From $2 Trillion to $20 Trillion: What Anthropic Has to Become Beyond Claude

0:08Good morning, good evening, wherever you are, and welcome back to the Enterprise AI Show. This is your host, Aaron. And on today's roundtable, Brian and Brandon play out a big what if. Assume the Anthropic IPO works. So what would it take to go from a$2 trillion company to a$20 trillion-dollar company. They ask whether Claude Code is just the beginning and whether Anthropic comes for finance, the ERP system, or even your tax return. And they debate a surprisingly bold idea where Anthropic stops selling to the companies it could disrupt and simply buys them or runs them with AI. And that is coming up right after this quick break.

0:58Today's show is sponsored by Nord Security. Your employees are using AI every day, but do you know which tools they're using, what data they're sharing, or where their AI agents are connecting? You can't govern what you can't see. Introducing Nord Layer AI Scout, the visibility layer for enterprise AI. AI Scout gives you visibility into AI activity across web and desktop apps, detects sensitive data exposure, helps control AI agent connections, and creates an audit trail for governance. so you can protect your data without blindly blocking innovation. See what your AI is doing, then decide what happens next.

1:33Connect with Nord Security to learn more about NordLayer AI Scout. Visit nordlayer.com slash AI dash scout slash enterprise AI and get started today.

1:48And we're back. And Brandon, a couple of things before we jump into this. So first off, good to see you, man. I know it's been a week or so. We, as we didn't do last week to talk about football, I do want to, for folks that were upset that we didn't talk about football, I will be going to a college football playoff adjacent game this weekend. I'll be going to the Tennessee-Kentucky game because my kids are at both those schools. So that's our football talk for it. Somebody will be watching football this weekend. I know you'll be watching. I'm going to be watching a little closer. But so what I want to do, we're going to do this as a two-part thing, and this follows on our last AI and Cloud News of the Month show that we did last week.

2:23We talked about the Anthropic S1. I think Anthropic is probably going to be the center of conversation for most AI stuff between now and the end of the year, assuming they might actually do an IPO. So I think people are pretty confident that might happen. Instead of saying, hey, will the stock go up? Will the stock go down? Is this a good company or not? I want to project into the future and do this as a two-part thing, where part one, which will be this show, is us basically going, okay, if they are going to be successful, let's just assume that, and they're going to grow to some number that makes people happy in the future, some exponential growth number, what does the anthropic stack look like?

2:59Because in order to be able to get to, to go from a$2 trillion company to let's say a$20 trillion company or something, 10X, they're going to have to be more than Claude. Maybe that's what it's going to look like. So I want to have an open-ended discussion of what do we think that might look like? Because I think the other side of this is people are saying, hey, look at AI is able to rewrite SaaS, which is obviously a huge, big usage in the enterprise and big revenue source and so forth. But I want to kind of start there with a big open-ended question. What do you think, if we're sitting here a year from now, two years from now, what has Anthropic gone after in order to make their stack bigger?

3:33Or maybe it's just Claude, Super Claude, Extra Big Claude, whatever it might be. What do you think it looks like? I like it. A Claude family size. That's right. Something like that. I think the place I was thinking about when you proposed this topic was like, we can always just start with what have they done, right? And then see where they could go to the adjacent there. So they've done cloud code, disrupted all of development. So the natural extension of that would be like, let's just call it back of house, right? Tools that professional white collar workers use all the time that could potentially automate as much work that has been done for, as we think of it as basically coding.

4:10So the obvious things to me would be something like tax, right? Like it's very rules-based. It's very close to code in the sense that it's a rules-based system. Ultimately, a tax return has to follow a certain set of rules. And most accountants are using some type of like pretty complicated, not complicated, but some type of automated way to create your return. And then they're mostly, I think most of the accountants are then trying to figure out, did you get all your deductions? Did you do everything right? But they're not really if you're adding the numbers anymore. So that seems, excuse me, domains that are adjacent to like rules-based, what do you call it rules-based structure.

