Oct 30, 2026 · 1h 23m
Anthropic’s growth tests the limits of the AI boom
Josh Brown: If This Is A Bubble, It’s Pathetic
Anthropic’s filing puts the AI market’s central tradeoff in view: extraordinary demand requires extraordinary spending, while durable profits and competitive moats remain unproven.
- 1Anthropic’s roughly 1,300% revenue growth is impressive, but customer retention and workflow lock-in matter more than headline momentum.
- 2Compute commitments can turn falling per-token costs into rising total bills, leaving AI companies exposed to capital-market conditions.
- 3Big Tech scale may strengthen market leaders during high-rate periods, even as AI concentration raises concerns about inequality and jobs.
Don't miss
The hosts debate whether Anthropic has a nonzero chance of becoming the world’s most valuable company within several years, despite its spending and competitive risks.
The brief
Josh Brown joins Ed Elson and Scott Galloway to examine Anthropic’s S-1, where explosive revenue growth collides with heavy losses, massive compute commitments, and an uncertain competitive moat.
The core question is not whether Anthropic can grow quickly, but whether customers will keep its models embedded in their workflows as rival systems improve and usage costs evolve.
The hosts debate whether Anthropic could eventually become the world’s most valuable company, weighing product advantages against infrastructure needs, financing risk, and competition from OpenAI and Big Tech.
The AI-bubble discussion widens into a neocloud stress test: falling financing availability could expose infrastructure bets even as major platforms benefit from scale, pricing power, and high rates.
The conversation then turns to Amazon, robotics, inequality, Netflix’s changing expectations, and Brown’s view that market gains could remain supported through year-end.
What was said on this episode
70 statements · 42 positive · 19 negative · 7 mixed · 2 neutral
Anthropic is on pace to exceed $30 billion in third-year revenue.
“This company is on pace to do something on the order north of $30 billion in its third year in existence.”
Listen at 5:50
Investors are more concerned about Anthropic’s revenue durability than its growth rate.
“I think what investors are more worried about, and maybe this is. the more reasonable bear take on Anthropic and AI in general is the stickiness of the platform and the durability of that revenue, more so than the growth rate.”
Listen at 7:56
“I think everybody would be okay with that kind of guidance”
Listen at 8:27
A slightly more advanced model could take Anthropic’s customers and workflows.
“The bigger issue is how easy is it for the next model that's 10% more powerful or advanced to snatch away those customers and those workflows?”
Listen at 8:31
Anthropic’s durable competitive moat is uncertain.
“Is there a moat?”
Listen at 8:54
Anthropic’s moat is uncertain, weakening confidence in a $2 trillion valuation.
“Is there a moat? Is there a moat? I wish I could tell you definitively one way or the other, but that's the thing that might give me pause in saying, yeah, you're right. This company's worth $2 trillion right off the bat.”
Listen at 8:54
During a land grab, pursuing scale can matter more than immediate accounting profitability.
“sometimes in certain periods of history, when companies have come to the realization that they're in a land grab, the more important thing than accounting profitability is grabbing as much land as possible so that you can eventually generate. the profitability that everyone wants.”
Listen at 14:08
Aggressive scale-building can precede and enable eventual profitability during a technology land grab.
“the more important thing than accounting profitability is grabbing as much land as possible so that you can eventually generate. the profitability that everyone wants.”
Listen at 14:16
Anthropic may not replicate the successful land-grab strategies of Malone and Bezos.
“And in both cases, it worked. It doesn't work all the time. It's not a guarantee that Anthropik could pull it off.”
Listen at 15:25
Anthropic may not successfully convert aggressive spending into eventual profitability.
“It doesn't work all the time. It's not a guarantee that Anthropik could pull it off.”
Listen at 15:27
Anthropic should not prioritize 25% quarterly profit margins during its land grab.
“if I were, and Dario gave a shareholder update and said, we are projecting. 25% profit margins in Q4. I would say, what the fuck are you doing?”
Listen at 15:44
Sam Altman likely views Anthropic-style spending as a race to scale and users.
“I think that's the way Sam Altman's thinking.”
Listen at 17:09
Anthropic is pursuing the strategy its shareholders expect.
“I think they're doing what the shareholders expect them to do.”
Listen at 17:28
Anthropic’s early objective should be strategic spending rather than profitability.
“This is like, this is year three. It could be a 30-year battle for supremacy. In year three, the goal is lose money but make it count for something.”
Listen at 18:04
Anthropic lacks SpaceX’s level of competitive moat.
“I don't think they have the moat that SpaceX has.”
Listen at 21:15
OpenAI will go public in the first quarter.
