Anthropic’s Enterprise AI Venture Buys Consultancy, The New Under-the-Radar Prediction Exchange

0:13Welcome, everyone, to The Information's TI TV. My name is Akash Pasricha. It is Thursday, August 20th. Today on the show, The Information has exclusive reporting that Anthropics, joint venture with big private equity firms focused on AI integration, has made an acquisition to grow its business. We'll talk about the progress that JV is having on the ground. We'll then bring on our crypto and prediction markets reporter who wrote a column this week about the predictions market exchange that you have never heard of, but who Robinhood is leaning on. And finally, to close out the show, we have a conversation with PwC about the turning point that big AI labs currently find themselves in.

0:55It's going to be a great show, so let's get right on into it. The information has exclusive reporting on an acquisition that Anthropics Joint Venture with Wall Street has made. I want to bring on our venture capital reporter, Julia Hornstein, who broke that story. Julia, welcome back to the show. It's great to have you here. Great to be here. Okay, so I want to get into this acquisition in a second, but very quickly, I mean, remind us, what is this joint venture here that Anthropics has formed along with these big private equity firms? The joint venture is a services firm that works to integrate AI, largely anthropics AI tooling, into businesses they see as ripe for AI transformation.

1:37And those are typically legacy businesses like accounting and healthcare companies. Sounds like a consulting firm, really. Is that fair to say? I think it's somewhat fair. I mean, it works with clients in a similar way that a consulting firm would work with clients in that it takes on projects from clients. And those projects, in this case, are AI transformation related. this means that employees of the joint venture which is called ode often tailor their ai projects in in solutions to the companies in a way that a consulting firm might help a company with a specific initiative and tailor solutions for them um and even sometimes they do work alongside anthropic engineers to integrate ai into their customers so i think that some of the the way that they engage with clients might be similar to a consulting company more like uh like you know instead of just selling a service or selling a product it's very much you know packaging these things all together and saying hey we'll you know we'll help you replace this entire function maybe with with an anthropic pilled uh version of the solution um let's talk about the news though that you broke so ode is the name of the company they made an acquisition sounds like Yep, they made their first acquisition since launching, and they acquired another AI consultancy.

3:03It's called Casper Studios. It's a four-year-old company with a dozen or so technical consultants, and they've done work for some pretty big names like Netflix and Pepsi. Okay, and how does this acquisition fit into the strategy? Are they just buying a team of engineers, or what is it? Casper's mission is pretty similar to OATS. It helps businesses develop AI apps. Ode, as it stood pre-acquisition, was focused on helping businesses, as I understand it, holistically transform with AI, whereas Casper's tooling and expertise will now let them build quicker AI solutions for specific parts of their clients' businesses.

3:44For example, Casper might help a business AI-ify their recruiting team or automate some financial reporting, basically helping Ode succeed in these quicker turnaround projects for clients. Do you have any sense from your reporting at Ode's business? I mean, do we know how much revenue they're generating, who the exact clients are they're working on? I mean, how are things looking on the ground? Yeah, I mean, a lot of that honestly remains to be seen. These initiatives were announced just a few months ago, so I think we're still really early innings here, but I'll be curious to see how things shake out as well.

4:25But Anthropic is not the only company, we should say, who has this kind of integration initiative. OpenAI also has its own, what's it called, like DeployCo or Deploy, what is it? Yep. And OpenAI has something called the Deployment Co, which also works with a handful of other PE giants like TPG being capital and Advent. I think it's a very similar business this strategy there. And then as I've reported earlier this summer, another example that's a little bit different from this is Thrive Holdings, which is an offshoot of Josh Kushner's venture firm Thrive Capital. They also engage in this kind of AI transformation work with these legacy businesses like accounting firms, but instead they own companies that take a meaningful ownership stake in those old legacy businesses.

5:13So they're actually investing capital into them and not only doing the service work, but then hopefully reaping the benefits later on as they hold on to stakes in those businesses. And we should say, I mean, one of the things you pointed out in your story, which I thought was a good point, is the private equity companies that are partnering with the labs on these initiatives. I mean, half of their intent is to get their own portfolio companies to use AI to reduce costs. And so in terms of who the clients are, I mean, If you just look at Blackstone's extensive list of software companies, I mean, it seems likely that they would point Ode's technology to XYZ HR software company in their portfolio or something like that to make it more efficient.

