Data Centers Are the New Railroads - DTNS 5137  ## Metadata - Author: [[Daily Tech News Show]] - Full Title: Data Centers Are the New Railroads - DTNS 5137 - Category: #podcasts - URL: https://share.snipd.com/episode/642907af-c79f-499f-bc90-1e9d80b7ff9c ## Document Notes ## Highlights - **Data Centers As The New Railroads** - Cloud infrastructure spending is surging as AI demand drives capacity builds across AWS, Google, and Microsoft. - Tom Merritt compares data center investment to railroads: costly to build but foundational for future services. Transcript: Tom Merritt So cloud is now king, as we saw in the earnings reports earlier this week from Microsoft and Alphabet, and further bolstered by Thursday's Amazon earnings. Amazon reported that AWS revenue was up 20% on the year, driving the overall 13% rise in revenue for all of Amazon. That's only behind advertising in growth. Ad, advertising rose 24%. AWS is growing at a pace we haven't seen since 2022, reaccelerating to 20.2% year over year. Our largest growth rate in 11 quarters. That's CEO Andy Jassy. Of course, he's right along with the others. Google's cloud revenue grew 24% last quarter. Microsoft Azure grew 40%. So 20% isn't even all that much, but AWS has been the leader. So you usually grow a little bit slower when you're that far ahead. Amazon also raised its capital expenditures forecast for the rest of this year and expect to have an even bigger number in 2026. They're going to spend a lot. Amazon also says that it's cost cutting. If you're like, wait a minute, ([Time 0:02:54](https://share.snipd.com/snip/4898d4eb-a452-49e0-96e7-12e08034333a)) - **Compute Demand Is Uncertain But Likely To Find Use** - Massive data center builds may create 'ghost' capacity if AI use-cases don't materialize as promised. - Yet rapid improvement in models suggests new, unforeseen uses will likely absorb much of that compute. Transcript: Tom Merritt You're right. When do we see all this money from data center build out turn into revenue? And I think they can point at these earnings and say, this is the beginning. This is the hockey stick starting to turn up. You know, we're going up the ladder and it's just going to go up from here. A lot of people are rightly skeptical that this is going to continue to rise as much as it is. A lot of people say that this is all circular. It's a bunch of people paying each other for something that eventually maybe no one will want. I tend to think that this looks like railroads. A lot of people went broke building railroads, but in the end you had tracks and people figured out how to make use of tracks. And you've got freight shipping all over the world on railroads as a result. A lot of times the companies that spent the money building the tracks were not the ones to benefit from the tracks down the road. And I wonder if that might happen here, too. Jenn Cutter As a Canadian, I always get jealous when Europeans talk about trains that actually go places. Well, that's passenger trains, right? Tom Merritt I'm talking about freight. No, but it's the same here. You have to stop and roll down your window because, you know, the huge procession of boxcars is going by. Jenn Cutter Yeah, well, we need more of those too. I do agree that like somebody is going to benefit from these data centers down the line. However, many of them get built, people will find a way to use them because they've already made the investment, whether it's selling them off or renting them off or any other ways to Move money around. But I think it's going to come down to a people problem. Same with how video games, they talk about like, oh, we've streamlined our development and gotten rid of some managers and a whole bunch of programmers. We're not going to see the effects of that for possibly three to five to seven years because of how long development processes take these days, especially for AAA. So right now there's 8 billion games and nobody has the time for it. But how does that look years down the line? Do these processes pay off? Is AI going to make a sustainable difference in the future? As you said, we're going to find out, and I don't know how fast we're going to find out. So that will affect the hockey stick. Tom Merritt Yeah, that would be the downside, right? If the promise of these tools is not even half as useful as people expect, then it might not make use of those data centers. You might have ghost data centers in the future, right? Where it's like, ah, they built this expecting the AI gold rush and it never happened. I expect that we will find uses for computing power. ([Time 0:06:41](https://share.snipd.com/snip/c8010535-f2ff-40a1-8c7c-4d5ab01edbbb)) - **Builders May Not Be Biggest Winners** - Companies that build infrastructure may not be the ones who profit most from it long-term. - Investment creates usable platforms that others can monetize, similar to historical railroad fortunes. Transcript: Jenn Cutter As a Canadian, I always get jealous when Europeans talk about trains that actually go places. Well, that's passenger trains, right? Tom Merritt I'm talking about freight. No, but it's the same here. You have to stop and roll down your window because, you know, the huge procession of boxcars is going by. Jenn Cutter Yeah, well, we need more of those