Why The Stock Market Is Booming — But Your Paycheck Isn't

Anthropic

Across the shows: 45 speakers in other episodes, 26 bullish and 47 bearish calls, past 30 days

• Ed Elson · Prof G MediaBearish

  Anthropic's growth is slowing and its adjusted profitability hides training costs, so unit economics are unproven.

  Ed says Anthropic's Q3 revenue of about $16.8B, up nearly 50%, is a sharp slowdown from the previous quarter, when revenue more than doubled. RAMP data also show per-employee AI spending among top spenders falling since summer, which supports the slowdown. He says the reported 85% gross margin excludes training costs, the company's largest expense, and the 11% adjusted operating margin depends on undefined adjustments. He still sees a massively unprofitable business and has seen no evidence that the unit economics work, though he credits Anthropic as more disciplined than OpenAI.

  “I still get the sense that this is a massively unprofitable company because of the training costs... I haven't seen evidence that the unit economics make sense here.”at 33:58

• Ed Elson · Prof G MediaBearish

  Anthropic's roughly $2T private valuation looks overstated; public AI multiples imply about $300-600B.

  Anthropic is reportedly valued at about $2T, which he says is supposedly the figure. On its most recent ARR it trades above 30 times sales, while chipmakers trade near 10 times sales. Applying the public AI multiples gives roughly $500B, or $300-600B. Private VCs ignore debt and profitability, but public investors will ask about long-term obligations, so he expects a rude awakening on listing.

  “If you would apply that to Anthropic right now, which based on its most recent ARR is trading at more than 30 times sales, then you're looking... you're looking at that company being valued at something like $500,000,000,000, between, like, 300 to $600,000,000,000.”at 41:01

• Scott Galloway · NYU Stern School of BusinessBearish

  At $2T, Anthropic's roughly 30x revenue multiple is hard to justify as growth slows and financing costs rise.

  At $2T, Anthropic is about 30 times revenue, for a company growing roughly 50% a quarter. Scott says the technology will outlast the valuations. Public investors may not pay a growth multiple if growth decelerates and the cost to finance it rises. Private investors are not dumb, he says, but they avoid accountability. He thinks management is anchored on the $2T number, which makes a lower IPO price awkward.

  “But at $2,000,000,000,000, that's about 30 times revenue for a company growing roughly 50% a quarter by these figures... The question is whether public investors will pay a growth multiple when growth is decelerating and the cost to finance that growth is going up.”at 45:44

• Ed Elson · Prof G MediaBearish

  Anthropic's IPO will not happen this year after already being delayed.

  Ed says the Anthropic IPO was reportedly set for November but has already been delayed, so he does not expect it this year. He calls this a hot take. He adds that if Anthropic or OpenAI do list, they may copy SpaceX by floating a very small share count so that a high price holds, rather than flooding the market at a lower valuation.

  “I don't actually know if if this anthropic IPO is gonna happen when they say it will. Apparently, it's gonna happen in November. Well, it's already been delayed. Yeah. I don't think it happens this year.”at 55:34

• Ed Elson · Prof G MediaBearish

  Anthropic will struggle in public markets and is likely to delay its IPO plans.

  Elson argues that public-market investors are a different group from the private investors who have funded Anthropic, and that Anthropic and OpenAI will learn this the hard way once they list. He expects Anthropic to delay its current IPO plan, and says that when it does go public the ride will be 'very shaky' and it will face a 'rude awakening'. The argument rests on the difference in investor base rather than on any specific financial figures.

  “I think Anthropic and OpenAI are gonna have a really tough time entering the public markets as they learn that this is a different type of investor.”at 1:17:41

Bumble

Across the shows: 2 speakers in other episodes, 1 bullish and 1 bearish calls, past 30 days

• Scott Galloway · NYU Stern School of BusinessBearish

  Bumble's $1.3B take-private rescues a broken dating-app model that rewards churn and keeps users swiping.

  Bumble's value fell from $14B to $1.3B, a loss of about 90% in five years. Scott says the business model rewards churn, because success equals churn, so the company is incentivized to keep users single and swiping. He argues that the new owners will have to reconfigure the model, because the current one does not work. He calls the deal a rescue, not a going-private.

  “Bumble's not going private. It's being rescued from the public markets. ... They're gonna have to totally reconfigure the business model because the business model doesn't work.”at 1:03:12

• Scott Galloway · NYU Stern School of BusinessUncertain

  Bumble should shift from selling romance to selling community, such as paid group activities and running clubs.