6:01So it very much is within the private equity frame of mind. Tell me, Julia, where do you think these businesses are going? But where do people you talk to, I mean, do they think that these businesses are going to become the next Deloitte or Accenture? Are they going to end up getting acquired by the big consulting firm? What do you see as the future of these businesses? Yeah, I mean, I think time will tell. I think that on the last point, you're exactly right, that private equity firms really stand to benefit from these projects because, you know, it essentially creates this flywheel for them.

6:37It gives Ode this customer flywheel. And in theory, the private equity firms can then refer their portfolio companies to Ode and transform the businesses in that way. I guess I'm kind of curious from there how these businesses stand to expand their customer base beyond the firm's portfolio companies that they partner with. So I think that that's a big question mark for me or whether they really need to. I mean, these funds have really extensive portfolios. So I'm kind of curious how the venture goes from there. Right, right. Julie, I want to ask you about another story that you published this week.

7:17So you and our NVIDIA reporter Phoebe Liu scooped that NVIDIA is in talks to participate in the funding round. Mercore is putting together the valuation of$20 billion. We've reported details about this round in some way, shape, or form previously. Just walk me through a little bit of the context here. I mean, does NVIDIA, do they rely on Mercore for training data? What's the story here? Yes. So the high-level context is we reported that NVIDIA is putting a check into Mercore's latest funding round that's coming together at, as we've reported, a$20 billion valuation. We've also heard that existing investor General Catalyst is in talks to lead that funding round.

8:04And NVIDIA is one of Merkur's many customers. We've reported that they've paid Merkur tens of millions of dollars in the last quarter, presumably for training data for their last to Nematron models. And I think that the investment just reflects an ongoing and tightening relationship between the chip developer and Mercore, which makes AI training data, sells and packages that training data for the labs to then pre-train and post-train their models. So I just think that we should keep an eye on this relationship going forward. Right. And it is yet another, it would be yet another example of NVIDIA very much expanding its venture capital portfolio into just about every pocket of the AI ecosystem.

8:53Julia, I want to thank you for coming on. That is Julia Hornstein, our venture capital reporter, here at The Information. Alshi and Polymarket are not the only predictions markets around, believe it or not. In our finance newsletter this week, my colleague Yueqi Yang, who covers crypto and predictions markets, wrote about another predictions exchange, Rothera, which Robinhood is leaning on along with other companies. I want to bring on Yueqi to share more about what she found in her reporting. Yueqi, welcome back to the show. It's great to have you here. Hey, Akash. Okay, Rothera, what is this new company we should be paying attention to?

9:31So Rothera is a new prediction market that's fast-growing, but also kind of low-profile and acting behind the scenes. And they're best known as the new prediction market exchange powering Robinhood. And that's because they are a joint venture under Robinhood and Sesquahuana International Group, which is a big trading firm. And who founded it? I mean, who are the people behind this company? Grothera was founded last year by a longtime financial infrastructure executive called Thomas Chippas. And he previously had experience in founding crypto companies and exchange businesses. He founded a crypto exchange back in the day called ArisX, which was acquired by SIBO later on in 2022, which became SIBO Digital.

10:25And this is his latest venture, which is to get into the prediction market space. And so just to confirm, I mean, it's literally, we could just think of it as a smaller, younger Calci. Is it effectively the same business or are there nuances here? There are nuances here. The main difference for Rothera is they're only acting as an exchange and a clear greenhouse. They are not going to be launching a consumer-facing app, unlike Calci or Polymarketed and many other prediction market startups. And the main reason for this is that they think they don't want to compete with their partners. They wanted to be able to execute the choice that their partners bring in, such as Robinhood and other brokers.

11:14And so they want to stay behind the scene as more of an infrastructure provider. So we'll talk about the Robinhood angle to all this in a second. but I mean, you had a chance to speak with the founder, Thomas Chippas. So tell me about that conversation. What did you ask him and what stood out to you about that conversation? So I asked him about what prompted him to found Rosera and why does he see a need for another prediction market in the industry. And I think what stood out to me is that they are keeping a relatively low profile and I think it's speaks to their strategy. They're not trying to build a consumer facing brand.