4:47So like accounting comes to mind, finance comes to mind. We've endless discussions been talked about all the different things that are done with spreadsheets on Wall Street and coming up with those types of stuff. But even putting that aside, forget about maybe Wall Street, just think about all the finance people and stuff like that. Is there a cloud for finance, a cloud for accounting that would be very similar in the sense of cloud code? Because what's nice about cloud code today is it is ultimately the developers, the consumer, But if you will, Anthropic isn't really, other than just making sure it runs, they're not really producing, if you will, end user facing software.

5:22It's like a tool that someone else uses to then produce work product that they then show off to someone. So that would be like, I would, I guess, adjacent to what they're already doing. And I think there's a ton of white collar work in the back office. So think of it, instead of getting Microsoft Excel and various access databases, you get Claude for Finance. And those are the tools you actually start to use. So I don't know. What do you think? Is that maybe because the way to push back on it would be like, it's just not like coding. Yeah, it's similar, but it's not. So what do you think? Right.

5:53No, I think it's a I think it's a good starting point. I think where the question kind of boils down to, at least for me, what I was thinking about this at a high level was that was my initial thought was like, do they focus on like IT type of tasks? Do they focus on front of house? So things like marketing, product development, things that that your customers are going to see or do they focus on back of house? So you threw out essentially what seems like, whether it's Claude for Finance or Claude the ERP, which seems like something that people would go, oh, we need this thing. It's a central heartbeat for all the activities within the business.

6:25It's also a place that we have tons and tons of data kept for a long time. I think that feels like kind of a good natural, logical place to go. I also wonder if that's what their thinking is, which is let's not try and create something brand new that people can't grasp what it is. Let's take something that lots of businesses know what this thing is. Today, let's just call it an ERP. And they go, what do we like about the ERP? It keeps all of our information. And they go, what don't we like about the ERP? And the list gets really long. They're like, we're entirely dependent on the consultants that have to always be involved with this thing.

7:00It's an ongoing, constant data integration challenge because we add a new thing, we acquire a company, we want to do a new process. Everything has to be rejiggered and reconnected and reintegrated. Yeah, that does feel like something that would happen. It also feels like something that it'll be interesting to see if Claude or Anthropic is willing to take a leap at something that big, right? Because in essence, you'd be like, oh, I have to go into SAP's territory or Oracle's territory or somebody like that who has huge amounts of data or storing data, has proprietary ways of keeping track of data, but then also has all of the context and logic and stuff in fairly proprietary ways, right?

7:43It's not just, oh, here's one gigantic SQL query that looks at the whole company. There is years and years of kind of cruft buildup around that. But maybe that is, maybe they do have to go after something that big as a starting point to set the tone of we want to help rethink your whole business. So what if they did this? Because we always talk about how maybe the AI CMO could do a better job promoting good things. What if you're anthropic and you're like, why don't we just stand up something that will create a 1040 easy tax return for anyone? Just if you want to use it, it's free. Just here it is.

8:16You can give us your information, upload your W-2 or even do it so it stays maybe on your machine and it just outputs a fully formed, complete 1040 easy. So for other people listening outside the United States, that's like the easiest tax form. And the vast majority of people in the United States will, they can usually file something like that. And just to say to people, it's like, you don't have to use it. You don't, you can check it if you want. Like you would print the whole thing out, be here's the return. Here's how I created it. If they want, they can take it to their accountant. They can read it themselves.

8:46And just to just kind of show here, we can do all of that for you, right? And I think that'd be an interesting proof of concept for the nation. It would also put us all to the test, right? Where people are like, I don't want to do that. But it's like, why not? Right? That's something that's, I would think I'm about to say the famous last words, it's like pretty easy, but it also has some stakes to it. It's like, you actually have to get it right. And then, and people could actually try it out. Those are the kinds of things that I think you could do on the consumer side. And then of course, and I guess that's the question though, kind of to the point you're asking, it's just, does Claude want to get in that business?

