In short
Squawk Pod Episode Notes: An NHL Deal with Kalshi CEO & AI with Airbnb CEO (10/22/25)
Podcast Overview Squawk Pod is a daily podcast from CNBC that curates top moments and discussions from the morning show “Squawk Box.” Anchored by Joe Kernen, Becky Quick, and Andrew Ross Sorkin, the podcast includes interviews, analysis, and lively debates.
Episode Summary In this episode, the hosts discuss a groundbreaking partnership between the NHL and prediction markets, represented by Kalshi and Polymarket. Tarek Mansour, CEO of Kalshi, elaborates on the implications of this deal for both prediction markets and traditional sportsbooks. Additionally, Brian Chesky, CEO of Airbnb, introduces the company's new features aimed at enhancing travel experiences and discusses the role of AI in the future of customer service.
Key Topics Discussed
- NHL Partnerships with Prediction Markets
- First major sports league to engage with prediction markets.
- Tarek Mansour highlights the significance of this partnership.
- Discussion on how this partnership may disrupt traditional sportsbooks like FanDuel and DraftKings.
- Regulatory challenges and legal considerations surrounding prediction markets.
- Kalshi's Business Model and Market Growth
- Kalshi’s rapid growth and expansion to $1 billion in monthly transaction volume.
- Seasonal trends in prediction markets: politics, economics, and now sports.
- Future collaborations with other major sports leagues anticipated.
- Airbnb's Innovative Product Launches
- Brian Chesky discusses new features aimed at enhancing social interactions through experiences.
- Introduction of AI-powered customer service aimed at improving response times and user satisfaction.
- Expansion plans focusing on specific cities like Paris and LA for service offerings.
- AI and Its Implications
- Discussion on the integration of AI into various sectors, especially customer service.
- Concerns regarding the accuracy of AI systems and the role of human oversight in technology.
- Market Trends and Economic Outlook
- Analysis of current market conditions, including drops in gold and silver prices.
- Updates on Netflix's earnings and challenges faced by Mattel.
- Overview of Meta's partnership with Blue Owl for a massive data center project and its implications for energy consumption and local economies.
Detailed Notes
Tarek Mansour on NHL and Prediction Markets
- Significance: Marks a pivotal shift in how sports leagues engage with betting and prediction markets.
- Competition: Challenges traditional sportsbooks by offering a regulated alternative.
- Regulatory Confidence: Mansour expresses confidence in their established regulatory framework and market integrity.
Brian Chesky on Airbnb Innovations
- Service Expansion: Introduction of social features for Airbnb experiences to foster community interactions.
- AI Integration: Enhanced customer service through AI, addressing complex user issues effectively.
- Future Vision: Chesky envisions a comprehensive service platform that goes beyond accommodation.
Market Analysis
- Gold and Silver Prices: Notable declines in precious metal prices, with gold experiencing a significant drop.
- Netflix Performance: The company reported lower than expected growth, attributed to external regulatory issues and competition.
- Meta's Data Center: A partnership with Blue Owl Capital to fund a $27 billion data center raises concerns about energy consumption.
AI and Technology Concerns
- AI Reliability: Recent studies indicate leading AI systems often misrepresent facts, highlighting the need for improved accuracy.
- Impact on Job Market: Ongoing discussions on AI's effect on employment and the need for transparency in AI applications.
Key Takeaways
- The NHL's partnership with Kalshi signifies a transformative moment for prediction markets in sports.
- Airbnb is steering towards a future of enhanced user experiences, leveraging AI for customer service.
- The ongoing volatility in market conditions reflects broader economic uncertainties affecting various sectors.
Episode Hosts and Guests
- Hosts: Joe Kernen, Becky Quick, Andrew Ross Sorkin.
- Guests: Tarek Mansour (CEO of Kalshi), Brian Chesky (CEO of Airbnb).
Conclusion This episode of Squawk Pod presents a blend of innovative business strategies in the sports and travel sectors, highlights regulatory challenges, and offers insights into the future of AI. With significant implications for both markets and consumer experiences, the discussions provide a valuable context for understanding current trends and developments.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Bring in show music please. Hi, I'm CNBC producer Katie Kramer, today on Squawk Pod. Sports books, look out. The first major sports league has inked a deal with prediction markets, Calci and Polymarket. Calci CEO Tarek Mansour on his partnership with the NHL and whether more deals are coming soon. What we have done here, and this is what's most important, we have set up a framework for us to collaborate with a big four league like the NHL. Then Airbnb CEO Brian Chesky on new products the travel site has launched to remake your vacation. We imagine what if you can Airbnb more than Airbnb? We launch services, we launch experiences, and since then we've been hard at work.
