Welcome, everyone, to The Information's TITV. My name is Akash Pasfritz. It is Wednesday, May 13th. We are kicking things off today with some exclusive reporting out of our Asia Bureau about a new AI lab looking to raise money at a big valuation. We also have new details about Anthropics customers who are prepared to eat the rising cost of AI coding models. We'll then get to our daily update from Rocket Drew, who was covering the Elon Musk OpenAI trial. We'll bring on a guest to discuss how investment firms are tailoring their software investment strategies in the AI era. And we'll end the show with a look at our weekly finance column and the big crypto bill that everyone is watching this week. It is a big show, so let's get right on into it. A former star researcher at Alibaba is starting his own company and looking to raise at a big valuation. That is according to an exclusive story from The Information. My colleagues Juro Osawa and Jing Yang reported that story. I want to bring on Jing to share with us what she knows. Jing, welcome back to the show. Who is this star researcher? His name is Jing Yang Lin. He spent three years as a lead researcher at Alibaba's Quint team. And during that tenure, he's built Quint into the absolute top of the class models in the open source domain. And then he's also been credited for building a team of researchers that are really dedicated to open source technology. So if he accomplished so much at Alibaba, then why did he leave? So his departure was actually very abrupt and in a very Chinese fashion, very public. So there's been a lot of speculations about why he left. What I can tell you is that in early March, he just all of a sudden posted on X that he's stepping down from Quinn. Bear in mind that that was before Alibaba executives had had any opportunity to actually announce his departure internally to staff. This is a very rare sort of public fallout from the Chinese context and the Chinese culture. You talked in the story about how much money he's looking to raise, and the valuation is pretty high. It's a $2 billion valuation. Do we have any idea what this startup is going to be or what his vision is at this price tag? We don't really have a lot of details right now. We just know that he's looking to do something related, again, to fundamental model development and research. I know a $2 billion valuation funded by a renowned researcher probably sounds like another Wednesday in the U.S., but in China, for a brand new company or brand new startup that has not had any product, to try to fetch that valuation is almost unheard of previously in the Chinese air land. And, you know, I do have a question about the valuation because you mentioned the story in the U.S., these researchers leave and there's sort of a question, well, maybe they'll get acquired, right? Maybe they'll get acqui-hired. We've seen all sorts of deals. It strikes me that based on the reporting that you did, that's not likely to be the case in China? Exactly. So in China, we simply have not seen the similar path, exit path for startups like we've seen so many times in the US. We simply haven't seen any of the tech incumbents, whether it's a Tencent, Alibaba, Biden, those guys out there actually acquiring or acquire startups or startup funders. That just simply hasn't happened in this wave of AI investments. And on top of that, Chinese new AI labs face the immediate, out of the gate, they face the immediate challenge of having to secure enough compute, given that the U.S. still controls exports of advanced chips to China. You know, as you say that, I can't help but wonder if Alibaba, if they were making so much progress, and if this researcher leaves, where does that leave this company then in terms of its AI pursuits? Do we have any sense? We do. So shortly, you asked me earlier why he left, but we don't know the exact reason. There's no official story given. But however, what we do know is that immediately after his departure, Alibaba started to reorganize the entire QAn team. And then shortly after that, Alibaba also started to prioritize more on making its models and the products attached to its models more proprietary so that hopefully that they can charge more revenue for it. But previously, Alibaba, like many others out here, was more dedicated to open source, making things cheap and good sort of route. And then so actually a few hours ago, Alibaba announced its first quarter earnings. And for the first time ever, Alibaba executive told investors that they expect that annualized revenue from Alibaba's model and application services will top $4.4 billion dollars this year i know that doesn't sound like a lot but bear in mind that this is a company that just started to really commercialize um their technology and this all happened uh in the last couple months after lean's departure jing before we let you go president trump is in china this week with a whole roster of tech executives uh what are you watching for in the discussions this week what What should we be paying attention to? Yeah, President Trump is traveling with really the who's who in the U.S. tagline. The CEO of Apple, you know, we'll have Jensel Han joining last minute. And then Tim Cook and Elon Musk joining. And also the executives from countries like Micron and Qualcomm, Cisco. And then there's also, you know, banking and financial big shots