Broadcom's Disappointing AI Outlook Drives Tech-Led Stock Rout

0:00Carol Massar:Get the news you need in just 15 minutes. Start your day with Bloomberg Daybreak, the podcast with a global view on the stories that matter. I'm Nathan Hager. And I'm Karen Moscow. Join us each morning for curated stories on current events, politics, business and foreign relations. Plus one conversation on the day's biggest developments, all in just 15 minutes. Subscribe to Bloomberg Daybreak for a precise, thoughtful take on the stories that matter. Listen to Bloomberg Daybreak each morning on Apple, Spotify or anywhere you listen. Bloomberg Audio Studios, podcasts, radio, news. This is Bloomberg Business Week Daily, reporting from the magazine that helps global leaders stay ahead with insight on the people, companies, and trends shaping today's complex economy.

0:50Carol Massar:Plus, global business, finance, and tech news as it happens. The Bloomberg Business Week Daily podcast with Carol Masser and Tim Stenebeck. on Bloomberg Radio. SpaceX has already received orders for more than the shares available in its$75 billion IPO. That's according to people familiar with the matter. It puts the company on the verge of setting the record for the biggest ever listing separately just crossing the Bloomberg terminal. Bailey Lipschultz leading this story too. A quiet effort by Morgan Stanley to prevent retail investors from placing multiple orders for shares of SpaceX in the company's massive IPO is facing pushback from some of the country's largest brokerages.

1:29Bailey Lipschultz joins us now. He's holding on the fort in Bloomberg headquarters in New York, the Bloomberg Interactive Brokers studio. Bailey, you lead our IPO coverage. We're going to talk about SpaceX and the fantastic reporting that you and the team have done around it. Before we do that, though, as we just heard from Charlie, we're seeing a major sell-off today with some of the chip names really leading the charge lower. Your view on this equity market, especially as tech has been leading the way, giving back some of these historic gains. I mean, Tim, I go to Italy for about two weeks and the market's at an all-time high.

2:04We missed you. On no news though, Tim. No news happens. The market goes up. No one bats an eye. We see two days of weakness and all of a sudden the sky is falling. Just want to call out the fact that yes, the Philadelphia Semiconductors Index is down more than 9.5 % right now. It's up 74 % year to date. If anyone told you that, that would be the case sitting here on June 5th. I think they would take that. NASDAQ 100, of course, down 5 % from its all-time high on Tuesday. It does seem like a bit of consternation. I think the big question when I talk to investors is, what's the next move for the Fed?

2:34And it does seem increasingly that the logical answer has to be a hike, whether that's later this year, whether that's early next year. Obviously, we'll see how the economic data play out. But it does seem like kind of a natural way to take your foot off the gas pedal going into a weekend. Obviously, we're still living and breathing everything that comes from the White House. But the big question at the end of the day is, is this a market that did nothing but go up for a number of weeks, if not months, for no real reason, and what kind of lies ahead. And obviously, there's going to be a big IPO next week that'll be interesting to track as well.

3:07Carol Massar:You know, having said that, Bailey, I do feel like this is a moment in time where there's just such a FOMO trade, especially when it comes to AI and tech and even some of these IPOs, which we're going to dig into in just a moment. We talked with John Flood over at Goldman at Sachs earlier, you know, and he tracks hedge funds and institutional investors and did talk about the hedging that these folks are doing so that they're not all in on everything. Do we see any of that? Do we see a rise in short positioning at all? Do we see anything that kind of says, all right, people are enthusiastic, but they're also playing this smart?

3:41I think there is kind of that sense when you look at some of the data that we see in others track, it does seem like dips are being bought in the sense that some of these memory stocks are being added to. Today, we did see people taking protection over the last few weeks as it relates to some of those higher flyers. But Carol, it still is an interesting market where you do see certain short positions rising, but in other pockets of the market, you're still seeing exuberance or even just shades of exuberance as it relates to what the future could hold. So we're seeing a lot continuing to play out.

