We learned, and a lot of people said, 'Oh my gosh, this is so great that Anthropic is not spending on capex,' but they admit to the problem that you're describing, which is that they are leasing compute from their competitors primarily, which is probably going to be an issue later down the line. That tells me they're about to build data centers and they're about to spend a ton on them. Is that right?
Right. That's exactly right. 'A ton' being the technical term here. Oh, this is also a fancy way of saying, or maybe a relatively opaque way of saying, that what you see in the past in terms of the structure of the business won't be representative of what it looks like in the future. And so that's what you're saying — saying the nature of what we spend in future not only is going to continue, but it's going to accelerate in terms of the capex, because we're in this strategic quandary of being reliant on our largest competitors, which tells me that the income statement is actually irrelevant. Right.
Welcome to Prof G Markets. I'm Ed Elson. It is October 27th. Let's check in on yesterday's market vitals. The major indices climbed as tech stocks got a boost from Anthropic's S-1. More on that in a second. The yield on 10-year Treasuries was flat ahead of the Federal Reserve's interest rate meeting. On Kalshi, the odds they'll hold rates steady are at 92%. Brent crude was volatile, peaking above $117 per barrel as President Trump said he was deciding whether to make a deal with Iran or, quote, 'finish it off.' And finally, Brazilian markets sold off after incumbent President Lula won re-election. Brazil's benchmark index, the Ibovespa, crashed as much as 17%, and its currency saw its sharpest drop in more than two decades. Okay, what else is happening?
One of the most anticipated S-1s of all time has officially arrived. Anthropic has filed to go public, and we finally got to see the books. In the first six months of the year, Anthropic brought in $16.3 billion in revenue. That is up from $1.2 billion in the same period last year. It reported a loss of more than $80 billion over those six months. However, $72 billion of that loss is a non-cash charge tied to its convertible notes and other fundraising instruments. Its operating loss, which is more important here, came to around $700 million. That is down from $2.1 billion a year ago. While the filing answered many questions about Anthropic's business, it also left out some crucial details. For example, the company did not disclose the total sum of its contractual compute obligations. It also didn't reveal who its biggest customers are. It didn't break out training costs either. And it didn't say when Anthropic will actually be listed. So, lots to get into here. We are breaking this down with Paul Kedrosky, managing partner at SK Ventures. Paul, great to have you on the show. Lots to get into here. I'm going to start with an open-ended question, which is: what did you make of this S-1? It's finally here.
I felt like it was like an MCU, like a Marvel movie, and I'd seen the trailer too many times. And so I'm like, 'Oh god, there's no spoiler. There's nothing left. It's all been spoiled.' So an awful lot of what was in this was pre-leaked, right? So some of it was obviously in a series of Reuters reports, Business Insider — it's been all over the place. And so I guess the surprising thing for me, for the most part, was how few surprises there were. That for the most part the major nuggets in terms of, you know, customer concentration, the deceleration in revenue this quarter — which was a little bit masked actually in the filing itself, but you nevertheless you can kind of back it out — those kinds of things had all been sort of pre-leaked, as I say, as long as I had seen the trailer and it was too long. So that sort of stuff was there. The other things, like customer concentration, I had expected a little bit more data on that. So like last year, I think their two largest customers were something like, from the disclosure table, I think was 25%. And this year they kind of masked out a lot of the names, and the S-1 showed that their largest customer is sitting at around, I think it was 11 or 12%, and then everyone else is kind of masked out, so you don't really know where it stands. So all we can assume is that customer concentration — it may be that there's 10 companies that are 25%. We just don't know anymore because they're only showing people over 10% share. I think the big question that we were sort of speculating about is how unprofitable will they be, and will they pursue some sort of accounting gymnastics to mask their lack of profitability. And I think that that is an open question, and we will get into that. But from the outset, when we look at the operating losses — and just to clarify for people, that net loss of $72 billion, that is related to the conversion. That is genuinely not relevant, because basically what that is accounting for is the fact that they now need to pay back investors because their enterprise value has gotten so large. So actually the fact that that net loss is so big reflects the fact that their valuation has gotten so high. So that's not the thing to focus on. The thing to focus on is the operating loss of nearly $700 million, which I thought was low, or at least narrow, compared to the six months compared to last year where it was more than $2 billion. That to me was a pretty good sign. I have some pushback to why that might be a good thing that we can get into, but your reactions just to that number — $693 million in operating losses in the first half of this year.
