Skip to content
Anthropic's Valuation Gap: Addressing Sales Runway, GOOGL & AMZN Concentration Fears artwork

Anthropic's Valuation Gap: Addressing Sales Runway, GOOGL & AMZN Concentration Fears

Schwab Network

October 26, 2026

MarpoleAI's Dmitri Maxim discusses Anthropic and the mounting questions surrounding the company's valuation after releasing its S-1. That same release showed a reported $80.7 billion in losses during the first six months of 2026.

Speakers Sam Vardis, Dmitri Maxim, Marlee Caden

TopicsInvestingBusiness

Sam Vardis (0:00)

Welcome back to Market on Close. I'm Sam Vardis alongside Marlee Caden. Joining us now to discuss the details of Entropix S1 is Dmitri Maxim, Forensic Financial Analyst over at MarpoleAI.

Dmitri, thank you so much for your time today. So we got some leaked information in recent weeks about this from the likes of the Journal and Voiters. So we did get some clues, but certainly this has shed more light. So what did you take away from what we're learning now about the strategy, the financials, and the risks here?

Dmitri Maxim (0:32)

Thank you very much for having me here.

I wanted to start with, as I did in the previous conversation, I wanted to start with a picture that would give you a total overview. Imagine a private card game, four players, Amazon, Google, NVIDIA, Anthropic. They're all passing chips, and every time they pass, the value goes up. In March 2025, the valuation was $61 billion.

In May 2026, it's $965 billion. So now, the valuation is how it's presented, it's $2 trillion. So there is quite a few questions that we wanted to discuss today. Like how this valuation come, what is based on, and whether this valuation has any solid background underneath of it.

Marlee Caden (1:32)

So let's dive in to those questions then, Dmitri. I mean, with this valuation, how do you conclude from the details that we have, whether this is actually an attractive investment, or if this is just another AI hype trade?

Dmitri Maxim (1:50)

In order for us to see whether it's a real valuation, or this is some kind of way of pumping the valuation, we need to look at who is doing this. So if you look at the filing, there is no any pricing over there. So there is no any trades, private, I mean public trades that would set the price on it. So as I mentioned before, Google, AMZN, NVIDIA and Anthropic, they invested in, I mean, they're all interconnected.

AMZN and Google, they own part of it and sell it computing power and almost half of last year's sales came through them.

Price goes up, they booked the profit. And by the way, AMZN booked about $53 billion in profit in the first quarter because of the reevaluation of Anthropic. And nobody outside of the table has ever paid the price. Now 22 banks ask the public for $2 trillion, which is double.

And the people who sit next, I mean, the strangers who join the table, they are asked to pay this price, whether it's a sustainable price or not. That's a very big question.

Sam Vardis (3:15)

Are there any concerns around the concentration risk when it comes to the top five customers here?

Dmitri Maxim (3:23)

Obviously, the more diversified the revenue stream, the better. Obviously, Google and Amazon, they're very big companies and they provide, I would say, enough assurance. But when we look at the forecast, we need to look also at not only the value that this company is bringing but also its liability. So if you look at the devaluation, so devaluation is done in such a way so that they took projected 2028 revenue and multiply it by 10

So it's like they projected it's going to be about $195 billion and they multiply it by 10, it comes to $2 trillion. But if you take today's sales of $65 billion, the devaluation of the company comes to $650 billion.

Same company, same day, only the number changes. So the gap between $650 billion and $2 trillion is $1.35 trillion of the price is based on the guess that two years from now.

Marlee Caden (4:45)

So Dmitri, if we're looking at guesstimations essentially is what you're throwing out here, how sustainable then is this quote unquote projected growth with the increased competition? I mean, we've got OpenAI, we've got Google, we've got other AI names popping up. We've got a global competition, of course, that needs to be discussed here. How do they have enough differentiation between their Claude models to sustain the kind of growth that they're projecting based on past growth where at certain points they only had one primary competitor?

Dmitri Maxim (5:22)

So let's look at this, the image that we get sold. Sales explode, so the profit will follow. Reality, the demand is real, like 386 million in 2024 and 4.6 in 2025, and 16 billion dollars in first six months of this year. So 12,000 customers pay over $100,000 a year, but almost half came from Amazon and Google.

The players at the table, the first real profit last quarter is $974 million.

So, but if we look at the number of investment, I believe a couple of 10 days ago, 10 days ago, Gartner, they issued a report that by 2028, I believe 2028, the AI industry would need to have like 8.6 billion, trillion dollars in order to cover all this cap expenditures. And they project that the revenue model of AI industry would reach only 3.6 trillion dollars. So there is almost like a 5 trillion dollar gap, which nobody knows how to close. So the sales proof, people want AI, but I'm not sure if they want the price. And if you take into consideration that these frontier models like OpenAI, Anthropic, they have huge investment into the infrastructure and we can discuss it if you want, the reason why it is so. But there is also models like say, Chinese models that are like say, three, six months behind and they're almost at par and they're offering their model almost for free. So it put a substantial in my kind of like view, if you have a competitor who's basically offering the same product, almost the same product for free, I would be very, very much worried about that.

Read the other 1 minutes of this episode here.

$15 gets you this episode and 99 more. Credits never expire. After checkout, one link brings you back here to read it.

Not what you expected? Full refund within 14 days.

Thousands of transcripts fetched by people building searchable podcast archives

Fetch transcripts with code

The same credits work from a script or an AI agent. The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000798247489