0:44And how he sees artificial intelligence taking over every corner of the internet. Right now it's like a gold rush. Plus Netflix's membership growth, sales of Barbie dolls are down, Meta and Blue Owl developing one giant data center. And AI chatbots are taking over. I'm not as nervous about that piece of it as you are. Because humans are going to be involved in this. Oh, good. But they maybe aren't that accurate all the time. No kidding. It's Wednesday, October 22nd. Squawk Pod begins right now. Stand Becky by in three, two, one. Cue it, please. Oh, yeah. Sorry. Good morning, everybody. Welcome to Squawk Box right here on CNBC.
1:27We're live from the NASDAQ market site in Times Square. I'm Becky Quick along with Joe Kernan and Andrew Ross Sorkin. And, yeah, just speaking of some of those things in AI, I mean, sometimes you get wrong facts from it. Facts. Facts that are wrong. It's one thing for opinions. It's being alternate facts, as Kellyanne Conway wants it. Okay, if you ask two of them what the 1929 adding average was for the Philadelphia Phillies. What year? 1929. You are so nice. Thanks. You believe that? She's running an ad for me. She's running an ad for me. Thank you. No, but with all of them, I've said, are you sure?
2:03You need to check. Oh, yeah. You're right. Sorry. They will cop to it and not even apologize. Oh, mine apologizes to me all the time. I'm sorry. I got that wrong. Well, that's because you don't yell at it. Joe calls it an idiot. Yeah. No, literally. No, it says I violate all the you're not supposed to because they have feelings, too, apparently. Yeah. The big news yesterday was what was happening with the commodities, the precious metals in particular. Gold down nearly 6 percent yesterday. I think it was 5 and three quarters percent. That was its worst day since June of 2013. So a dozen years since it's seen a drop like that.
2:38It's indicated down a little bit weaker once again this morning, down to 4 ,079. But if you look at the drop yesterday, it really just put it back to where it was earlier this month because gold has had such a tear to the upside. The weird thing was there was no real news that premeditated any of this. It was just kind of moving on sentiment or something else. A lot of people viewing this drop as a long overdue pullback for gold. It has run up 20 percent just in the last two months, and it's still up by about 55 percent this year. Silver had a big drop, too. It was down by even more. I think silver fell 7.1 percent.
3:16It settled at 47.04. That was its worst daily performance since just April 4th of this year. And then if you're looking at crypto prices as well, it looks like Bitcoin right now is off by about three and a half percent to one hundred and eight thousand. Now it's correlated. Right. Well, what is it? What's it correlated to gold? Now it's correlated. Right. Right. Crude oil prices at this point, if you take a look at things, you'll see they're up by about one point three percent, sitting right at fifty eight dollars a barrel for WTI. President Trump saying that India's prime minister has assured him that the country will ease purchases of Russian crude oil.
3:53Probably good news just in terms of what we're trying to do on a defense perspective, on defense of Ukraine, but not necessarily great news for the world oil market because obviously oil is fungible. Also, the Energy Department announcing plans to buy a million barrels of oil to start refilling the nation's emergency supply of crude. Good news at these low levels. Right. Let's talk about Netflix. Joe mentioned that at the top. Netflix missing third quarter profit estimates citing an ongoing dispute with Brazilian tax authorities. Now, revenue matched expectations rising 17 percent of the quarter in line with estimates that was driven by membership growth, pricing adjustments in what the company said was its best ever quarter for ad sales.
4:37Netflix's full year revenue growth estimate of 16 percent was in line with previous expectations. Now, Netflix reduced its full-year forecast for operating margin to 29%. That had been at 30 % prior to that. Now, the company's still riding, of course, the success of June's release of K-pop Demon Hunters. Its fourth quarter content slate expected to include the fifth and final season of Stranger Things, a new season of The Diplomat, and Frankenstein is going to be there. Guillermo del Toro's Frankenstein, which I saw. You've seen it. I did. Do you like it? It was long. I like the guy who played Frankenstein.