4:10The big thing has been this is a market that when I look back over a few days ago, even I expected to see a sharper pullback. And coming into today, we were down, you know, 89 % of the NASDAQ. Sorry.

4:23Carol Massar:Yeah, no, don't be sorry. It's like a lot coming at us. We're looking at, if we look at the chart, kind of a slope down on all these major equity averages, and we are now down about four and a half percent or more than four percent on the NASDAQ 100. So, we continue to see selling into the close here, Tim. Yeah, Bailey, does it change? Does it, I know one day does not make a market, and especially given the rapid move higher going back to the end of March that we've seen, especially with tech. As I mentioned earlier this week, the NASDAQ 100 was up more than 30 % just from March. The Sox was up more than 90 % earlier this week, still up, as Carol mentioned, more than 70%.

5:02Does it change? I'm not going to say does it change the timing, but does it change the environment that SpaceX goes public in? I think anyone who says that the NASDAQ 100 down almost 5 % in a single day doesn't matter is probably lying to you. But I do think, listen, this is, I'll give you the summary that I've been giving people. It's been five months ago, we and others reported that SpaceX wanted to raise$75 billion at a 1.75 to 1.8, maybe$2 trillion valuation. That was five months ago. Here we are in June and it's happening. They set a fixed price a few days ago for this IPO. If there ever was a deal that has been as transparent or the worst kept secret, it's this one.

5:47So anyone who's now all of a sudden saying, you know, the NASDAQ 100 is down 4.5%, I actually can't cut the check to SpaceX that I thought I could a week ago. Well, where were you when the market was in a bit more turmoil? And I think that's just one thing that I think people have to keep in mind, that we've been talking about this for months. This isn't a deal that came out of left field. that's not a deal that started with a price range that was north of$2 trillion and they have to sell investors on that notion. This is something that bankers in the company have been, at least through our reporting and others, been pretty open and very clear about.

6:20The demand that we're already seeing right now for this IPO next week, how would you characterize it? It's expected. I mean,$75 billion. Nothing off the charts, nothing surprising? No, I think this is, again, we know that the company did one-on-one meetings at Starbase with some of the biggest portfolio managers in the world back in January. Does it need to do that? Everybody sees what the rockets do. Everybody knows who Elon Musk is. Tim, I'm going to direct you to some of our reporting and say that this is a company that a year ago was a space monopoly, and now they're pitching a$26.5 trillion AI market.

6:58This is a company that when you look at the sell-side models that we were able to get our hands on. Analysts are penciling in AI to bring in$755 billion in sales in 2031, according to Evercore ISI research analysts. That was 3.2 billion last year. So there's a lot of excitement. There's a lot of hopium. But this is a company that before the XAI merger was strictly a space company. And now we're seeing the magic of Elon Musk, or at least people being excited about what AI could bring and what orbital data centers could bring. And I tuned into that conversation Musk had with Jamie Dimon yesterday talking about making Star Trek reality.

7:35Yes, this is all Elon Musk, but this is something we saw play out with Tesla. And we're going to see it play out with SpaceX.

7:41Carol Massar:Well, as people reminded us at the Bloomberg Technology Summit yesterday, Bailey, never count Elon out. Just about 30 seconds. I think it was going to be... Yeah, it was Trey Stevens at Anderil with his conversation with Ed Lovell. He said, don't bet against Elon. Exactly. What will be key, though, is how this stock, once it IPOs, trades in the months right after. Just quickly. Yeah, it'll be critical. And I think even day one will be closely watched. If 30 % of this does get allocated to retail north of$20 billion, we've never seen that. So it's going to be key for markets and for the company, hands down.