9:15Because basically what we're stating is, Do they want to take Intuit, one of the largest online tax preparation softwares? Do they want to take them on head on? Because that would be a huge change. Or would they prefer that they just do a big enterprise deal with Intuit, who behind the scenes, they use Claude, and they're the ones that do the double checking that the tax return is good. But to me, this is where I think if they're going to realize these massive returns, right, I think they at least have to be ready to take a couple of these industries on, or I guess they do it maybe in a more quiet way.

9:48They just, and we kind of do some of this, it's like you just quietly have people and users discover themselves that Claude can do this thing that used to pay software to do it. And maybe that's what they're counting on. It's like people just figure it out on their own. Yeah, yeah. No, I think you're right. I think they do have to figure out, do they want to be, do they want to be, two things. I think they figure out, do I stay mostly in enterprise and government, which is by itself very large markets. Although consumer, you think about like where spend is, consumer tends to be more, but it's things like ads and other stuff.

10:18So I think that's probably top of mind. They haven't really shown an interest. I don't want to say they haven't shown an interest. They haven't shown a starting point in the consumer side of things too much. Like most people, if you're writing software, you're probably doing it for something having to do with your business. And we've seen them talk about having really big customers. But yeah, I think the other thing that you highlight, which is really important, is are they taking something that looks like AWS where you start off with, we're just going to provide you all the building blocks of the business.

10:46We're going to give you primitives of the business. And again, this is where you'd go, okay, what does that exactly mean? Does that mean a finance department? Does that mean a legal department? Does that mean a marketing department? And then over time, you just start to every year at Claude Reinvent, whatever their event is going to be called, they go, oh, hey, here's the new features. And then people are like, oh, they just picked off startup. They just picked off industry one, industry two, industry three. And those are all now going to be part of the Claude suite or something like that. I think those are interesting because it was one thing for AWS to be like, hey, we added some new databases.

11:21And you're like, oh, okay. Those are just kind of extensions of something people do. It's a whole other thing to be like, we're going to try and take out an entire industry, which involves consultants and involves just all sorts of stuff. But yeah, no, I think that's it. The other thing I wonder is, do they go at this with a very sort of frontal approach? We're just going to put out there that we are going after this industry. So perfect example is they threw out a thing that said, hey, we could just rewrite all the mainframe software code. And all of a sudden, IBM stock went in the tank and people were like, oh my gosh, they're going to replace IBM.

11:54That's a big chunk of the industry and all sorts of stuff. but then something never really happened it never really materialized it just floated out there or do they do it in more of a stealth mode where they sit behind the scenes the equivalent of the kid with the unlimited token budget who goes like I could rewrite this whole thing but I'm gonna have to see the whole thing first I'm just gonna have to go live in the shadows of what's going on there so I'll be I'm curious do you feel like they are more of a company who's gonna be head first in terms of announcing where they're taking their fights or do you think they're gonna be doing a lot of stuff behind the scenes.

12:28It's like, oh, we've just noticed that like a lot of finance people are using Claude. And so we're just scraping all of their stuff, trying to figure out how to create like a super financier. And then we create a product. I think it will be. I think it will be mostly emergent, right? I think, of course, the things they're going to do first are the things that they discover themselves, right? Because that's really where Claude Goode came from, right? They themselves started to use Claude or some version of it. And so I think that'll lead them mostly first. I don't know. So today I'm trying to think, as much as we've talked about SaaSpocalypse and things like that, I'm trying to think of, has there been a takeout of any major piece of software?

13:03You know what? I guess it is. I was about to forget. What's the question answer site? I just forgot the name of it. That used to ask the coding questions on this. Oh, yeah, Stack Overflow. Stack Overflow. So Stack Overflow, I think they were acquired for a billion dollars. And I think now the question, I think they're getting 30 questions a month. This is after getting whatever, however many millions. And so that's probably the best example of a business that's been completely taken out. But I was trying to think of, I don't know of anything else. Now to be the question kind of going forward is, do they even try for this or does it just happen almost by accident?

