
Nov 20, 2026 · 1h 5m
Listen from 36:21
Listen at 36:21
AI valuations collide with tougher tests for growth investors
Debating An Anthropic Buy; Nvidia's Bad Vibes Get Worse; SoFi, Nu, and Digital Banking Tailwinds
The discussion connects AI economics, IPO timing, portfolio assumptions, and fintech competition to the harder question of what deserves a premium valuation.
- 1Anthropic’s debut raises questions about optimistic growth forecasts, cloud costs, margins, and whether strong product usage predicts investment returns.
- 2Nvidia’s valuation, market-share outlook, cyclicality, and leadership controversy sharpen the risks of treating AI infrastructure demand as durable.
- 3MercadoLibre, SoFi, and selected payments companies illustrate how deposit growth, ecosystem strength, and business quality can outweigh cheaper multiples.
Don't miss
The hosts use historical IPO returns to challenge the instinct to buy immediately after a first-day pop, emphasizing lockups, insider selling, and staged entry.
The brief
Anthropic’s IPO becomes a test of whether personal enthusiasm for Claude and ambitious ARR forecasts can justify a demanding public-market valuation.
The hosts debate whether AI productivity gains translate into durable investor returns, while cloud costs, accounting choices, and defensibility complicate the economics.
Nvidia’s drawdown brings valuation, cyclicality, future data-center share, and leadership questions into focus; the discussion warns that severe portfolio stress can signal excessive position size.
A historical IPO pattern of first-day strength followed by weaker six- and twelve-month returns supports patience, lockup awareness, and staged buying in risky growth stocks.
The conversation widens to MercadoLibre, SoFi, housing stocks, Sea, Grab, and payments companies, favoring durable business quality over superficially cheap multiples.
What was said on this episode
52 statements · 29 positive · 19 negative · 1 mixed · 3 neutral
Ryan is unusually inclined to buy Anthropic based on its business.
“But as far as businesses go, I've never been more inclined.”
Listen at 6:20
Ryan spends approximately 90% of his work time using Claude.
“my work life, it's like 90% time spent in Claude.”
Listen at 6:33
Anthropic stock could crater if it misses the $200 billion ARR expectation.
“Expectations, though, are for $200 billion by the end of next year. That's saying like, all right, maybe returns will be slightly higher if they get there. But if they miss that, the stock's going to crater.”
Listen at 6:58
Claude substantially reduces manual financial-modeling work.
“it has taken out a ton of the grunt work”
Listen at 9:13
Claude reduces a multi-day email campaign task to roughly 30 minutes.
“Prior to Claude, I would have gone through and I would have done all that manually. it would have taken me four or five days and would have been an absolute headache. Plug an MCP in from Beehive. It's done in 30 minutes. It's such a time saver.”
Listen at 10:40
Locked-in workflows will drive higher usage and support high prices.
“Because usage will go up. And there are some things that are so locked in, like some workflows that are so locked in. people will pay outrageous prices for them.”
Listen at 11:02
Brett would avoid buying NVIDIA even at one times earnings.
“I don't even know if I'd buy it at one time's earnings. I'm serious. That is the absolute worst time you want to buy a cyclical.”
Listen at 11:26
NVIDIA earnings could collapse.
“The earnings could collapse.”
Listen at 12:15
NVIDIA’s upside outweighs downside risk at one times earnings.
“The upside 100% outweighs the downside at one times earnings.”
Listen at 12:27
Buying a cyclical company at peak earnings is especially dangerous.
“For a cyclical, that's when it's the most dangerous to buy”
Listen at 12:37
Anthropic’s reported gross margin may be overstated by revenue presentation.
“I think the gross margin, come on, it can't be as good as it is then.”
Listen at 14:59
Anthropic’s margins at scale are highly uncertain.
“no one has any idea what margins could look like at scale.”
Listen at 15:56
Anthropic could be attractively priced if it reaches 40% operating margins and sustains growth.
“If Anthropic gets to 40% operating margins, yeah. You know, buying a 10, what do we add? If you go on ARR, it's something like 15, 20 times. If you believe the growth will continue and they can get to really fat margins, yeah, it's a good price here.”
Listen at 16:11
Newly listed stocks often decline as IPO lockups unlock additional supply.
“stocks drop after the IPO because there's so much supply that gets unlocked.”
Listen at 16:52
The AI competitive landscape could look entirely different within a year.
“a year from now, we can look up at the AI situation could be entirely different”
Listen at 17:44
Claude coding users are unlikely to switch unless its model substantially worsens.
“If you're regularly coding on Claude, I imagine unless the model gets substantially worse, I can't see any reason why you'd switch.”
Listen at 18:06
AI products will eventually become commodities.
“The thing is, it will end up being a commodity.”
Listen at 18:57
Investors should buy risky hyper-growth stocks in thirds.
“buy in thirds for a hyper-growth stock that you think has some sort of risk that it could go down 50%”
Listen at 19:41
A position causing severe stress after a 20–30% drawdown is too large.
“if you have a position in, say, NVIDIA, or any other position, and a 20%, 30% drawdown like this is making you extremely stressed, the position is too large.”