like Jim Fraser, the CEO of Citi. However, I want to say that this is really a milestone visit. This is the first US presidential state visit to China since 2017. And in 2017, the president that visited China was also Trump, because Biden didn't visit China during his term. And then given the delicacy of US-China relationship right now, based on analysis and the people I've talked to, the things that are going to be high on the agenda would probably be Iran's streets of homes and Taiwan tariffs. Anything related to tech, AI, and semiconductors will hopefully be discussed, but probably not really be high on the agenda. But I guess at least from our end, we'll be watching in the next two days as Trump makes this historic visit. kind of agreement or deal or goodwill handshake will come out when it comes to AI and the semiconductors. Right. Well, Jing, I want to thank you for coming on. That is Jing Yang, our Asia Bureau Chief here at The Information. Anthropic's latest pricing model change means that some customers could end up paying more for AI, especially for power users of the company's models. A new exclusive story from my colleagues Laura Bratton and Aaron Holmes shows that despite all that, customers are in many cases willing to eat the cost. I want to bring on Laura to share more with us what she knows. Laura, welcome back to the show. Remind us, what was the pricing model change that Anthropic implemented recently? Yeah, so recently Anthropix switched from charging Claude Enterprise customers $200 a month, which included AI usage, to charging a flat fee of $20 per user per month and any additional usage of computing power that those companies use. So it was a change towards usage-based pricing, basically, instead of a fixed model. And so how are customers reacting to this? Yeah, it's really interesting. So far, customers are eating the costs that I talk to, and they're definitely worried about those costs increasing but it not changing their behavior yet For example I talked to a smaller software company called PagerDuty and even though Anthropic has recently changed its pricing model and it will affect this company in particular, they're still rolling out Cloud Enterprise to 1,200 employees and really think that it's going to make them more efficient and help users basically do things faster by accessing all of their apps through Claude rather than going into each app separately to do their jobs. Did you get any kind of a sense from PagerDuty or from any other customers, I mean, the extent to which prices are increasing? I mean, is this like 10, 20%? Is it like double, triple? Did they offer any clarity there at all? They didn't offer clarity, but from what we previously reported, companies could see costs double or triple. I did talk to one small customer that has about 30 employees. And now I should specify that the usage-based pricing does primarily impact companies that have more than 150 employees. But even smaller companies are noticing that their costs are going up as their users hit usage limits. So the smaller company I talked to called TelAid, it's based in Connecticut, said that the 30 users that it has of Claude Enterprise, they've seen costs from those users triple from March to April, which I thought was a really stark example of how companies of any size can see costs really go through the roof. And I am curious, the reason behind their willingness to pay more for these products, Are all these companies just in love with Claude Code? Like, you know, they just can't get enough of it? Or what was the core reason they said they're willing to do it? Yeah, it's interesting. PagerDuty in particular is just an experimentation mode. And they were really transparent about the fact that costs are going to be lumpy and that they don't know exactly what the return on investment is yet. I think ServiceNow is maybe a better example of efficiencies seen from Claude at the same time as there's a lot of concerns over how costs are rising. ServiceNow's CIO told me that she recently met with the CFO to figure out how they're going to manage costs. But she told me that the efficiencies they're seeing is, you know, salespeople are using Claude to help prepare for customer meetings. They have an AI agent that they built with quad code that's helping them reduce the time they spend preparing for customer meetings by about 95%. And then they also launched a new customer tool in 20 days that they said would have taken them months to do previously. But I think that we're starting to see the rubber meet the road where it sort of begs the question of will the return and will these efficiencies match the actual cost companies are starting to have to pay for these tools? What about Anthropic? What has been the impact of the pricing model on their business? Yeah, I mean, they've seen revenue grow at an unprecedented rate, which we've reported on. And I think that this is all good for Anthropic. And as we see, you know, computing costs rise, we recently reported on a GPU crunch. I think that, you know, it's in the best interest of model providers to charge based on usage or charge based on, you know, tasks an AI tool completes rather than charging a flat fee. But I think we're going to start to see some customers potentially pull back their usage of AI tools because of this new cost they're seeing hit their financial statements. And