How to Build a Data Center Without Making Enemies

(upbeat music) Welcome to Tech News Briefing. It's Tuesday, September 8th. I'm Bell Lynn, a reporter for the Wall Street Journal Leadership Institute. The social media company Snap is gearing up for the launch of new smart glasses later this month. And at a price tag of nearly $2,200, they're not cheap. We explore why Snap is making such a big bet on the glasses, even though not everyone is bought in. Then, the good news about data centers is that there are ways to build them with people and communities in mind. The bad news is that's not happening right now. We dive into some of the ways that tech companies can develop data centers that won't alienate communities and why they're still such a long road ahead. But first, even as Snap struggles with profitability, the company's CEO, Evan Spiegel, is doubling down on its foray into smart glasses. That has led to concerns from investors and Snap employees that the move may not pay off. WSJ Deputy Editor Jessica Tunkel is here now to discuss the turmoil inside Snap and why its upcoming glasses launch is such a big test for the company. So Jessica, why Snap's CEO, Evan Spiegel, making such a big bet on smart glasses? So Evan really sees these specs as like the future after the phones. These are better than phones, he believes. He thinks they allow users to interact with the physical world in a way that you can't with a phone because you're looking at it or you have it up to your ear. He describes his kids playing in the backyard and being able to play laser tag while they're wearing these specs. So he thinks it's just a more interactive, natural way of being present, but also being able to do all these other things. So it's a big bet for him and for the company. And it's interesting that he personally believes in the sort of future of smart glasses, but then he's also organized his company around this big gamble as well. How has he done that inside Snap? So he has been spending more time as we've reported and he's allocated more resources. He allocated about 500 employees, which is about 10% of the company to work on specs according to our sources. And he recently set up a subsidiary just for specs. So a lot of people believe that ultimately if this takes off, he could spin it out to be a separate company. But right now Snapchat is funding specs. So a lot has to happen for that to work out. And it seems like there's no shortage of critics and doubters out there. Why are some of Snap's investors, analysts, and even some of their own employees concerned that this bet won't work out? I think if you take a step back and look at Snap as the company, the stock is down more than 30% this year. It's down around 90% since it's pandemic highs. It has not had a profitable year yet. So with that backdrop, a lot of people are asking, why are we going into this whole new thing? Why are we launching this new hardware business? This is a very expensive product. It's $2,200. Are people going to pay for it? So there's just a lot of questions. If this is the bet that Speaker should be making now, given everything going on. Internally leaders as well as other executives within the company have told Evan, they've expressed their concern that he's been spending more time with specs at the cost of Snapchat, which is the core product. And there are people who believe that the Snapchat product could be more competitive if he's spent more time there. So what's really at stake for Snap? I know you just described the company's resources being poured into this area. But does Snapchat continue to fund the company if specs doesn't work out? What's really the big challenge here? We reported that they had looked at outside funding to fund specs that didn't happen. I don't know why that didn't happen, didn't work out. But the question is, how long can Snapchat fund this business? If it's not going to make money anytime soon, Evan Spiegel himself has said he doesn't think there will be mass adoption of smart glasses for a decade. So there are a lot of questions for this business. And what about you, Jessica? Would you wear the Snap specs? And do you think they'll eventually catch on? I don't think I'm the target audience for specs. So I don't want to say that I am the ideal customer. There are very few things I would spend $2,200 on. I'm sure my kids will be asking for them though. So we're just going to have to wait and see how well these do. That was WSJ Deputy Media Editor, Jessica Tunkle. Would you buy Snap's new smart glasses or any of the others out there? If you're a listener on Spotify, leave us a comment with your thoughts. Coming up, is there a way to build data centers people won't hate? Our tech columnist has compiled a list. We'll get into it after the break. (upbeat music) Tech companies are quickly realizing that if they want to build all the data centers they've promised to, they're going to have to do it in ways that don't lead to massive community pushback and political opposition. 