevons ParadoxPrivate Equity

Summary

Hosts and guest Matt Witheiler discuss the resignation of NVIDIA's CEO, noting the resulting market concentration risks and the stabilization of the AI debt ecosystem following CoreWeave's earnings. Witheiler highlights Anthropic's upcoming IPO as a critical 'pure play' for assessing AI demand, citing 10x revenue growth and positive operational cash flow despite accounting losses on convertible notes. The discussion contrasts Anthropic's enterprise-focused strategy with broader tech giants, emphasizing the Jevons paradox in model consumption and the long-term validity of the AI investment thesis amid short-term macro volatility.

Episode sections

  1. 0:00 – 0:33

    Sponsor Read: Electronic Payments Coalition

    Paid advertisement for the Electronic Payments Coalition regarding credit card mandates.

  2. 0:33 – 4:03

    NVIDIA Leadership Shakeup and Market Impact

    Discussion on Jensen Huang's resignation, NVIDIA's loss of top market cap to Apple, and management structure risks.

  3. 4:03 – 7:11

    CoreWeave Earnings and AI Debt Ecosystem

    Analysis of CoreWeave's earnings, secured financing, and stabilization of the AI debt sector.

  4. 7:11 – 7:20

    Anthropic IPO and Market Structure

    Overview of Anthropic's IPO as a pure-play AI model provider and its role in assessing demand-side dynamics.

  5. 7:20 – 10:45

    Inflation and Macro Context

    Discussion on hot CPI data, 6% inflation, and expectations for a December Fed rate hike.

  6. 10:45 – 13:21

    Guest Introduction: Matt Witheiler

    Introduction of Matt Witheiler from Wellington Management to discuss Anthropic's private market performance.

  7. 13:21 – 18:27

    AI Supply vs. Demand Dynamics

    Witheiler argues demand-side signals from model providers show no weakness despite supply-side focus.

  8. 18:27 – 28:11

    Anthropic's Financials and Profitability

    Analysis of Anthropic's 10x revenue growth and accounting-driven net income losses in the S-1 filing.

  9. 28:11 – 30:12

    Revenue Concentration and Jevons Paradox

    Speaker 3 notes 80% of revenue from recent models and applies the Jevons paradox to AI consumption.

  10. 30:12 – 31:35

    Market Positioning and Portfolio Role

    Discussion on Anthropic's unique value as a pure-play AI company distinct from infrastructure providers.

  11. 31:35 – 33:14

    Enterprise vs. Consumer AI Strategy

    Clarification that Anthropic focuses 100% on enterprise AI, excluding the consumer segment.

  12. 33:14 – 35:13

    Long-Term Valuation Horizon

    Argument that long-term valuation should rely on perpetual discounted cash flows despite short-term sentiment volatility.

Assets discussed

AssetKindStanceContext
NVIDIANVDAequityneutralIndeed, indeed. We first we of course had Jensen Huang, the CEO of NVIDIA, resigning incredibly, due to extramarital affairs, including one with a client who may have indeed been a spy for China. Crazy, crazy news.
AppleAAPLequityneutralNvidia is not just one of the most valuable, not actually now the most valuable stock in the market anymore. Apple has overtaken it, but but was for a long time the most valuable stock in the market, still around a four trillion dollar valuation.
SpaceXprivate companyneutralexemplified by SpaceX recently, where you had a company that went from not available in the public market to being valued at north of a trillion and a half dollars in public market cap um and living its entire journey until that point in the private markets.
Anthropicprivate companyneutralSo as we contemplate watching Anthropic make that transition from very highly valued, already crucially important to our economy and the market's private company into a public company, it's happening at a time when the markets have been going
OpenAIprivate companyneutralhey, these tokens will be consumed or not. Instead, the public market is living month to month trying to guess what anthropic's ARR or OpenAI's ARR RR or one of the c token consumers uh demand for tokens are. And so I think once we start getting some of these businesses out, the market will better understand
AlphabetGOOGLequityneutralestablished ad business, but also have the AI potential demand upside. And I think about the way I use, you know, AI, which is not in a very sophisticated way, but I kind of think of it as a glorified search engine. That's sort of how a lot of consumers use it, I think. And so where do I go for search? Well, I go to Google, which is where everybody goes. So in that sense, maybe Alphabet wins. What do you say to that?

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

Claims worth verifying

Jensen Huang, CEO of NVIDIA, resigned due to extramarital affairs, including one with a client who may have been a spy for China.

Unidentified speaker (UNKNOWN) ·0:38

How to verify: Check NVIDIA's official press releases or SEC filings (Form 8-K) for the resignation announcement and any subsequent investigations or statements regarding the reasons. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Apple has overtaken NVIDIA as the most valuable stock in the market, with NVIDIA around a four trillion dollar valuation.

Unidentified speaker (speaker_1) ·1:23

How to verify: Check real-time market data from financial news sources (e.g., Bloomberg, Reuters) or SEC filings for market capitalization rankings as of the release date (2026-11-15). For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

NVIDIA stock was down seven percent mid-morning on the day of the news and fifteen percent for the month.

Unidentified speaker (speaker_1) ·2:17

How to verify: Check NVIDIA (NVDA) stock price history on financial data platforms (e.g., Yahoo Finance, Bloomberg) for the trading days surrounding the resignation announcement. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

CoreWeave reported better-than-expected earnings.

Unidentified speaker (speaker_2) ·4:03

How to verify: Review CoreWeave's earnings press release and SEC filings (if public) or reputable financial news reports covering the earnings call. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

CoreWeave is junk rated and has unsecured bonds with yields up to 15% in recent weeks.

Unidentified speaker (UNKNOWN) ·4:20

How to verify: Check credit rating agency reports (e.g., S&P, Moody's) for CoreWeave's rating and bond market data for yield levels of its unsecured debt. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

The top 10 names in the S&P 500 accounted for 40% of the index's value, with NVIDIA alone at 7%.

Unidentified speaker (speaker_2) ·3:12

How to verify: Calculate the weightings from the S&P 500 index constituents as of the date of the podcast or the referenced article by Miriam Gottfried. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Anthropic has grown revenue 10x year-over-year for three years.

Unidentified speaker (UNKNOWN) ·18:42

How to verify: Review Anthropic's S-1 filing with the SEC for revenue growth figures and historical financial data. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Anthropic has over a hundred billion dollars on its balance sheet and may raise up to another hundred billion in the IPO.

Unidentified speaker (speaker_3) ·21:53

How to verify: Check Anthropic's S-1 filing for balance sheet cash positions and the prospectus for the IPO offering size. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Anthropic has forward commitments to buy approximately eight hundred billion dollars of compute over the next decade.