it'll be really interesting to see how their behavior changes. Right. And see if it's sustainable, really, because presumably there's a ceiling with which people are willing to pay. Laura, I want to thank you for coming on. That is Laura Bratton, author of our Applied AI newsletter, here at The Information. The Information's Rocket Drew is covering the Elon Musk OpenAI trial from the courtroom all week long. Here is his latest update. Sam Altman finally took the stand in Elon Musk's lawsuit against OpenAI. Altman's testimony has been highly anticipated because his own actions are central to the allegations that are made in the case. In response to friendly questioning from his own lawyer, Altman began by laying out some familiar arguments. That he didn't know of any promises or conditions that were attached to Musk's donations in the early days of OpenAI. That he has disclosed conflicts of interest to the board. and he said he even informed Musk that Microsoft was going to invest, both before they invested in 2019 and also 2021. The real fireworks, though, came when it was Musk's lawyer's turn to question Altman. Musk's lawyer has been pretty aggressive in his questioning of some witnesses in the trial, but with Altman, he took a different tack. His tone was almost soft and soothing when he started asking questions, but he asked a series of ferocious questions about Altman's trustworthiness. He talked about allegations from witnesses, including past colleagues, board members, executives that have testified in the trial, saying that they found that Altman was not consistently candid with them and that he lied or put words in other people's mouths. So Musk's lawyer asked Altman to address these things. He also asked why the jury should believe anything that Altman has to say. Altman defended himself. He said that he believes he's an honest and trustworthy business person. So, both sides were able to score some points. It seemed like, ultimately, Altman's testimony favored OpenAI. For example, if you look at this polymarket that's tracking the odds that Musk eventually wins the whole lawsuit, over the last few days it's been hovering around a 43% chance that Musk eventually prevails, and over the course of the day's proceedings, it fell to about 35%. Now, maybe some of that is a testament to the ways that Altman supported Opening Eye's side of the case. It could also be that he didn't end up supporting Musk's side of the case. There was some question that if Musk was really going to pull out some silver bullet argument that we didn't know about, it would happen during Altman's testimony. Given that that didn't happen, you could imagine that Musk's odds look a bit less favorable as a result. So there's just a bit of time left in the trial before the jury starts deliberating. The jury will decide whether OpenAI, Altman, and Greg Brockman are liable, and then the judge will make a final determination about the remedies to Grant, if any. So that's what's coming up last in the must-be-Altman trial. FTV Capital is one of the many investment firms tailoring its software investment strategy to the AI era. To talk more about the software landscape at large, I want to bring on Brad Bernstein, managing partner at FTV. Brad, welcome to the show. It's great to have you here. Thanks so much for having me. It's a pleasure. So just remind us here, FTV, are you private equity or venture capital? Are you both? So we are a growth equity firm, so we only invest in high growth businesses, Our average company that we invested in about a 50% growth rate last year. And we specialize in those high growth businesses that are demonstrated and validated their markets, their products, but are still in high levels of growth. And we don't use much debt. So we're not leveraging up these companies. We're really leaning into operational leverage. Got it. Got it. Okay. So I want to get your take then on a couple of the hot button issues that we've been reporting on here at The Information. And look, the last reporter that just came on before you, we were talking about the pricing model changes that Anthropic has embarked on and how customers are eating those costs. And we haven't really seen a diminishing in demand yet. You work with dozens of software companies in your portfolio. How are they dealing with the issue of pricing model changes at Anthropic and also pricing model changes broadly around the AI ecosystem? Jim O'Shaughnessy- It's such a great question because right now is really a period of massive experimentation. And so people are less focused on the cost and more about figuring out what they can do with this technology especially when you comparing it to the cost of additional headcount So the ROI is there and people are trying everything But there's also real risk that there's false positives in the market where people look at the demand or the usage and think it's sustainable forever or think it's just going to keep growing forever. But to your point, as pricing is changing, we are going to see people evolve what they do, how they do it. You're going to see more interest in open source, smaller LLMs. You're going to see smaller LLMs that serve purposes in a more cost-effective way. Even at FTV, we have certain things that we encourage people to use chat GPT for because it's cheaper. And then we have other things that are more complex