13:35You know what I'm saying? It's, and I don't, it doesn't feel, cause this is where I always go back to, I think we talked about on the show, there was the Palantir CEO that kind of had that major, what do we call it? Kind of crazy speech on CNBC rant. Let's just call it, he had a long rant. And he's very much of the mind that the AI companies are going to soak up all the domain expertise of the SaaS companies and recreate them all. And I don't feel like that's a stated strategy. I feel like maybe that could be an outcome. It's just the AI just gets so good that it's like it just understands how to do these things.

14:08But also another one I just want to throw out because I think sometimes we talk about the enterprise so much. One of the things I thought about was just anytime you have a point of sale terminal, is that a place where you could just potentially have a clawed talk? And I think about this at fast casual restaurants. Sometimes you can take a Panera Bread or something. You can talk to the cashier or you could just go to the kiosk. Another day I was over here in Austin, Texas. I was at Torchy's Taco and I was clicking. I was clicking on my, what was I getting? The Trailer Park Chicken Taco. And then of course there was like probably 15 screens.

14:36And I was like, there's a lot of screens here. There's a lot of upsells and stuff like that. But what if I just walked up to this thing. It's like, yeah, give me the Taylor Park chicken taco. That's it. And I was like, that seems completely doable. Right. And then it would just kind of show me here it is. Here's your invoice or your receipt rather. And it just kind of goes, makes me think of, is there a version of that becomes a really generic thing that Claude just becomes really good at, right? Here's my menu. Here's the thing. And you could just talk to it. And it doesn't, and you're like, Brandon, what's the big deal?

15:04Can't you just type on, click on the things? You can, but it does take It takes a lot longer. Have you noticed that? It takes a lot longer. So that's one thing I just wondered. Point of sale, check-ins. Why don't we just talk to this thing? That's possible. Obviously, we've seen voice mode with certain stuff. I know a lot of developers are starting to like voice mode. The other thing I wonder is we think about something like bending spoons. So I think we've talked a little bit. I know you guys have probably talked about it on Software Defined Talk. Bending spoons is this combination of it's not really the SaaSpocalypse where it's like, hey, why don't I take some software and just completely rewrite it so that I'm unbound by some license to it, or I just don't like, I don't like Adobe anymore.

15:40So I'm going to rewrite it as whatever. But bending spoons feels like it's a little bit of a variation of it. It's like, I'm not necessarily like rewriting the experience for you, but I am going, hey, the ongoing maintenance of it, and then maybe all the backend operations of all these different companies that are probably really disparate. Could I unify those in some ways? Could I run AI behind the scenes on these things? Maybe it's for new feature development. Maybe it's for understanding signals in the market. Maybe it's how backend operations works. And I wonder if they create the concept of something like bending spoons.

16:14Go to somebody's CIO, CEO, and say, hey, look, how much are you spending on IT these days? Or how much are you spending on SaaS, whatever? What don't you like about it? What if I streamline that whole thing for you, sort of bending spoon style? But behind the scenes, it's sitting there and it's trying to figure out how your things work, what you like, what you don't like. And it goes, instead of operating this is like 47 different silos. I just take over that thing. I take care of it for you. And you're not having to manage it yourself. I can work on making the licensing be much more streamlined.

16:45You can update it once a year, whatever it is. I wonder if they're looking at what those guys are doing and going, okay, that's a big opportunity. If you could do that and go to large organizations that have thousands of applications and be like, I could do some sort of magical transformation. I think that would be a pretty, it'd be an interesting discussion, if nothing else. I don't know how much people would want that much change at any given time. I feel like bending spoons might be a precursor to not just like what private equity looks like in the future, but maybe like what enterprise IT looks like in the future.

