How Anthropic Beat OpenAI to Wall Street

0:00Kate Clark: AI adoption is growing worldwide.

0:01Berber Jin: AI is redefining what's possible.

0:03Kate Clark: The future is AI. We've all heard it, and somehow your AI still isn't delivering. No context for how your business actually runs. Broken integrations, stalled pilots, enough. Enterprises deserve AI that works. Not someday, not theoretically, today. Real outcomes, already running in real enterprises. Publicis Sapient, AI that's built to deliver.

0:33Berber Jin: you're like a huge skeptic of AI. I don't want it to take my job. Hey, welcome to Tech News Briefing for Friday, October 30th. I'm Bob McMillan, a cybersecurity reporter here at the Wall Street Journal, and I'm with Berber Gin, and we have Kate Clark over in New York. Why don't you guys introduce yourselves? Kate, who are you and what do you cover? Yes.

1:00Kate Clark: Hi, Bob. I'm Kate Clark, and I cover startups, venture capital, and the AI industry.

1:06Berber Jin: And I'm Berber. I cover AI, OpenAI, Anthropic, all those companies. Well, that's great, because we're going to talk about Anthropic. They dropped an S1 on Monday, and everybody's talking about it. It's a very long document, and it's an interesting one as well. I've never really seen an S1 quite like it. Berber, what kind of company is this? Yeah, Anthropic is a very weird company. I don't think it's like any company we've seen before in the history of Silicon Valley, mainly because they say they're building a technology that has a very high risk of posing catastrophic risks to society. And you see that in a very long set of disclosures around the risks behind building AI.

2:00Berber Jin: But at the same time, they're asking investors to put in as much as$100 billion to accelerate the development of the technology. So I think the central paradox of this company, you can see it spilled out all over the S1. Kate, why do they need$100 billion?

2:20Kate Clark: The short answer to that question is infrastructure. They need computing power in order to build and develop these AI models. And although they are generating significant revenue,$16 billion in the first six months of this year, up from just$400 million two years ago, in order for this to eventually one day become a profitable business, they need to continue to raise a lot of money to fund that compute build-out. Kate, I just wonder what to you stood out as just this is not a normal company kind of in the S1. Yeah, I mean, I think, you know, we all know that Anthropik's stated goal is to develop this AI as safely as possible.

3:00Kate Clark: But at the same time, as you've both alluded to, they have 90 plus pages of risk factors that border on science fiction. I think perhaps one of my favorites or one that stood out to me the most was that it said that its technology may cause an incident so catastrophic that the entire insurance industry might lack the capacity to cover the damages. I think, you know, and that's just one of many across dozens of pages of, frankly, frightening risk factors. If you're an investor, what is the risk you're most worried about, though? Oh, I think for the investors, they're most worried about, will this company have the infrastructure it needs to continue growing at the rate?

3:40Kate Clark: And I think investors are also worried about the company being able to sustain the pace of revenue growth. I think they know, of course, it cannot continue at this pace forever. I mean, going up 1 ,300 % from last year in revenue is not something you see every day or ever. But it's definitely a concern because the market is getting a lot more competitive and you have the rise of much cheaper open source Chinese models that are definitely going to be eating into Anthropics market share. I mean, how do they fend that off, Berber? It seems to me that sometimes when I look at these companies, these frontier companies, they remind me of Unix companies from the 1980s and 1990s where they're building these super systems that are just better than everything else.

4:27So people are going to pay for something that's better than everything else. But, you know, we've seen in just even in the last few months, a lot of these Chinese open weight models have become pretty competitive. Like I write about cybersecurity and they're being used all the time by the cyber folks. What's their moat? Like how does Anthropic stay ahead?

4:46Berber Jin: So I think this is like the billion dollar question. And I know you're very skeptical of the defensibility of the Anthropic business model. I think we have to see how it plays out. I mean, what Anthropic would say is that any big company will always want to pay money for the most advanced frontier model. And obviously, Anthropic, they can package better security, cyber controls, all that type of stuff in a way that if you're just using open source, you can't really do. But I guess we'll have to see. I mean, they list open source as a big risk to their business model. But from the numbers they've disclosed, like Q1 to Q2, I mean, they grew like an astonishing rate.

5:38And you expect that to keep up in Q3? Like that's by every indication that's not going to slow down with the Chinese threat?