Unidentified speaker (speaker_3) ·22:13

How to verify: Review Anthropic's S-1 filing for 'Risk Factors' and 'Management’s Discussion and Analysis' sections detailing long-term purchase commitments. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

80% of Anthropic's revenue comes from models that are six months old or newer.

Unidentified speaker (UNKNOWN) ·28:11

How to verify: Check Anthropic's S-1 filing for revenue breakdown by model vintage or product line; verify via SEC EDGAR. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Anthropic is the first pure-play AI focused company available in the public markets.

Unidentified speaker (speaker_3) ·30:32

How to verify: Review Anthropic's S-1 filing for business description and compare against other recent IPOs or public filings of AI companies. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Anthropic's mission is 100% focused on enterprise AI.

Unidentified speaker (speaker_3) ·31:55

How to verify: Check Anthropic's S-1 filing under 'Risk Factors' or 'Business' sections for stated focus areas; verify via SEC EDGAR. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Consumer AI was not mentioned in Anthropic's S-1 filing.

Unidentified speaker (speaker_3) ·32:32

How to verify: Search Anthropic's S-1 filing for the term 'consumer'; verify via SEC EDGAR. For recent events, verify 2026-08-15 through 2026-11-15. Use another period only when the speaker explicitly names it.

Primary source

Do your own research

  • Verify NVIDIA's corporate governance structure and board composition via the latest DEF 14A filing to assess key man risk.
  • Review CoreWeave's S-1 or earnings release to confirm the ratio of secured vs. unsecured debt and the terms of compute agreements with Meta.
  • Analyze Anthropic's S-1 filing for detailed revenue breakdowns by model vintage and customer concentration.
  • Check the latest CPI report and Fed meeting minutes to assess the probability of a December rate hike and its impact on AI valuations.
  • Compare Anthropic's enterprise-only strategy against Alphabet's or OpenAI's public disclosures regarding consumer AI initiatives.
  • Review the accounting treatment of convertible notes in Anthropic's S-1 to understand the impact on reported net income.
  • Assess the concentration risk in the S&P 500 by reviewing the latest index methodology and holdings data.
  • Investigate the historical performance of 'pure play' AI stocks versus diversified tech giants during previous market corrections.
Full transcriptexpand
0:00
Unidentified speaker (speaker_0): Even before a new Congress steps foot in Washington, corporate lobbyists want lawmakers to pass the Durban Marshall credit card mandates. This radical bill reduces access to credit, jeopardizes your rewards, and raises fraud risk, all to pad megastore profits. Community banks, credit unions, military lenders, airline workers, Republicans, and
0:19
Unidentified speaker (speaker_0): Democrats all say no. Know. Make your voice heard. Tell Congress, hands off our cards. Visit Guard Your Card to take action. Paid for by the Electronic Payments Coalition.
0:33
Unidentified speaker (UNKNOWN): Hi,
0:33
Unidentified speaker (speaker_1): Miriam. Hi,
0:34
Unidentified speaker (speaker_2): Tellus. So it's been a totally crazy week in the world of AI. We
0:38
Unidentified speaker (UNKNOWN): had
0:38
Unidentified speaker (speaker_1): a century's worth of news.
0:39
Unidentified speaker (speaker_2): Indeed, indeed. We first we of course had Jensen Huang, the CEO of NVIDIA, resigning incredibly, due to extramarital affairs, including one with a client who may have indeed been a spy for China. Crazy, crazy news. But then we had some positive news from Corweave,
0:59
Unidentified speaker (speaker_2): which we've talked about on this podcast, that brought a lot of AIS stocks up along with it. And then against this volatility and this crazy backdrop, we have anthropics Anthropic still planning to go public uh this coming week. So I think we have a lot to talk about. We're gonna dig more into the Anthropic IPO. That's the subject of our big conversation
1:19
Unidentified speaker (speaker_2): today. But let's turn to NVIDIA. This is a crazy story.
1:23
Unidentified speaker (speaker_1): Nvidia is not just one of the most valuable, not actually now the most valuable stock in the market anymore. Apple has overtaken it, but but was for a long time the most valuable stock in the market, still around a four trillion dollar valuation. And so uh the company and its founder. CEO Jensen Wong have been uh it's it's hard to
1:43
Unidentified speaker (speaker_1): overstate how kind of central they've been to this whole story. And for NVIDIA in particular, Jensen Wong's resignation is really impactful. I mean look, any company that has a founder, CEO, kind of charismatic type is going to feel it. But um NVIDIA was a unique company. It did not have a chairman of the board. Uh
2:03
Unidentified speaker (speaker_1): the management structure is very flat, dozens of people reporting directly to Jensen, and uh thus far not a super clear um succession plan. I mean the
2:14
Unidentified speaker (speaker_2): He apparently didn't believe in succession planning. That's what he said.
2:17
Unidentified speaker (speaker_1): Well, uh right now the CFO has been named the interim CEO, but uh NVIDIA stock was down seven percent uh kind of mid morning as we're talking the day of the news. It was already down fifteen percent this month. So big impact on the the stock that has defined this era.
2:37
Unidentified speaker (speaker_2): I mean, I think he has been such a figurehead for this company. He is truly like its salesman in chief. He's done deals directly with the US government, he's developed this relationship with President Trump. Yeah, And
2:50
Unidentified speaker (UNKNOWN): it's better. for or for worse.
2:52
Unidentified speaker (speaker_2): But I mean, these things can't easily be replaced. And I think, you know, it redefines, you know, key man risk, as as they say. And I think it also really highlights something that I've been writing about, which is concentration risk. Investors, as we've talked about, are so, you know, concentrated in the names
3:12
Unidentified speaker (speaker_2): of these big AI companies. A lot of people have made a ton of money off of owning NVIDIA over the past decade. And as I wrote in my story, if you just own the S P 500, the top 10 names accounted for 40% of the value. And NVIDIA alone was 7%. So I mean that's huge
3:32
Unidentified speaker (speaker_2): concentration risk um just in what is supposed to be the broad index. Yeah,
3:37
Unidentified speaker (speaker_1): even if you're not like a, you know, super rich, you know, kind of deeply invested in the markets person, if you just looked at your retirement, you'd probably be, if if you haven't been following, like pretty surprised by how much of it is staked on NVIDIA. That said, that said, I don't want to overstate that this news is necessarily going to
3:56
Unidentified speaker (speaker_1): totally rock and derail the market and the AI trade. That's not what we're seeing in the markets. market so far.
4:03
Unidentified speaker (speaker_2): That's true. And that's partially because there was some good news that emerged this past week for the AI trade. Core weave actually reported better than expected earnings and its stock did really well. And that took a lot of the market with it. And now remember, core there were a lot of questions hanging over core weave
4:20
Unidentified speaker (UNKNOWN): and
4:20
Unidentified speaker (speaker_2): its health because it's seen as being like maybe one of the more tenuous players in the AI ecosystem. It has a lot of debt. and it's junk rated, um, and it relies on a lot of others, you know, for its business to work out, including NVIDIA, which owns a big stake in it.
4:36
Unidentified speaker (speaker_1): Right.
4:37
Unidentified speaker (speaker_2): Um, but Core Weave was sort of able to rise above the fray and dispel some of the deeper concerns about things breaking down in that trade.
4:45
Unidentified speaker (speaker_1): Yeah, its earnings report and and and subsequent analyst call accomplished a couple of things. One is that it showed the business. that was on track. They continued to add, you know, more compute power to the business and expect to add more. Um their kind of profit margins, the way they measure them, were kind of hitting some milestones even faster than they had originally