where we might use Claude. So, you know, there's already starting to see this. And I think that's going to be a bigger and bigger issue as time goes on. When you say false positives, expand more on that. What do you mean? Absolutely. The reality is that right now, everyone looks at the growth in the AI market, and they just see it as long-term, permanently sustainable. But take myself for an example. I have three LLMs that I subscribed to last year. This year, I'm down to two. I have a feeling that over time, I might go to one. At a certain price point, I'm willing to have multiple LLMs, but as the price goes up, I'm likely to reduce the number of LLMs I want to pay for. So I think there can be a lot of early adoption where people are experimenting and where the costs are subsidized and low that drives demand that may ultimately go away. Right. I want to ask you a little bit about your strategy at FTV insofar as keeping talent at the portfolio companies that you're working with. And the reason I ask is because the talent wars is a topic that comes up time and time again in an era where all these application layer companies, AI startups, even AI labs are hiring like crazy. What is the pitch that you're making to folks working at your portfolio company? Some of whom are software companies have been around a lot longer. to say, no, no, no, no, stay with us. You know, don't go to the AI coding startup. I imagine the pitch is getting a bit tougher, right? Well, look, all of our companies, the management teams have significant equity. And in many cases, the people we have at our companies are not necessarily AI experts. So we're leaning into the training and helping them develop skillsets. And they're pretty excited about the equity they have. You know, if you look across our portfolio, we have very high growth businesses, proven businesses, where they see a lot of opportunities. So the key is to demonstrate that we're committed to innovation, we're committed to growing the companies, and that we've given them fantastic equity packages. And very honestly, while there's fantastic opportunities at many of these frontier labs and so forth, the work lifestyle balance at some of those firms is pretty terrible. And you see some churn because of that, even with the high economics. And if you're coming in at these very high valuations, it's much less clear about what the future holds or how much higher you can go when your options are struck at close to a trillion dollar value. So that really hasn't been an issue that we've seen with our talent. Even with the SaaSpocalypse taking a beating to public market valuations of these software companies, you haven't seen concerns from employees saying, hey, I have this equity, but I don't know how much it's going to be worth, given that Circus Now is down 50% in the last year. That's a really key point. So let me unpack it a little bit. So where I think you do see the kind of issues you're talking about is people that were doing these large, mature software company deals where they have very low growth rates. They're using leverage. In some cases, the leverage is making it hard for them to pivot and invest aggressively in R&D. And they're really worried that the buyout was done at too high a price. It's a value trap, and they don't really see the path to an exit. I think you're right in that segment of the market. Where we sit, we've always been pretty disciplined about value entry and making sure that even post-SASpocalypse, our entry prices still make a lot of sense. And we're really investing around the fundamentals of exiting on an EBITDA multiple that makes sense in any environment. It's sort of a cash flow fundamentals basis. So when you have businesses that are growing 50, 40, 30% a year, and you have the opportunity to take AI as an accelerant and enhance the growth opportunities, and everyone can see that the model still makes sense and the exits still make sense, you don't have those kinds of churn issues. I wanted to ask you about the partnerships we've seen recently between AI labs and consulting firms. I mean, we saw OpenAI has now embarked on their own sort of consulting venture with a couple of other firms involved. They're also partnering with firms like PwC. You know, the role of the consultants is something that has fascinated me, certainly in an era where the labs have their own forward deployed engineers. I mean, you know, it's sort of intriguing to me is this competition. Which one is better? Do the consultants have a role if the labs have their own FDEs? What do you think of that? I think it's really interesting. The reality is we just don't have enough skilled people to be forward deployed engineers or to do the consulting work as it relates specifically to AI. So, you know, I think it'll be interesting to see can these consulting firms that are coming together with multiple PE firms actually execute. The vision sounds great, but it also sounds like there are going to be a lot of cooks in the kitchen and there's going to be a lot of questions about whose portfolio companies get prioritized, why one company versus another, what if the management team of a portfolio company has a different point of view on which LLM to use. I can see a lot of issues for conflict between the different shareholders at the private equity level, the portfolio companies, and even finding