17:18I don't know. So I really, as you were speaking there, it made me think of two things. It's like, they can either, as you were describing, show up, and this is really where the floor deployed engineering kind of roles are. They can either show up and be someone that wants to help. Go in there, please use our software. Let me show you how it can help your software. But then there's the ultimate skin in the game where it's like, you know what? Rather than trying to teach everyone, and what you said I think was really important earlier, you said it's not sure what people want because there's going to be resistance and people may not want it and things like that.

17:50What if they went the other way and just be like created another, if you will, company and just basically we're a company that's going to go out and buy existing companies and we're going to bring them into the fold. And instead of being private equity, it's like we're going to be clawed equity, right? We're going to run this company and we're going to put people in charge of it that understand AI head to toe and that's all they want to do and they run it. So think of, I think Meetup was bought by Bending Spoons a while back. Right, right. And so instead of Bending Spoons, no, they just go, we're going to buy Meetup and they probably would actually get rid of the employees and they're like, this is now going to be managed by maybe a couple people and 100 % AI.

18:29Much, I think we talked about it on Software Defined Talk that they're rewriting one of the, I think it was one of the JavaScript or Node libraries, completely rewrite the whole thing in Rust, right? And it was like, rather than try to convince everyone, let's just do it and own it. So I think that would be like, I don't know, that I think talk about controversial. That would be very controversial. But we talk so much about digital transformation on all these podcasts. And we always talk about it's really hard. The people don't want to change. None of us people want to change. And it's like, why not just go in there and be like, you know what?

18:57I'm just going to own it. Rather than trying to tell everyone, it's like, I'm going to buy it. I'm just going to run it this way and see what happens. I don't know. Is that too far? Do you think people would like just revolt against that? I think it's an interesting strategy. It feels like kind of a complicated strategy. And as you were thinking about it, I was trying to think about, I was like, okay, let's suppose, for example, they go and raise, I don't know,$100 billion from this IPO, right? And if you're reading, as I've been listening to a lot of people try and dissect the S1 and everything, they're like, hey, it is becoming more obvious that most of the costs, the big costs, the big ongoing costs are in the R &D side of things, the building of the model, the improving of the model, less about the inference side of the model.

19:36And so it would be very interesting to be like, okay, they're going to take on$100 billion. Maybe they take on$200 billion. And instead of them being like, we're going to put the vast majority of that into compute and building models, they were like, here's what we're going to do. And this is diabolical. but let's start picking off the most, the systems that enterprises are most dependent on, like SAP, like an ERP system, a CRM system, a whatever system, right? Collaboration system. And they start going, okay, first thing we're going to do is we're going to take some of that money. We're going to build the SaaSpocalypse version of that thing.

20:11And we're going to start going around to customers and be like, hey, if you would like to just include the equivalent of Claude SAP ERP in your system, we can essentially give you that for next to nothing. as long as you're buying the enterprise Claude, Claude for all your people. And that starts getting around, the price point of SAP falls down as a stock price. And then they just go and snatch them up. And they're like, hey, I have the option to either own, I can just own the SAP thing going forward, or I can give you the replacement thing going forward. And I wonder if there is somewhere, there's a strategy that says we should have a buy or build strategy that is either there to take the value out of these people that are incumbent in the enterprise for the next 10 years that we'd like to get rid of them, but that's going to be really hard.

20:53Or we just swap them for something that becomes this massive lock-in to our AI tool. And I think that's a little bit of a variation of kind of what you're talking about, which is on the back end, yes, you run it in some fully automated, fully AI automated operational way and so forth. But yeah, it gets me thinking about being like, how far would they go? How much revenue do they have? They're not going to just go buy Oracle, they're not going to go buy SAP or they're not going to go buy VMware or something. But how cheap could they build the equivalent of it and just go to customers and be like, we could basically do this for free, do this for free, as long as you're using Cloud for all of your employees, for a bunch of other tactics or something like that.

21:31I think we look at something they announced here, because when we talked about Cloud Code, so what's the next most closest adjacent would probably be GitHub source control. We'll put a link in the show notes here on October 13th. there's a whole post here, redesigning source code, source control at Anthropic, right? So basically they're coming out and this has been real popular that source control under GitHub, GitHub is straining and it's anyone that does something with a reasonably large team and has a bunch of agents, it's not really built for that, right? So this is a good example of what at some point, do they just make their version available?