5:05
Unidentified speaker (speaker_1): been anticipating. So
5:07
Unidentified speaker (speaker_1): it helped resolve the question that I think, and again, we've we've talked a lot on the show about
5:12
Unidentified speaker (UNKNOWN): at
5:12
Unidentified speaker (speaker_1): least thus far, the the is there demand. for AI question hasn't changed. Yeah. There's still a lot of people who want compute, who are lined up to buy it. There's not enough in the market to supply the need for it. So from that point of view, it's sort of calmed things. And Core Weav did something else too, which is it said that it didn't actually need to tap the
5:32
Unidentified speaker (speaker_1): high-yield High yield debt markets to keep growing for the time being. Um, so what's been happening in the debt markets is that core weave bonds, unsecured, that means that like when you're buying them, there's just, you know, you're you don't get paid back, you're not gonna get something in exchange for that.
5:48
Unidentified speaker (UNKNOWN): You're
5:48
Unidentified speaker (speaker_1): you're you're just kind of
5:49
Unidentified speaker (UNKNOWN): there's no
5:50
Unidentified speaker (speaker_2): collateral.
5:50
Unidentified speaker (speaker_1): There's no collateral.
5:51
Unidentified speaker (speaker_2): Yeah.
5:52
Unidentified speaker (speaker_1): Uh those core weave bonds had been the ones that had really sold off a lot. Their yield had gotten up to 15% or so uh in recent weeks. But core Corweav's weave's other financing, they have loans that are secured by their agreements with big players such as Meta to buy compute from them in the future. Computing power. Sorry, Miriam.
6:12
Unidentified speaker (speaker_1): Computing power.
6:12
Unidentified speaker (speaker_2): I know. I've just been ignoring
6:15
Unidentified speaker (UNKNOWN): it.
6:15
Unidentified speaker (speaker_1): Those loans have actually held up okay. They haven't sold off the way those bonds have. So what Corweave said is that look, we think we can continue to grow our business by using this kind of secured financing, and we don't necessarily have to tap. the junk bond market to keep raising financing. And what that kind of did was say, okay, everyone's concern that rising
6:35
Unidentified speaker (speaker_1): interest rates and sort of higher hurdles for debt investors to continue to back players like Horweav, okay, maybe that's not actually gonna kind of mess with their near
6:44
Unidentified speaker (UNKNOWN): term
6:44
Unidentified speaker (speaker_1): business plan. You know, people will still debate. I mean, their stock has not totally recovered, and I'm sure people can still debate about whether or not they. really like this business five
6:52
Unidentified speaker (UNKNOWN): years
6:53
Unidentified speaker (speaker_1): from now or ten years from now, or the core economics of it. But at least for the short term. And the market rallied a bunch on, you know, as those earnings came in. It seemed to reassure people. And then you were hit with this NVIDIA news. But overall, now, you know, I would say that that for some companies, this week hasn't actually been that bad.
7:11
Unidentified speaker (speaker_2): And against this crazy backdrop, we have Anthropological. Anthropic still planning to move ahead with its IPO as of what we know right now. Oh,
7:20
Unidentified speaker (speaker_1): and by the way, don't forget, in the century's worth of news, we also had the CPI, the consumer price index inflation report. I know. I'm thinking like, weren't we supposed to talk about something else? Yes. Oh yeah. We had a hot inflation report that showed that and I I think this has an AI kind of relationship to it, that that the price pressures that have been building are
7:40
Unidentified speaker (speaker_1): broader than just gasoline. And we didn't have a year over year comparison for the CPI because we're now a year f out from the period of time when the it was shut down and CPI was not doing that. So um but if you look at it on an annualized basis over the last few months, you know, inflation's running at like five 5.7%, point seven percent, almost almost
8:00
Unidentified speaker (speaker_1): 6% six percent for the last few months that in on an annualized basis. So and seeing those broad pressures, including in areas like computing and software, which, like I
8:10
Unidentified speaker (UNKNOWN): kind
8:10
Unidentified speaker (speaker_1): of hinted at, like I think is probably you can link to AI. I mean, I'm sure economists can pick that apart, but but I would argue that it at least raises the the concern that like this giant build-out is bringing broad price pressure. You did see that the market really locked in expectations for a December rate hike. from
8:27
Unidentified speaker (UNKNOWN): the Fed.
8:27
Unidentified speaker (speaker_1): You saw Treasury yields really react to that. So that's playing a role here too. So the fact that AI companies actually some have done okay and rallied at
8:35
Unidentified speaker (UNKNOWN): times this
8:35
Unidentified speaker (speaker_1): week, despite the fact that those interest rates that everyone was worried about two weeks ago are going up does show you that that maybe people are not in total panic mode despite all this wild news.
8:46
Unidentified speaker (speaker_2): Maybe it's those retail investors buying the dip.
8:49
Unidentified speaker (speaker_1): Maybe it's those retail investors buying the dip. Um, but I interrupted you. You were about to talk about the other giant news
8:55
Unidentified speaker (UNKNOWN): that's
8:55
Unidentified speaker (speaker_1): happening in the world of AI and markets this week.
8:58
Unidentified speaker (speaker_2): Yeah. A totally crazy week of AI leads to potentially another totally crazy week of AI when we have anthropic planning to do its public market debut. It seems like they're not deterred by, you know, what's been happening. It's unclear what exactly the relationship between NVIDIA and Anthropic would even be. And I
9:18
Unidentified speaker (speaker_2): think that's part of why we need a company like Anthropic to be publicly traded, because we don't really know how the market is going to treat it versus the other AI players. We don't have a pure play model provider, frontier model provider as a public company yet.
9:33
Unidentified speaker (UNKNOWN): And
9:33
Unidentified speaker (speaker_1): just to be clear, like in the Anthropic IPO IPure Prospectus, NVIDIA is in there as one of the companies that is also going to be. buying some stock in Anthropic as concurrent with the public market offering. No indication that that has changed or anything. So, you know, thus far, it seems like, you know, despite the fact that that these companies all have a close relationship,
9:53
Unidentified speaker (speaker_1): people are still interested in it. And you make a great point about anthropic does provide something a little different
10:01
Unidentified speaker (UNKNOWN): to the
10:01
Unidentified speaker (speaker_1): market in terms of the type of AI company it is. And to get at that question, kind of like what is anthropic? And importantly, what did we learn about it from its S1 that we didn't already know about the company? We wanted to talk to somebody who's been thinking about Anthropic for a long time as an investment. And we found a great person to do that. His name is Matt
10:20
Unidentified speaker (speaker_1): Withheiler. With Heiler. He's the head of late stage growth at Wellington Management. And what he does is he looks after investments. in pre-IPO companies, companies that are already on the way to being very big companies that will someday be public. He invests in them while they're private, including companies like Anthropic, which Wellington owns. So he has a really interesting perspective
10:40
Unidentified speaker (speaker_1): on not just what do we learn about Anthropic, but but kind of what exactly is this business?
10:45
Unidentified speaker (UNKNOWN): I
10:45
Unidentified speaker (speaker_2): learned a lot from this conversation. That's coming up in a bit.
10:53