It was narrower than I expected as well. So again, it's hard to know. I mean, having been on the other side of this, and not to be completely cynical, there's a lot of work that goes on in terms of preparing a company's numbers for an IPO. So the thing that they wanted to show, probably show their best face on, was what that operating number looked like. And so that's not particularly surprising, I guess, that it was a little better than expected. But nevertheless, like, full credit to them. I think this is nowhere near profitable, and it's probably going to get worse in future because of a bunch of reasons we can talk about. But the number they presented there was — I was thinking we'd probably see a number around 1.1 billion, on that order of things. And so we're about 400 million shy, in a good way, of that number, for all kinds of reasons they can control. So it's not necessarily something — this isn't written as a law of physics that it has to stay at these levels, but nevertheless it's slightly narrower than expected. There were other things that I thought were interesting that they did that are not nefarious. They're just things you need to keep an eye on, like non-cash things, like adding back stock-based compensation. That was a material item in there. Anthropic is very, how shall I say, generous in terms of stock-based compensation, and that will probably continue post-IPO. So it's something that investors will have to keep an eye on because that's dilutive, obviously, and as a public company, one of the consequences of having such high stock-based compensation is that you'll have to buy back a lot of stock to avoid dilution of your institutional — well, all investors, but institutional investors in particular. So that item actually is consequential going forward, because in terms of your sources and uses of cash, it's not all going to be going out the door again for funding compute commitments. You obviously have this very large number that they're going to be having to deal with in future in terms of trying to continue to protect investors, if you will, from future dilution, given how aggressive they are in terms of stock-based compensation. Credit to them for not giving us a completely, you know, non-GAAP adjustment where I remember you and I were speculating: are they going to strip out the training costs? Are they going to say, 'Oh, don't worry about what we have to do in terms of training these things'? They didn't do that. However, they didn't break it out cleanly, but they didn't try to do something non-GAAP in terms of hiding it.
That's exactly right. Yes. They didn't try to say, 'Don't worry, it doesn't exist. It doesn't matter.' But so I was looking for — okay, well, maybe this is going to work. Maybe compute costs are actually like they're on a path to profitability, which seems, you know, plausible and, to be honest, imminent given what we saw. But then I go through the document. I want to get your reactions to what I make of this. And I get to the liquidity and capital resources section. This is like page 161. It's way further down. And they go through some of their future compute agreements, and they show us that they have an agreement with Broadcom with $161 billion and Google $115 billion. And you add it all up and you realize that they are about to spend — on a timeline that they don't tell you — but they're about to spend more than $800 billion on compute.
That's correct. That's right.
And none of that is included in the income statement.
And so I'm looking at the income statement now and I'm thinking, does this actually tell me anything? Because it seems as though they've left all of their compute costs at the bottom as a footnote, not because they're trying to pull a fast one, I don't think, but because those costs haven't actually been reflected in their income statement yet, because they're about to happen. So what do you make of that part?