8:11Carol Massar:Got to say, it is a fun time to be covering markets, right, Bailey? Yeah, don't go back to Italy next week. I came back and all of a sudden everything was still going crazy. Yeah, I agree that you can kind of disappear, come back and crazy still exists. Bailey Lipschelz, we love you. Bloomberg News Markets Reporter covers the IPO market and so much more at Bloomberg. Stay with us. More from Bloomberg Businessweek Daily coming up after this.

8:40On June 10th, Bloomberg Invest is back in Hong Kong.

8:43Carol Massar:We look at the role Hong Kong plays between China and the world as major powers compete and markets realign. As global investors rethink risk, we'll explore the forces driving Asian demand and the future of private capital. Catch exclusive interviews with top newsmakers, plus a live recording of Bloomberg's Odd Lots podcast. Visit bloomberglive.com forward slash invest Hong Kong to learn more. Supporting sponsor Deutsche Bank. You're listening to the Bloomberg Business Week Daily Podcast. Catch us live weekday afternoons from 2 to 5 Eastern. Listen on Apple CarPlay and Android Auto with the Bloomberg Business App.

9:21Carol Massar:Or watch us live on YouTube. A force behind the event, Tom Giles, he's Senior Executive Editor for Global Technology here at Bloomberg. He joins us from our San Francisco Bureau. Just steps away from us. Tom, first of all, congratulations on a great event. You had some awesome conversations. We're going to talk about some of them. Andrew Feldman over at Cerebris, Dan Schulman of Verizon, Ali Goetze, presidency of Databricks, Hocktan of Broadcom. That's kind of where I want to start because Broadcom's now having its worst two-day stretch ever. Shares getting crushed yesterday, down more than 12%.

9:55The company forecasts for AI. Chips disappointed investors down again today. Really timely conversation. What did you want to hear from Hocktan at Broadcom? Well, the biggest thing, I mean, you know, yesterday on the news was the fact, as you said, its shares were falling, dragging down ship stocks. It's having another bad day today. And that's on the heels of basically a lighter than expected forecast for revenue in the current quarter. And remember that that is on the heels of a huge run up, adding hundreds of billions of dollars in market capitalization in a matter of days. What that speaks to is just how many, how high the expectations are for chip makers like Broadcom in the AI era.

10:40And if you don't meet or exceed expectations, and sometimes if you don't exceed expectations by a lot, there's going to be disappointment. What people are worried about and a theme that we kept coming back to during the conference yesterday is whether end user demand will be sustainable. And that's demand for chips and infrastructure needed for these large language models that are basically changing the way we live and work and operate. And then whether there will be end user demand for the actual outcomes, the actual services that are being created by OpenAI, Anthropic, XAI, and the other large language models that are investing so much money into this infrastructure.

11:29well let's let's so tom i want to jump in because i want to go to part i want to go to part of that interview that you had with hock 10 over at broadcom yesterday this is him there was a lot to choose from this is him just weighing in hock 10 at broadcom on the ai super cycle and the hype cycle the enthusiasm yes we are kind of like ai in a very surreal environment frankly i don't think about it. No, it's hard not to, but no, just focus on fundamentals, create value, and stop thinking about your stock price. Trouble is very hard to do that. Yeah, I would imagine it's very hard to stop thinking about the stock price.

12:10That was Hawk Tan on the Broadcom AI hype cycle from your conversation yesterday at the Bloomberg Tech Summit. I mean, it is hard to walk around San Francisco and not think about AI because it is everywhere. I mean, every single ad, every billboard from the legacy tech companies talking about how they're harnessing AI agents to startups that this point I've never heard of. I mean, this is what this is the environment that we're in right now. Yeah, I'm so glad other people are experiencing it who live outside of San Francisco. It is surreal. And we are in something of a bubble. And I don't mean a bubble in terms of inflation of stock prices, although that may be the case.