5:14I don't know why they did not cite my favorite one, though. Black Rabbit. Oh, God. I thought it was so awesome. You watched the final? I have not watched the final episode. Don't tell me what happens. I don't know. I saw it all. No, it's really not. You shouldn't have given him that. No, Jude Law is good. You're baiting him. You've got to suspend all disbelief. No one could be that stupid as Jason Bateman and do so many stupid things over and over and over again. He's an actor, the character. Oh, yeah. Oh, yeah. Okay. All right. I'll give you that. Because he's stupid in real life, too. Although his sister is really smart.
5:52Texas Instruments, yeah? The view through the political prism. Exactly. Is that where we're going here? Texas Instruments shares down sharply. The company forecasting fourth quarter profit and revenue below Wall Street estimates. That's a lack of clarity over semiconductor tariffs plaguing the sector. TI's CEO says when he speaks with industrial customers about investing and building, there's a bit of what he called a wait and C mode. And that stock cut down about 8 percent. Mattel missing third quarter profit and revenue expectations as tariffs continue to dent the company's North American sales.
6:29About half of Mattel's toy sales come from the United States. The company says that North American revenue fell by 12 percent in the third quarter. Global Barbie sales were down 17 percent year over year. Fisher price sales were down 19 percent. Mattel issued full year guidance after pulling targets back in May. Those new projections call for yearly net sales to increase between one and three percent. Earlier in the day, Mattel, Hasbro and Netflix announced that they were partnering to bring toys from the smash hit K-pop Demon Hunters to store shelves. Didn't help Mattel shares that news that they gave us about the guidance down by 5.6%.
7:05And remember, guys, when people were worried about the tariffs and what they would mean, toys were one of the often cited areas that they talked about. And remember, the president at that point said, well, maybe people have to go with one less Barbie doll at Christmas time. And you said Netflix, of course, on the K-pop phenomenon. I had no idea. The K-pop demon thing? That was huge. Yeah, it was big. I know. I didn't watch it, but I know of it. Yeah. Nobody in my house was watching it, but I do know what it is. But it was everywhere. How strange can things get, too? Five seasons of that? I mean, I saw the first one, I think.
7:41The first season of K-pop Demon Hunters? No, Stranger Things. Oh, Stranger Things. That's how strange. He's saying five seasons later how strange it is. I mean, that's pretty strange. I saw the first two seasons. You did? Yeah. Things were very strange in the first season. Yeah. They got stranger in the second season? I didn't see the continuation. Let me tell you something that's even stranger. Can I tell you something that's even stranger? Better? Yeah. That's your segue? I'm trying here. I'm trying. This is strange and interesting and all sorts of things. Meta forming a$27 billion joint venture agreement with Blue Owl Capital to fund and develop the social media company's Hyperion Data Center in rural Louisiana.
8:16Now, Blue Owl is going to own 80 % of that venture. Meta will own the rest and oversee construction. It's slated to be completed in 2030. It's being built on a site. It's roughly 1 ,700 football fields big. It's kind of remarkable. Local utility energy company telling CBC earlier this year that the data center could consume about twice as much electricity as the city of New Orleans on a peak day. Now, what's interesting to me about this is so Blue Owl is effectively in the private credit space. We always talk about sort of where's private credit? Is this leveraged? You know, is the world of AI leveraged?
8:50Well, yes, it is. That's what it is. but it's interesting it's getting now moved off the balance sheet to some degree of these tech companies as they try to form these partnerships to find people who will help them pay for these things i mean i'm trying to we just can't get our head around 1700 football fields or the electricity piece i mean it's the electricity but by the way i don't know if you know how big that is are you just going to gloss over that what is 1700 football fields what's 100 football fields look like it's is it bigger than manhattan bigger than breadbath i wonder might be bigger than Manhattan.
9:20I think when they announced this, it was an unbelievable thing. But I was going to say, there's a number of these data centers that Google and Microsoft and others have tried to open, or at least talked about opening places, and they've effectively shut them down or stopped doing them because the communities don't want them. And their communities don't want them because they think their electricity prices are going to go up, which they are. When I asked them 1 ,700 football fields, how big is that? And And they said, it's approximately the size of a complex, like the one being built by Meta in Louisiana.