11:58Tom himself hasn't really done much press since he launched Rothera. So I got to ask him exactly how did he find this company? And then he said last year when prediction market became hot, he and his co-founder Matt were looking at the market and they felt like there's this gap, which is for a company to be a pure play exchange provider. And that's where Rosara came in. He said that they were fortunate to be able to bring investment from Robinhood and Susquehanna. And as a result, they don't need to worry about where their customers will be coming from on day one, because Robinhood obviously is an enormous distribution engine.

12:43Did you get a sense for how big the company is at all in terms of revenue or trading volume? So by trading volume, they're only launched in June this year, just in time for World Cup. And their trading volume is ranked among the top five prediction market exchanges right now, mostly because of the volume that are coming from Robinhood. And as we talked about before, at Robinhood, prediction market is now one of their biggest business line. The revenue for Robinhood coming from prediction market is now higher than the revenue they generated from either equities trading or crypto trading. So the company itself has about 50 people.

13:27They're based in Chicago. And Rosera is also trying to hire more people to expand the team. And do we have a sense for why Robinhood went with Rosera? I mean, what is so good about Rosera? Is there something that is very differentiated about its technology? Do the founders know each other? you know and and also i mean um we've talked about this a little bit you and i here in the newsroom i mean do you think this partnership is one that will last or do you think robin hood might over time uh start to step away because i mean even calci uh you know i think you've been on the show previously they've walked back a little bit away from calci right so uh what's so special about it and where do you see it going so if you think from robin hood's perspective Obviously, they saw prediction market is a fast growing business.

14:20It's becoming increasingly more important for them. So they will want to try to monetize it more and have more control over this business line. Initially, Robinhood sent most of his orders to Kaoshi, which means that they're heavily reliant on an outside party. And with Rosera, which they took a 45 % stake in, they're able to have better control over this business. And you might also want to ask that why isn't Robinhood launching their own prediction market exchange then completely in-house? And my understanding is that in the traditional financial market space, usually a broker does not try to also own an exchange.

15:05It's the same thing as how when you trade on Charles Schwab or Morgan Stanley E-Trade, These companies are not running or owning the New York Stock Exchange or NASDAQ. And there are a few reasons for this. They want to avoid conflict of interest and they want to avoid the heavy regulatory burden that comes with running an exchange. And most importantly, if every single broker runs their own exchange, it will really fragment liquidity. So it's actually better for the industry to have a few exchanges where everybody sends their orders to, so that the liquidity is concentrated at a few venues. So that's why there's this joint venture that Robinhood eventually decided to create with Sescahuana and they made the investment into Rosera, which is the entity that is the joint venture.

16:01Although, so correct me if I'm wrong though, I mean, Calci and Polymarket, have they occupied both roles here as both the exchange and then the customer-facing broker? I mean, they kind of occupy both roles, right? Yeah. Yes, they do. Which is unique. And then, I mean, Coinbase, too. You know, in crypto, I'm just looking for parallels here. Coinbase also occupies both the broker role and the exchange role. so it's not like it's not been done before it just hasn't been done before you know for the the equity trading companies I guess that's a great point in crypto it's very common to have what they call vertical integration which is exactly what you mentioned coinbase is both a broker exchange a clearinghouse for crypto trades but in traditional financial markets usually these functions are separated.

16:58Right. For now. Who knows what will come of it. I do want to get to another story that you covered this week. But I mean, let me just ask you, Yuichi, you know, do you expect more of these types of under the radar predictions, markets, exchanges to become the backbone of these, you know, the movement of traditional finance moving into production markets. Where do you see the story going? I think there's room for a few players in the industry because usually for brokers, they want to send their orders to multiple exchanges rather than just relying on one single exchange for a variety of reasons.

17:46They'd want to have redundancy. If one exchange breaks down or has some kind of outages, they want to make sure that they can continue to allow their customers to trade at a different exchange. And so I think Calci Polymarket will continue to be important, especially since they have such a strong retail brand now. People do know about Calci and Polymarket, and they go to their website or app specifically to trade on it. But I think for some of these other exchanges, such as Rothera, they are betting that institutional investors that may want to trade prediction markets in the future, they might want to go to a pure play exchange such as Rural Era, which is more closely aligned with how traditional market structure is.

