China’s Moonshot Unveils AI Model, Fueling Tech Rout

0:00So there's a lot of noise about AI, but time's too tight for more promises. So let's talk about results. At IBM, we work with our employees to integrate technology right into the systems they need. Now, a global workforce of 300 ,000 can use AI to fill their HR questions, resolving 94 % of common questions. Not noise. Proof of how we can help companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business. IBM. Amazon Health AI presents Painful Thoughts.

0:34Carol Massar:Why did I search the internet for answers to my cold sore problem? Now I'm stuck down a rabbit hole filled with images of alarmingly graphic sores in various stages of ooze. I can clear my search history, but I can never unsee that. Don't go down the rabbit hole. Amazon Health AI gets you the right care fast. Health care just got less painful. The official language of football is trash talk. Late night group chats, memes, and unbelievable highlight clips. That's why Boost Mobile brings you our new global connection plan. The first plan ever made for WhatsApp. Get unlimited data, talk and text, international roaming, and calls to over 100 countries for just$40 a month.

1:21The$40 price includes$5 a month auto pay discount. After 40 gigabytes of premium high-speed data, speeds will be lowered. Coverage not available everywhere. Visit store or boostmobile.com for details.

1:33Bloomberg Audio Studios. Podcasts. Radio. News. This is Bloomberg Businessweek Daily. Reporting from the magazine that helps global leaders stay ahead. with insight on the people, companies, and trends shaping today's complex economy. Plus, global business, finance, and tech news as it happens. The Bloomberg Business Week Daily Podcast with Carol Masser and Tim Stenevek on Bloomberg Radio.

2:03Carol Massar:All right, so let's get to it. There's a lot of stuff going on. As I said, I'm obsessed with some of the news out of China. A surprise breakthrough from Chinese AI startup Moonshot rippling through global markets. We're seeing that play out. It started overnight in Asia. semiconductor stocks. Charlie just talked about it sharply lower. We're kind of thinking, is this another deep seek moment? You had President Xi Jinping also overnight, his first appearance at China's premier AI summit, lending his support to the country's push to become a global leader in AI. And you've also got alphabet tanking.

2:36Carol Massar:There's another story kind of at play here. Right. Shares of alphabet tanking as Google is months behind on schedule on delivering Gemini 3.5 Pro, its most powerful flagship AI model. It's trying to improve its capabilities, particularly in codings. And then the credit markets are also reacting to this AI trade. We got to talk about credit, Carol. The bond sold by hyperscalers to fuel their AI ambitions have become a drag on investor portfolios with falling prices and wider spreads to negative total returns. That is certainly your world. I feel like it's a lot, but it is always when it comes to what's going on with AI.

3:08Carol Massar:We wanted to check in once again with Mandeep Singh. He's Bloomberg Intelligence Global Head of Technology Research. Mandeep, I know we, I feel like we have the same conversation with you over and over again, but there's stuff that keeps coming at us. SK Hynix ADRs too are bouncing around, the Cosby route. Feels like nervousness continues to grow increasingly out of this AI trade. Is it fundamentally based? Is it valuation based? Is it just a little bit of fear of people being on the wrong side of the trade? You watch the fundamentals. You always remind us what matters. So how do you see it? Yeah, I would agree with your assessment that it is fundamentals that are still holding up.

3:49And it's the valuations and the nervousness around positioning is what is driving this sort of volatility. And look, we know AI being expensive has been raised as a concern by a lot of companies lately. So it doesn't surprise me that, you know, a model like Kimi 3, which is an open rate model, much cheaper to use, is gaining traction. And there are companies out there that have explored using this model, notably Cursor, which is a company that XAI and SpaceX bought. You know, they use Kimi 3 under the covers. And so from that perspective, open source will have some sort of adoption. But fundamentally, nothing has changed.

4:37Everyone wants to use LLMs. It's just that they were expensive. That's why we had concerns around memory pricing. And, you know, now we are talking about open source adoption. But I don't think fundamentally anything has changed. It's just a function of how folks were positioned when it comes to the valuations and certain sectors have gone parabolic within the tech space. And you're seeing a drawdown over there. Mandeep, tell us about this new Chinese AI model, because as Carol and I were discussing just a few minutes ago, it wasn't too long ago in markets that we had that deep seek moment, that scare when a new model from a different country comes out and spooks the market.

5:18Is this just a deep seek 2.0? Is there something different about moonshot? And when you look at the market right now, is it really reflecting the same type of fear that we saw when the deep seek model was released a little over a year ago? Yeah, I mean, look at how far we have come since that DeepSeek moment. We had a company like Anthropic grow from, you know, 9 billion revenue run rate to 47 billion run rate in six months. So that's despite DeepSeek. And so, look, open source gaining traction is something that has carried forward. I should stand corrected. It's open rate, not open source. But all these companies are looking for ways to reduce their, you know, inference pricing, which is very expensive for frontier models.