9:54Thank you for that comparison. That's where they got the seven. They looked up. But think about that. People are worried about AI taking their job. And now they're going to start worrying about them taking their electricity. So, you know, there's going to be a whole sort of political and policy implication of just how quickly this gets developed. As prices have gone up, and there was a Bloomberg report just in the last month or so that said that for consumers that live near these data centers, their electricity prices have gone up by 2.7 times versus over the last five years since these things have been built.
10:32So the increases that they've already seen, and you talk to the electricity guys, they'll tell you it is a large chunk of it. You talk about this on Squawk on the Street all the time. They've read this map that they've been using for a while over it. So that's how big, like Central Park, think about the immensity of Central Park, and then look how this spreads beyond that. But all of these governors who are begging these companies to come to their states, I think in five years from now, they're going to be wishing this way they never did it. Well, the governors may not be there. Yeah, the governors may not be in charge anymore.
10:58You need maybe some longer-term thinking, but it is interesting what it could mean for the political calculus at that point and the pushback if consumers feel like they are getting the short end of the stick on this. how big is the the biggest when you're dealing with major biggest solar field how many acres i don't know we'll we'll ask chat gpt that too and then we're right we'll be needed for yeah so we can just ask of things yeah and wind fields i mean i see you know wind farms you see those those are take up a lot of space as well i mean i'm still impressed by amazon like data centers when you're driving along.
11:33The largest solar power plant in the world is in China, which covers 235 square miles. That's big. Yeah. That's bigger than 1 ,700 football fields. Much. Some cautionary news about the reliability of AI. New research shows that leading assistants misrepresent news content in about half their responses. Not to worry, though. That's according to the European Broadcasting Union and the BBC. Researchers studied AI assistance, including ChatGPT, Microsoft Copilot, and Google's Gemini, for accuracy, sourcing, and the ability to distinguish opinion versus fact. And 45 % of the responses had at least one significant issue.
12:20Sourcing was a major problem for Gemini. My overall issues of accuracy were found in 20 percent of responses from all assistants. So, I don't know. In the news business, it's even worse if you rely on something. That would be a problem. That would be a problem. And I've said anecdotally I knew this was true because I've gotten answers that I know are not right. And then I say, are you sure? And they go, oh, you know, I wasn't sure. I've actually looked a little more closely, and it's the opposite of what I told you. It's a little human. It's a little human. Maybe for you. And after all that, the answers are wrong half the time.
13:02Hey, I got this new drug I designed for you. This is going to be awesome for you. Oh, sorry. That was for some other disease. This is going to kill you. I mean, we've got to get this together quickly. Can we? It's not Six Sigma. 30 % is not Six Sigma. I'm not as nervous about that piece of it as you are. Really? Okay. Why? Because the language models for the scientific ones are... Because humans are going to be involved in this. Oh, good. But, look, I do think that the large language models have... You're going to have small language models that are more destined to do jobs just like that for the science applications on some of those things, too.
13:43That's not going to take a broad vacuum of every bit of information that it can. Grok the other day. That's why I argue with people on Twitter, as you know. So, you know, someone's saying, oh, I hope you like your tax cuts, fat boy. You know, and it's like my tax cuts, my taxes soared under Trump. And, you know, I say and I said to Grock, figure this out. Here's here's X. What happened to those? And I was amazed that it took them like 10 minutes. But the assumptions and the research and everything that they did, they put I don't know why they put property taxes in there. But all they need, it was much easier.
14:21Because property taxes are no longer, if you have a self-adduction. Right, right. AMT, they had to do all that. But just to prove to some schmo that that's obviously not why, or else I would be a Kamala. That's obviously not why you're second up. I'd be a huge Kamala bro if it was based on my own situation. Is Grok better, Sorkin, than some of the others, or could they have all done that? Would I have gotten a similar answer? I like Grok for news. The Crockford News. Yeah, because the Twitter feed is part of it. So you use, how many are you using? I've got, no, I've got three. I've got too many.
14:58Anthropic. You use anthropic, anthropic too. Yeah, Claude. Yeah, I understand what I mean. Okay, I only got chat, GPT, perplexity, and grog. Right. I know we're going off-piece and off-topic. You want to hear just something very funny? Yeah, I do. I was talking to a lawyer the other day. That's the show off-topic. No, I was talking to a lawyer the other day who was telling me that they'd marked up a contract or something, and then they got an email back from somebody else who said, well, here are the changes from Claude. And he writes back. Who's Claude? Who's Claude? Did we hire? Is there a new associate named Claude?