  Galloway argues that dating apps' core promise of love is not working, and that Bumble should move to a subscription model built on community. He suggests a paid service (about $20 a month) that gets people out of the house and into the company of strangers through running clubs, athletic leagues, and events like shelter-dog walks, with points for attendance. He ties this to a broader public need for in-person 'third places', which he says are disappearing (40% of UK pubs have closed since COVID). His view is a strategic recommendation rather than a forecast of Bumble's results, so the stance is uncertain.

  “I think it needs to move away from the promise of love to the promise of community and roll up running clubs, roll up athletic leagues, and just say, pay us $20 a month, and we're gonna get you out of the house and in the company of strangers.”at 1:13:14

• Ed Elson · Prof G MediaBullish

  Bumble should double down on community-based, shared-interest matching, since it can spark romantic connection too.

  Elson argues that a setup where people meet through a shared activity, in the same place and with a shared interest like running or climbing, produces more real connection than a dating-app setup where both people arrive looking for a partner. He says this 'works for friendships' and also 'can work for actual romantic connection', so Bumble should lean into the community model rather than only romance. His view supports the pivot Galloway describes, so the stance is bullish on that strategy.

  “It works for friendships, but I also think it can work for actual romantic connection, and they should double down on that.”at 1:14:50

CoreWeave

Across the shows: 33 speakers in other episodes, 10 bullish and 30 bearish calls, past 30 days

• Scott Galloway · NYU Stern School of BusinessBearish

  CoreWeave's debt dwarfs its equity, so a modest AI selloff could wipe out shareholders.

  CoreWeave has a market cap of about $37B but an enterprise value of about $90B because of roughly $52B of debt, so most of its enterprise value sits in debt. Scott says that if the AI trade cools and the company's enterprise value falls by a third, the equity would be close to zero. His caveat is that some debt is already trading at about 80 cents on the dollar, but the bond yield touching 15% signals real doubt about repayment. Equity holders are the 'tail of the whip', absorbing most of the downside.

  “CoreWeave has a $37,000,000,000 market cap, but it has a $90,000,000,000 enterprise value because it has 52,000,000,000 in debt.”at 34:44

• Scott Galloway · NYU Stern School of BusinessBearish

  Rising borrowing costs will shrink CoreWeave's equity value as it keeps funding expensive capex.

  Interest rates have accelerated, but the market has not priced this in. CoreWeave has about $52B of debt. If its debt service rises from 10% to 15%, that adds roughly $200M a month in interest. Capex has become much more expensive, which raises borrowing costs and cuts enterprise value. Since the company will not cut capex, equity value falls. Companies heading into IPOs are not disclosing this.

  “Because if you're a [CoreWeave], just to use the public, and we know they have 52,000,000,000 in debt, when their debt service payments go from 10% to 15%, 5% on 50, that's an additional $200,000,000 a month in interest payments.”at 56:01

OpenAI

Across the shows: 33 speakers in other episodes, 19 bullish and 29 bearish calls, past 30 days

• Ed Elson · Prof G MediaBearish

  OpenAI's supposed $1.5T valuation is where public-private dislocation is most dangerous.

  Ed says the valuation dislocation sits with Anthropic and OpenAI, which are supposedly valued at $2T and $1.5T. Private investors do not care about debt or profitability, but public markets are now skeptical of that. He argues the companies will face public scrutiny of their long-term obligations once they list, and that the adjusted metrics they show will not satisfy it.

  “Here's where I do get worried, Anthropic and OpenAI. That to me is where the dislocation is.”at 40:04

• Ed Elson · Prof G MediaBearish

  OpenAI will struggle in public markets, facing a very shaky ride after it lists.

  Elson argues that public-market investors are a different group from OpenAI's current backers, and that OpenAI will be in for a 'rude awakening' when it enters the public markets. He expects the transition to be shaky, based on the investor-base difference rather than on specific financial numbers.

  “I think Anthropic and OpenAI are gonna have a really tough time entering the public markets as they learn that this is a different type of investor.”at 1:17:41

NVIDIA

Across the shows: 40 speakers in other episodes, 31 bullish and 12 bearish calls, past 30 days

• Ed Elson · Prof G MediaBullish

  Nvidia at about 16x forward earnings is reasonably valued, so he is not worried about it.

  Ed argues public valuations have contracted from last year and are becoming more sensible, so he is less worried about the big names. He cites Nvidia trading at about 16 times forward P/E, which he says is okay. His concern sits instead with the private AI labs.

  “I'm not so worried about Nvidia, for example, which is trading at, you know, 16 times forward PE. ... That's okay for me.”at 39:41

Space Exploration Technologies

Across the shows: 16 speakers in other episodes, 22 bullish and 8 bearish calls, past 30 days

• Ed Elson · Prof G MediaNeutral

  SpaceX's extremely low float undermines price discovery and makes its stock easier to manipulate.