6:09And we have been compute constrained. Those are the things that reflect the fundamentals. Everything else around, you know, nervousness around deep seek and new types of models just shows that there is a lot of investment being made. Companies are exploring how to reduce their costs, but they want to consume more capacity and deploy as much of LLM functionality as they can. So from that perspective, the fundamentals are still good. And that's what's reflected in the revenue run rates of these companies and the demand that we have seen in terms of the backlog. And we'll hear more from hyperscalers when they report over the next two weeks in terms of how they're guiding around their CapEx numbers.

6:52Carol Massar:All right. And we're going to talk a little bit more, too, about the China angle with a member of two of our AI team in just a moment. But I want to talk about one of the hyperscalers. And as you said, we've got to watch and see what they say on earnings. But we've been watching shares of Alphabet. Stock has been down almost, well, 3.7 % of its lows today, down right now, still about 3%. Talk to us about this news we highlighted yesterday late, the delay of that Google Gemini launch. What do we need to know about this? What does it indicate for Alphabet more broadly for the AI story? Yeah, I mean, this is such a fast moving space that if you have a delay in the release of your model and your competitors have rolled out, you know, new models that are getting a lot of attention, like Kimi 3 did with, you know, some of the new functionality in terms of coding agents and established use cases.

7:46That is where Alphabet, because maybe they lost some of their key researchers on the LLM side, that may have had an impact. I don't know the reason for the delays, but for a frontier model company like Alphabet, that has a lot going for them in terms of the chip side and the cloud business. Delay of the model is certainly a setback. I do think they will release something soon. But for now, I think it's right to see that getting reflected because you can't afford to, you know, not be on the frontier if you're a hyperscaler right now.

8:22Carol Massar:All right. So, Mandeep, a good buddy, you're watching the World Cup this week and turns to you and says, Mandeep, what the heck's going on? Like, should I pull all my AI investments? What would you say? No, I think you just got to write out these volatile phases. And look, there may be some changes in terms of what's going on in the memory complex. If Apple is get to like they're looking to get CXMT qualified for use, that is a big deal. Now, those are the type of things that have a bearing on the market. But on the demand side, when it comes to LLMs and the need to deploy LLMs and inference tokens, that I think will hold good because everyone knows that, you know, there are established use cases around these LLMs.

9:09Carol Massar:All right. So I'm going to read that as Mandeep Singh saying, take a chill pill for a little bit, everybody. All right. Mandeep, thank you as always. Have a great weekend. Mandeep Singh, he's Bloomberg Intelligence Global Head of Technology Research. We lean on him so much. Stay with us. More from Bloomberg Businessweek Daily coming up after this.

9:51collagen sparkling water. Your daily glow up now in three fresh flavors. Strawberry blossom, lemon lime, and blood orange. Improved skin health in as little as 30 days thanks to collagen peptides. Cheers to that. Or go with our classic collagen peptides. So you can stay vital, stay you. Visit vitalproteins.com to learn more and where to buy. These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease. They're back. Live across North America. It's your fault.

10:25Guns N' Roses Tour 2026.

10:31Get tickets now at LiveNation.com.

10:41Don't miss Guns N' Roses. You're listening to the Bloomberg Business Week Daily Podcast. Catch us live weekday afternoons from 2 to 5 p.m. Eastern. Listen on Apple CarPlay and Android Auto with the Bloomberg Business app. Or watch us live on YouTube.

10:58Carol Massar:We're going to stay on Tech, Emily and I, and the AI bill, trade, and threats and concerns. I want to go a little bit more into what is going on when it comes to China and AI. I'm like perplexed, and I know Mandeep just addressed it. Let's get to a member of our team on the AI Watch. That is Seth Fiegerman. He is Bloomberg News AI editor. Seth, good to have you here with us. We just talked with Mandeep about the trade today, the news out of China, President Xi kind of being all in, no surprise, right, when it comes to building up China's presence in AI and moving along. The AI race and nervousness from you and the team's reporting, what's your take on all of this?

11:37Carol Massar:What are you hearing from those in the industry? Yeah, you know, I think a lot of the narrative for the last year and a half post-Deep seek was that we were seeing Chinese rivals put out credible models for lower costs that can undercut top US AI firms. But we weren't really seeing any of these firms compete what they call the frontier, you know, the absolute cutting edge of AI technology. And so I think part of what's jarring about this announcement, at least at first blush, is whether Kimmy and Moonshot are now at that level. And I think there's mixed reactions in the industry so far as we continue to report this out.