15:33Yeah, that would be. Yeah. And, of course, it was. Yeah. I thought that was kind of funny. And it wasn't a joke. So you need a lawyer. What for, Andrew? What happened? No, I don't need a lawyer. You said you were talking to a lawyer the other day. Yeah, no, I was talking to a lawyer. Oh, no, no. I actually went and spoke about the book to a whole group of lawyers. You're not like motor vehicle or something? No, no, no. Lawsuit? No, no, no. Plagiarism? Not yet. Not yet. Thanks. Yeah, exactly. When you need a lawyer. Yeah. When you need one. Yeah. It's good to have a good one. That's all I have.
16:08Better call Saul. Better call Saul. Exactly. Please. Cheese will be next. Coming up next on Squawk Pod, the NHL goes where no major sports league has gone before, a deal with prediction markets, Calci and Polymarket. What happens to sports books like FanDuel and DraftKings now that the majority of volume on these prediction sites is sports? Calci CEO and co-founder Tarek Mansour discusses it all. I think there is seasonality to prediction markets. So last year, at the end of the year, our most popular category was elections. In mid this year, we had a lot of volume around when there's a lot of market volatility, financials and economics.
16:44And now we're seeing a lot of volume around sports. We'll be right back.
16:53This is Squawk Pod from CNBC. Here's Becky Quick. The NHL is striking an official partnership deal with Kalshi and Polymarket. This is the first major sports league to partner with the prediction market. And joining us right now to talk about it is Tarek Mansour, who is the CEO and co-founder of Calci. And Tarek, thank you for joining us this morning. This is a big deal. It's the first of the major sports leagues to actually sign a deal, sign a partnership like this. And I guess I wonder, how did this come about? Well, I totally agree. You know, this is a big deal. A big four league like the NHL signing a partnership with Calci is a testament, I think, to our market leadership.
17:35our astronomical growth this year. And I would say the golden standard of market integrity and customer protection that we have established over the last few years. So, you know, it is, I think it's a seminal moment for prediction markets and Cal sheet because, you know, a league like the NHL partnering with us is a strong sign that prediction markets are here to stay. There are already people looking at this. The Wall Street Journal wrote this up and suggested that this is a real assault on the sports books like FanDuel and DraftKings. Do you see it the same way? I don't really think about that in that way.
18:12I mean, look, I think there is a business model disruption here. And we've seen it, you know, with, for example, Airbnb and hotels. We've seen it with Uber and taxis. And so when there is fundamental technological disruption, a technology that transforms an industry, I mean, there are sometimes critics or dissidents that come about, and that's part of the course. What we're really focused on is, you know, building really good product, building it fast, and improve the market diversity for our customers, which, you know, we're leading on all three, and we're going to work hard to keep doing that.
18:48There have been challenges because while the sportsbooks are not allowed throughout the nation, you all are being offered right now everywhere in America. There are sports contracts in all 50 states, even those that restrict the betting on the standard sports books. And there have been some lawsuits that have come from this, an outcry from the gambling industry groups, legal challenges coming from state regulators, and some senators who are taking up against this as well. How assured are you that this can continue in all 50 of those states? We're very confident in our regulated structure and our regulated setup.
19:29You know, we've worked for years with the CFC to regulate prediction markets and, you know, develop a framework for customer protection. This idea that on a federal basis, it has been legalized by the CFTC, but there isn't, that's not to say that there aren't politics there. If there was a change at the CFTC, that could make it more complicated. How confident are you that this can remain something that you can legally do throughout the country? Yeah, you know, we are confident and comfortable with our regulatory setup. So we spent years working with the federal government to regulate prediction markets.
20:04And, you know, when we think about the announcement today, you know, the NHL deal is really about that. It's essentially a validation of the fact that we have established the right set of customer protection and the right set of market integrity measures to protect our markets, but also the game, you know, to enable this product to thrive in all 50 states. And that's why we're seeing the sort of type of customer growth that we're seeing on Kalshi. And, you know, today we're at a billion dollars of monthly transaction volume. And so we're excited to see that keep growing. Yeah, it's been growing rapidly.