18:38Right, right. Okay, very quickly, I want to pivot to another story this week that you've been covering. So the SEC has new proposed regulations out around crypto. This is, I guess, the first time that they've proposed some of these regulations. Help us understand why this is the first time they've done it and what exactly they're proposing. So the SEC made its first proposal on crypto regulation and specifically is focusing on how companies can raise capital by issuing tokens. And they provided some exemptions from securities laws and disclosure for companies that are seeking to issue a certain limited amount of tokens to raise money.

19:27And this is coming out right now because in Congress, the Crypto Law, the Clarity Act is currently being delayed. In this summer, the lawmakers failed to make progress on the law. And as we run into midterm, it looks like the chances of the law being passed is increasingly dropping. So the regulators are coming out to say that we're going to step in to provide some kind of clarity for the crypto industry because the law is currently being stopped. And this is the first proposal that's coming out from the SEC addressing. And so what would change tactically about what crypto companies can and can't do under this proposed regulation?

20:10So crypto companies, once this proposed regulation is enacted, which will take a while, it's not immediate because the SEC will still need to get public comments, they need to deliberate on the proposals. But once it's passed, crypto companies will be able to make use of it to launch token and raise funding without fear of being investigated by the SEC later on. And this is something that crypto companies have always done. They've always launched tokens and they raised funding. But in the past, there's always this gray area about whether they're issuing unregistered securities. And now they have a pathway to do so.

20:53Although I would say rulemaking by agencies are not as long lasting as legislation, because when there's another administration, regulators could potentially overturn or modify rules. that are being passed by the previous regulator. So in other words, it's really just the first time the SEC has come out and said, all that stuff that you've been doing for years, you can keep doing it and you have the official green light from us. And so tactically, it feels like maybe not a whole lot will change, but at least from our regulator perspective, it's perhaps the clearest green light that crypto companies have gotten to keep crypto-ing, I guess, as you could say.

21:39It is a great story, and it's one that we'll be watching a lot more. Yuichi, I want to thank you for coming on. That is Yuichi Yang, who covers crypto, fintech, and prediction markets here at The Information. Our next segment is with our partner, PwC. As OpenAI and Anthropic grow to astronomical heights, both have had to reckon with the tremendous responsibility that comes with that. For more on how Fortune 500 companies are thinking about the AI labs right now, I want to bring on Dallas Dolan, tech, media, telco, industry leader at PwC. Dallas, welcome to the show. It's great to have you, Yak.

22:16Always great to be here. Good to see you. So what do you make of this pushback right now that we're seeing against these big AI labs? I mean, part of it is they're growing so quickly. We've not seen companies grow this fast. They're going to IPO at some point in the next year or two. how are you thinking about the discourse around them right now?

22:37Dallas Dolen:Yeah, well, I don't know if the term is pushback, Akash. I think you have some market conditions that especially, including the last couple of weeks of conversations on open versus closed models, what does it cost to run them? What's the benefit of having a more efficient model? That's been a major discussion point. Do you get a lot more out of an efficient motor than an inefficient motor? And is it easier and maybe even more secure? right? We talked about that before, too. I think what you're seeing, though, is, you know, CIOs are under a ton of pressure. CFOs are under a ton of pressure. And CEOs are, you know, candidly also under pressure and sort of pushing its way down, you know, from the board saying, hey, like, where is the benefit?

23:16Dallas Dolen:And it's no longer an ideological, you know, pursuit necessarily to say I'm using AI, but there's the economic pursuit of what am I actually showing from, you know, using AI, both on the product side, what's getting into the product and, you know, what's allowing me to do more with it, but then also on the back office side, the adoption rate within a finance team, a treasury team, HR, and the like. So I think what we're beginning to see is that concept of, where do I get the biggest bang for the buck? And perhaps construct, if you will, of verticalization and the benefit related there too, that probably will benefit a lot of companies, especially as the SaaS providers get better at integrating a lot of these products into their platforms.

23:58So it sounds like, you know, regardless of what posts on X get circulated most widely or whose essay gets reposted, which CEO is loudest in the media, sounds like on the ground for Fortune 500 companies who you work with, it really is just a matter of ROI right now. and which model they can afford, which model they feel like they can get the most out of. It's a very mathematical calculation in some ways.

