You will own Anthropic whether you want to or not

You're going to own this whether you want to or not because you're going to own it in your index funds. >> Can you say whether Anthropic is a good investment or not at the IPO? >> We don't know what the price is. So, it's impossible to say, right? So, I can't I can I can say whether it's a good business or not. And what we know about the business um because, you know, if they price this thing at a million dollars a share, it's not a good investment. If they price it at a dollar a share, it's probably a good investment. >> It'll be somewhere between those numbers. Um, we also want to know what they do with the share account and things like that. But fundamentally, what we see is a business that is really, really popular with businesses. Um, and there's a lot of different ways to cut it. But what we see is a business that is profitable. Roll the headlines and say it's not profitable. This company was profitable in the last quarter in which they have earnings in this quarter. It was profit. It's fantastic gross margins and we see a really steady and and kind of jaw-dropping progression in revenue growth in gross profits and then now in operating profits. You can see it creeping, not creeping, rushing towards a a a slope, a hockey stick-l like slope. Again, not just in revenues, but in profitability and in free cash flow generation, let alone cash operations uh generation, but now free cash flow generation, which is to say everything is pointing in the right direction. And they're trying to go public at this inflection point when everything has turned positive. Corey, let me maybe push Julie's question one step further, which is once we get past the IPO, is there a risk for investors to not own Anthropic? Is there is there a risk in sitting out? Again, it's all about the price. But look, uh you're going to own this whether you want to or not because you're going to own it in your index funds and you're going to own it in in uh any any fund that's trying to compare itself to what the larger funds are doing. uh in the same way that SpaceX got itself kind of jammed into um all of the big indices um right after the IPO regardless of the price, regardless of the quality of the business, the size of the business alone and the new rules in the indices and the the sort of passive movement of investing which is to say that funds even that aren't trying to mimic the index are trying to be close to the index. You're going to own some of this whether you want to or not. a cororates proc you know you better you better own it because if the if the uh clawify thought you didn't buy the IPO it'll punish you as as was mentioned in the [laughter] risk factors of the perspective but just real quick just kind now digging now into the actual um S1 itself you know we see that first half of this year versus first half or versus last year revenue the just revenue skyrocketing right compared to a year ago but also are you kind of curious cory why we >> 100% or so I mean big numbers but are you a little curious we didn't get the Q3 numbers there in this in this perspectus. >> No, I I'm I I I I was disappointed. Of course, I like more numbers. Uh and in fact, it's kind of it's kind of a bummer for an Excel junkie like me to not give more numbers in these deals, but the quarterly numbers, I expect that'll probably come with an amendment. Um I think they're required to give those before the end of November. So, which would be 45 days after the end of the quarter. So, there will probably be an amendment. It'll probably have those Q three numbers and we'll be curious to see what changes. One of the things that we're looking into um as we write our research report about this thing is is what is happening with the competition and we understand that uh open AI chat GPT is has seen a new wave of usage on a corporate level. I'm talking to a lot of companies. I'm here in in Las Vegas at the Oracle AI world and there's a lot of conversations about model routing, which is to say because claude has become so expensive and so popular, sometimes companies are writing uh elaborate prompts that take the cheap part of the prompt, dump it into an openweight uh LLM for China or something and the thing that really requires specialization or something and running that against a fable or mythos if you will. and so in clog. So this model routing is going to be a topic we're going to hear a lot about and we're going to look into some of those numbers. Finally, we have to look at a lot of those numbers outside of this S1 to understand how companies are are really putting uh AI to work. But what we see in these numbers is really impressive. The largest customers are using cloud more than they were even last quarter or the beginning of the year or the end of last year. Um the the most revenues are seem to be coming from partners like Google and Amazon. Uh again the growth is coming from a corporate basis. You know a year ago it wasn't uncommon for us to sit around and talk about are people really going to use AI? Are they using AI in business? Is there any actual tangible proof that anyone's doing anything with AI? What we see in these numbers is that major corporations and small corporations and startups and everybody on the business side is using AI and using claude to change the way they do work to get answers that they didn't have before and to create businesses that didn't exist before. Corey, um, what you were talking about though sort of gets to the sustainability question, right? And the reason that a company like this reports, you know, something like ARR, right? Because it it can be lumpy from quarter to quarter. We don't know if if OpenAI is going to take away some market share. We don't know if these openweight models um are going to take more market share. So, how do you like what are you looking to answer to get at that sustainability question? You know, I the