20:45In fact, Calci and Polymarket saw notional trading volume rise above$2 billion for the first time last week, in the week that ended last week. Most of your most popular trading categories at this point are sports, correct? Back in 2024, it was mostly politics. Now, in 2025, the last list I saw has something like 1, 2, 3, 4, 5. I think 8 of the top 10 are sports categories, led by the NFL, and then secondarily, NCAA football. Look, I mean, I think there is seasonality to prediction markets. So last year, at the end of the year, our most popular category was elections. You know, mid this year, we had a lot of volume around when there's a lot of market volatility, financials, and economics.
21:24And now we're seeing a lot of volume around sports. It really depends on what's top of mind and what's trending in people's mind and what people have opinions about or what people want to hedge. Actually, this last week, we've seen quite a bit of volume on sports. And you may have seen the announcement from Underdog, you know, who is going to be hedging some of their positions. So leading sports, sportsbook hedging, laying off risk on CalShit, which is, you know, very much the purpose of these markets. But we're also seeing a rise in volume and things around policy. For example, the government shutdown, where we see startups such as Arrived Homes also hedging some of their risk on the government shutdown.
22:03And what is the latest on the government shutdown? What does the prediction market tell you in terms of when this might end? So we're forecasting 41.5 days, so 42 days for the government shutdown, which would make it the longest in history. Wow. Tarek, I have to ask you, if the NHL is doing this and this deal has been out there, I assume you're in talks with some of the other major sports leagues as well. Is that a correct assumption? It is a correct assumption. I think, you know, what we have done here, and this is what's most important, we have set up a framework for us to collaborate with a big four league like the NHL.
22:42And really, they're kind of taking a market leadership position here in how we develop data sharing to combat fraud, to ensure customer protection, and to preserve market integrity. And I think that same framework that CalShit has developed with the NHL can be replicated across the other leagues. So be on the lookout for more announcements soon. And you're paying a licensing fee. Will you tell us what that licensing fee is? We're not sharing specifics of the deal, but really what we're getting out of this is a little bit of a makeover on the product. So if customers log on to Calci today, what they'll see is, you know, we're using the marks.
23:16So the product looks a little bit more beautiful than it looked yesterday. And we're very excited about that. And what the NHL is getting is our whole suite of customer protection and market integrity measures that will help them preserve the important integrity of their game. Tarek, I want to thank you for joining us this morning. We appreciate your time. Thanks so much for having me. Next on SquawkPod, Airbnb changed the way we travel, how its new products change our experiences when we go around the world. Well, some of us anyway. Is there a feature where I can guarantee I won't meet anyone when I'm on?
23:52Services you can have all to yourself. Squawk Pod is back right after this.
Read the full transcript
24:01You're listening to Squawk Pod with Joe, Becky, and Andrew. Up and Andrew, Q. Welcome back to Squawk Box. Airbnb rolling out a series of product updates, including some new social features and expansions to its AI-powered customer service tools. We've got a lot to talk about. First, right here on CNBC is Brian Chesky, Airbnb co-founder and CEO. It's great to have you. Thank you very much. Let's talk first about some of the new products that you're rolling out, what this all means. We talked when you first announced the sort of revamp of the app. Yeah, so last May, we completely re-managed Airbnb.
24:36We imagined, what if you can Airbnb more than Airbnb? We launched services, we launched experiences, and since then, we've been hard at work. We had 65 upgrades, and we launched three things, or three big ones. Number one, people are telling us when they go on Airbnb experiences, they want to meet other people. And so we thought, what if we built features for people to see who's going? You can keep in touch after and you can message them. So now experiences are much more social. We made it much easier to find Airbnbs. We've completely revamped our maps. And then finally, with AI, we have AI customer service.
25:07Now you can have your problem answered in like six seconds. This is really, really hard to solve, AI for customer service. So the team worked really hard on this. How much are people using some of these new services in terms of, I remember, you know, trying to find somebody who will either cut your hair or do something for you, get your makeup, or all of those kind of things. One of the things we learned is you still got to go city by city. Airbnb started in here in New York City. I know Uber, when they launched, they launched city by city. So we're really focused on a couple cities, Paris, LA, and it's really, really starting to take off.
25:35So I think in the next year, you're going to see a lot of traction with these products. There's no Amazon for services. People say Amazon's the everything store. It's actually the everything in a cardboard box store. And so we think there's a service platform that can be really big. And experiences, we think that could be a multi-billion dollar business. Is there a feature where I can guarantee I won't meet anyone? Services you can have all to yourself. Every service is totally booked yourself. So you can prevent me from meeting. Exactly. I remember when it first happened, I think you and I had a conversation about whether people would use it to find somebody great and then go off and try to create sort of a third-party one-on-one relationship with them.
