Nvidia Challenger Cerebras’ Funding, TikTok Update, Navan’s Signal About Tech IPOs | Sep 22, 2025

0:12Welcome everyone to the information's TITV. My name is Akash Pasricha. It is Monday, September 22nd. We have got a great discussion planned for you today about the state of the IPO market. And we're going to break down the IPO prospectus for Nivon, which filed on Friday. We're also going to talk to the CEO of another company challenging NVIDIA in the AI chip space. Andrew Feldman of Cerebris is coming on the show. But I want to start with the latest on the TikTok saga, which made big news over the weekend. There was a flurry of updates and also some news that the Murdoch family could be getting involved.

0:48To break it all down, I want to bring on Sylvia Varnum-Oregan, who covers all things Washington, D.C., for the information. Hey, Sylvia, how you doing? Hey, Akash. I'm doing great. How are you? Good. It was a pretty busy weekend for you, I take it. I mean, gosh, there's a lot going on in the land of Washington, D.C., and tech, not the least of which was all the TikTok stuff. Yeah, it was a whirlwind weekend, that's for sure. So, look, I want to get into the chronology here, but last we spoke to you, you know, it was just before President Trump was going to speak with President Xi over the phone on Friday.

1:21And, you know, we kind of left it on a bit of a cliffhanger because we didn't quite know what would come of that call. So let's start on Friday and then walk me through what happened over the weekend. Right, sure. So on Friday, as you say, Trump had this much anticipated call with President Xi of China. And afterward, he said in a post on Truth Social and in a subsequent press conference in the Oval Office that President Xi had approved the deal or a preliminary deal at this stage. So that was quite significant because this was a big holdup for a long time. The question of whether China, the Chinese government, would approve a deal for ByteDance, which is TikTok's parent company, to sell its U.S.

2:04operations to a cohort of investors. Right. Okay, so that was Friday. So China gives the green light saying, okay, fine, you have this deal, you know, we give you our blessing, you know, you can move ahead with it. Then we get a little bit of detail on the weekend as to what this deal could look like with board seats, the Murdoch family. Talk to me about what happened. Right. So on Saturday, the White House press secretary went on Fox and offered a little bit more detail about this proposed deal. She talked about how there would be a seven-member board, for example, and six of those seven members would be Americans.

2:41Now, the last board seat is going to be designated by ByteDance, which is, again, TikTok's Chinese parent company that will hold a stake of just under 20 % in the new company that will own TikTok US. I should also add in an interesting detail that the Wall Street Journal reported as well that the US government will receive a multibillion dollar fee for effectively brokering this deal, which was another sort of fascinating bullet point in this whole saga. The deal-making administration. I mean, gosh, President Trump making sure that he gets paid his investment-making fees for all this looks like.

3:20I should say, you know, we'll get into what came on Sunday, but just on the board. So we know that six of seven will be held by people in America. I guess we don't have any idea as to who will sit on the board as of now, right? That's right. We don't know the identities of the individual board members, just that six of them will be American. Right, right. Okay. We also had this detail about sort of the algorithm. I mean, talk to us about what we learned about, you know, the sort of moving ahead of the U.S. getting a copy of the algorithm, you know, how the licensing agreement will work. Where did we land there?

3:58Right. Well, in that same interview on Fox, the press secretary, Carolyn Levitt, said that the algorithm will be controlled by America. I think that that is a sort of interesting detail to unpack because it's not quite as clear cut. Some of the reporting that I did yesterday, which we published this morning, shows that what will actually happen is that ByteDance will create a copy of its algorithm and license that technology to TikTok. And then to TikTok US, I should say. And then TikTok, specifically Oracle, stay with me, people, which is one of the investors in this deal, will be responsible for retraining and sort of overseeing the security and functionality of that algorithm.