5:45Berber Jin: We know that they've told investors, they grew that a lot from Q2 to Q3, but not as fast as Q1 to Q2. And what is interesting about the S1 is that there are a lot of mentions of how unpredictable and unknowable this business model still is. I think Anthropic, the execs, would be the first to tell you that it's very difficult to forecast revenue around this business. And so I guess we'll just have to see whether in two or three years it can still get to like 100 plus billion in revenue or something crazy like that.

6:24Kate Clark: And on the defensibility and moat question, I think this is something that people are asking themselves every day. But these companies are trying to create like an ecosystem around them. So for one, you use it at work, your company uses it, so then you choose to use it at home. Two, it has memory, it remembers you, it knows you, so then you are even more locked in over time. And I'm sure a lot of people listening can relate to that and what they're telling JatchyBT. And also these companies are moving into the world of agentic AI where they can actually complete tasks on your behalf. And I think just that the effort to create those ecosystems is what they think will keep users over time.

7:01Kate Clark: but I think it is still like TBD whether people do feel loyal to any one of these AI companies, especially as it gets more competitive. They have to be chasing the next frontier. I mean, coding definitely was the first for them. It was like a huge breakthrough. Cybersecurity, I would argue, is number two. What's the third? Kate, you want to take that? Yeah, I mean, I think Anthropic is making it very clear that healthcare, drug discovery, I mean, they even said in their S1 that there's potential that their models could cure major diseases. like cancer. And I think, you know, we've seen open AI go in interesting directions with robotics and hardware and devices.

7:37Kate Clark: I think while that is happening, Anthropic seems to be going down the pathway toward healthcare, drug discovery, pharmaceutical, all of those kinds of interesting things. So I think that will probably be the next big chapter for them that we'll be paying attention to over the next couple of years. One thing that's like not 100 % clear to me, Kate, is that Anthropics products are good to a certain percentage, right? Like there's still, even in the realms where they perform the best, there's still these like sort of edge cases where they do things that are not expected, where they kind of, they need like human guidance, you know, to go in the right direction.

8:19Is it clear that they're going to be good enough in every domain? or I kind of like the skeptical part of me is wondering if we're going to hit some kind of winter somewhere where they, you know, they're sort of superficially are doing well, but then, you know, when it comes down to actually like people shelling out money for this, these new categories, they're going to hit a wall. So is AI good enough, I guess, is my question. That's an interesting question

8:49Kate Clark: and it's hard to answer. I think that, yes, Yes, it's getting better every single day. And I think we've all are experimenting with it. I can see how much, how quickly it is improving. I think that what's going to happen is that it becomes commoditized and it's just a lot harder for these companies than to justify the high price tags. And that's why they need to kind of diversify and think of other business lines. because like you have both said, paying significant prices for AI coding tools just isn't probably going to be the norm once all of these companies have caught up and kind of gotten to the same place.

9:29Yeah. I mean, we saw this in the cyber domain, right? Like where there are these people who can find bugs and when they're married with AI tools, the latest AI tools, they can find bugs that no one could ever find. But in that domain, there's still a need for humans. There were a lot of fears, Berber, of the job apocalypse a little while ago. I don't feel like it's played out. I mean, there's still pretty good employment in the United States, but Microsoft just offered buyouts to like 60 ,000 people this week.

10:06Berber Jin: Yeah, I'm not a macro economics guy, but it seems like the job reports are still pretty good. And like the fears of like the near term impact of AI causing mass disruptions to the workforce have been overplayed. And I do feel like there is this dynamic of companies also using AI as an excuse to lay people off. The market likes when CEOs say they're experimenting with AI and the markets are rewarding companies that are also laying staff off. So I don't think we've seen it yet. But, you know, I think Anthropic, they believe that there will be massive economic transformations that come with AI. And to the point on unlocking new use cases, I think they also have an incentive to do that because right now what we're seeing with AI is it's in the past few months, right?

11:07Berber Jin: It's concerns about AI automating away the workforce and AI going rogue and causing massive cybersecurity incidents. So neither of those use cases are incredibly confident. They're not very inspiring use cases, I think, for the general public. And so I think part of the push also to move specifically into drug discovery and medicine, they also know there's a PR advantage to doing that.