15 states have weighed moratoriums on data center construction and at least 100 localities have approved one according to Brookings. Methods to make data centers better do exist. They're just not all being deployed. Here now to discuss what a good data center looks like and how tech companies can build them is WSJ columnist Christopher Mims. Something I found interesting from your reporting is that data center builders nowadays are more inexperienced than they were before. Is that a relatively new development and how does that contribute to communities mistrust? What you're seeing now, which absolutely didn't exist even two, three years ago, are what I call data center wildcatters. And that's a term borrowed from the oil patch. Those were folks who would just go out and drill a well and hope for a gusher way back in the day. And what has happened since the take off of JTBT and AI in general is there's been a lot of speculation. So folks will go get a title to land and try to essentially flip it or figure out if they can connect power and fiber in order to sell it on to maybe a NeoCloud or somebody else. You're seeing data centers built in a speculative way and it's the Wild West in some ways. So it seems like because you have these speculators who don't really know what they're doing, they might be giving a bad name to some of these data center builders adding to people's mistrust overall. Not just the speculators. I think we got to be clear that some big, more experienced builders of data centers, specifically SpaceX AI, have done things like install many unlicensed, small natural gas turbines, for example, in their Colossus data center campus in and around Memphis. They have acknowledged that these are temporary turbines. They say that they have begun shutting them down and eventually they're going to replace them with a 1.2 gigawatt natural gas power plant, a more conventional one that's cleaner. So in the race to stand up new compute as quickly as possible, you're seeing even experienced data center developers do some things which local communities object to, like setting up these temporary power plants which are louder and more polluting than what's been used in the past. - So one of the other issues that a lot of the public is very cautious about is that their utility bills might go up because of these data centers. Is there anything new or any new techniques that tech companies can do to mitigate these potential utility rate hikes? - Notably a year or two ago, the idea that these tech companies would pay for 100% of the required grid upgrades to get power to these data centers was a little bit of a radical idea. Now it is standard language. Another thing they can do is build power on site like natural gas. And then that's also driving these more radical future technologies which really aren't here yet and those include small modular nuclear reactors which again don't exist yet, but companies are investing in their development. So there's a lot of ideas out there. - One other interesting example from your reporting is a Google data center in Finland which you bring up as an example of how a data center can be both specific to the region where it's built and also use less water. Is there a bigger lesson here for other tech giants? - The big lesson from the Hamina data center in Finland which uses seawater to cool itself and the waste heat is pumped into a district heating system which heats the local town. Is that data centers have to be built in a way that we're seeing. to respect their local environment. - You make this kind of overarching argument towards the end of your story that kind of boils down to two simple things that if these builders slowed down the data center build out and spend more money, then they could solve a lot of the issues we just talked about. How would that work? - These data center bills are already spending more money. Every single hyper-scaler wanted to tell me about their community first approach. They're doing things like buying school lunches for all the local kids for the next two years or contributing $80 million to local community development funds. Those two things were Microsoft and OpenAI respectively. And they have the money to do that kind of thing. They could spend even more to try to shift to, for example, cleaner and less noisy sources of energy. This is already happening Oracle in New Mexico. That a ton of pushback on a giant data center complex called Jupiter. And they switched how they're going to produce power from conventional natural gas turbines to fuel cells from a company called Bloom Energy, which are much quieter and much less polluting. In terms of slowing down, it's a real prisoner's dilemma. They're all locked in what they see as an existential battle with each other to capture the market for AI. And they're being slowed down by regulators and not just the usual ones whom you would expect in the usual states. It's really notable that Texas, which was very welcoming to data centers until recently, has enacted a temporary moratorium and said, you cannot build a data center here unless you satisfy this long list of requirements. So whether or not they want to slow down, it's clear that at the state and the local level, they're going to be forced to do. That was WSJ columnist, Christopher Mims. And that's it for Tech News Briefing. If you're listening on Spotify, be sure to leave us a comment. Today's show was produced by Julie Chang with supervising producer Melanie Roy. Logging off, I'm Bell Lin, a reporter for the Wall Street Journal Leadership Institute. We'll be back later this morning with TNB Tech Minute. Thanks for listening. [BLANK_AUDIO]