4:46So it will be under the purview of these new investors, but it's a licensing arrangement at the end of the day. And the thing that I sort of find a bit, not mysterious, but it's a big question mark for me is, okay, so you license the copy of this algorithm. I mean, the algorithm, it's not like it's a static thing, right? I mean, you have software teams that are working to improve it every day, every week. And so I guess the question I have is, okay, so you get a copy of this algorithm. Is Oracle going to be as good as ByteDance and improving it as time goes on? Will it remain competitive? I mean, these are questions that I guess we don't really have answers to insofar as what TikTok US looks like on a go forward basis.

5:32Yeah, well, I think you're touching on a really important point, which I think is something that we're kind of grappling with in the newsroom, which is that there are so many unknowns still here. How do you split off a business like this? How do you continue to operate this algorithm when you've got a duplicated version of an existing one? What does that maintenance, what does that retraining look like? I'm sorry. I do think that these questions are quite important and thus far unclear. Right. So now let's go to this week. We have, you know, another week ahead in the land of TikTok. What are we expecting in terms of, you know, is there going to be things that the White House is going to sign?

6:19Can we expect more details to come out? What should we be looking for? Right. So more details came out over the weekend about this deal. As you touched on in the intro, Trump alluded to the idea that the Murdochs could be involved, for example. And I learned through reporting that if an investment like that is made, it will be Fox Corporation rather than Rupert and Lachlan individually. So we are learning more every day, but I think this week will be significant because the White House is due to issue an executive order, which will essentially outline some of these details about this proposed deal.

6:56deal, it will also include a sort of, it will buy the group of investors more time, which I think is significant because it sounded for a second there like things might be moving quite quickly. And Carolyn Levitt said over the weekend that she expected the deal to be signed in the coming days, for example. I don't think that's quite right. I think this could actually take quite a long time. In this executive order that's coming out later in the week, the White House is going to provide for another 120 days to sort of stave off a law coming into effect that would - It's like more buffer room, essentially.

7:36Law buffer, right. Because remember, this whole thing started because a law was passed that essentially you have to divest from ByteDance or Facebook. So this is providing more time for that to not come into effect, which effectively allows the investors more time to finalize all the legal and financial paperwork and details. So, but it will be a significant step that this executive order comes out because it's essentially the White House offering its approval of the contours of the deal and creating a little bit more insight into the path forward. Right. Well, look, I think the key thing you mentioned in there was that whichever deal track you're looking at, it's just more time to figure out what exactly will happen.

8:18And so I do suspect that this is going to go on a lot longer than perhaps the last few days even suggested. Sylvia, thank you so much for coming on the show. That is Sylvia Barnum-Oregon, our Washington, D.C. correspondent at The Information. Okay, well, it has been a fun time for IPO watchers since tech companies are finally going public again. The latest company to push ahead is Navon, the company formerly known as Trip Actions. The company filed its S1 on Friday. To break it all down, I want to bring on CJ Gustafsson, who runs a great newsletter called Mostly Metrics. He's going to help us break it all down.

8:51CJ, welcome to TITV. It's great to have you. A-Kosh. First time, long time. Thanks for having me on. That's a great sign you got in the background, man. Wow. It's like you're coming on live from a movie theater, it looks like. I'm trying to step it up. Sometimes we get my dog in back, but I think he's downstairs begging for treats. Okay. So let's talk about, let's start narrow and then we'll go zoom out a little bit. So Navon is the latest company that released their S1. You did a great teardown of it over the weekend. First of all, for those who aren't familiar with the company, what exactly is the business model for Nivan?

9:29What does it do? How does it make money? Just let's start there. Definitely. So like you alluded to, it's the artist formerly known as Trip Actions. And so they started as a travel booking company. So there are all these different APIs that go on in the background for you to be able to book a trip. It's called a global distribution system. And basically what they did is they found a friendlier way to work on what you call T &E at work. So it's when you travel for work, for your hotel, for your rental car, for your flights. And as you can imagine, they hit some bumps during COVID. And they came out the other side with this pretty awesome layer cake strategy where they went from just being a travel platform to a full expense management platform.