26:10Yeah, that can happen in marketplaces because we're travel and the demand and the supply come different cities and they don't usually go to the same place again this not as frequent but yes any recurring service in this a local market would happen tell me about the ai piece of this right now because yeah so we that's the conversation everybody's having yeah i mean obviously we're pretty pretty knee-deep in ai and we're not deep in the at this research level we're deep in application layer and so basically customer service is one of the best uses of ai now a lot of people like with AI, they start with agents.
26:43One of the most common ideas is a travel agent. We thought instead of helping you plan, let's do the hardest problem. I'm locked out. I have a problem. I need to alter my reservation. I feel unsafe, like the hardest problems. So we've used 13 different models, trained on tens of thousands of conversations. Our AI customer service agent has handled over a hundred thousand conversations. It's reduced the need to contact an agent by 15%. So now we're rolling it out globally. Our AI agent is more personalized. We've developed a custom interface where it's a user interface where it can take actions for you.
27:13So it's really useful. I will say the first thing I ask whenever I get on one of these chat situations is, am I talking to a real person? Yes. Do you answer honestly? Yeah, yeah. We are very clear about when you're talking to an AI agent. If I want customer service, I basically want an agent. And you can go right to a person very quickly. And the AI is really good at escalating you to the right person very quickly as well. Okay, let's go. Okay, so here's one for you. You have not integrated with your pal Sam Altman's ChatGPT. No, not yet. So why not? Funny enough, I was talking to him about this idea, and we talked about integrating different applications.
27:47I said, if you want to do this, you need the SDK, Software Developer Kit. And the iPhone, they developed a great software developer kit. I said, we don't all want to just be data layers. We want to be able to have robust services. So I'm very close to Sam. I talked to him. I was aware of the development of this product. And we'd love to be a part of it. The SDK wasn't quite robust enough for the things we want to do. When you join Airbnb, it's not a commodity. It's a community. and they've got to develop the SDK a lot more. What does that mean? It has to give you more control over it? Yeah, like when you go to the iPhone, they have this really robust software developer kit and you can basically create whatever user interface you want.
28:21It's really robust. Right now, it's just like these little modules and I think it would be a little bit, it just needs to be a little bit more robust. Is that because you want to maintain control of your community and not basically share that with anybody else? You'll use their stuff. Well, that's a separate topic. That's a separate topic and that's a business decision of just kind of like, do we want to put inventory on Google? But then does Google disremedia us? And we chose not to put our inventory on Google search to make it indexable, you know, like through like the hotel finder. We didn't put it there, but we do we do pay for SEM.
28:51So that's a business question. I think we would probably want to be a part of ChatGPT and other platforms. We just want to make sure it's almost like you go to a retail store and there's a brand in the retail store. You want to make sure you're not just a product in an aisle. You have your own little. Well, that's what I was going to ask. So, you know, ChatGPT just came out with a new browser. I don't know if you've been playing with it. I've been playing with it in just the last 24 hours. It's fascinating. But it could have, ultimately, I would think, it could just intermediate yourself because it's seeing all the things you're browsing on all the time.
29:20So I'll tell you what I'm saying. At some point, you're going to hit the chat, and you may never even go to the website. Yeah, that's a couple things to say. Number one, one company can't just run the global economy. That's just not going to happen. And so the idea that, like, one chatbot, one company is going to do everything. Well, let me ask you a question. how come Apple didn't create every app in the App Store? Because one company just can't create Airbnb and Uber and Spotify. There's a lot to build, a lot to do. And if one company did create everything, you're introducing a huge amount of bureaucracy.
29:48So we're all going to have to work together. I think AI is going to lift up a lot of companies. And so they can be distribution. But if they want to vertically integrate every single thing, that's going to be very, very difficult. So I don't think it's going to happen. I think it would draw the attention of regulators. Yeah, beyond the fact that probably regulators don't want one company to own everything. But again, I think we're at the very beginning. We're also in this period where if you go to the top 50 apps in the app store, number one is Sora, number two is Chagipiti, number three is Gemini.