  Ed says SpaceX was cut in half as he predicted and then recovered. He calls its low float bullshit. Nasdaq had to change rules to include it because the float was so small. Only a small fraction of the roughly $2T of shares trades, so a few players can control the price. He presents SpaceX as the model that Anthropic and OpenAI might copy.

  “one of the things that is going on with SpaceX, which I think is honestly bullshit, is the fact that the float is so extraordinarily low.”at 48:42

Constellation Brands

Across the shows: 1 speaker in other episodes, 2 bullish and 0 bearish calls, past 30 days

• Scott Galloway · NYU Stern School of BusinessBearish

  Constellation's beer and spirits rebound is a bounce, not a boom, against a secular decline.

  Scott welcomes the rise in Gen Z weekly bar-going, which went from 26% to 32%, and wine and spirits sales that were up 17%. But one good quarter does not reverse the secular decline. The stock is down 37% over five years, versus the S&P up 86%. Diageo is down 50% and only AB InBev is up, 48%. He says the business is seeing a bounce, not a boom.

  “So one good quarter doesn't reverse the secular decline. We're seeing a bounce, not a boom.”at 1:04:38

US consumer

• Ed Elson · Prof G Media

  Rising consumer spending is misleading; it reflects higher prices and falling savings, not household strength.

  The administration cites consumer spending growth of 6.1% as proof the economy is fine. Ed says the figure is inflated by prices: spending rises because prices are up. It is also being funded by savings, as the savings rate has fallen from 5.7% under Trump to 4.1%, the lowest in nearly four years. He adds that spending is increasingly driven by the top 10%, who account for about half of consumer spending, so the aggregate figure hides strain among ordinary households.

  “One, spending is up because prices are up. So, of course, we're spending more... Two, spending is also up because savings are down. When we look at the savings rate since Trump took office, it was 5.7%. It's fallen down to 4.1%”at 15:19

• Scott Galloway · NYU Stern School of Business

  Stock-market gains boost top-decile wealth, which keeps aggregate consumer spending looking strong.

  Scott says the top 10%, now responsible for about 50% of consumer spending, are driving the data. The stock market has boomed, driven mainly by hyperscalers and the AI trade, and even after a pullback it remains well up. That creates an upward spiral: stock gains raise the wealth and confidence of the top decile, which sustains the spending that makes the consumer look healthy while the rest of the economy struggles.

  “It's the ultimate upward spiral because the stock market has boomed, mostly driven by hyperscalers and the AI trade... The stock market is way up. That has an enormous positive benefit and psychological benefit on the top 10% now responsible for 50% of consumer spending.”at 18:22

US–China AI & chip race

• Scott Galloway · NYU Stern School of Business

  China is deliberately flooding the market with cheap AI inference, pressuring American AI economics.

  Scott says he is doubling down on an earlier prediction that China is attacking the American economy by pushing down AI inference costs. The mechanism is that Chinese providers dump cheap LLM compute, which drives token prices down. That lowers what US AI providers can charge, so the debt they took on to buy GPUs is less covered, and the stress shows up in weak credit and equity values for firms such as CoreWeave.

  “I think Xi is coming for the jugular of the American economy and is dumping AI, compute inference, LLMs, whatever you wanna call it, and is reflected in the plummeting price of tokens”at 36:06

• Scott Galloway · NYU Stern School of Business

  Economic war with China is the biggest war, and IP theft and corporate espionage are rampant.

  Scott argues the biggest ongoing war is an economic war between China and the US. IP theft and corporate espionage are huge, and the amount of corporate espionage dwarfs real espionage. He cites an espionage case involving OpenAI. The US built its manufacturing by stealing technology in the nineteenth century, and IP theft is the most lucrative economic model.

  “the war that's ongoing and is probably the biggest war is an economic war between China and The US. And IP theft and corporate espionage. Huge.”at 53:28

AI infrastructure & data centers

• Ed Elson · Prof G Media

  CoreWeave-triggered neocloud selloff signals stress across the whole AI infrastructure trade.

  After CoreWeave's debt yield spiked, CoreWeave fell 26% and Nebius fell 22% in a week. Memory makers fell 14-18%, Nvidia 7%, AMD 12%, Intel 13%, and Oracle and the neoclouds about 9%. Ed says the market is pricing the risk across the entire trade, not just one company, and that this news reflects that.