“OpenAI will go public in Q1.”
Listen at 22:44
OpenAI, Anthropic, and Gemini could enter a severe price war.
“They could rip each other to shreds for the next three years and Gemini in the mix. Like, it could be really – it could be ugly. It could be a price war.”
Listen at 23:00
Anthropic’s software improves through recursive learning from user interactions.
“It is now undergoing. this process of recursive learning. It's teaching itself to get better.”
Listen at 25:04
Anthropic’s software learns from hundreds of millions of users and personalizes its behavior.
“It is learning the habits of hundreds of millions of individual daily users and conforming the way it works to those habits on an individualized basis.”
Listen at 25:12
Anthropic retained about $112 billion of the $141 billion it raised.
“So they've raised $141. They're still holding on to $112.”
Listen at 26:43
Anthropic has committed to approximately $800 billion in future compute spending.
“they have already signed up to spend $800 billion”
Listen at 27:56
Anthropic is better off overcommitting to compute than lacking sufficient capacity.
“if they get caught without enough compute, it's probably worse. than over committing.”
Listen at 28:33
Anthropic may acquire a distressed neocloud at a low price.
“Maybe once... all of the wreckage of the neoclouds has been sort of piled up. Maybe they'll go in there and buy a neocloud on the cheap.”
Listen at 29:57
Anthropic’s Claude Enterprise business is more advanced than ChatGPT’s enterprise business.
“the enterprise business at Anthropic, the Claude Enterprise is much further along.”
Listen at 31:01
Per-unit compute costs are falling sharply.
“The per unit cost of compute is going down. It's actually crashing.”
Listen at 32:39
Total compute spending is rising despite falling per-token costs.
“the dollar cost of compute is not falling. It's rising.”
Listen at 33:19
Falling customer spending per token will slow Anthropic and OpenAI revenue growth.
“If that per token spend is also going down, which it is, then that is also going to be a problem because it means that their revenues are going to start to slow.”
Listen at 34:21
Anthropic could become highly valuable but faces a significant bankruptcy risk.
“it could be an incredibly valuable company, but I think they could also get caught in a very bad situation. And there is significant risk that they don't keep up with their costs and eventually they could go bankrupt.”
Listen at 35:19
Anthropic could fail to control costs and eventually go bankrupt.
“there is significant risk that they don't keep up with their costs and eventually they could go bankrupt.”
Listen at 35:27
Enterprises will impose company-wide budgets on AI usage.
“there's going to be some sort of budget on AI use company-wide.”
Listen at 36:51
Open-weight AI models may create effectively zero switching costs for enterprises.
“the free weight, the open weight models, the free models. Just this idea that there actually are no switching costs”
Listen at 37:18
Enterprise customers face no switching costs between AI models.
“there actually are no switching costs”
Listen at 37:24
Enterprise model-routing tools will improve AI-use efficiency.
“all of a sudden, you start seeing companies get smarter about the efficiency with which they utilize these tools.”
Listen at 38:05
Enterprises will increasingly optimize AI usage and control costs.
“companies get smarter about the efficiency with which they utilize these tools.”
Listen at 38:07
Enterprise efficiency and cheaper models threaten premium frontier-model providers.
“That's obviously bearish if you're trying to sell the premium. furthest out on the frontier model at a premium price.”
Listen at 38:14
Anthropic posted positive adjusted operating income.
“this is a company that, as far as I can tell, posted positive adjusted operating income.”
Listen at 42:25
Anthropic could become a major cash generator if growth persists.
“if they can navigate this landmine and continue to show a fraction of the growth they're recognizing right now, this thing becomes a cash spigot”
Listen at 43:53
Anthropic’s valuation reflects both growth and existential risks.
“I think we're sort of doing both at the same time.”
Listen at 45:23
Anthropic would reach a $2 trillion valuation even after a $1.5 trillion IPO.
“If it gets cut down to one and a half, it'll get to two.”
Listen at 45:32
Anthropic’s enthusiasm will peak around its IPO.
“I think the enthusiasm around this company will be peaking right as it goes public.”
Listen at 45:35
Anthropic’s IPO timing should support strong initial trading.
“I think this company is coming public just as the excitement around it is peaking. And that's a good thing. That bodes well for the first few days of trading.”
Listen at 46:29
Anthropic could easily raise $100 billion through its IPO or private market.
“So do I think they could raise $100 billion? Yeah, like with their eyes closed.”
Listen at 47:15
Anthropic could raise $100 billion easily in public or private markets.
“I think they could raise $100 billion? Yeah, like with their eyes closed.”