10:11And since then, they kind of have these three distinct layers of how they monetize. They have the travel booking platform, which is a take rate model. They make about 7 % on everything that's booked. They have a subscription expense management platform, which is done over multi-year deals. So CFOs, I'm formerly a CFO like myself, can see what's going on and try to get all the expenses in one place. And then they also have a corporate credit card program, which allows them to essentially double dip if that's the card in the background where they're booking trips. So I wouldn't call them a pure play fintech, but they're kind of a combination of like a marketplace middleman with a lot of corporate card elements.

10:47And which of those three business lines you talked about is the biggest for the company right now? Yeah, great question. So 90 % of their revenue is the usage-based take rate model on travel. And that's kind of the edge of the wedge. That's how they land their initial customers. And ideally, they add on the other products with that. Got it. Okay. So they've got this usage-based bookings platform. They take a percentage of the bookings that they sort of help their customers book. I mean, my numbers here, we're looking at$613 million in last 12 months revenue, growing 32 % year over year. Tell me about your reaction to that.

11:26I mean, you actually, if I read, well, I did read in your in your newsletter. You were in the travel space for a little while. You tried to build a travel company. I mean, help us put that number into context. What are they doing right? And what could they be doing better in your opinion? Yeah. So this is not Expensify 2.0. Okay. I think it's important to say that off the bat. This is a much larger company and a company with its kind of hands in different parts of the ecosystem that is expense management. But to your point,$613 million, if I was to look on a relative basis to other IPOs out there.

12:03I benchmarked the last 14 companies that have been out in the last two years. That is definitely testing the smaller end. So the median LTM revenue, so that's the last 12 months of revenue that you have when you IPO, has been$813 million and growing at 31%. They're in the growth zone, like you said, growing at 32%, but they're about$200 million short of that. So what's exciting to me is it tests the market to see, are we ready for some of these smaller IPOs? And what market cap do you need to self-sustain once you are public? So how do you think it does when it eventually does go public? Well, I'm not a stock picker or anything, but their last Series G round was$9.2 billion, right?

12:46It had some awesome investors on there, including A16Z, PremG. These are smart people. They probably wouldn't want to get into this if they were going to lose money. but if you were to forecast out and say well let's say they're going at 32 now let's just be i don't know let's call it 25 next year so companies are usually valued on a next 12 months revenue basis that would put them around 740 million and if you multiply that out and say close to a fintech multiple then pure play sass it's it's probably in the ballpark of eight to 10, 10 on the high side. So once again, who knows? Figma went out, blew doors off 175 % up first day trading at like 45, 50 export revenue.

13:31Now they're down towards 25, but it is a smaller size to go out. It's like, what size of a rowboat do you want to be in if there is a storm? Right. And one of the things I wanted to ask you about is sort of the profitability side of the story, because you put out this video on YouTube. It was a 20 minute mini documentary on how the sizes of IPOs have evolved over time. I watched all 22-something minutes of it. It was a great, great video. But you talked a lot about top line, and I'm with you. The take-home was companies, the threshold for going public from a revenue perspective, the threshold is getting higher and higher.

14:07What we didn't talk about, though, was profitability. And the thing that I wanted to ask you about is how you've seen profitability of these IPO candidates or IPO companies changing because a couple, you know, two years ago, you know, there was a window where people say, hey, you know, you have to be at least break even on cash flow, or you have to at least shown a couple quarters of profitability sort of be convincing as a public company candidate. I feel like that threshold has come down a bit over time. I wondered if you'd studied that at all for the last sort of 10 IPOs and what you're seeing there.

14:42Yeah, out of the last 13 or so IPOs, the majority are making money. And there's no beating around the bush, Akash, with this one. They are losing money. Even on an LTM basis, they lost$80 million net operating income, right? And you say, well, what's the rule of 40? If I know your growth rate, 32%, they lost 13 % of revenue. So they're well below the rule of 40. They're somewhere in the 20s there, depending on if you want to use free cash flow, adjusted EBITDA. But like profitability, I'm wondering, will this test if it still matters? Does efficiency matter more than we thought or is the market super exuberant?