the resources. So I'm rooting for their success, but the execution sounds challenging from my seat. Last question for you. We've reported on the GPU supply crunch and how, in some cases, venture capital firms, I'm thinking of General Catalyst, will help to negotiate these larger deals en masse for their portfolio companies to access chips and access compute. I'm not sure the extent to which the companies in your portfolio are actually involved in building their own models, but are you doing any of these kinds of mass negotiations on behalf of your portfolio companies? We are not. We've built a lot of partnership relationships with the LLMs and other key players in the ecosystem. We feel it's very important and valuable to be providing that connectivity for our portfolio. but we are not having any kind of GPU crisis or compute problem across our portfolio today. Great. Well, Brad, I want to thank you for coming on. That is Brad Bernstein, managing partner at FTV Capital here on TITV. Crypto is at a critical junction as the Senate Banking Committee is expected to approve a key bill this week. The Clarity Act would still have to move through Congress, but our senior finance editor, Ken Brown, wrote in his weekly column that the crypto sector may not for a long time see this window again to get this extent of regulatory clarity. I want to bring him on for all of his thoughts. Ken, welcome back to the show. It's great to have you here. My gosh. So let's start with the news. Where are we in this story? What's happening this week with the Clarity Act? So it's going to get voted on by committee, the Senate Banking Committee. This has been dragging on and dragging on. And so this is a big moment. And it also sets the bill up for, you know, what some version of this will be voted on by the Senate. Okay. And so we think this is ultimately going to get to the Senate in the next few months, right? Yeah, this is going to go from this committee to the Senate. And, you know, they'll pick it up. I mean, there's a bunch of stuff going on, so you don't know when they're going to pick it up. But there's a lot of urgency down in Washington because there is, this is an election year and so people are distracted it's also getting close to summertime so you know really there aren't that many windows where you're gonna actually get stuff done in the Senate and so they're really they're pushing to get it get it down to the down for a vote but it just unclear when that going to happen Right OK What in the bill So the bill is this big crypto legislation that will put a lot of regulatory controls on crypto which the industry wants. It will it will make things a little more stable for them in the sense of you'll know who your regulators are. You'll know what the ground rules are a little bit. And so what crypto wants is stability and reliability. And what they don't want is what they view is what happened in the Biden administration where they were just being, the government was just going after them for a bunch of stuff. So they want surety in that. And frankly, the finance industry broadly wants that because it would allow more investments in crypto, which everyone is trying to figure out. And so having legislative certainty and legal certainty will really help. And can you, you know, we've talked about this a little bit in bits and pieces on the show, but can you just give us an example of, you know, regulatory clarity, just an example of, you know, a clause or something that, what exactly is the rule that is being proposed? You know, the SEC regulates the exchange. What is it? Well, that's right. So, yeah, I mean, the so just as an example, it's really not been clear who the regulator is for crypto. It's a Securities and Exchange Commission or the commodity, the CFTC, the Commodity Futures Trading Commission. And so the question is, who regulated what? And so the SEC is going to regulate things that are securities, which is typically its role. So if you issue a token, say, that would be run by the SEC and the commodities are going to be commodity. The CFTC will regulate the, you know, the trading of things that are, you know, sort of non-securities, the equivalent of commodities. And, you know, there is this is hundreds of pages long. And so but there are things like that. There are rules around, you know, trading. There are rules around custody, things like that, how exchanges work. So it just stabilizes and gives a lot of stability and surety for what's going to happen in crypto. OK, so who is for it and who is against it? Well, I mean, it is a fight. So there is there's typically, I mean, broadly speaking, Republicans are for it. You know, Trump is for it and the Republicans have embraced crypto. And then typically, and this is not consistent everywhere, but the Democrats are against it. They've been worried about people losing money and all that stuff. But then the biggest split is between the crypto industry and the banks. And the banks are strongly against this bill for, you know, in one specific area. And that is the banks are very, very powerful in Washington. And they have a lot of constituents there because there are thousands of banks in the country. And every congressperson has a bank in their district and the banks lend to the small businesses. And so that's a big constituency. So this is a big fight going on. And the banks are