12:14And there's a joke sort of that, you know, it takes the worst time to review a new model as the day it comes out. You know, you need to kind of live with that and see how it kind of socializes and gets used day by day. But it's clear that the capabilities of this model from what they publicly disclosed match or exceed those of models from U.S. firms that were released as recently as a month or two ago. So suddenly the question is, is China six to 12 months behind, as some had thought, or is it really just one or two months behind? And the difference could be pretty monumental. And talk about just what makes the product so competitive, because it's not just competitive on price, right?

12:48They're going for more of a premium, kind of high-quality model and looking to attract buyers who maybe aren't as price-sensitive? Yeah, this is, as you said, price like a premium model, I think is the most expensive model that we've seen come out from China at this point, certainly among the open weight sides. It's comparable to some of the mid-tier offerings from Anthropic. With its startup models, this is a little bit different than what we'd seen with DeepSeek and others, where it's really fully a pricing game. What this is is more saying this model can compete on the more advanced tasks, the agentic tasks, doing coding, for example, which has become really the lucrative and key selling point for AI technology right now.

13:28And again, it comes at a moment when we're hearing more and more that there's cost consciousness emerging among developers and businesses of how much they're spending on AI models. And I think there's a growing awareness that maybe you don't have to bet entirely on one model. Maybe you kind of have a little bit of a buffet of models that you rely on and you can use different ones for different tasks. You may not need the absolute top of the line ones for every task. That way you can save costs. So if it prices at a premium way and starts to draw away customers who might have turned to top of the line offerings for those AI agentic tasks, that could be a meaningful hit to some of these U.S.

14:02firms.

14:02Carol Massar:Seth Figerman, you guys follow this stuff inside and out, but I keep thinking about the EV race and the EV industry and how China is now what we are talking about, kind of nonstop, high-end, low-end, more affordable. And so I do wonder, China's lead when it comes to EVs, is that kind of a lesson for the U.S. and kind of for everybody who's looking to stake their place when it comes to AI? You know, it's funny that you mentioned, because again, I think that the sense has been that China is following a playbook that it's used in other industries, that it's flooded the market with more affordable models.

14:37And at least until today, the thinking was it didn't have to have the best model. It could still siphon away meaningful market share and potentially destabilize or undercut US firms who are trying to make as much money as they can to offset their heavy investments here. But with this model, I think the question becomes not just whether they undercut you on price, but whether they actually are able to create comparable capabilities that compete at the cutting edge tier of AI development. And we're still not quite there, but we could be heading there sooner than we thought. What is that going to look like in practice?

15:07How are investors going to be able to judge when they look at the various LLMs that come out of China? Is China just catching up to the U.S., or are they actually surpassing the U.S. in the technology? I don't think anyone yet thinks that China has surpassed the U.S., but the narrative until today have been that we thought China was a bit further behind than they might have been. Now, again, this is early days. This model just came out. People need to play with it. And one of the challenges throughout AI is that there's not always agreed upon benchmarks and leaderboards to really say this one is definitively better or worse than that one.

15:42But the initial tests that they publicized certainly suggest that it ranks near the top of models out there, but still behind the most recent offerings from opening an anthropic that came out in the last couple of months. So, again, China has not surpassed the U.S., but it is certainly competing at a closer neck than we might have thought.

15:58Carol Massar:Listen, the team, though, you know, reminding us that Airbnb runs a mix of Chinese open source models along with Western vendors. Pinterest touted their utility. Even Microsoft, one of OpenAI's biggest backers, CEO Satya Nadella saying he's considering using DeepSeq models in its co-pilot, co-work product. So, yeah, I get it. And I feel like there is now the discussion of where are cheaper models. And you're right. Not everybody needs the top of the line because it's expensive. Having said that, are there security issues? Where's the U.S. government? Do companies need to be concerned that the government is going to maybe clamp down on this?

16:37Yeah, I mean, I think there's a couple of different ways to answer that question. One is that the Trump administration, if they do feel that China is on the cusp of overtaking the U.S., might decide that that in itself is a national security concern and weigh whether it makes sense to have more restrictions there. The other is, and I think this comes more from the AI labs in the U.S. and certainly companies like Anthropic, that a lot of this year has been defined by models that are more advanced at, say, cybersecurity capabilities and what the risks are there. And the belief has been, well, if Anthropic is the market leader, they can help kind of set precedents around what steps should be taken, what safeguards should be offered.

17:12And ultimately, they can deploy their models to trusted partners first before other bad actors might exploit them. If we enter a world where China has one or more models that are more advanced or equally advanced than what we have from saying Anthropic, would they follow those same precautions? And if they don't, what might be the societal impacts of having a widely available, more advanced open source model through a Chinese firm that hasn't gone through that same rigorous safeguarding process? We're not there yet. Those are the questions I imagine folks would be asking.