22:05And does that start to become the thing that replaces GitHub? And then you kind of get into, because I think the part And people would say, why would you buy any of these things? It's like, listen, sales and marketing awareness is still very hard, right? So it's like you could see over time to be like, you're introducing a new approach. And at the right price, if you can get the brand or the biggest brand in that area to kind of replace them, that's a really powerful combination, right? Because you kind of have existing contracts and people using to it. And just the knowledge of people go to GitHub.

22:34That's where most people go when they're looking for social controls. So this is a good example of they're on to something better. at least they think they are. And probably they are because they, who else would know better how to use force control and Aegis and Anthropic. And so it's becomes one of those things where you're like, yeah, this seems like a good vertical. So I don't know. It's almost like a prediction, which I must be like, will they ever, if you will, release something that is actually a full on replacement. And also too, like maybe for them, is it worth displacing a whole business or do they just kind of make it for free because they feel like they're getting some other benefits for it.

23:05So I don't know, maybe it's like everything else, maybe it's already happening And we're just like, we're just now catching on to it. Yeah. Yeah. Now, I think we've hit on a bunch of the things, at least related to the enterprise government side of things, is that they're going to go down a strategy that's either going to be try and make a replacement for a lot of the things that already exist. And we've seen some beginnings of that, right? Sort of the SaaS apocalypse part of it. That obviously, they've got much more control of it. Obviously, the downside of that very much becomes you have no brand awareness.

23:33You have no, people may have to make just tons and tons of changes and they're locked into stuff. Or do they do some sort of strategy that looks like a combination of buy up things that are strategic and then start to build, rebuild the ecosystem completely around that? Like you said, take software development and be like, okay, we're going to rebuild what GitLab, GitHub do or something else around that. So it's, is it, what was the, I'm trying to think what the old thing that people used to talk about, like when they were doing cloud native apps, it was like you wanted to land and then you wanted to start dissecting everything around it.

24:02Yeah, the whole land, expand strategy. Yeah, land and expand. And then there was another one that we used to talk to about like how you decomposed applications into things you could turn into microservices and then you could start to pick away at those sort of things. But yeah, no, I think that feels like sort of what their initial plans would be. I think the really interesting part becomes when you've taken that much money, like how big are the expectations going to be? Like are people going to be fine with – because the last couple of years, you look at their growth rates and they've been like 10x, 14x, whatever.

24:30are people going to be fine with you being like, yeah, we grew 50%, which would be phenomenal, bigger than any of the cloud providers at this point in time. Or are they going to have to have 150 % growth, 200 % growth, 300 % growth in the first couple of years? And if that's the case, replacing GitHub isn't, it's like a fly on a horse's back. It's not a big thing. Yeah. I think we hit on probably what the logical things are. I don't know if we hit on the like gigantic big bets that they could do, but yeah, I'll be curious if people have any thoughts about what those massive bets could be to get there to do that.

25:03So anyways, let's wrap it up there. I think we've exhausted what we think maybe their initial strategy will be. The second show we're going to do, the one that'll come out, so this one will come out Wednesday. Yeah, probably Wednesday. And then another one will come out on Sunday. We're going to put ourselves in the shoes of a general manager of one of these groups that might go after these customers and like what that strategy might be, what that approach might be to go be successful in those things. We're going to wrap that up there. We'll jump into the next part of this and you'll be able to hear it on Sunday.

25:27And kind of, you can play along as well. You can put yourself in the shoes of these folks and see how you would grow what's going to look like maybe one of the fastest growing software companies in the world. So do that. We get back. Thanks for listening. Thanks for telling a friend. And we'll be back to you next week. Thanks for listening. Check us out at theenterpriseaishow.com for past shows, newsletters, and all things Enterprise AI.