Listen at 47:15
The current AI market is an anti-bubble rather than a valuation bubble.
“It is an anti-bubble.”
Listen at 48:46
AI stocks are not experiencing a valuation bubble.
“You cannot say it's a valuation bubble.”
Listen at 48:59
Tech earnings are estimated to grow 65.6% year over year this quarter.
“The year-over-year estimate for tech earnings growth is 65.6% for the quarter that we're in now.”
Listen at 49:13
Technology earnings growth excluding semiconductors is estimated at 24.9%.
“If you exclude semis and semi-equipment, it falls to only 24.9%.”
Listen at 49:30
No semiconductor stocks have made 52-week highs since July.
“There has not been an above average reading, which I would say is like 70% or higher names making new highs. Since July, there are 0% 52-week highs for the semis.”
Listen at 50:09
SanDisk’s PEG ratio is 0.08.
“Sandisk has a price to earnings growth or peg ratio of 0.08.”
Listen at 50:29
Micron’s PEG ratio is 0.10 and Western Digital’s is 0.13.
“Micron, 0.10. Western Digital, 0.13.”
Listen at 50:41
Micron trades at five times earnings.
“Micron is selling at five times earnings.”
Listen at 50:45
NVIDIA trades at 15 times forward earnings.
“NVIDIA's forward PE is 15 times earnings.”
Listen at 50:57
NVIDIA’s forward-earnings valuation is historically low.
“This is the lowest valuation, I think, in history on a forward earnings basis you have ever had the opportunity to buy NVIDIA.”
Listen at 51:01
AI-related markets will experience a correction or setback.
“my expectation is that we're going to see some sort of correction. We're going to see some sort of drawback. There'll be some problem.”
Listen at 53:06
Many neocloud companies will run out of money.
“for a lot of them, they will.”
Listen at 53:23
A neocloud debt shakeout would benefit AWS.
“if you think there's going to be a shakeout in the neoclouds that are spinning up all this compute and are doing it with debt capital, and you think that that debt capital will be less available or unavailable, and that is a worst case scenario for that part of the market, who is it good for? Well, if you're AWS”
Listen at 53:49
AWS’s backlog will grow if neocloud companies fail.
“Your backlog will only grow.”
Listen at 54:46
Neoclouds’ inability to raise capital below 15% benefits Microsoft and Google Cloud.
“how is it bad if the NeoClouds can't raise money at less than 15%? That sounds great to me.”
Listen at 55:04
Amazon’s robot-as-a-service business could become significant.
“it could be big.”
Listen at 1:01:07
Higher compliance and energy costs create moats for the largest companies.
“Those become moats for the biggest businesses because they are the only places that can appropriately staff with the experts.”
Listen at 1:02:44
Economic difficulties can increase large companies’ comparative advantage and market share.
“sometimes, quote unquote, struggles in the economy actually end up being tailwinds for the largest companies in each sector because they actually take on a comparative advantage and they take market share from everyone else.”
Listen at 1:03:06
Anthropic will create thousands of multimillionaire households.
“Anthropic is going to create what all is said and done. Thousands of multimillionaire households.”
Listen at 1:04:51
Anthropic’s compute commitments could create millions of jobs.
“When they commit to something like $800 or $900 billion in compute, that all has to get built. You're talking about millions and millions of jobs.”
Listen at 1:09:22
Netflix is disadvantaged by comparisons with YouTube’s viewing-time metric.
“I think the big problem for netflix they're being compared to something that they can't do anything about”
Listen at 1:11:51
Netflix earns approximately $3 billion per quarter and is highly profitable.
“Netflix is hugely profitable. They're earning like $3 billion a quarter.”
Listen at 1:12:46
Netflix should not imitate YouTube with podcasts and video games.
“What's not going to work is trying to be YouTube light and just throwing podcasts on the platform and video games.”
Listen at 1:13:47
Netflix should prioritize higher-quality programming.
“I think they should, they should do what Paramount's about to do, which is double down on the quality of the programming”
Listen at 1:13:55
Netflix needs must-have programming to strengthen viewer demand.
“You need to have must-have programming that people get pissed off when everyone's talking about it and they didn't see it.”
Listen at 1:14:40
A double-digit midyear S&P gain implies an 89% chance of a December high.
“When the S&P is up double digits. through the midpoint of the year, you have like an 89% chance that the high of the year will take place in December.”
Listen at 1:16:51
The market and Anthropic could be higher by Thanksgiving.
“Come Thanksgiving, people are sitting around looking at their portfolios. They're up 15%. And anthropic price as well. Like, that's the setup.”
Listen at 1:18:10
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.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
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.