30:16Those three apps are AI apps. Imagine AI is like electricity. Those companies got this magical new technology. They're electrified. Apps four through 50 don't yet have it. Over the next three years, you're going to see everyone having it. And everyone's going to have access to basically the same models because they're all available via API. So suddenly the playing field will start to get a little more level. And over the next three years, you're going to see a lot of consumer apps. Right now, it's like a gold rush. Intelligence is the gold. And everyone's just making picks and shovels. I'm on the board of Y Combinator.
30:45Almost every startup is an enterprise company. There's not a lot of consumer companies yet. AI doesn't change the world until it changes the daily life of regular people. And that doesn't happen until most consumer apps are AI apps. Okay, so what about the economics of all this? So one of the news stories we've been talking about this morning is Meta is building this$27 billion data center down in Louisiana. They have effectively partnered now with a private credit firm called Blue Owl to try to offset some of the risk for them. They take sort of these short-term four-year leases. They guarantee some of it, I think, over 16 years.
31:19But this is the future of how these things are going to finance. But there's also a lot of leverage in that. Do you think the economics of this makes sense? Is there an ROI that you think you could, that people in the Valley can point to, or is this a religion? I would say this. Will this technology generate trillions of dollars returns? A hundred percent it will. There's no question. I'm looking at how daily life is going to completely change. The more money everyone raises, the sooner they're pulling that timeline up. And it's not going to pay back in just a few years. And the challenge with some of this infrastructure investment is the GPUs depreciate.
31:56So the DPUs depreciate, you need to have a payback within that depreciation window. And I think that we're pulling the windows up faster the more money we invest. So I think we need to be careful about excitement turning to euphoria, turning into mania. So I think if it's excitement, we should be feeling good. Where are we in that at the moment? We're leaving excitement. We might have left excitement. While we have you here, can you tell us about what's going on just in the travel business? The YOLO economy. Yeah, I think, you know, it's funny. For all the tariffs, for all the instability in the world, people still travel.
32:32And, you know, a few basis points here and there, currency exchanges change it. I've been shocked the last six years with the pandemic, with, like, wars, with massive change. Like, travel's been pretty stable. It's one of the things people don't want to give up. Now, travel changes. When the economy changes, people maybe don't do that international trip. But they get in a car and they just still travel. But are you seeing either international trips change? I remember, I think you were one of the first people who were saying there were people who had changed, like, where they were going. But not just where they were going, the days they were going.
33:06People were leaving places on Thursdays, coming back on Monday nights because they didn't have to go to the office. Yeah, a lot of people, like, over Thanksgiving, people used to go away on, like, a Wednesday, come back on a Monday, and now they'll go away on that previous Monday. Like, they'll make longer trips. People are doing more near-term destinations. You know, Europe is doing really well over the summer. But trips are longer. People are traveling more with other people. And they're traveling more nearby as well. Brian Chesky, thank you. Thank you very much. Really interesting. Always. Always, always.
33:41Thanks for coming in. Thank you. That is Squawk Pod for today. Thanks, as always, for listening. I've been off for a few days. It is so great to be back here with you on the pod. Squawk Box is hosted by Joe Kernan, Becky Quick, and Andrew Ross Sorkin. Tune in weekday mornings on CNBC at 6 Eastern. Get the best of our TV show, the latest headlines, and newsmaking interviews right into your ears when you follow Squawk Pod wherever you listen to podcasts. We'll meet you right back here tomorrow. We are clear. Thanks, guys.
34:19Thank you.
From the publisher
The NHL has inked an official partnership with Kalshi and Polymarket, marking the first time a major sports league has teamed up with prediction markets. Kalshi co-founder and CEO Tarek Mansour discusses his platform’s sports volumes, potential regulatory issues, and competition with sportsbooks. Airbnb has launched new products and features to remake user vacations, and CEO Brian Chesky is just getting started. Chesky discusses the future of his company and the impact of AI. Plus, Netflix streamer numbers missed expectations, Mattel’s Barbie sales are down, and Meta has partnered with Blue Owl on a data center.
Tarek Mansour - 19:20
Brian Chesky - 27:39
In this episode:
Tarek Mansour, @mansourtarek_
Joe Kernen, @JoeSquawk
Becky Quick, @BeckyQuick
Andrew Ross Sorkin, @andrewrsorkin
Katie Kramer, @Kramer_Katie
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