Listen at 21:11
A company growing from hundreds of billions to trillions faces potential 30% drawdowns.
“any business in that scenario, a 30% drawdown is in the cards.”
Listen at 23:56
Companies in full control generally do not need to reassure investors verbally.
“When you are in full control, it usually does not need to be said. You don't need to tell investors we got this.”
Listen at 24:43
Jensen Huang resigned and stepped down.
“Jensen Huang did resign. He did step down.”
Listen at 25:52
NVIDIA’s board has no intention of reinstating Jensen Huang.
“they have no intention of bringing him back.”
Listen at 26:22
NVIDIA’s board will not bring Jensen Huang back soon.
“the fact that the board did not, tells me it's not happening.”
Listen at 30:28
Bullish Anthropic investors should wait six months because IPOs often fall 30%.
“If you're bullish in Anthropic, just wait six months. Almost every IPO bleeds like 30%.”
Listen at 31:25
Ryan presents return data for major IPOs across one-year periods.
“This is returns for the largest IPOs from day one to six months to one year.”
Listen at 32:05
Most large IPOs were negative one year after listing.
“the vast majority were in the red.”
Listen at 33:19
Anthropic may follow typical IPO weakness unless it greatly exceeds expectations.
“my guess here is that this probably looks similar unless they just outrageously beat expectations.”
Listen at 33:45
Anthropic employees should sell some newly liquid employee equity.
“Everyone should be selling some of that today.”
Listen at 35:00
Ryan used Claude to build models for every portfolio company.
“I asked Claude to build models for all the companies in my portfolio.”
Listen at 36:21
MercadoLibre and Amazon ranked first and second in Ryan’s projected portfolio returns.
“number one, Macaro Libre. You mentioned it. Number two, Amazon.”
Listen at 38:55
Adyen ranked third in Ryan’s projected portfolio returns.
“number three, you're going to like this one, Adyen.”
Listen at 39:01
Adyen is an underrated wide-moat stock trading cheaper than perceived.
“I think it's one of those underrated wide-mode stocks out there. And it's much cheaper than people think.”
Listen at 39:16
Ryan’s model projects a 24% annual return for Adyen through 2030.
“My model said it expects a 24% IRR or rate of return until 2030.”
Listen at 40:10
Payment processing relationships are highly sticky for businesses.
“payments processing in general is just so sticky for a business”
Listen at 40:23
SoFi deposits are growing rapidly in the United States.
“SoFi's are booming in the United States.”
Listen at 48:45
SoFi’s deposit growth indicates an opportunity for Nubank in the United States.
“I think it also indicates the opportunity with Nubank. in the United States”
Listen at 48:46
Nubank’s U.S. checking account offers a promotional 4% APY.
“the checking account earns the full 4% APY, the promo rate.”
Listen at 48:58
SoFi deposits could increase tenfold over the next decade.
“there is a massive runway to 10x that over the next decade.”
Listen at 49:53
Traditional banks will grow deposits much more slowly than SoFi and Nubank.
“they're not going to grow at the same rate or anywhere near it as SoFi, Nubank.”
Listen at 51:25
SoFi has a steady long-term growth tailwind, not explosive billion-user growth.
“that's just a steady growth runway for the SoFi. It's not going to be Instagramming it to a billion users in five years. But there's a steady, steady tailwind at your back.”
Listen at 52:21
SoFi set a new net-member-additions record in each of the last four quarters.
“the last four quarters, they have broken their net member additions record every single time”
Listen at 52:44
Builders FirstSource could generate its market cap in cash within five years if housing normalizes.
“they, I think, could generate their entire market cap in cash within five years if the housing market kind of starts ticking up or normalizing a little bit.”
Listen at 54:12
U.S. home affordability for median earners is at its worst level ever.
“affordability on the average home price for the median income earner has never been worse in America.”
Listen at 54:45
Brett prefers housing-cycle-levered stocks such as Builders FirstSource or Real Brokerage.
“I'd rather go for something a little more levered to the housing cycle, such as a builder's first source or the real brokerage.”
Listen at 56:28
MercadoLibre has a more durable model than Shopee.
“I think MercadoLibre has a more durable model”
Listen at 58:03
GTA 6 will hurt other gaming stocks during the next quarter.
“That's going to ruin every other gaming stock for the next quarter, basically.”
Listen at 58:54
Remitly, Wise, and Adyen have attractive models in an oversold payments sector.
“I think both. I like the business models of those three companies, and I think the industry is, the sector is way oversold.”
Listen at 1:00:55
Digital remittances continue gaining market share.
“digital remittances have continued to just take share.”
Listen at 1:01:16
Toast has a strong, restaurant-specific business model and leading product.
“Toast, I like. We just did an episode on them. good business model, very vertical specific. They've built number one product for restaurants.”
Listen at 1:01:38
Ryan considers Block mismanaged and will not own Jack Dorsey-led companies.
“Block mismanaged. I will never own anything that Jack Dorsey is running.”
Listen at 1:01:47
Investors should not choose a much lower-quality business for a small P/E discount.
“Don't just go for a slight PE discount for something that is a way worse quality business.”
Listen at 1:02:20
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.