  “you have a lot of pain that is being reflected and triggered on the news of a single company because it says something about the entire trade itself.”at 39:06

• Scott Galloway · NYU Stern School of Business

  Trump will likely announce a bailout-style backstop for hyperscalers before the midterms to prop up markets.

  Galloway argues that the stock market is the only bright spot left for Trump, given failed Iran policy, tariffs, gas prices and inflation. He expects Trump to respond before the midterms with something that looks like an investment but works as a bailout, propping up the hyperscalers and reassuring stock investors. He says Trump may only threaten it rather than enact it, but he expects an attempt along these lines from Trump or the Treasury, with the hyperscalers as the target.

  “I believe between now and the midterms, he's gonna announce something that resembles a bailout cosplaying an investment.”at 1:16:53

Venture capital & startups

• Ed Elson · Prof G Media

  Private AI valuations have almost no price discovery; a few lead investors choose the numbers.

  Ed says private rounds are set by a handful of kingmakers, such as Sequoia choosing a number that others follow. Investors join because a lead investor is in. With few shares trading, the game is controlled by a small group, so the price is not discovered in a real market. Private investors also care about founders and technology, not debt or profitability.

  “There is no private... There is no price discovery in the private markets. Like, it's just Sequoia's leading the round. They chose a number.”at 47:28

• Scott Galloway · NYU Stern School of Business

  US private capital formation is unmatched; similar Brazilian firms trade at a tenth of San Francisco valuations.

  Scott cites an analysis of private tech companies in Brazil with similar growth, revenue run rates and infrastructure. They trade at one-tenth the valuation of San Francisco companies because US capital formation is extraordinary. Anthropic founded in Europe would be among its three or four most valuable companies. A Series B of $50-100M happens only in a few US cities. China forces companies to go public, which he thinks is healthier.

  “very similar companies growing at the same rate with the same revenue run rate... trade at one tenth the valuation of companies headquartered in San Francisco. And that's because capital formation in The US is just extraordinary.”at 50:56

Wealth, inequality & billionaires

• Scott Galloway · NYU Stern School of Business

  Inflation is a regressive hit: asset owners gain while young renters lose, widening wealth gaps.

  Scott cites a generational split: since 2022, wealth for Americans 75 and up has risen 37%, while wealth for those 35 and under has fallen 23%. He calls inflation a regressive tax. If you own assets it is a tailwind, but if you are young and renting it is a 'mugging'. He argues the anger among young people is justified and that the economy has been rewarding asset owners while the cost of living squeezes everyone else.

  “Since 2022, just literally four years, wealth for Americans 75 and up is up 37%. For people 35 and under, it's down 23%.”at 12:41

US inflation

• Ed Elson · Prof G Media

  Energy shocks are feeding into core inflation, so gas and diesel prices matter beyond headline CPI.

  Scott argues gasoline prices matter mostly for political perception. Ed disagrees: they are important for real inflation. Diesel prices have risen close to 70% since the war began, and diesel is an input for fertilizer, which feeds food prices, and for freight, which moves all goods. He says this pass-through is already showing up, with core CPI approaching 3%, so energy costs are spreading through the wider economy.

  “diesel prices rise close to 70% since the beginning of the war. And as we started to learn, it actually does trickle through to the rest of the economy because diesel is used for fertilizer, which is used to grow food.”at 14:09

US jobs & labor market

• Ed Elson · Prof G Media

  Real wage growth is the key economic number, and inflation-adjusted wages have been flat since Trump took office.

  Ed argues the most important indicator is real wage growth, not headline spending or GDP. Average hourly earnings grew about 3% while CPI rose 3.7%, so real earnings fell about 0.7% year over year, the sixth straight monthly decline. He says inflation-adjusted wage growth for the average American has been roughly zero since Trump took office, so households are standing still even as the headline data look strong.

  “I think the most important number is real wage growth. I think that's the thing that matters. If prices are up, but wages are rising higher, that's fine... since Trump took office, the inflation adjusted wage growth for the average American has been zero.”at 15:53

Earnings & corporate profits

• Ed Elson · Prof G Media

  Corporate earnings are booming on AI and productivity while workers have captured none of the gains.

  Ed says S&P 500 earnings have grown almost 40% since Trump took office, while real wages have been flat. He links this to AI adoption, rising productivity and GDP growth, which he sees flowing to corporations. He calls it a K-shaped economy on steroids: the gains are concentrated among shareholders and older, wealthier households, while people who depend on salaries are falling behind.

  “S and P 500 earnings in that same time frame have grown almost 40% since he took office... The economy is roaring only for corporations and somehow... literally zero of the gains have accrued to American workers.”at 16:41

US rates & the Fed

• Ed Elson · Prof G Media

  The Fed's 2% inflation target is no longer credible given years of running well above it.