Unidentified speaker (speaker_0): Even before a new Congress steps foot in Washington, corporate lobbyists want lawmakers to pass the Durban Marshall credit card mandates. This radical bill reduces access to credit, jeopardizes your rewards, and raises fraud risk, all to pad megastore profits. Community banks, credit unions, military lenders, airline workers, Republicans, and
11:12
Unidentified speaker (speaker_0): Democrats all say no. Make your voice heard. Tell Congress, hands off our cards. Visit Guard Your Card to take action. Paid for by the Electronic Payments Coalition.
11:28
Unidentified speaker (UNKNOWN): Alright,
11:28
Unidentified speaker (speaker_1): welcome back. We're here with Matt Withheiler of Wellington. Matt, welcome to the show.
11:33
Unidentified speaker (speaker_3): Thanks so much for having me, you two.
11:34
Unidentified speaker (UNKNOWN): I'm
11:34
Unidentified speaker (speaker_1): excited to have this conversation, Matt, because in a previous iteration of my job. here, I covered the IPO market and also a lot of the pre-IPO companies. Back when we first started talking about unicorns, and I know this is a
11:47
Unidentified speaker (UNKNOWN): I
11:47
Unidentified speaker (speaker_1): feel like a common talking point today, but but at some point it dawned on all of us, you know, reporters who cover these markets that you're just missing some of the biggest, most important companies in technology And if you're just looking at what's publicly traded. Because you had companies like Facebook, Uber. et cetera, companies that were making it to massive valuations and obviously had huge roles before they
12:07
Unidentified speaker (speaker_1): even went public. So finding investors who were thinking about them as private companies, but at the scale of public companies was really important.
12:15
Unidentified speaker (speaker_3): Yeah, that trend of companies staying private longer and ultimately approaching the public market when they have more scale, when they have more predictability of revenue, when they're further along in kind of their corporate journey, that started I think if you roll the clock back, that probably started 15 or 20 years ago. And if anything, it's just accelerated. Um, I think most
12:35
Unidentified speaker (speaker_3): exemplified by SpaceX recently, where you had a company that went from not available in the public market to being valued at north of a trillion and a half dollars in public market cap um and living its entire journey until that point in the private markets. Yeah,
12:50
Unidentified speaker (speaker_1): that's way beyond. unicorn isn't it when you get to a trillion.
12:52
Unidentified speaker (UNKNOWN): What's the name for that? I
12:53
Unidentified speaker (speaker_2): don't even know.
12:54
Unidentified speaker (UNKNOWN): That's what's rarer than a
12:55
Unidentified speaker (speaker_2): unicorn.
12:56
Unidentified speaker (speaker_1): Flying unicorns, Pegasus.
12:59
Unidentified speaker (speaker_1): So as we contemplate watching Anthropic make that transition from very highly valued, already crucially important to our economy and the market's private company into a public company, it's happening at a time when the markets have been going
13:13
Unidentified speaker (UNKNOWN): pretty
13:14
Unidentified speaker (speaker_1): pretty heavy haywire. What is your story of what has happened in the AI trade over the last few weeks? What's on the market's mind?
13:21
Unidentified speaker (speaker_3): I mean, it's been a roller coaster, so it's a little hard to figure out what's going on. But if I take a step back and think about what is the market trying to figure out here, I think in general, the market is saying, hey, we've got questions around this whole AI thing, but what are those questions specifically? I think looking at the public market, what
13:41
Unidentified speaker (speaker_3): the public market is struggling with is this supply side of the AI equation, the data center build outs, the need for the infrastructure associated with these data center build outs, and ultimately, can all the dollars that have been poured into this, primarily recently in the debt side, can those actually be money good or not?
14:01
Unidentified speaker (speaker_3): So what you're seeing in the public side. today, I think, is a lot of angst around this data center supply build-out part of the component. But I strongly believe that's only one part of the whole AI trade.
14:13
Unidentified speaker (UNKNOWN): Aaron Powell
14:13
Unidentified speaker (speaker_2): And how does anthropic itself figure into that? Is it a victim or is it a beneficiary of the market thinking that way? Aaron
14:22
Unidentified speaker (UNKNOWN): Powell
14:23
Unidentified speaker (speaker_3): Well, I think Anthropic serves or participates in a different part of the AI ecosystem. Because, like we just talked about, I think the challenge in Anks has been primarily on the supply side of AI. Who's going to be making and delivering these tokens for people? And where Anthropic plays, as well as other model providers and other AI
14:42
Unidentified speaker (speaker_3): users, is on the demand side. How much tokens can I actually need? And I think what the the challenge in today's market is the market is being kind of indiscriminate around this AI trade and saying that everything AI is all the same. And I actually think that there is the supply side that we talked about and then the demand side, the anthropics of the world. And I think when you
15:02
Unidentified speaker (speaker_3): look through at the signals of demand, whether it's through anthropic or through some of the other model builders or any of the companies innovating in the technology space with respect to AI, what you see. is no sign of demand weakness that you can point to. Aaron
15:17
Unidentified speaker (UNKNOWN): Powell
15:18
Unidentified speaker (speaker_1): So the main business of Anthropic is the way I think about it, um, developing its frontier model. And Claude is a brand that we know. Um and it's similar to what other companies in the market are doing today. You have uh Google developing Gemini, Mini, uh Facebook now has the Muse Agent, but those are also
15:37
Unidentified speaker (speaker_1): vast companies that do lots of other different things. What's interesting about Anthropic is that they are focused on that. That is their core business. And so in that way, is it a bit of a different bet on AI than if you were buying shares of Alphabet or Meta or any of the hyperscaled companies today?
15:56
Unidentified speaker (speaker_3): I mean, I think undoubtedly, right? And you think about why that's the case. At the onset of this conversation, we talked about the fact that companies are staying private longer, right? And that trend has been continuing for the past decade plus. And if you think about how long we've been in this AI journey, we had the chat GPT moment four years ago. And so we're not really
16:16
Unidentified speaker (speaker_3): long enough in the AI journey to have these native AI businesses like an anthropic be available in the public market today. And I think that that is part of the challenge of the angst the market has been having around this whole AI ecosystem is that there isn't a good measure of demand on the public market to say,
16:36
Unidentified speaker (speaker_3): hey, these tokens will be consumed or not. Instead, the public market is living month to month trying to guess what anthropic's ARR or OpenAI's ARR RR or one of the c token consumers uh demand for tokens are. And so I think once we start getting some of these businesses out, the market will better understand
16:56
Unidentified speaker (speaker_3): what the demand side looks like and therefore what the supply side of this equation should be in a couple years.
17:01
Unidentified speaker (speaker_1): Yeah, I think that that ARR number, that uh annual recurring revenue number for anthropics specifically and and for OpenAI has been like a market moving number even before the companies were public. So whenever there was a new rumor or announcement or anything in the press about those things, the market would really move, even though on the
17:21