competition is going to be a big part of that, but I think it's hard to think about these markets the way they exist right now. Julie, you've known me a long time. I'm not a very bullish person. I'm not a I'm I'm I'm overly cynical at most times, which has not always served me well. But if we, you know, if we want to think about the total addressable market, we want to think about how big this could be, right? The kind of Adam and Eve moment of what what changed right here. You can think about all the people in the world with computers and phones. You can think about all the business of the world and how big they might be in two and five and 10 years. But it's bigger than that. Most of the uh traffic, most of the hacking traffic that Palo Alto Networks reports comes from agents. It comes from bots. It's not from humans, which is to say there's more activity on the network than there are humans on the network. I think what we're starting to see and certainly that's a big conversation here uh with the Oracle people and they're talking a lot about data security as well. They talk about data because it's Oracle and get databases to sell that most of the traffic on networks is coming from non-human sources. What does that mean? If if an agent is economic activity, then most economic activity happening in networks is starting to be more than human more numbers. It's not just how many workouts can you design, how much food tracking can you do. If every human in the world is using AI 10 times a day instead of one time a day, how big is that market? What if there are thousands of agents doing economic things compared to every for every human person? This market gets really really big and you can see it in the numbers of this S1 that this market might be bigger than how many humans there are in the world. And that's what you have to start to try to figure out when you want to figure out what this deal is worth. >> Yeah. And it's it's a difficult concept to wrap your head around. >> It really is crazy. >> I mean, we were talking about you probably heard us talking ahead of time. we were talking about fitness tracking and and using it for that. Um Corey, as you go through and as you went through the S1, are are there any unanswered questions for you besides the, you know, the third quarter, which as you say, we'll probably get an amendment, what did you want to know that you didn't find out? >> I I look, some of it has to do with the fact that it's only it hasn't been out for 24 hours yet, right? So, we're still kind of trying to figure it out. Closing in on 24 hours here, looking at the clock. um I need to understand better when certain um lease agreements hit. So what we saw was a business that generates a lot of cash flow and generates a lot of profit and a lot of cash flow and even free cash flow. They're going to need that money to pay for the compute that they're using. Right? So we you saw this big deal with Broadcom that they're talking about here. They're trying to get more comput. They're trying to get more compute. Everyone I talk to in this industry from the smallest startups to the the biggest companies in the world are fighting to get more comput. I was talking to Oracle people last night about the building of data centers and they wish they could get more. They wish they can get them done tomorrow because they've got customers for all of that. So everyone's rushing for comput. So we look at the financing and we say geez you've had these companies that have always had free cash flow like Google and Amazon and Meta Facebook whatever even back then they had free cash flow. these companies willing to put their free cash flow aside to dump money at this issue of building out more chips, building out more networking, building up more data centers. They see this opportunity. They see this demand. And so what I want to understand is, you know, there's a moment in in these deals where, you know, Oracle contracts with Blue Owls uh subsidiary Spark to build the biggest data center ever built or one of the biggest ones ever built in New Mexico, Project Jupiter. And uh the minute it's completed, it's turned over to Oracle which instantly turns it over to the customer and then the customer has to start paying that this case that customer is OpenAI. There are moments at which that moment when the customer has to start paying in this case anthropic. What are those moments? How much has to be paid? I'm going to I it may I'm pointing at my S1 which is here next to me and well marked up. You know me, Julie. I want to know when those moments are when those payments have to be made so we can kind of chart >> right because it seems those goals when it happens like those payments haven't been made like I thought that was a really surprising part of the S1 I don't know about you guys how much cash this company still has supposedly it's spending handover fist but it hasn't done that spending yet I thought that was really interesting >> you got hundred billion dollars of cash >> it has done the spending it's done a ton of spending and there's still more spending to Uh yes, they have an enormous amount of cash on hands that you I have to say the first thing I looked at when I saw that number I'm like why do this IPO? You have so much money but I think they see how big this opportunity is and they really and I think they also you know it's amazing that this business has gone from nothing in 2021 we were we don't all remember where we were in 2019 kind of pre-COVID 2021 we all remember where we were. We were wearing masks. We're trying to figure stuff out. We were taking the masks off. We were going out. We were going back in. These guys were starting a business that is now one of the biggest businesses in the world. It is the fastest growing business in the history of the world and it's growing at the fastest rate. It grew faster in the second quarter than it did in the first quarter of this