12:47I just mean a bubble in terms of the fact that this is on everybody's mind. It's the topic of conversation. And there is a big conversation happening right now about whether we've experienced paradigm shift from where the old values, the old adages and the old truths about how markets operate, boom bust cycles, and whether we're breaking out of it. And there's a lot of people who are very pessimistic about the idea and that look we are in a bubble that there will be a crash that demand will not meet these expectations um there's also people who say we've kind of entered a new phase and i just want to give one case in point when i was talking to broadcom broadcom uh when i was talking to hawk tan i asked him about this is a company that has built itself in many ways on acquisitions hawk is known for big bold acquisitions not all of them came together but he made some really really big ones.

13:44And that was very much a trademark of the way he managed things. Now in the generative AI era, he has obviously embraced AI chip making, providing an alternative, a competitor to NVIDIA, which is the far and away leader, right? In designing these AI accelerator chips. And he's basically said, we're kind of entering a post M &A phase where I don't need to do that to add growth, what I'm doing now is doubling down on chip development and working. And a couple of examples are the deal that he has with Google through 2031 to work on his TPUs. Another one is the work that they're doing with Anthropic, one of the biggest LLMs.

14:31Carol Massar:Well, and you know what's interesting, like kind of playing off of that, so many people, Tom, as you know, just talked about the incredible demand and trying to keep up with it. There was another conversation that Shireen Ghaffari had with Anthropa co-founder and president Daniela Amadei. And she too talked about just the incredible spend that's needed. Let's just listen to a snippet of that conversation at Bloomberg Technology yesterday. Speaking for ourselves, and I think probably really for the AI industry more broadly, it's a very capital intensive business to train AI models. I think the sort of core set of companies that are working to advance the frontier are just going to need access to capital.

15:14And I think the public market is very well suited to that.

15:18Carol Massar:And that, of course, is Daniela Amode of Anthropic. It's interesting on a day where Meta is weighing, raising tens of billions of dollars in a new share sale. That came from FT. But everybody seemed to talk about that you've got to keep spending to build, Tom. Right. Every couple of days, we hear of another mega multi-billion dollar deal whereby a company, whether it's the OpenAI, Anthropics side of the fence, or the Googles and the Metas and the Microsofts of the world, these so-called hyperscalers, they all are looking for ways to raise the capital that they need to keep developing their models and to keep ensure that they have access to the computing power.

16:07It's very expensive. It's very capital intensive, as you said. You know, every couple of days there's another, you know, multi-billion dollar deal. And, you know, we're talking about we're on the cusp of three major IPOs, another big theme of the conference, another big theme out here in Silicon Valley. is SpaceX, Anthropic, OpenAI, all of them looking to raise tens of billions of dollars in the public markets. And, you know, just when you thought it was safe to, you know, kind of get back into, you know, IPO mode, here comes Google, which IPO decades ago. They, too, are going to tap the public markets through an equity offering, again, to raise tens of billions of dollars.

16:48But to get back to Daniela, one of the one of the nice things about that interview was getting her to weigh in on the IPO. There's not a lot that they have said about their IPO plans. There's not a lot that they can say publicly. They did file confidentially and they really need to be careful to not run afoul of SEC rules. But she did, to our delight, talk a little bit about the IPO and just admitting like, yeah, we need this is another source of capital for us to buy those chips and those data centers. Yeah, you know, it's kind of a good segue to talk about an interview that you did, too, this week, Tom, with a company that did go public a few weeks ago.

17:35Carol was there for this one. It was incredible. She's been raving about it. We're talking about the interview that you did with Cerebris's CEO, Andrew Feldman. Let's take a listen to what he had to say about the idea of bubbles. I think historically, bubbles were characterized by a notion of if you build it, they will come. What is unusual about AI right now is the builders are so far behind the demand, it's absurd. We have a backlog of more than$25 billion of demand. That there are none of us, not us, not AMD, not NVIDIA, that can keep up with the demand that your employees are driving. And that's sort of, in a lot of ways, the opposite of a bubble.