Unidentified speaker (speaker_1): results of Anthropic, even though it wasn't a public company yet. So it's it's really interesting to see it now going to be, you know, at some point, if if the IPO pulls off, in the public realm. Do you think that will kind of change the dynamic a little bit around around around questions of AI demand? Will that maybe take out a little bit of the volume? volatility? Because we'll we'll have some more
17:40
Unidentified speaker (speaker_1): clarity on that number?
17:42
Unidentified speaker (speaker_3): Yeah, I think it'll it'll start and it'll be the first one that we'll be able to measure on at least a quarterly basis how much revenue or ARR is being added, how much demand is there for the frontier models. And I think over time, as we see subsequent AI native businesses, whether it's uh OpenAI or maybe a Databricks or companies like that go public, the market will
18:02
Unidentified speaker (speaker_3): just have more and more. data points at which to understand here's where this AI trade is going, again, on a long-term basis, and therefore what should the supply side of the equation look like?
18:13
Unidentified speaker (speaker_2): So we've had Anthropics S1 sitting out there for a little bit now. We wanted to get your take. Uh, what were some of the biggest takeaways for you from that document? And, you know, what were some of the things that you learned from that that we didn't already know?
18:27
Unidentified speaker (UNKNOWN): I
18:27
Unidentified speaker (speaker_3): mean, I think the most striking takeaway was initially the fact that this company has gone now assumed to be three years of growing 10x year over year. And I think that was jaw-dropping to a lot
18:42
Unidentified speaker (UNKNOWN): of people.
18:42
Unidentified speaker (speaker_3): And I think probably the most insightful piece in the S1 was the fact that Anthropic was able or is able to deliver this. growth and they've been able to do it profitably while generating cash flow. I think that was the aha moment for a lot of people. Because I think there was this popular opinion or belief that, oh yeah,
19:02
Unidentified speaker (speaker_3): this AI thing is really cool. They're generating a ton of revenue, but it's profits profitless prosperity. Like they're generating revenue, but it's all at negative margin and they're always going to lose money forever.
19:12
Unidentified speaker (speaker_1): I need you to clarify a little bit by what you mean by profit profitability.
19:15
Unidentified speaker (UNKNOWN): Because
19:15
Unidentified speaker (speaker_1): if you look at it just on a net income basis, actually, Anthropic was hugely unprofitable uh at points this year, right? You know, reporting losses in the tens of billions, not necessarily related to its day to day business, but but so explain to me what you're looking at when you talk about its profitability. Because I think a lot of people would say, wait a second. It's not
19:34
Unidentified speaker (speaker_2): profitable.
19:34
Unidentified speaker (UNKNOWN): It's not profitable.
19:35
Unidentified speaker (speaker_3): I think it's this uh and don't shoot the messenger. It's this uh adjusted net income or adjusted cash flow statement. And the reason the adjustment here makes sense, and we're not back into a community adjusted EBITDA for those people
19:48
Unidentified speaker (UNKNOWN): who remember the we were thing.
19:50
Unidentified speaker (speaker_3): The reason we are not reliving that is because the adjustment for anthropic is an accounting charge. And that accounting charge that's really dragging on net income is how much the notes that the company has raised, so the dollars the company has raised from investors that convert into equity when the company goes public, how much those notes are worth. And
20:10
Unidentified speaker (speaker_3): since Anthropoc
20:11
Unidentified speaker (UNKNOWN): has
20:11
Unidentified speaker (speaker_3): done tremendously well, the people who gave them money that are going to convert into equity, but today recognized as a note, the value of that note is really high. And
20:21
Unidentified speaker (UNKNOWN): so
20:21
Unidentified speaker (speaker_3): that's what's driving the large losses on a net income basis.
20:25
Unidentified speaker (UNKNOWN): The
20:25
Unidentified speaker (speaker_1): value of that liability increased. and that flowed through to the to the net income basis.
20:29
Unidentified speaker (UNKNOWN): But isn't
20:29
Unidentified speaker (speaker_2): that a real expense? Aaron
20:31
Unidentified speaker (UNKNOWN): Ross Powell
20:31
Unidentified speaker (speaker_3): It's a dilution expense,
20:32
Unidentified speaker (UNKNOWN): but
20:33
Unidentified speaker (speaker_3): it's not an expense with respect to the company's ability to generate positive earnings
20:38
Unidentified speaker (UNKNOWN): on
20:39
Unidentified speaker (speaker_3): a revenue basis. Aaron
20:40
Unidentified speaker (UNKNOWN): Powell
20:40
Unidentified speaker (speaker_1): Let's talk about how the market's current concern about the cost and ability to to continue to build out AI could play into Anthropics IPO. because one of the risk factors they talk about is that they won't be able to secure the compute that they need to meet this demand. And you know, already today I think people are very sensitive to the
21:00
Unidentified speaker (speaker_1): price of tokens. People who are using these things are like, oh I, you know, my work isn't giving me enough tokens to do what I need to do. And you know, the the office manager says, well they're they're expensive. We can't just give it you know so so they they need more compute in order to be able to fulfill the demand that that that is out there, it seems. So if it is much more costly for them to raise debt, or
21:20
Unidentified speaker (speaker_1): as Miriam said, you know, the real cost of issuing notes and other
21:23
Unidentified speaker (UNKNOWN): financing
21:24
Unidentified speaker (speaker_1): goes up, does that slow them down? Do you think that that will be something that people in the market will be really focused on as they discuss the IPO at at whatever price it might be offered at?
21:33
Unidentified speaker (UNKNOWN): Well,
21:33
Unidentified speaker (speaker_3): I think people will be focused on yes, how much compute can anthropic get over time? to drive the revenue growth that they're generating? And I think it's a it's a totally valid question. Now, I think with respect to do they need a ton of debt to go do this and will they be able to get it? I think if you look at the balance sheet, they have over a hundred billion on the balance
21:53
Unidentified speaker (speaker_3): sheet today. Uh in the IPO, it's rumored that they could be raising potentially up to another hundred billion of uh of uh cash as part of their um raise. So then you're looking at and saying like, hey, we got two hundred billion of cash. And in the in the S one there was a lot of talk around the forward commitment uh to to buy, I think order of magnitude
22:13
Unidentified speaker (speaker_3): around eight hundred billion of compute. And when you think about the two hundred with respect to the eight hundred, you say, Oh man, that's scary. But
22:20
Unidentified speaker (UNKNOWN): I
22:20
Unidentified speaker (speaker_3): think if you start looking through and thinking through how these contract lengths are, the eight 800 hundred is not, hey, we gotta we're going spend to spend 800 billion in the next year. It's, hey, we're going to spend 800 billion in the next decade plus. And so it's the pacing of that. And as we talked about before, the company does generate cash. They do generate profitability. And therefore, can
22:40
Unidentified speaker (speaker_3): they use the the uh operating income of the business to continue to fulfill the forward obligations that they have for compute?
22:47
Unidentified speaker (speaker_1): Yes, some of that is due. I think they said in the S1 uh that they're
22:50
Unidentified speaker (UNKNOWN): not going to be able to do
22:50
Unidentified speaker (speaker_1): that some some amount of that $800 plus billion is due in the next year, but it's not $800 billion. It's a it's a much small I think it was under a hundred billion dollars. So so to your point, you're saying like they've got the resources to do that today. And that's kind of the point of these IPOs, isn't it? It's to raise cash and also make it easier for you to access cash in the market through
23:10
