Meta’s AI Spending Starts to Spook Investors

[00:00:00] This episode is sponsored by HPE. Today's networks must evolve from simply providing connectivity to harnessing AI to deliver exceptional connected experiences. At the break, HPE's Rami Rahim shares how self-driving networks can make that happen. Welcome to WSJ Tech News Briefing. It's Friday, July 31st. I'm Isabel Busquets, a reporter for the Wall Street Journal Leadership Institute.

[00:00:28] Young adults are looking to AI chatbots to help them figure out what to say in social situations. We're diving into why they're doing it and what the risks are. Then, Meta's stock stumbled this week after the company said it anticipated $137 billion in capital spending this year on AI infrastructure. We're looking at how Meta is paying for all of it and why investors are starting to lose patience.

[00:00:58] But first, do you ever get stuck in a conversation and not know what to say? For a growing number of young adults, that's no longer a problem. And it's all thanks to AI. 20-somethings are now asking chatbots for advice on what to say to friends, family, even potential romantic partners on dating apps.

[00:01:18] And while it's helping some ease their social anxiety, psychologists warn it could be damaging interpersonal relationships and teaching young people that they can't trust their own judgment. Natalie Kaufman, who reported on this for the Wall Street Journal, spoke to our colleague Imani Moise. What was the most surprising or extreme example that you came across of someone outsourcing their social life to AI?

[00:01:42] Definitely someone who uses AI in real time to help her reply to someone in a tricky or otherwise kind of uncomfortable situation. So if she didn't know the subject of something having to do with politics or a book or article that someone mentioned, she wanted to seem smart and well-read. So she would just ask AI for a quick generated summary of the topic and use that as sort of her little script or note cards to know what to say next.

[00:02:09] So we know that AI can make mistakes at work. Can it also backfire in social situations? For sure. Someone I spoke with said she was trying to figure out what to say to her friends in person. And AI gave her just this terrible response that backfired on her. And she says she has lost or at least weakened some of her relationships by trying to outsource her thinking and her emotions to AI because it doesn't come across the way sometimes you think it will. It comes off as kind of robotic and cold.

[00:02:38] She has done this with dates as well. So before she goes on a date, she'll prepare these quote unquote note cards as sort of like talking points. And she's gotten to the restaurant or the movie or whatever, and it just doesn't come out the way she anticipated it would and kind of lands with a thud. You also spoke to someone who was on the receiving end of this where she was dating someone and then all of a sudden suspected that he was starting to use chat GPT or some type of AI in all of their communications. How did she know? What was a tell?

[00:03:07] Yeah, so this was her friend, actually. She was seeing this guy at the time and his texts were coming across with really poor grammar and spelling and just kind of rambling. And all of a sudden the texts became really polished and clean. And she said the way that she and her friend knew that these texts were AI generated was that before he kind of had a hard time figuring out the difference between there, there and there. And all of a sudden he's an expert. He's perfect at grammar. Yeah.

[00:03:36] That extreme aside, you spoke with several people who use AI this way. Did any of them tell you they felt more confident over time or more dependent on the technology? So it definitely varies. Some people felt empowered by the AI and had it sort of be a sounding board for them. And others found that they used AI sort of more as a crutch, something where it made them second guess everything unless they put it through AI. What do experts say about that dependence or checking AI a lot?

[00:04:03] Yeah, a lot of them warn against using it so much to the point where you lose faith in your own voice and your own ability to regulate your emotions. But one friendship and relationship coach I spoke with actually advises her clients on sort of healthy use of AI to not script conversations, but to prepare going in so that you're not as worried.

[00:04:27] And she mentioned that for the socially anxious, it really helps unlock conversations that they might have otherwise shunned or avoided. She does recommend to limit use for a big event just the day before is plenty just to get some talking points. And then for recurring events, just once a week. But it really just depends on the person and what they're comfortable doing. That was the Wall Street Journal's Natalie Kaufman speaking with our colleague Imani Moise.

[00:04:55] A reminder, we want to hear from you. Are you worried AI is looking more and more real? How do you make sure posts you're seeing are real? Send us an email to tnb at wsj.com or leave us a voicemail at 212-416-2236. Or if you're listening on Spotify, leave us a comment.

[00:05:19] Coming up, Meta isn't the only tech giant spending big on AI infrastructure, but it's having a harder time justifying some of those costs to investors. That's after the break. Here's Rami Rahim, HPE Executive Vice President, President and General Manager of Networking. A self-driving network is a network that essentially configures itself.

[00:05:48] And what used to take months, maybe up to a year to deploy, can now take weeks, if not days. A network that optimizes itself. And very importantly, a network that heals itself. Meta had a tough week after failing to ease investor concerns over its massive AI spending plans.

[00:06:13] During its earnings report on Wednesday, the company said it's anticipating capital expenditures of $137.5 billion this year. A total that analysts expect to put it into negative free cash flow territory in the latter half of the year for the first time since its IPO in 2012.

[00:06:34] Investors have been somewhat patient with tech giants' massive AI spending in the past, but Meta's uniquely heavy reliance on its ad sales business to support all that spending now has some spooked. WSJ's Asa Fitch joins us to break down why Meta is in a different position than some of its big tech peers, and whether it plans to back down on its big spending plans anytime soon.