24:29Dallas Dolen:Yeah, and I think that there's a lot of benefit, candidly, to the model builders out there to show that ROI, right? And there's probably two different ways that they're going to go about doing it. One of them is going to be continuing to work with their teams, their sales teams, and also the FDEs they have on staff and the SIs they work with to build things out. I think the other one that we're going to see coming into the fold as we go closer to the end of 26 and into 27 is going to be how do they really begin to do two things. One of them is actually show benefit to integration and what is it you can do with a platform, for example, like an agent force, you know, the Salesforce or others that makes the underlying system of record a better system of record.

25:13Dallas Dolen:I think that's going to be a really significant play that we'll see coming into the fold. And what it's going to create, and I think this is a really important point, especially for those who are looking very carefully at ARR of the model companies, which is just a fantastic conversation to be had. It's also one that drives probably the most posts and the most clicks and the most responses on X, too. The thing that's going to drive consistency there is going to be having consistent buyers. It's not going to be, you know, the random consumer at home who's, you know, working with their, again, their Mac minis or what have you and building things.

25:49Dallas Dolen:It's going to be the large enterprises and medium-sized enterprises who are finding value from the SaaS platforms that have integration of AI into it. That's going to create, I think, a lot more consistent revenue flow for the models. And it's also going to, you know, create a more consistent experience as it relates to how do I measure ROI and what am I getting? It may pressurize some of the profits, if you will, in SaaS land, but it will definitely push them to the situation when they need to drive efficiencies there. Well, so let's unpack. There's a couple of things in there that I want to get into.

26:23So relationships that the model companies have with the enterprise software companies. You used Salesforce, AgentForce as an example. So help me understand this. Is it your view that the model companies will end up selling their products to businesses through these enterprise software companies or that revenue from the enterprise software companies will inevitably be the revenue that ends up driving their business?

26:56Dallas Dolen:I think it could be a little bit of both, but I actually see the business model being the latter. So it's the integration within the SaaS platforms themselves where, again, the system record is the thing that people rely upon the most, certainly the thing that will allow even the compliance functions of these companies feeling good about whatever it is that they're doing, whether it's tracking revenue or tracking their employees or tracking the sales teams and what's happening there and tracking obviously the metrics associated with for reporting purposes. it's that integration concept where you have some level of control as well as you know a very high degree of confidence in what's happening the push to actually get roi from those integrations which you know we've been seeing a little bit of that but you know if you go back a couple of years you know where you had uh you know for example the erps with the basis layers the cloud companies that was one motion but that was largely just to operate the platform on top of it it wasn't creating necessarily a lot more efficiency, although there were some things that maybe at the end of the line, the IoT world was benefiting from.

27:58Dallas Dolen:If you think about the ability to put agentic at that level all the way down for the end user within the system of record and sitting on the most efficient and perhaps also the fastest cloud, I think you start to be able to compound the capabilities and compound the benefit associated with it. What then ends up happening is that's the value proposition that the SaaS providers I think will be going out and saying, hey, actually, our value proposition is that we can consolidate a lot of things that you, the CIO or the CRO or the CFO are trying to do and put it all into one spot, maybe mitigating the concept of the individual contributor who's building their own agent on their MacBook or on their ThinkPad, but actually where everyone's using it in a consistent fashion.

28:43Dallas Dolen:That makes enterprise life a lot better and a lot safer too. but then the second half of what you said earlier was then the the profit margins for the software companies because using models from Anthropoc from open AI they're they're not cheap and and sure they have open source open weight alternatives available to them but what is there a new you know if 80 % was the old gross margin that software companies loved I mean do we get used to something closer to like a 70 or a 65? What ends up happening? Yeah, well, I think we've already seen the, you know, the margin progression happen in software and margin has migrated itself over the last couple of years into the hardware space, you know, as is.

29:25Dallas Dolen:And that's going to continue to happen, right? You see that migration of profit going to the memory companies, obviously the chip makers and to the infrastructure providers. That's probably, you know, at a static point, I don't see that, you know, changing a whole bunch, although it'll probably rest in that space because the demand is still outstripping the supply. So just basic, you know, macroeconomics say they continue to, you know, to attract profits. What ends up happening in the software space, though, is I think you do have the ability for them to command some level of, you know, margin enhancement, if you will, if they're providing a better outcome.