I think that's exactly right. And this is a deep structural problem that I think Moody's pointed to a few weeks ago in terms of the way that all of these compute commitments trickle back to and are consolidated in the financial statements. Because, well, they're non-cash obviously now because they're future commitments, but they don't even show up necessarily in the form of orthodox debt, because these are future commitments that aren't actually debt instruments. They're more like a lease that's being discounted back to their present value today. And they're all obviously highly consequential, and investors would need to think about it as you consider what the actual cash flow nature of the business looks like going forward. Because in particular, it's not just the size of the commitments — they're all non-cancellable, right? This is really important. This is not something where I say, 'You know what, it's not working out the way we wanted, we're not going to have that commitment.' Oh, these are all non-cancellable capital commitments, with the only exception being that there's a growing capital commitment in terms of doing compute at SpaceX, which I think works out to about $3 billion a year. So much smaller in the great scheme of things than these other non-cancellable compute commitments. And that I think is cancellable on a 90-day basis, 90-day notice, which speaks more to what SpaceX is trying to do, I think, than anything else — that they're desperate to have this cash flow coming back in because they in turn consolidate it as growing AI revenues when it's really not. Obviously, this is just compute you're providing to someone else on a rental basis. So it's important, I think, to remember that the preponderance of the commitments are in future. They're massive and they're not all non-cancellable.
And even if you were to get rid of SpaceX, you're close to $700 billion in compute spending.
That's exactly right. It's immaterial in terms of the great scheme of things. And I think that's by far the most important factor here that they're looking at. And in turn, those compute commitments are even more important because in a sense, you're working through distributors. And they say this in the S-1 itself. And so they're distributing their large language models through competitors, right? Microsoft is one of the prime examples, Google and so Amazon. And this is a really unusual relationship because you're distributing through companies who also sell on your behalf but also take a rake off the top as they do it. So my analogy is this is a little like as if Spotify, which takes a rake from the music companies of about 30% of their revenues, but instead of it being 30% of revenues, it was a fixed percentage — it doesn't matter how much you sell. Not only that, all of your competitors also have music streaming services and they're perfectly happy to kick you out. So this is the deep fragility and the nature of the distribution business. And I use the Spotify example because it shows you how a company that's been penalized for this kind of relationship is actually in a much more favorable position than Anthropic is. So this brings us to what I think is the most illuminating part of the filing, which is that we learned — and a lot of people said, 'Oh my gosh, this is so great that Anthropic is not spending on capex,' which has obviously been a huge expenditure for the hyperscalers, i.e., building their own compute, building their own data centers. And everyone said, 'Look, this works because they don't have this capex issue,' but they admit to the problem that you're describing, which is that they are leasing compute from their competitors primarily, which is probably going to be an issue later down the line. And so I'd like to just read you this quote from the S-1: 'Our increasing efforts to build out our compute and infrastructure require substantial capital expenditures and are changing the cash flow profile of our business. Whereas the compute purchased through our CSP agreements is generally paid as the compute becomes available to us, we expect an increasing portion of our compute and infrastructure investment to take the form of capital expenditures and other commitments, including lease obligations that require significant upfront payments well in advance of the related compute becoming operational.' That tells me they're about to build data centers and they're about to spend a ton on them. Is that right?
Right. That's exactly right. I should be using the technical term here. So this is also a fancy way of saying, or maybe a relatively opaque way of saying, that what you see in the past in terms of the structure of the business won't be representative of what it looks like in the future — that we can't maintain these lease relationships with our competitors in terms of passing product through. So we're going to be forced to make duplicate capital investments so that we're not beholden to them, both in terms of strategically, but also in terms of this rake off the top in terms of what we have to give up from what's called marketplace sales, where we're actually selling products through the marketplaces on these other services. And so that's what you're saying — saying the nature of what we spend in future not only is going to continue, but it's going to accelerate in terms of the capex, because we're in this strategic quandary of being reliant on our largest competitors, which tells me that the income statement is actually irrelevant, right? I mean, if that's not even close to a reflection of what this business is about to look like, why does it matter?