  Ed says inflation has been well above 2% for several years, with no sign of a return. The Fed keeps repeating its 2% target, but Ed says he no longer takes it seriously because the gap has persisted for so long. He treats the target as a rhetorical position rather than a credible policy anchor, which in his view means inflation expectations and policy credibility are at risk.

  “We've been way off 2% for several years now... it's almost a joke that the Fed's target is 2%... They keep saying it, but I just don't really believe it at this point.”at 21:27

AI pricing & unit economics

• Scott Galloway · NYU Stern School of Business

  Token prices have fallen about 90%, so AI compute demand must rise tenfold just to stand still.

  Scott says demand for compute and tokens is still rising fast, but the price of tokens has fallen about 90%. Because providers finance GPUs with debt and earn revenue per unit of compute, a collapse in unit prices squeezes their economics. Unless demand grows around ten times, revenue only stays flat. He links this to the debt stress at neocloud providers like CoreWeave.

  “demand for compute or demand for these tokens is still skyrocketing, but the cost of tokens has been cut by about 90%. So unless the demand is up 10 x, you're running in place.”at 35:49

Power demand & the grid

• Scott Galloway · NYU Stern School of Business

  Power build-outs could overshoot, since most drawn-up data centers never get built, risking a power glut.

  Scott asks whether power stocks, which have surged, could face a glut. If chips become more efficient and demand falls, the power build-out could overshoot. He notes that 70-80% of data centers on the drawing board never leave it, which could cut demand for power companies such as Constellation Energy. He frames this as a question, not a firm call.

  “70, 80% of the data centers on the drawing board never leave the drawing board, and all of a sudden, you know, Constellation Energy sees its demand go down.”at 37:29

AI bubble & tech valuations

• Ed Elson · Prof G Media

  Memory chipmakers rallied hard but trade at only 3-5x forward earnings, so he is not worried.

  Ed acknowledges that memory makers such as Samsung, Micron and SK Hynix have had huge run-ups. On a multiple basis, though, he says they still trade at about three to five times forward earnings, which he does not find worrying.

  “I'm actually not terribly worried about the memory makers as well. Yes. They've had these huge run ups, but on a multiple basis, they're still trading at, like, three, four, five times forward earnings.”at 39:50

Software & SaaS

• Scott Galloway · NYU Stern School of Business

  The SaaS selloff is a short-term brushfire, since most SaaS spend goes to client relationships, not tech.

  Scott says SaaS companies spend at most about 20% of revenue on actual tech. The other 80% goes to UI, customer service, debugging, events, and insurance, which builds client relationships. This makes them more resilient than people think. Last week Shopify rose 5%, SAP 7% and Salesforce 9%. He says terminal values should be discounted, but the selloff is more of a short-term brushfire.

  “The data that always struck me that made me believe the SaaS apocalypse was more like a, I don't know, a SaaS brushfire, if you will, at least in the short term.”at 44:03

IPOs, M&A & deal activity

• Scott Galloway · NYU Stern School of Business

  Public company count has halved, and private valuations avoid accountability, so more firms should list.

  Scott says the number of public companies has been cut in half over 20 years, through acquisitions or staying private longer. He thinks more public companies are needed. Private markets are opinions, while public prices are facts. He argues private markets have become where valuations go to avoid accountability. He cites recent public investors rejecting an IPO priced on private-market vibes.

  “Private markets are essentially opinions, and public prices are facts.”at 45:14

Open-source AI models

• Ed Elson · Prof G Media

  Chinese subsidized open-weight models are deliberately eroding US AI businesses and hurting the US AI trade.

  Ed agrees with Scott's theory of Chinese AI dumping and says he is nearly certain it is intentional. He argues China is subsidizing open-weight models, which deteriorates the sustainability of US AI businesses and causes real pain in the US AI trade. He says the pain is already reflected in public AI stock prices.

  “I feel as close to certain that your theory of Chinese AI dumping is true, that it is intentional, and that we are seeing the sustainability of these businesses deteriorated by the hand of China who is subsidizing the shit out of these open weight AI models”at 54:35

Consumer & retail

• Ed Elson · Prof G Media

  Beer sales rebound may signal young people turning from dating apps back to in-person socializing.

  Ed links Constellation's strong beer sales at bars and restaurants with dating apps like Bumble and Match Group falling sharply. He wonders whether people are turning away from swiping and meeting in person. He hedges heavily and says the link may be overreading.

  “Maybe the tide is beginning to turn. People might be turning away from living their lives online and swiping. Maybe they're actually getting together in person.”at 1:02:47