against this because of what portion of the bill are they are they really concerned about specifically? So the big the big fight is over stable coins. So stable coins are this crypto that is always pegged to a currency. It's typically always worth a dollar. One stable coin is always worth a dollar, unlike the other stable coins that are, you know, fluctuate wildly. And stable coins, the crypto industry wants to issue stable coins and to pay a yield to the investors in the stable coins. Like you get in your bank account and like you get in your money market account, you get a yield. The banks hate that idea because they feel like all these deposits will leave the banking sector and people will go buy crypto, buy stable coins. The problem with that is banks take your deposits and my deposits and they lend them out to businesses. And that's how banks make their money and that's how the economy runs. And so banks are fearful that and regulators are fearful that money will leave the banking sector and it could really hurt the economy and particular small businesses. So that's the fight. So there was a stablecoin bill last year, a law that passed last year that was supposed to prevent stablecoins from paying yield. And it's in the law and it's all there. And then the crypto industry figured out a loophole and they've been paying what they call rewards. And that made the bank industry really mad. And so now this is their chance to close that loophole and they're really not going to let it go by. And we're confident that, I mean, I guess the question I'm asking is, whether or not this passes or not, what's the likelihood that there will be more loopholes that come up for people to do whatever they want either way? Well, you know, I would have to say the banking industry and the crypto industry are very, very good at finding loopholes in laws. They have lots of lawyers. And so, yeah, of course, that's totally the risk. But it is a huge issue for the banks. I mean, the banks kind of got a little duped in the stablecoin bill last time. They thought they had won this rule, and clearly they didn't. And so they're really not going to let go. So they're trying to unleash all of their backers against Congress, the small business owners, the farmers, the local banks. So they're doing that. And the crypto industry, which has a ton of political money to donate, is countering with hundreds of millions of dollars in donations. And so that's the fight now. Well, and so this is also, I mean, you and I were chatting this morning here in the newsroom. I mean, this is a scenario where you have the banks and the Democrats who oppose this together, which is not a common pairing, right? Right. No, the Republicans were always the party that would back the banks. And Democrats would be, you know, on the other side. I mean, I'm generalizing, but. Yeah. And so this is that strange bedfellows. The banks are sided with the Democrats. And actually, the AFL-CIO came out yesterday and said they don't like this law. And so it's now the labor, the banks and the Democrats, which is a weird mix. And then on the Republican side, it's Republicans and crypto. And that, you know, we'll see. We'll see where that goes. But, you know, the banks, I mean, both crypto and the banks could end up playing a role in this election because they both have a lot of power in Washington. And and someone's going to be really unhappy at the end of this. And do we have any sense, Ken, for like, I mean, this will go to Congress. But I mean, is the expectation that this is going to get passed based on the existing makeup of Congress or we don't have any idea right now? So, you know, I think support for crypto has gone down in Congress. It was pretty high, you know, last year. And so the House actually passed this law last year. And then it had been sitting in the Senate. And this is a different version. And the House will have to pass this version if the Senate passes it. And so the question is, a bunch of Democrats voted for it in the House. And it seems unlikely that that many of them will vote for it. And then, yeah, the question is, do all the Republicans vote for it, too? Because it is a it is contentious and, you know, public support for crypto. I mean, the other thing was last year when crypto that they passed the law, crypto was booming and the market was way up and people were diving in and all this stuff. And this year, crypto is just dismal. Right. The market is down. You know, people have lost a ton of money. People who jumped in last year have lost a ton of money. There's been all kinds of, you know, typical crypto scams. The administration just went after the crypto exchange, Binance, for allowing money to flow from Iran, you know, against sanctions. And so this is like, you know, crypto is not this is not crypto's greatest moment. Right. Well, Ken, I want to thank you for coming on. That is Ken Brown, our senior finance editor here at The Information. That does it for today's show. A reminder, we are on this stream Monday through Friday at 10 a.m. Pacific, 1 p.m. Eastern. If you can't make it then, episodes are available on theinformation.com, on our YouTube channel, or wherever you get your podcasts. Make sure to follow us on social media on X, Instagram, and TikTok. I'm already excited for our next show tomorrow. Have a great rest of your Wednesday. Bye-bye for now.