Yeah. And I think that's the thing that people are missing is that in a sense, we all know that income statements and financial statements in general are backwards-looking documents, but this is backwards-looking at a particularly sharp inflection. And the sharp inflection is from a business that was largely relying on distributors to a business that's increasingly going to be trying to go direct. This is the transition that's underway. And in going direct, it has two consequences. One is massive current capex for doing that. And second — and this is the thing people miss all the time, and kudos to Michael Burry and others for pointing this out — is that it creates maintenance capex. You don't have to spend it just once. You have to maintain these things. And so you have a future capital obligation that's not just the amount you spend to build, but also the amount you have to spend to maintain these very expensive facilities. But they're in this Hobson's choice. They have no choice but to do that, because otherwise it's to be reliant on their largest competitors who could, you know, like Microsoft did back in its heyday, you know, cut off their air supply at any moment.
So here's a little conspiracy theory that I'd like to see if you think is perhaps correct. Is it possible that this is why they wanted to go public now? Is it possible that they wanted to take this one moment in time where operating profitability looks pretty good right before it's about to look quite bad? Because that seems like that is what is about to happen. They're at this inflection point.
I was arguing that earlier today multiple times with people. So I think that's exactly what's going on. This is as good as it looks from the standpoint of the business as currently structured. And now you go through this inflection that makes it much more capex-intensive. So you're essentially changing jockeys and changing the underlying structure of what's happening. And so I think this was probably their — and they know this — this is the best time to do it because the future business is going to be very different. They're essentially reduplicating AWS, Amazon Web Services, in both directions, building up and building down in terms of both the client customer-facing as well as the cloud-facing stuff. It doesn't mean it's impossible. It could work, but it does mean that the way the business is characterized in this current statement is not representative of the future.
What would it working look like in your view?
Well, I think it would have to look like you'd see, for example, one of the statistics they want to tout in the document is the declining compute intensity of token production — that it's been coming down, something like it's fallen like 50% or something like that. And so what the business would have to look like is collapsing cost in terms of compute intensity of the underlying business, and that's the only way that the margins are supported and could work, because otherwise you just simply are going to eat the business on the basis of the compute costs associated with it. So that has to continue to decline. The more troubling piece is that there's a revenue deceleration underway. If you looked at the leaked Q3 numbers in the context of everything else, they're now looking at something like 50% revenue growth, but arguing that in Q4 they'll see a spike, it's implied, but back to about 64% or so. So you have to see that reversal in the fortunes of the business. So one is declining compute intensity, one is a reacceleration of growth. And that a company coming like this to market needs already to reaccelerate growth just blows the top of my head off.
Final question. This is set to go public. The reported valuation is going to be $2 trillion. There has been some debate as to whether that will actually materialize.
Yeah, 1.5 or so, I keep hearing. Yeah.
Yeah. What do you think will happen here? Do you think this goes out at $2 trillion, $1.5 trillion? Does it go out at all?
I think if it goes, it goes at 1.5 or so, but I reserve the right — and this is my long-standing claim — is the US won't have a functioning government for the balance of the year. So who knows what will actually be able to come public and what won't.
It's not very helpful, Paul.
It's not my fault that we don't have a functioning government. Blame politicians. But nevertheless, five assuming we have a functioning government, zero if it doesn't.
Paul Kedrosky is managing partner at SK Ventures. Paul, always appreciate your time. Thank you.
We'll be right back. And if you're enjoying the show, be sure to subscribe to the Prof G Markets YouTube channel at the link below.
Support for the show comes from SoFi. Here's the thing: most traditional banks pay almost nothing in interest, so your money ends up just sitting there twiddling its thumbs, doing nothing. And if your money is doing nothing, it's not doing its job. You work hard for every dollar, so why not put those dollars to work for you? That's where SoFi comes in, with the everything app for smarter money. With SoFi checking and savings, you can easily track your spending and save towards your goals while earning nine times the national average savings rate. You'll get paid up to two days early with direct deposit, and you can access additional FDIC insurance up to $3 million on your deposits through a network of participating banks. You'll pay no account or overdraft fees, and you can also earn a cash bonus when you sign up with eligible direct deposit. You also can organize your savings into vaults for the goals that matter most. Small decisions like this can create big momentum over time. You can open your SoFi checking and savings account today. Head to sofi.com/profg to get started. That's sofi.com/profg. SoFi checking and savings is offered through SoFi Bank, N.A., member FDIC.