11:37Kate Clark: Yeah, definitely.

11:38Berber Jin: All right. I got to ask you about one other thing in the S1, which I'll start with Berber, the images. Mm-hmm. So if you look at the S1, there's like these cool images. You were interpreting some of them. Like there's a Petri dish, I think, in there. There's some kind of a shell maybe. What's going on with the images? It's a great question. It definitely fits the Claude aesthetic. Like I feel like if you look at how they market Claude in ads, it's like elevate your intelligence, pursue new domains, you know, augment the human consciousness, stuff like that. So they're kind of these allusions to discovery, science.

12:22I did get a little bit of these vibes. Like when somebody sends you like some AI-generated content, it's always like 400 pages or something like that. It's always really, really long. It's maybe got some weird stuff in it that you don't know how to interpret. So, yeah, I don't think it's AI-generated.

12:37Berber Jin: And there weren't that many M-dashes. Oh, you checked? I think there were, maybe correct me if I'm wrong, but I didn't see that many M dashes. Yeah, I think that's right. Okay, well, this is probably a good spot to just take a break here. When we come back, we're going to be looking at Anthropics' chief rival, OpenAI. That's after the break.

13:05Berber Jin: AI adoption is growing worldwide.

13:06Kate Clark: AI is redefining what's possible.

13:08Berber Jin: The future is AI. We've all heard it, and somehow your AI still isn't delivering. No context for how your business actually runs. Broken integrations, stalled pilots, enough. Enterprises deserve AI that works. Not someday, not theoretically, today. Real outcomes, already running in real enterprises. Publis is sapient, AI that's built to deliver.

13:37welcome back so open ai like this i really want to talk to you guys about this part because open ai is run by sam altman former ceo of y combinator you know they kind of blew the world's mind with chat gpt and yet uh anthropic is first to ipo berber why is that so this is a

14:00Berber Jin: very interesting question that I'm sure will be the center of some business school case study, because OpenAI is an older company, they had a lead for the long time in the AI race. And OpenAI as a company has much more of that Silicon Valley, you know, revenue maximization, you know, move fast and break things DNA than anthropic. Yet somehow they're in this position where they've lost their lead in the race and now they're not going to be the first out of the gate with an IPO. And I think there are a lot of different reasons you can boil it down to, but I think fundamentally they kind of put their chips in the wrong basket.

14:44Berber Jin: For a long time they thought consumer was going to be the big business that made money for AI and they kind of wrote off enterprise. Anthropic did the opposite and had this moment where coding really took off starting at the end of last year. And it was like a second ChatGPT moment for the industry. And it just disproved a lot of the assumptions that OpenAI had about how to grow revenue. And I also think OpenAI got a little bit lost. Like they, you know, chased image and video generation and had all these press releases to build their own chip. And so they were kind of just pulled in a million directions.

15:30Berber Jin: But they've really banded down the hatches now. And I think history is long. You know, who knows what the state of play will be in a year or two. So they got a wake up call in November of last year with Cloud Code. Oh, for sure. Yeah. And it came at the same time ChatGPT kind of plateaued the growth. So kind of a double whammy. Kate, what do you think? I mean, is this a business school case study here?

15:53Kate Clark: Yeah, I mean, it really comes down to execution. And, you know, the answer here is really cloud code and the success of it. But I'm like of two minds about what it really means that OpenAI is following Anthropik to the public markets. I mean, I think a lot of us, if we had to guess, would still say OpenAI will IPO early next year. Some people will tell you that it will be difficult for the public markets to absorb a third mega cap IPO following SpaceX and Anthropic. Others will tell you that the markets are insatiable and there is so much demand for all three of these companies, which is to say that OpenAI will have still a massive, hugely successful IPO.

16:32Kate Clark: So really, I'm just curious to see how this does all play out and whether OpenAI will indeed go public in Q1 or if we're going to be waiting around for the better part of next year, kind of wondering. So is OpenAI still a consumer company, or are they more enterprise-y than they were a year ago? Well, they're definitely leaning into the enterprise side now. I think that they are trying very hard to steal many of the customers that Anthropic has. Yeah. I mean, the classic example would be like Amazon, right? So when they created AWS, a lot of people thought Amazon is like a consumer company. They have no credibility in the enterprise.