Unidentified speaker (speaker_1): a debt offering, though I don't know, given the way that the market is treating kind of AI debts and and we don't know whether Anthropic would be an investment grade or not investment grade rated company. You know, that whether or not they actually get a lot more access to financing, that's an open question. I feel like that's something that people will probably be thinking about when they are thinking
23:29
Unidentified speaker (UNKNOWN): about
23:29
Unidentified speaker (speaker_1): valuing the IPO.
23:30
Unidentified speaker (speaker_2): I also, I mean, it is pretty rare to have long-term long term commitment to be on the hook for this amount of money, even if it is 10 ten years from now. I mean, who knows? So much could change over that period in terms of their, you know, competitive landscape, their access to the needed computing power that, you know, that they that they're trying to
23:50
Unidentified speaker (speaker_2): to secure in order to do their business. It does seem like that will continue to be like sort of, you know, these commitments will continue to be something that people are watching very carefully.
24:00
Unidentified speaker (UNKNOWN): Yeah, and to
24:00
Unidentified speaker (speaker_1): put a finer final point on it, like what's the visibility into the the revenue side of things? Like do they have commitments from people that are going to buy
24:07
Unidentified speaker (UNKNOWN): compute and use
24:08
Unidentified speaker (speaker_1): compute? Like how do you think people will think about that that balance?
24:11
Unidentified speaker (speaker_3): Well I mean I think the way the company talks about it is they are seeing tremendous appetite for their specific use cases in coding, right? So they're seeing a lot of demand for people who want to use their models, use them in environments where they can make develop developers more efficient. Aaron
24:27
Unidentified speaker (UNKNOWN): Powell
24:28
Unidentified speaker (speaker_2): So in a business context, basically. Aaron
24:29
Unidentified speaker (UNKNOWN): Powell In
24:30
Unidentified speaker (speaker_3): a business context. And I think if you read through the S1, the way that the team talks about the subsequent what they call S curves is we have the first S curve, which is this coding S curve, this business use case S curve. And the next S S-curve curve that the team is hyper focused on is this biology S curve. And the
24:50
Unidentified speaker (speaker_3): way the team talks about it is we had rapid adoption in uh this coding use case because the person who buys the model for coding is an engineer. So they're used to taking risks and buying new stuff. And they have a lot of pricing power at their companies because they add a lot of value. And so the acceleration of coding has been steep and is
25:09
Unidentified speaker (speaker_3): continuing to be steep. And the coding or the biology. S curve, the next S curve they talk about, they talk about it potentially ramping slower because there's more, there's more inertia uh friction in the biology use cases. But when they talk about
25:27
Unidentified speaker (UNKNOWN): Is that
25:27
Unidentified speaker (speaker_2): like healthcare and pharma?
25:29
Unidentified speaker (speaker_3): Healthcare, particularly pharma. And when you hear the company talk about it, they talk about kind of biology and drug discovery. And when you think about the time associated with getting that S-curve up and running, it may be longer than the coding one. But when you think about the market opportunity, it's at least as big, if not bigger, than
25:49
Unidentified speaker (speaker_3): the coding market opportunity. And so we're talking about gigantic portions of GDP that is now addressable in the second S curve. And then they talk about the third S curve being this physical AI. The we've got robots out in the real world at some point. What is going to allow those robots to be able to interact in a day to day environment, the physical
26:08
Unidentified speaker (speaker_3): application of AI? That'll be the third S curve. And so the demand piece that we just talked about is really going to be predicated on how much can they keep growing in coding? How quickly can they come up on the biology S curve? And when does that S curve around physical start taking off?
26:25
Unidentified speaker (UNKNOWN): Well,
26:25
Unidentified speaker (speaker_1): and also it seems like it's still
26:28
Unidentified speaker (UNKNOWN): the
26:28
Unidentified speaker (speaker_1): the cost of of of computing and buying, you know, uh the the brain power from Claude or whoever to to to do big innovative stuff like that is is is still fairly high. Um how much do you think that the cost of compute needs to come down to incentivize people to to do these giant
26:48
Unidentified speaker (speaker_1): projects, like, you know,
26:50
Unidentified speaker (UNKNOWN): put
26:50
Unidentified speaker (speaker_1): put these models to work in curing cancer and and and things like that. And what does that mean for a business like uh anthropic? I know people talk about the Jevons paradox of oh, the as the cost come down, comes down, your revenue goes up because people start using it even
27:05
Unidentified speaker (UNKNOWN): more.
27:05
Unidentified speaker (speaker_1): How important is that dynamic to this business? Do you think
27:07
Unidentified speaker (UNKNOWN): that's a good thing? Well,
27:08
Unidentified speaker (speaker_3): I think the Jevons paradox is definitely important, but going back to this question around how do costs need to evolve in order to to enable the biology use case, I actually think that in those scenarios, your willingness to pay is really high because the potential outcome is incredibly lucrative. And therefore, you you don't need
27:28
Unidentified speaker (speaker_3): to be as cost focused in these highest value tasks. And where anthropic plays, which is at the frontier, the frontier of model building, the most cutting edge, those frontier models should be able to charge the most for their capabilities because of the tremendous value they're able to deliver. That being said,
27:48
Unidentified speaker (speaker_3): it will also be accelerated, the revenue at Anthropic, by this Jevons paradox thing, which is every time Anthropic releases a new model, you know, they just release the new uh um mythos and fable models. Um every time they release new models, the cost of prior models comes down. And on one hand, you may say, oh, that's a
28:08
Unidentified speaker (speaker_3): bad thing for the company.
28:09
Unidentified speaker (UNKNOWN): Because they're
28:10
Unidentified speaker (speaker_2): undercutting themselves.
28:11
Unidentified speaker (UNKNOWN): They're
28:11
Unidentified speaker (speaker_3): undercutting themselves. At the same time, if you look at what they said in the S1, what they said in the S1 is that um I think it was 80% of revenue comes from models that are six months old or newer. So the vast majority of revenue is still coming from
28:27
Unidentified speaker (speaker_3): the newest models, even though the cost of those models is starting to come down. And that's the Jevons paradigm.
28:33
Unidentified speaker (UNKNOWN): And it's probably
28:33
Unidentified speaker (speaker_2): just me at home with my Claude subscription using the older models.
28:37
Unidentified speaker (UNKNOWN): Exactly.
28:37
Unidentified speaker (speaker_3): And creating uh PowerPoint presentations
28:39
Unidentified speaker (UNKNOWN): or whatever
28:39
Unidentified speaker (speaker_3): you
28:39
Unidentified speaker (UNKNOWN): do
28:39
Unidentified speaker (speaker_3): in your
28:39
Unidentified speaker (UNKNOWN): screen.
28:40
Unidentified speaker (speaker_2): I'm just experimenting. I accidentally signed up for like a seven-day free trial and then forgot to quit. So, you know, I I'm keep
28:47
Unidentified speaker (UNKNOWN): using it.
28:48
Unidentified speaker (speaker_3): company.
28:48
Unidentified speaker (speaker_1): I
28:48
Unidentified speaker (UNKNOWN): mean,
28:48
Unidentified speaker (speaker_1): I would hope that when I ask like my agent, like which toothpaste is best for sensitive teeth? Like, it's not using the most cutting-edge computing power that could be used for cancer curing. Like
28:58
Unidentified speaker (UNKNOWN): that, I
28:59