We're back with Prof G Markets. Markets are beginning to feel a chill in the air. The Nasdaq 100 fell nearly 3% on Thursday, its worst day since June. As worries about how big tech will pay for the AI buildout hit tech stocks, the same fear showed up in the corporate bond market. Corporate bond yield premiums surged to their widest levels since May 2025. Meanwhile, investors moved back into Treasuries, pushing the 10-year yield below 5% for the first time in a month. Now, stocks did rise yesterday, but the S&P and the Nasdaq are coming off of two weeks of losses, with big tech earnings and a Fed meeting still ahead this week. So the question for investors: is winter coming for the markets? Joining us to discuss this, we're speaking with Robert Armstrong, US financial commentator for the Financial Times and author of the Unhedged newsletter. Rob, good to see you. You wrote that last week's, quote, 'epic Treasury rally and the accompanying rout in tech stocks had the feeling of a turning point, the sudden arrival of winter after a long summer.' Very general question that you posed in your piece: is winter coming for stocks?
I think it's really important that the 10-year and the 30-year yield went up last week when stock markets were faltering. Because if you remember, it was hardly a week ago when all bad things were being done by Treasuries and Treasuries were trash. The US fiscal system is unsustainable. Inflation is out of control. These things are monstrous. They only go down. And with that as the context, things got spooky enough in equities that people started buying the crap out of Treasuries. Anyway, like, you know, we had like a 19 basis point day on the 10-year last week. I can't remember the last 19 basis point day when the yields went down that much. People were really buying the stuff. And by the way, the bad feelings about the bond market may have been overdone, right? When you can get, in real terms, in inflation-adjusted terms, you're talking about 2 to 3% from the bond market, I'd say that's pretty good. And so it's not like there was an irrational rush into bonds, but this fast change of mood, I think, really told you something. And that was, you know, and that's extremely striking. That's point number one, and I'm rattling on here, but I'll get it all out and then you can respond. Point number two is that cracks had really been appearing in the equity markets for quite a while before this happened. So, as we've discussed on this show before, I believe the Nasdaq and the S&P 500 have been heading in a herky-jerky way sideways for six months, since May. But below that, we have seen the market behaving in a way that is consistent with the end of a cycle and is consistent with higher bond yields putting pressure on them. So we've seen things like — one of the examples I pointed out last week was there was this group of industrial stocks like GE Vernova and Caterpillar was one of them, and Comfort Systems. These were all the industrial stocks that sort of benefit from the AI boom. They like put the cooling and electronic systems into and the construction part of the data centers. Great. Those had an unbelievable run. A lot of them doubled. Well, they've been doing like crap since May, right? And you know, the small-cap stocks have been doing terrible. A lot of the interest-sensitive parts of the S&P, utilities, real estate, those have been doing badly. So it was like, you know, more and more work was being done by the big tech stocks. And last week those started to crack, and we saw the response in bond yields, and it's like, okay, something is different now. It's not all unicorns and rainbows and hard candy at this point. You know, it's like something feels different. So it doesn't mean we're going for a crash, but the euphoria is broken. There's a much more realistic, much more wary set of fundamental price action going on in the stock market now, and that remains true today, even though we had a pretty decent day today, markets, I would argue.
It seems that the 10-year is really the crucial thing to look at. And just to be clear for everyone, you know, beginning of the month we were at 5.3%, and I'm going to take my little victory lap here, which is that on October 5th, I said that this was the end of the selloff, and my prediction was we're going to see the yield come back down and we'll see bonds rally. That is what's happened. We're now down below 5%.
Yeah, that is a huge move. Just for readers who don't follow the stock market, that's like a 15% move in the stock market would be an equivalent size. This is a shift.