[00:06:59] First thing, Meta is not the only big tech player projecting massive spending on AI infrastructure right now. But they do seem to be the most financially stretched. Why is that? Meta has a great core business. It sells ads, and it does that really well. But they also are taking on lots of debt. They have a lot more obligations than they had in the past. They basically had no debt, if you can believe, before 2022. And now they have over $80 billion of debt.

[00:07:27] But they also have a lot of off-balance sheet obligations that aren't even included in their financial statements. They're doing these huge data center projects in Louisiana and Texas. All this stuff adds up to just a lot of obligations for Meta that it didn't have in the past. And it's really stretching the company increasingly financially. Of course, Meta thinks that these are good things to do because they'll generate good returns. So exactly how much are they spending to build out all this AI infrastructure? Pretty much any money they can find.

[00:07:58] I mean, they're spending this year, it could be as much as $145 billion. Next year, analysts are forecasting. The consensus is around $174 billion. But some of the analysts believe they're going to spend upwards of $200 billion, perhaps even as much as $280 billion. So the spending spigot for Meta is still very much flowing. And, you know, it's just going to get bigger next year. And that's going to put more strain on the company.

[00:08:24] I know you mentioned the ads business, but walk us through where they are getting all the money to do this. From the ads business, that throws off a ton of money. But they're also borrowing tons of money. They're issuing a lot of bonds. And they also have these financial structures they've entered into with various sort of players like Blackstone, etc., that will give them money to do this stuff. And in the future, they could potentially raise equity. They haven't really said they're going to do that.

[00:08:52] But it's something that Alphabet did fairly recently and fairly successfully. So it's something that Meta could do. But they also have plans to get into new business ventures and do things with their social media complex that make them more money, too. Yeah. What are these other potential sources of revenue for the business on the social media side? And how real or promising are they? They have a few of them.

[00:09:15] I mean, they have subscriptions for their social media, like Instagram and Facebook and so on and so forth, where you can get quote-unquote premium features if you pay them a little bit of money every month. I think the prospects of those kinds of things driving large amounts of revenue are pretty slim. Another thing they're trying to do is bring in business customers. That's a big focus now. You have a lot of small businesses. You have a lot of even large businesses that want chatbots, that want other AI-linked services.

[00:09:40] And the idea for Meta is those people can come to Meta and do it on Facebook or whatever and have their customers interact with, let's say, a chatbot that kind of acts as a customer service agent for them on Meta. That could be appealing to a lot of, let's say, small businesses that can't afford to go out and train a model or just license something from Anthropic or whoever it is. They're hoping they'll get money through their paid API for their advanced AI models. That started to roll out earlier this month.

[00:10:10] And another big one for them is the cloud computing business. They hope to sell effectively spare computing capacity to others to use. One thing to note about these things, though, is that these are basically all effectively businesses that are making zero money today. So broadly, it seems like in the last couple of years, investors have been at least somewhat comfortable with a decently high level of AI infrastructure spending. Is that starting to change? Yeah, I think there are many signs of that.

[00:10:39] I mean, you saw Alphabet last week when it announced earnings. They raised their CapEx range by $10 billion. And that scared investors. People are worried that these tech companies, despite having very healthy core businesses and a lot of money, are overextending themselves. And they're punishing them. I don't think investors really think that AI isn't going to materialize or become something real. That's not really the question. The question more is, does it make sense financially? So then what's next? Is Meta going to pull back on some of this CapEx spending?

[00:11:08] Or what are we hearing from executives? Meta, because it's led by Mark Zuckerberg, who is a true believer in AI and has put a lot of money behind that conviction, is pretty unlikely to pull back on AI spending anytime soon. However, the CFO of the company, Susan Lee, characterized their spending in a way that suggested that they could pull back in 2028.

[00:11:32] They're focusing on capital spending, getting things in place, doing these big data center projects, and so on and so forth this year and next year. And then, you know, she said after that, there's some flexibility, there's some optionality. They have some ability to adjust to whatever the environment is there. But what we know for certain is that 26 this year at 27, it's going to be a lot of spending. And we're just going full blast on it. That was Heard on the Street columnist Asa Fitch.

[00:12:01] And that's it for Tech News Briefing. If you're a listener on Spotify, be sure to leave us a comment. Today's show was produced by Julie Chang. I'm Isabel Busquets, a reporter for the Wall Street Journal Leadership Institute. Additional support this week from Pierre Bien-Aimé and Imani Moise. Jessica Fenton and Michael LaValle wrote our theme music. Our supervising producer is Katie Ferguson. Our development producer is Aisha Al-Muslim.

[00:12:29] And Chris Sinsley is the deputy editor of audio for the Wall Street Journal. We'll be back later this morning with the TNB Tech Minute. Thanks for listening. Here again is HPE's Rami Rahim to explain why it's critical for modern enterprises to use AI-powered self-driving networks.

[00:12:50] Gone are the days where humans can keep up with the complexity and overcoming the cyber attacks that are happening using elbow grease. You must leverage agents and artificial intelligence in order to deliver a seamless connectivity experience to everyone and everything. It's not just people that are connected to the network today. It's AI agents. It's billions of things.

[00:13:17] In fact, you can't do this manually. It must be done using artificial intelligence. And everything from the deployment of the network or speed matters to the identification of issues that inevitably happen to the remediation of those issues needs to be done without human intervention. Find out how a self-driving network can help you deliver exceptional user experiences at HPE.com.