15:21Do you think the threshold comes down at all again, you know, back to that 100, 200 million dollar revenue range that we used to see, you know, something like 12, 15 years ago? Do you think IPOs get smaller again ever? I think of like when Twilio went public, right? Like they were at a billion dollar valuation in 2016. They raised a hundred million dollars. That was blockbuster. Yeah. And now we're looking at it. Like if you're raising less than 500 million, it's on the small side. Yeah. I don't think it comes back because you have these large institutional investors who need to be able to get in or out.

15:55And like a market cap below 5 billion makes it super difficult to do that. So like, could you go public and have a really great company at a hundred million in revenue, of course, but like, I don't know, it's, it's, it's a risky proposition to want to be out there. And, um, like I said, like these bigger investors need to, they probably need to hold more than, you know, two, three, 4 % of your company to get out of bed in the morning without triggering like regulations and filings and stuff. So being smaller is not your friend, but once again, this, this IPO is exciting because it tests the lower bounds of size and efficiency.

16:30Yeah. And you know, the reason I asked questions is because the next guest we have coming up is the CEO of Cerebris, an AI chip company. And they are one of those, quote unquote, smaller companies that has talked about going public. They put a filing out there. It's$136 million in the first half of 2024 was their revenue. So you double that. I mean, look, the company is growing faster. The question I'm going to have for him is when the company is actually going public, because there's been some delays there. But again, the question is, could smaller companies, maybe in different pockets of tech, could they make for compelling IPO candidates?

17:05We'll have to see how it all shakes out. CJ, thank you so much for coming on the show. I love the teardown. And next time we get another one, we'll be sure to have you back on. That is CJ. Thank you, sir. The author of the Mostly Metrics newsletter. Okay. Well, if you're an avid reader of the information, chances are you might recognize the name Cerebris. The company is developing AI chips it says can challenge the likes of NVIDIA. It filed to go public last year, but those plans have since faced significant delays as a result of national security reviews. In the meantime, it has been looking to raise$1 billion in private funding, according to our own reporting.

17:40I want to bring on the CEO of the company, Andrew Feldman, to talk about this moment for the company. Andrew, welcome to TITV. It's great to have you. Well, thank you for having me. How are you doing today? I'm doing well. So look, there's a lot to get to. It's kind of an exciting time for the business. The company planned to go public. This was back in 2024. It faced delays. I just want to get a sense here. Where are we at with those plans? I appreciate you inviting me on today. We aren't talking about our IPO plans, but I think we are talking about how well the business is going. I think the business continues to go forward at an extraordinary rate.

18:21Today, we are announcing a new data center here in Oklahoma City. Right. And this is our fifth U.S. data center. This is a data center that is step one in a multi-step process here in Oklahoma. We've invested significant resources here. And it is part of our ongoing plan. You know, we manufacture in the U.S. and we have five data centers in the U.S. And so we are very much committed to growth domestically. Right. Well, so I want to talk about the data center. I want to talk about the business. I mean, I just want to get a sense here, though. I mean, on the IPO side of things, what is the status of – I mean, is the ambition still to go public?

19:10Where are the national security reviews at? I mean, just talk about that a little bit. So we were, we encountered a CFIUS review in the Biden administration. And that was resolved in March, as we have previously stated. And so there are no such impediments remaining. Okay. And so what about the$1 billion in funding? We reported that the company is looking to raise a billion dollars. Axios put out a story of getting closer to raising that money privately. What's the status of that funding round? Well, we haven't commented on either of those. We appreciate your ongoing interest in our fundraising.

19:49But I think it's inappropriate at this time for us to comment on that. But I think where you have extraordinarily rapidly growing companies with world-class technologies, those are the elements that make raising money less challenging. Okay. So let's go to the growth of the business. I do want to talk about the customers that you've signed on because there are a lot of customers since we may have last heard from the company that have come on to use the product. But, you know, just talk to me about, you know, the customer mix. I mean, you know, when you guys had said you were going public initially, more than 80 % of the revenue was coming from one customer.