17:14They basically provide an operating system to all their customers, and they have a very consumer-friendly user interface. So you can just get your operating system up on the cloud running very easily. But they don't provide all the enterprise support that is traditionally thought to be required. So they leverage that consumer interface into becoming an enterprise company.

17:35Berber Jin: Yeah, and I think that's an argument for why OpenAI could, it's a business strength, right? Like, the people who use ChatGPT could help bring in enterprise businesses. But the same logic also applied to Anthropic, where kind of your average coder in Silicon Valley just started using Cloud. So then the companies they work for were like, okay, all of our software engineers use Claude, so we might as well purchase an enterprise contract with Anthropic. I mean, I think another dynamic that's interesting is that right now OpenAI is in the middle of this price war with Anthropic, where they're basically trying to sell codecs for a cheaper price than Claude.

18:18Berber Jin: But I think the risk there is OpenAI burns a lot more money than Anthropic. They're a less efficient business. Why are they less efficient? Is it because of the consumer play or is it just like sort of data center pixie dust magic kind of reasons? This is a really good question. I mean, I think they have more than a billion users using ChatGPT. and that means they have to subsidize a lot of compute for those users because the vast majority of ChatGPT users do not pay for it. But this is a question I think also confounds investors because they're pretty similar businesses, but at least for now, Anthropic has a much stronger financial profile than OpenAI.

19:08Well, I wonder about this, Kate. I can remember when Uber rides were very, very cheap and everyone was enjoying the benefits of these subsidized businesses. I don't know if we're getting that right now. Just looking at the Anthropic S1 and just from what you know about OpenAI, are we paying what it costs for these services or are we going to just all get addicted to them and then see the prices get jacked up?

19:35Kate Clark: Yeah, they're heavily subsidized. The two companies together have raised many, many, many tens of billions of dollars from different types of investors. This will surely change one day. And it very much is similar in some ways to the Uber and Lyft dynamic. And even more than that, I mean, if you're someone who's like experimenting with smaller companies who are developing interesting new AI products, I mean, many of those are completely free right now, again, because they're subsidized and they want to lock you in. But I mean, I wonder myself sometimes just how expensive some of these products will become.

20:10Kate Clark: But I think we can count on the prices going up over time. Do you see the products changing, too? Like, I wonder about OpenAI and ChatGPT and just sort of, you know, there are other ways that they could be making money. We definitely saw this with Google as, you know, the Google.com page was like, you know, and their search results page was like unchanged and very consumer friendly for a long time. And then, you know, these ads started getting bigger and there's just all this other stuff that got thrown in there. Now there's AI in there. I mean, just from a consumer experience perspective, do you think we're likely to see changes with these chat interfaces?

20:51Kate Clark: I think so. I think they're going to change. I mean, Anthropik has indicated that they don't want to have ads. But I would predict that advertisements do become part of these broader AI business models. I would just add, too, that we're in this moment where agentic and AI agents, meaning AI that can really execute tasks on your behalf, we're really in a moment where the companies are very excited about those. And OpenAI recently announced that they have their own AI agent product called Dots. All these companies are experimenting all the time with new products that might be able to get people really excited, lock them in.

21:28And I think, yes, we're going to see these things become more expensive or they'll have ads or, you know, it's hard to predict right now, but I don't see another – I don't see how that can be avoided. Does OpenAI need to – like, is there pressure for them to IPO? Like, how much money do they have? How long are they going to last? You know, like, how strong is the pressure for them to sort of follow Anthropic right now?

21:52Kate Clark: That's a good question. I mean, the pressure was quite high, but they recently went back to investors and started raising another mega round. They're raising tens of billions of dollars, and that gives them a lot more wiggle room. It also shows if that round is successful, and they have not announced yet what they've raised. But if they were to raise the amount that they were looking for, it shows there is still capital in the private markets for them. And if there is still capital for them to raise in the private markets, they can continue to delay. But ultimately, they need to go public because they need capital from the public markets in order to keep, for the same reasons Anthropic does, in order to keep funding compute for infrastructure to expand the business.

22:30Kate Clark: But I will say, like, I think we've been all continually surprised at how much money there is in the private markets to fund these companies. It's like, where is it coming from at this point? But it continues. And I think until that, until the music stops, they can probably remain a private company, which also comes with a lot of perks for them. The S1 for Anthropic describes a corporate structure, a public benefit corporation, where decisions are basically controlled by the founders, right? Is that going to be a differentiator, do you think? Or is that something that investors are going to push back on?