too. I do agree that like somebody is going to benefit from these data centers down the line. However, many of them get built, people will find a way to use them because they've already made the investment, whether it's selling them off or renting them off or any other ways to Move money around. ([Time 0:07:43](https://share.snipd.com/snip/eec051bf-5fe9-4087-859e-594e4ebeda6b)) - **Don't Slash Staff Prematurely** - Be cautious about claims that AI will immediately replace large numbers of workers. - Companies should avoid over-cutting staff until real productivity gains are proven, Tom warns. Transcript: Tom Merritt It's not yet AI related. I noticed that he's like, well, at least not yet. So I think he, he, yeah. Andy Jassy is still indicating that maybe he thinks they might have to reorg again once they start figuring out what these tools are good for. But he's being very cautious, of course, because he's talking to investors when he says that. And I appreciate that. I think there's a lot of other executives out there who are saying, we're going to have to lay off a bunch of people in the future because these AI tools will take over. ([Time 0:10:24](https://share.snipd.com/snip/73c456e8-e211-4601-b31e-749473529a33)) - **Sora Needs Rights Holders To Scale** - OpenAI monetizes Sora by selling extra video generations and seeking licensed IP deals. - Success depends on rights holders joining and whether Sora can reach viral cultural adoption. Transcript: Jenn Cutter Free users can generate 30 videos a day, but can now pay $4 to get 10 more. Paid subscribers have higher limits, but can buy more generations if they max out as well. OpenAI is also exploring deals with rights holders to let them sell the ability to include their intellectual property in video generations. Tom Merritt So this is where the rubber meets the road on Sora. The most optimistic think Sora could become its own kind of social media platform. Some even think it would replace TikTok. I don't think it's going to be that. I think it'll be a different kind of thing, just the way TikTok didn't replace Instagram. I think this is a good gamble for them to offer this to rights holders. I don't know if rights holders are going to bite on it. It will take one or two brave souls to risk the ire of their own employees and clients to do so. And, and this is the big and, also see a huge benefit from it, like a huge promotional uptick from it, then I think the floodgates with Open. There's a lot going on on the Sora platform that's really fun, that's really promotional, that makes people, you know, bond with brands the way that they would on other social media Platforms. And I can see that being a potential revenue stream for open AI. ([Time 0:19:30](https://share.snipd.com/snip/4f0aae62-a5a8-4a02-aea7-45004d441607)) - **Sora's Viral Moment Is The Big Hurdle** - Sora may become a distinct social platform but faces hurdles: safety, user base, and content rules. - A few high-profile adopters could break the hump and drive broader adoption. Transcript: Tom Merritt But if they can get a big enough name to do it, and I think they will, they'll get a big name or two to say like, you know what, let's try it. Let's see what happens. Whether they can actually get a benefit from it is the big question. I have been of the opinion that Sora is a novelty that won't sustain forever. And I'm having a lot of fun using it, but I still am skeptical of whether it will be a long-term product of its own. But this is the kind of thing that would have to happen to make it that, right? Is somebody going like, oh my gosh, this movie, let me remix trailers, et cetera, et cetera, et cetera. And it was so fun. And we saw all the memes and it just takes off. I don't know if Sora has the user base for that to pay off yet. ([Time 0:21:28](https://share.snipd.com/snip/34c53ac6-d1d7-492d-a9e1-33726c886731)) - **Keeping Untitled TextEdit For Notes** - Tom Merritt keeps multiple TextEdit documents open as simple scratch pads, showing preference for predictable, minimal tools. - Andy Beach and Tom describe returning to single-purpose editors over feature-rich cloud apps for simple tasks. Transcript: Jenn Cutter As AI makes our apps more complex and our screens more crowded, some users are retreating to the simplest software of all. Andy Beach explains why that matters. Tom Merritt Andy, do you know that I have a long running text edit document called Untitled 11? Andy Beach I didn't know that, but I inherently knew that about you. Tom Merritt I guess at some point after opening 10 instances of text edit, I kept the 11th version open on my desktop just as a scratch pad. I have right now six text edit documents open where I've just taken notes. So this story speaks directly to my heart. Tell me about this piece from Kyle Chaka. Andy Beach Yeah. So this is a New Yorker piece called The Quiet Rebellion of TextEdit. And it is one of those everything old is new again, almost in a rebellion to sort of the complexity and the AI-ness, the cloud-ness of all of the apps that we're feeling. There is this drive right now for really stripped down and simplified applications that do a very simple, simple task purpose, much like TechSedit was. Tom Merritt I understand where they're coming from, where Kyle is