20:35How has that been diversifying over the past six to 12 months? I would make several observations. First, we do have some customer concentration, but we are in a business of whales. NVIDIA has 20 % of their business with one customer. They have 50 % of their business with four customers. So in our industry, there is no way to avoid concentration. That's the first observation. The second observation is that we are really proud of our relationship and our strategic relationship with G42. They are a global powerhouse in AI. And we together, we are solving extraordinarily difficult problems. worldwide.

21:21And so we don't shy away from that at all. I think they were the first to see our value at scale dove in. But since then, since we last spoke, we've won major deals at Meta, at IBM, at the French AI lab, Mistral, with leaders like AlphaSense, with pioneers like Figma and Notion. These are extraordinary growth on customers that are across the range. We do a big business with Galaxia SmithKline with Mayo Clinic. So we have customers in the largest of enterprises with the U.S. government in both military and civilian applications. And we have large customers in the rapidly growing AI native category.

22:14So where revenue at one point, you know, where more than 80 percent of your revenue was coming from G42, what percent of your revenue now is coming from them? Well, I prefer not to say. I think they will continue to be a very, very large customer. Our targets are continuing to buy an enormous amount of equipment. And I think that's what you want. You want customers who buy hundreds of millions of dollars worth of equipment. And I think it is only in sort of the twisted world of finance where that's a bad thing, right? We want customers to buy hundreds of millions of dollars. We want customers to sign up for multi-year engagements with us.

23:07And it's only on Wall Street where this could somehow be twisted into a negative. last question for you talk to us a little bit about why your technology is better than or poses a significant threat in your opinion to a company like NVIDIA I think it poses a significant threat because it's 20 30 40 times faster and speed is the fundamental driver in inference and so I think there is no debate in the industry we are the fastest bar none and NVIDIA's published benchmarks, third-party published benchmarks, all of them point to the fact that the GPU, by architecture, cannot deliver the performance that we can deliver for individual users.

24:01And so that is the threat. The landscape right now is making performance more important. Real-time workloads, agentic workloads, reasoning workloads, all are the type of work in which if you ask your customers to wait for an answer, they leave you and they replace you with competitors. And in these applications, we are fastest bar none. Right, right. Well, I want to thank you for coming on the show. Last question for you before we let you go. I do want to bring it back to the IPO. So is there still an ambition to go public this year in 2026? What can we expect? Look, I think first, going public is a milestone.

24:50It's not a goal line. When you're trying to build a great company, you have to be clear on what the differences are. Right. An IPO is a fundraising event. And if you can raise the money privately, then why go public at all? Well, there are certainly, you should ask Sam that, right? open AI, you should ask the guys at Databricks that. I mean, I'm asking you that because it looks like we've got a billion dollars in cash coming in from private markets. I mean, you know, maybe you don't want to go out. There's a reasonable argument for that. But that's actually not my view. I think historically, there were real benefits to being a public company.

25:30There are real costs, too. Right. It is their cost on the company in terms of administrative overhead and burden. So it sounds like you don't really want to go public anymore. That's not what I'm saying at all. What I'm saying is that I don't really want to tell you about my plans, despite you asking me two or three times. That's what I'm saying. All right. Well, that's fair enough. Look, I'm just here to have the conversation. conversation. That's awesome. I appreciate that. And I think our plans are to build an extraordinary company and to finance it in the way appropriate for extraordinary growth.

26:13And we are pursuing that approach relentlessly. And I believe in the future that will include an IPO, but reserve the right to change my mind. Great. Well, Andrew, thank you so much for coming on the show. It's an exciting time for your business. And I do hope that you'll come back on when there is more news, maybe about a private fundraise or maybe about an IPO. Please do come back on the show. We're happy to come on and talk about any number of things. So we're happy to come back at a time appropriate. Okay. Well, Andrew, thank you for coming on the show. That is Andrew Feldman, the CEO of Cerebris.

26:46Well, that does it for today's show. A reminder that we are live on this stream Monday through Friday at 10 a.m. Pacific, 1 p.m. Eastern. I want to thank Amazon Web Services, who is our presenting sponsor for this production. And I want to thank you for tuning in. We really do appreciate your viewership. I'm already excited for our next show tomorrow. And so until then, bye-bye for now.