Unidentified speaker (speaker_1): I don't want to.
29:00
Unidentified speaker (UNKNOWN): You're cool with
29:00
Unidentified speaker (speaker_2): that.
29:00
Unidentified speaker (speaker_1): I I'm okay if it takes if it takes the model like an extra minute to think about the answer if it's busy doing something important like that.
29:06
Unidentified speaker (speaker_2): We're going to
29:07
Unidentified speaker (speaker_0): take a
29:07
Unidentified speaker (UNKNOWN): quick
29:07
Unidentified speaker (speaker_0): break,
29:08
Unidentified speaker (speaker_2): and when we come back,
29:08
Unidentified speaker (speaker_0): we'll have more with
29:09
Unidentified speaker (speaker_2): Matt Withheiler of Wellington Management.
29:17
Unidentified speaker (UNKNOWN): This
29:17
Unidentified speaker (speaker_0): episode is brought to you by Hard Lessons, a podcast from Morgan Stanley. Some investing lessons only become clear after you see how a call plays out. On Hard Lessons, iconic investors sit down with Morgan Stanley leaders to go behind the scenes on the critical
29:31
Unidentified speaker (UNKNOWN): moments, both successes and
29:33
Unidentified speaker (speaker_0): setbacks that shaped who they
29:35
Unidentified speaker (UNKNOWN): are today. Watch or
29:36
Unidentified speaker (speaker_0): listen to Hard Lessons wherever you get your podcast podcasts.
29:48
Unidentified speaker (speaker_1): So as people think about, you know, in in a in a choppy market, as they think about, okay, do I need another giant um AI company in my portfolio? What role does a company like Anthropic play in portfolios and in AI that gives it this value that
30:08
Unidentified speaker (speaker_1): that goes well beyond the things that we're talking about in the here and now?
30:12
Unidentified speaker (speaker_3): I mean, I think the two fundamental things that it offers are one, growth that, as we talked about before, nobody has ever seen. Not not in our lifetime, like ever anybody has ever seen. So you've got unprecedented levels of growth. And the second thing that it offers is a differentiated exposure to this AI ecosystem. The fact that
30:32
Unidentified speaker (speaker_3): Anthropic is really the first, will be, the first pure play AI focused company available in the public markets versus the infrastructure providers versus the hyperscalers, where they don't have that same direct in-the-token flow exposure that Anthropic
30:50
Unidentified speaker (UNKNOWN): has.
30:50
Unidentified speaker (speaker_3): I
30:50
Unidentified speaker (speaker_2): did want to ask you one other thing, which we've talked about how Anthropic is a pure play AI company and it's kind of like the unique pure play demand side company. But we've we've also talked on this show about how it can make sense for an investor to hedge their bets a little bit with that demand and to own something like Alphabet, where you have an
31:10
Unidentified speaker (speaker_2): established ad business, but also have the AI potential demand upside. And I think about the way I use, you know, AI, which is not in a very sophisticated way, but I kind of think of it as a glorified search engine. That's sort of how a lot of consumers use it, I think. And so where do I go for search? Well, I go to Google, which
31:30
Unidentified speaker (speaker_2): is where everybody goes. So in that sense, maybe Alphabet wins. What do you say to that?
31:35
Unidentified speaker (speaker_3): I would say that there are a variety of places that AI is impacting the economy. And I think the consumer one is one place. And perhaps that gets impacted by accelerating revenue growth at Google. Perhaps that gets impacted by OpenAI taking share from Google. But separate from the consumer play is really what Anthropics focused on, which is on this
31:55
Unidentified speaker (speaker_3): enterprise focused segment of AI and AI adoption. And that's really ever since the onset, when I sat down with Dario at this point over three, almost three and a half years ago for the first time, the mission was 100% clear. We are going to own enterprise AI and we're the winners for that space.
32:13
Unidentified speaker (UNKNOWN): Yeah,
32:13
Unidentified speaker (speaker_1): I kind of wonder if enterprise, I mean, to me, that almost feels like a more likely thing to play out over the next few years versus consumers. I But feel like
32:21
Unidentified speaker (UNKNOWN): consumers will
32:22
Unidentified speaker (speaker_1): be late adopters in some respects with Ada. If we I mean it's not changing my life. that much yet, I'm doing very light things.
32:29
Unidentified speaker (UNKNOWN): They weren't
32:29
Unidentified speaker (speaker_2): mentioned on the S that you talked about. Yeah,
32:31
Unidentified speaker (UNKNOWN): that's exactly right.
32:32
Unidentified speaker (speaker_3): Consumer
32:33
Unidentified speaker (UNKNOWN): was not mentioned on the S curve. And
32:35
Unidentified speaker (speaker_3): that's purposeful.
32:36
Unidentified speaker (UNKNOWN): Um
32:36
Unidentified speaker (speaker_3): and I believe Anthropic's point of view is other people can go have that, whether it's Google or OpenAI or some nascent player that we are not talking about.
32:44
Unidentified speaker (speaker_1): So do you think that anything has changed enough in the last few weeks that will radically change how people might perceive this company? Is the market going to struggle to see it as the company that it seemed to think it was a few weeks ago? Or do you think that, you know, the way people think about IPOs and approach IPOs
33:03
Unidentified speaker (speaker_1): and these long-term kind of plays, that there's something that people will look through the current volatility and say, yeah, this is still gonna be one of the market's biggest and most important companies?
33:14
Unidentified speaker (UNKNOWN): I
33:14
Unidentified speaker (speaker_3): believe that. the market will take a long horizon and operate on that information to trade it as the best proxy for what its future discounted cash flows are in perpetuity, which is ultimately where a company should be valued. And the oscillations around
33:34
Unidentified speaker (speaker_3): sentiment towards AI are gonna going to go up and down, but at the end of the day, maybe may on a day to day basis impact stock price, but over a long term horizon, shouldn't matter. Um, and so I I do think that, yeah, versus a quarter ago, is there more noise around this AI trade? Is there more noise that anthropic is
33:54
Unidentified speaker (speaker_3): going into the public markets with? Yes. But is that noise sufficient to say that, hey, this AI thing isn't a thing, and therefore we don't want to make any more AI infrastructure? investments. I don't believe so. Well,
34:04
Unidentified speaker (speaker_1): Matt, this has been a really interesting conversation. Thanks for joining
34:07
Unidentified speaker (UNKNOWN): us.
34:07
Unidentified speaker (speaker_3): Thanks so much for having
34:08
Unidentified speaker (UNKNOWN): me.
34:10
Unidentified speaker (UNKNOWN): And
34:10
Unidentified speaker (speaker_2): that's everything you need to know to take on your week. The show is produced by Alexis Moore and Michael Laval. Michael Laval is our sound designer. He also wrote our theme music. Aisha Al-Muslib is our development producer. Chris Sinsley is our deputy editor. Lee Tal Malad is our senior director of shows. and Samantha Hennig is the Wall Street Journal's head of
34:30
Unidentified speaker (speaker_2): multimedia. For even more, head to WSJ.com. I'm Miriam Gottfried. And
34:34
Unidentified speaker (speaker_1): I'm Telas Demos. Until next
34:36
Unidentified speaker (UNKNOWN): time.
34:42
Unidentified speaker (speaker_0): Even before a new Congress steps foot in Washington, corporate lobbyists want lawmakers to pass the Durban Marshall credit card mandates. This radical bill reduces access to credit, jeopardizes your rewards, and raises fraud risk, all to pad megastore profits. Community banks, credit unions, military lenders, airline workers, Republicans, and
35:02
Unidentified speaker (speaker_0): Democrats all say no. Make your voice heard. Tell Congress, hands off our cards. Visit Guard Your Card to take action. Paid for by the Electronic Payments Coalition.
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 WSJ's Take On the Week