coming from with this. However, I'm definitely not doing it right because I have Gram on correcting things in my text edit. And there's some spellcheck and stuff that you can do in text edit. You can turn it off, of course. But I get where he's coming from, even though I do that, because one of the reasons I use text edit is because it's simple. Because I know it's not going to change a feature on me. It's been unchanged for decades, and it's always going to act predictably. Andy Beach Yeah, I think this goes back to another argument that we saw, particularly around sort of the recent redesigns we've seen in macOS and iOS around sort of the return and the reemergence Of skeuomorphic design, which we haven't seen in a while. And that's it, which that return of interfaces that sort of look like physical objects. And so I think there's a little bit of that combined with just the desire for there to be more minimal pieces. I do think there is a balance. So to your point, I think you can add in the elements you want now that get there, but you can do it as plugins. You're not relying on an app to necessarily have to integrate every piece as a feature for you. So you can find that balance of how much you want either AI or a new feature to interrupt, but you can still have something that is simple and straightforward and takes up very little processing Power to run and just sits there and does what you need it to do. Tom Merritt Yeah. Everything old is new again, right? And for Windows fans, Notepad is perfectly respectable. I use Notepad in the same way. I just don't write as much on my Windows machine as I do on my Mac, but it's the same kind of thing. Do you use a notepad or a text edit, Andy? Andy Beach So I have gotten slightly away from them over the years. And I would say I made an early jump into like the Evernote world that sort of got me into using something that I could use on any device, which then became OneNote. And I have slowly moved back out of that because I'm now back in a world where I have both Mac and PC much more predominant in my life. And I find myself using sometimes just the simple notepad inside of iOS because to your point, most of the writing I'm doing these days is, is on my Mac cause it's my portable, uh, versus My PC that is more of a desktop. So it tends to be more of like a production machine, not a writing and brainstorming machine. Yeah. Tom Merritt Well, and text edit, uh, now syncs through iCloud. So I can actually open untitled 11 on my phone. Should I, should I need to. But for that situation specifically, I actually use SimpleNote, which has been around forever. I used to use it with an app that the name escapes me that always saved to SimpleNote so that I could get it across machines. And then I just started using SimpleNote in the browser and stopped using the app. But it's that kind of need for just simple text editing. I'm glad to see is perhaps even making a resurgence here. Who knows? Andy Beach Yeah. And I think this is a positive indicator, even in the world of AI, that there's still always going to be room for those simple sort of single function apps out there in our lives. They don't all have to be Swiss army knives and that's totally okay. ([Time 0:27:40](https://share.snipd.com/snip/36ae1f68-9f28-4f1e-a2e8-26052ce20a49)) - **Simplicity Rebounds Against Complex Apps** - There's a growing appetite for minimal, single-purpose apps as a reaction to complex, AI-heavy software. - Andy Beach suggests plugins can add features while preserving a simple core experience. Transcript: Andy Beach Yeah. So this is a New Yorker piece called The Quiet Rebellion of TextEdit. And it is one of those everything old is new again, almost in a rebellion to sort of the complexity and the AI-ness, the cloud-ness of all of the apps that we're feeling. There is this drive right now for really stripped down and simplified applications that do a very simple, simple task purpose, much like TechSedit was. Tom Merritt I understand where they're coming from, where Kyle is coming from with this. However, I'm definitely not doing it right because I have Gram on correcting things in my text edit. And there's some spellcheck and stuff that you can do in text edit. You can turn it off, of course. But I get where he's coming from, even though I do that, because one of the reasons I use text edit is because it's simple. Because I know it's not going to change a feature on me. It's been unchanged for decades, and it's always going to act predictably. Andy Beach Yeah, I think this goes back to another argument that we saw, particularly around sort of the recent redesigns we've seen in macOS and iOS around sort of the return and the reemergence Of skeuomorphic design, which we haven't seen in a while. And that's it, which that return of interfaces that sort of look like physical objects. And so I think there's a little bit of that combined with just the desire for there to be more minimal pieces. I do think there is a balance. So to your point, I think you can add in the elements you want now that get there, but you can do it as plugins. You're not relying on an app to necessarily have to integrate every piece as a feature for you. ([Time 0:28:30](https://share.snipd.com/snip/0fdf2b3e-fd1a-4272-8e0d-e49ca2c8629f))