18:25Andrew Feldman, Cerebers' CEO, talking to you, Tom, earlier this week at the Bloomberg Technology Summit here in San Francisco. The demand question, I think, is a really important one right now. He pointed, if you're just listening to us on radio, he pointed to the audience and he said, your employees are driving. And he's referring to the executives in the audience there. But at the same time, we're hearing from some companies that they want to crack down on token usage at their companies. We had an executive at Uber this week essentially saying that. We've had critics come on our program that say, well, the idea of all these employees using all these tokens and us spending so much money, we're not actually seeing a return on that.

19:08How are you thinking about that part of the conversation right now? Yeah. You know, we talked to several executives about that, including the Verizon CEO. And there's two schools of thought. There is the people who say, look, yeah, we do need to impose limits. We don't want people, you know, just willy nilly using this computing power, which is so expensive. And so, yeah, they're being a little bit more choosy. They're putting limits, they're metering, they're throttling, whatever verb you want to use. They want to know that there's going to be a return on that investment. So in the earlier days of the generative AI phase kicked off, you know, late 2022, roughly, there were a lot of companies who were telling their employees, go experiment, build, create chatbots.

19:58And now we're talking more and more about agents. That's still happening. But I think people are starting to get and businesses are starting to get more strategic about how they're doing that, how much you can spend. And so it does make sense for people to be a little bit more choosy, be a little bit more, impose some restrictions, put some parameters around that. Um, uh, yeah. So, um, I think, yeah, the other thing, well, go ahead, Carol.

20:26Carol Massar:No, you know what I'd love to is when you talk to, um, Andrew Feldman and he said, we're going to, you're going to see people shopping, I think for AI, right. In different models or usages, like we go to Costco, like it was a really smart kind of analogy of to how you think about it. Yeah. I remember him talking about buying a big tub of mayonnaise and, you know, You just don't need it. So you got to be strategic. The other thing that we're starting to see, too, is that as engineers, software programmers, have the option to choose between different LLMs, you don't have to be wedded to one particular one.

21:04So they're going to start shopping around more and more and being strategic. And what is interesting is that there are certain tools that we can look to that open router, for example, that can give us an indication of token usage and who's ahead and which LLMs are sort of gaining momentum at any particular time. And we've seen over recent months when people have the choice, when engineers have the choice, when they can choose between a different LLM for particular, you know, one thing or another, one task or another, they're gravitating toward the lower cost LLMs. And increasingly, the lower cost LLMs are coming from China.

21:47That's something that's really interesting. Now, there are certain companies, there are businesses that place limits, maybe don't let their employees use, say, something created by Alibaba or DeepSeek, for example. And then there's others where if you've got access to them, people are using them. And obviously, different LMs have different strengths. Anthropic has really leaned into tailoring its tools for businesses. And so people, for certain tasks, people say, nobody can beat Anthropic. Mythos, their tool for cybersecurity is something that comes to mind. People say that's significantly ahead of comparable tools by other companies.

22:29On the other hand, if there is some kind of a task that you can accomplish using a cheaper LLM, people are going to do that. And that does have implications for, again, lower cost models coming from China. They just have done a better job of doing things more efficiently. And that's something that's, you know, should resonate with us at a time when the U.S. is engaged in this war with China over who's ahead in AI. We have done, the U.S. has done a lot to try to handicap China's ability to get access to chips, for example, advanced chips, to sort of ensure that the U.S. stays ahead. Right. And, you know, in many ways, Chinese based LMS are catching up.

23:19Carol Massar:Yeah, I mean, highly, it's just amazing, like all the conversations and what's going on. And you can't help but feel that the CFOs of companies are going to start looking at that ROI and looking at the cost and what the benefits are to an organization. Tom Giles, you rock. Stay with us. More from Bloomberg Businessweek Daily coming up after this.

23:42Carol Massar:you're listening to the bloomberg business week daily podcast catch us live weekday afternoons from 2 to 5 eastern listen on apple carplay and android auto with the bloomberg business app or watch us live on youtube