23:03I mean, they seem historically very content to give founders a lot of control in these companies because they are the visionaries. They see where the future is. They're going to go there and achieve it.

23:13Kate Clark: Yeah, that's very unusual. And I think just from all of my conversations with investors, there is a lot of excitement for this IPO. And that does not seem to be discouraging people from investing in this$100 billion plus offering. I feel like we're just in this moment of AI euphoria that is continuing and is propelling Anthropic. It's the new industrial revolution. And if you look at the revenue per employee, it's pretty crazy high, right? Oh, yeah. I mean, that was one of the most interesting things, I thought. And it really speaks to this era we're in where companies can be much smaller and generate so much revenue still per person.

23:55Kate Clark: I think it was about$6 million in revenue per employee, and they have 5 ,300 employees. It's pretty insane. Yeah, and growing at this crazy rate, too. I can, yeah, it's going to be a nutty year ahead.

24:14Berber Jin: Well, yeah, we definitely are not generating 6 million each for Dow Jones. Berber. I told you not to mention that. Well, I'd just like to get from both of you just sort of predictions for the year ahead. I mean, what are you going to be watching in the next year, I guess, for Open AI, the Anthropic? Well, I think it'll be interesting to see how Anthropic trades in the next year, because as they've said in the S1, it's very hard for them to forecast revenue. They're kind of paving the path for a business that we just don't have a lot of data points. We don't know what direction it's going to head in.

24:55Berber Jin: Is there going to be a slowdown in growth? Is the business model going to collapse? Are the assumptions behind how to make money in AI going to change? And when Anthropic becomes a public company, investors will be reacting day to day with the stock price. And so is Anthropic in a year from now going to 5x its valuation? Or is it going to be like a volatile shoot up during the IPO? And then there'll be moments of time when it kind of collapses, when investors get nervous about AI or when their earnings don't come up to speed. That'll be really interesting to see. And I think every analyst and investor will be waiting with bated breath for every quarterly earnings disclosure that Anthropic makes, because that will be the clearest bellwether for the AI trade and how sustainable it is.

25:55Kate Clark: I agree with Berber. I think it's going to be a volatile stock. And I think that will, I mean, it's just given what we've seen over the last few months from the AI doom narratives and all the things that would certainly have caused some fluctuation in the stock price had it been public. I think that will be really interesting. But also just more broadly, not necessarily about the rivalry, but I think that it's going to be really interesting to see what kind of impact a$2 trillion IPO and SpaceX IPO has on the private companies that are all preparing to go public. Like what I've heard is a lot of these kind of$100 billion companies are waiting until OpenAI and Anthropoc go and they all want to then follow.

26:33Kate Clark: So I think we're going to see a lot of still very, very large tech IPOs that people are going to be very excited and interested in. And I think it will be very interesting from a reporting perspective to kind of see how next year plays out. Well, we're going to keep reporting on all of this. Check it out on wsj.com. We really hope to help folks understand what's going on in this crazy, crazy story. And we'd love to hear from listeners of this podcast. What questions do you have about Anthropic going public? Drop us a comment if you're on Spotify or shoot us an email at tnb at wsj.com. We would love to hear from you.

27:14Kate and Berber, thanks. It was fun to chat with you guys about AI in the future.

27:18Berber Jin: Thanks. Thanks for hosting. And that's it for Tech News Briefing. If you're a listener on Spotify, be sure to leave us a comment. Today's show was produced by Julie Chang. I'm your host, Bob McMillan. Jessica Fenton and Michael LaValle wrote our theme music. Our supervising producer is Katie Ferguson. Special thanks to Sarah Krause and Sam Schechner. Our development producer is Aisha Al-Muslim. Chris Sinsley is the deputy editor. Lytel Mollad is our senior director of shows. and Samantha Hennig is the Wall Street Journal's head of multimedia. We'll be back Monday morning with a new episode. Thanks for listening.

28:02You're either all in on this model, or maybe you're building with another. You're either speed, or is it security? Or you're AWS AI. Don't pick a side. Pick them all. AWS AI is how.