17:44Carol Massar:A lot of questions, right, with everything and anything to do with AI still as we move forward. Hey, Seth, thank you so much for jumping on with us. Seth Fiegerman, he is Bloomberg News AI editor joining us here on Bloomberg Business Week Daily. You're listening to the Bloomberg Business Week Daily podcast. Catch us live weekday afternoons from 2 to 5 p.m. Eastern. Listen on Apple CarPlay and Android Auto with the Bloomberg Business app or watch us live on YouTube. Hard to believe the World Cup wrapping up this weekend. I kind of can't get my head around it. Who are you rooting for? Are you allowed to say?

18:19Carol Massar:No, I'm not going to say. Sebastian who should I be rooting for Spain give Spain a chance you already won just recently it's been fun to be around the newsroom people are watching everywhere they're watching their phones streaming networks you name it and as AP reports every match of the tournament has been available for US viewers in English on Fox or FS1 or Telemundo or Universo in Spanish and there are streaming options such as Fox1 or Peacock and what caught her attention from this AP story. Thank you, Sebastian, for pointing it out for us. That the American Spanish Language TV network owned by NBC Universal, we're talking about Telemundo, points to Nielsen ratings to show that roughly half the World Cup viewers in the United States have watched at least some portion of some matches in Spanish.

19:06Carol Massar:We wanted to know more. So joining us to talk about it is Miguel Lorenzo. He's Senior Vice President of Sports Content at NBC Universal Telemundo Enterprises. He joins us from New Jersey. Also with us is our Vanessa Perdomo-Maglione. She is Bloomberg News sports business reporter, co-host of the Business of Sports podcast. She joins us right here in our studio. Welcome, welcome, Miguel. Great to have you here. Sounds like it's been a great run this round of the World Cup. Tell us about the ratings. Who's watching? And how do you know that roughly half of the World Cup viewers in the U.S. have watched at least some portion of some matches in Spanish?

19:44Carol Massar:Talk to us about that. Sure. Well, thank you for having me. The audience in this World Cup has been an amazing story. And I think it speaks to many things. It speaks to certainly Telemundo and Peacock's coverage of the World Cup across many different platforms. I think it also speaks to the World Cup tournament itself, whether it's the expanded field where there's more teams and there's more games, the legends like Messi and Ronaldo, the stars of the World Cup. and we can't discount the fact that it's a World Cup taking place in our backyard, in our communities. And a World Cup that is a world away is still interesting to all audiences of soccer.

20:24But when it's taking place in your country, it's hard to replicate that excitement, knowing that whether there's matches in your city, in your communities, watch parties, it's another level to it that we haven't seen since the last World Cup here in 1994. for and certainly whether it's on social media or through our coverage, the response from audiences with these matches have been amazing. Miguel, obviously one of the things we can't ignore when we're talking about Telemundo is the iconic voice of Andres Cantor. I mean, everyone knows that. Goal! You know, everyone knows what that is. Love time.

21:04Do you think that's one of the reasons why people are so invested to tune into Telemundo? because even people who don't speak Spanish are tuning in. Absolutely. You know, Hispanics are about 20 % of the U.S. population. Yet, as you mentioned, nearly half of the World Cup audience in the United States has been on Telemundo and Peacock. And so the excitement and the passion that our talent and that our production brings, Andres Cantor certainly being one of them, it's certainly one of the reasons why audiences are watching on Telemundo. So, you know, Andres is a very important part of it. And certainly, you know, his calls of these various Argentina comebacks have been very viral.

21:46Oh, the last one, he almost lost his voice. I love that. We joke with him like he needs to really make sure he's taking care of us, you know, drink some honey and some lemon. We got to save you. There's a lot of games to go through. But, you know, it's been him. It's been other talent on the team. I mean, we've had some major names join our team, whether it's Andres Guardado, Carlos Salcido. These are big names with the Mexican national team that have played with that team. And we saw when Mexico beat Ecuador, the Mexican team came over and started throwing them up into the air in celebration.

22:21So there have been so many great moments like this. Maxi Rodriguez is another one. Messi has given him like a celebratory hug before every game. It's worked out. It's been good luck because obviously Argentina has made it this far. So when you look at all of our talent and all the excitement that they bring and the passion into that coverage, and that's not just during the game, but everything before and after and everything happening across our platforms, it's no surprise to us that audiences are gravitating towards our coverage and watching these matches on Telemundo. Hey, Miguel. So the Spain versus Argentina match, this is featuring two Spanish speaking teams facing off in the World Cup for the first time World Cup final since 1930.