WSJ's Take On the Week

Can Everyday Investors Beat the Pros in an AI Selloff?

Episode discussion: Retail investors show sophistication and resilience amid AI selloffs, acting as market buffers.

Speakers’ claims: Scott Rubner of Citadel Securities and Hannah Erin Lang discuss how retail investors have evolved since 2020, shifting from speculative behavior to professional-like strategies. Rubner notes that retail investors often buy when institutions sell, providing liquidity during market dips. Lang attributes this surge to zero-commission trading, mobile apps, and pandemic stimulus. The hosts highlight that retail investors are now allocating to bonds and ETFs, mimicking institutional diversification. Rubner also points to strong Q3 earnings growth and AI productivity spreading beyond mega-cap tech as key market drivers.

Retail InvestingMarket StructureAI Stocks
30 min2 assets10 claims to verify
WSJ's Take On the Week

Tech Stocks and AI Bonds Are Tumbling. Is the AI Trade in Trouble?

Episode discussion: AI credit sell-off driven by rate fears, not demand collapse, as CoreWeave faces leverage scrutiny.

Speakers’ claims: Speakers argue that the recent sell-off in AI-linked bonds, particularly CoreWeave, is driven by rising interest rates and financing costs rather than a fundamental decline in AI demand. They highlight CoreWeave’s high leverage and reliance on cheap capital, contrasting it with hyperscalers like Google and Microsoft that are developing proprietary chips. While noting significant concentration risk in AI bond issuance, the speakers view the downturn as a temporary correction caused by monetary policy, not a bubble burst. They point to successful equity raises by companies like Meta and Anthropic as evidence of continued capital availability and underlying strength in the sector.

AI InfrastructureCoreWeaveInterest Rates
35 min5 assets15 claims to verify
All insights