Hey, did you hear about those elections in the US this past week?
You can listen to the latest episode of The Kettle right here on this page, as well as on Spotify, Apple Music, or YouTube. Those platforms also let you subscribe to The Kettle, so you are always notified when the latest episode goes live.
The Democratic party has seized control of Congress and state governments around the country. It's easy to see this entirely as a rebuke of Trump and his policies, but blowback against AI, datacenters, and the tech industry played a role too, if the past week of chaos in markets and Silicon Valley is any indication.
Join host Brandon Vigliarolo, , to talk about what the winds of political change mean for the tech industry.
A lightly edited transcript is below:
Hey everyone, and welcome to the latest episode of The Register's Kettle Podcast. I'm your host, Reporter Brandon Vigliarolo, and hey, chances are you've heard about those elections we had in the US this week. The Democratic Party dealt a pretty hearty rebuke to Donald Trump, Republicans, and the MAGA movement, with Dems winning elections in states that haven't gone blue in decades. Now it's easy to see this as a blowback against Trump and Republican politics. There's a lot of indication that the tech industry was kind of on the ballot this year, too.
After all, they cozied up hard to Trump in his first two years in office in a bid to push AI and promote datacenter projects, and people have been none too happy about any of that. With me to discuss the elections and what they mean for Silicon Valley is Reg editor-in-chief Matt Rosoff and a special guest this week, senior reporter from The Verge, Elizabeth Lopatto. Thank you both for coming on, and welcome, Liz.
Thank you. It's lovely to be here.
I was so glad we could make this work. So let's start off by dissecting the degree to which AI datacenters and attitudes toward the tech industry in general sort of played a role in this election. I mean, were those major elements in the Democratic sweep or just part of the equation?
I think they were part of the equation. I think first and foremost people were pissed off about the corruption. And second, people were pissed off about ICE. Having people snatched from your community is the sort of thing that kind of sticks with you. So I think those were probably the two biggest issues. But the corruption is not totally separate from the tech industry, given the way that the tech industry has been cozying up with Donald Trump, the fact that Elon Musk and DOGE were cruising through the federal government, destroying stuff. So those things are related.
And then obviously there's the datacenter stuff, which is universally unpopular, not the only bipartisan issue in the US right now, but certainly a big one. So it's in there, I think.
I thought it was interesting that of something like 47 anti-datacenter initiatives on local ballots, 45 of them passed and only one failed. So that seems to me to be some indication that people at least don't like whatever these things are being shoved down their throats, which is funny because two years ago nobody knew what a datacenter was.
I think that there are sort of two things happening here, right? The first is that a lot of these datacenters have been popping up in people's communities without a robust public comment period. There have been a lot of NDAs being used and backroom deals. And I think most people don't like that stuff anyway. But add to that how frustrated I think a lot of people are with the tech industry right now. And suddenly this is the avatar of the tech industry doing a bunch of things that people don't like. And they can't get rid of Elon Musk. They can't pry Sundar Pichai out of his CEO role. There's not a lot you can really do, but you can stop the datacenter.
They can't get rid of AI results in search.
Yeah, unfortunately.
I think that's just an expression of that frustration, right?
Yeah, and I think the industry at least has to see it a little bit, right? Because I think it was maybe the day after the elections, OpenAI came out right away and was like, "Hey, you know what we're going to do? We're going to really try to be better civic citizens in our communities. We're going to stop signing NDAs." And then they're also trying to rebrand datacenters as "compute campuses" to make them more attractive, as these aren't just big warehouses where we're doing AI processing. We're trying to do all this other stuff there, too. It just seemed like there's very much an awareness that, "Oh wow, this may have been a bit of a blowback against us, or at the very least, something we're going to need to adapt to in a new political cycle."
I found it really interesting: I saw a study yesterday, maybe the day before, that said 41 percent of active datacenter projects are in states that Democrats took control of this election cycle, including Texas and my home state of Michigan, where Democrats didn't just capture federal offices, but state-level legislatures and offices as well. That could potentially mean that a lot of datacenter projects are going to be slowed or stalled completely in the next few years. Is there any indication that that could happen, you think?
Well, maybe. I think that the frustration with datacenters has really been growing over the last two years. And you've seen things like Project Jupiter in New Mexico where the locals didn't even know who was coming in, right? There was total secrecy. And then as soon as people figured out what was going on, they made the moves they could to start blocking access to power.
And so I think anybody who's politically ambitious picks up on that and says, "Okay, this is a way to really channel some populist frustration. I can probably ride this." So there's that part of it. But I also think these people got elected a couple days ago. And in some cases, I think this was a race in Iowa, the Democrat had dropped out because he was sick and got elected anyway. So there's a certain amount of just pure rage being expressed here.
I've heard a lot of people saying that it wasn't even necessarily a vote for Democrats in many cases as it was a vote against Republicans and their policies.
Yeah, so I think that using that as a way to consolidate power is a smart move, right? So you see even before election night's out, there are these statements like "lawyer up." And I imagine investigating DOGE is going to be super popular because a lot of people lost their jobs, right? Separately from all of the people who died because USAID stopped. I think that there is just a real political will for investigations, for some of this stuff becoming public. And specifically in Texas, you have these emails between Elon Musk and Greg Abbott that have been really heavily redacted. You want a fast and easy political win, you unredact those emails.
And I can't imagine that the new governor of Texas isn't going to — leave it to the Democrats to take something like this and completely squander it by sitting on their thumbs and trying to compromise. So part of me is like, "Wow, what a mandate you have. Let's see how badly you mess this up." I'm hoping that isn't the case. I'm hoping that we see some actual drive from Democrats to do things like unredact those emails. Give voters a reason to actually say you did something good, right? It's easy, like we were talking a couple of minutes ago again about datacenters kind of being a scapegoat for the tech industry right now. And Elon Musk is easily that as well, right? He's the person you can point to and say, "Look at this asshole," and all the things he's been doing, especially in Texas, right? I'm hoping that they do take that and run with it. He's been—
I think it's interesting. Texas didn't just elect a couple Democrats. It was a Democratic slate from top to bottom. And I think a lot of Trump voters showed up specifically to vote against Republicans this time. So I don't know if that's all—some of that is related to Paxton being on the ballot and the sort of known history of corruption that he's shown and the fact that the MAGA movement backed him so boldly. But some of it is, I think, Elon being so loud and so belligerent and charging in and saying we're just going to do whatever we want and get out of the way. And there has kind of been an arrogance among the tech industry and the tech executives for the past few years, but I think it's really become loud in the last couple of years. And I think to some degree, datacenters are a backlash against that and some of the ire directed towards Elon Musk is a backlash against that.
We all live in the same society, right? We've been hearing Sam Altman out here like, "Well, some people are going to die because of AI, and we're just going to have to accept that as a cost." And it's like, well, Sam, is that you? Are you planning to die for AI? Because it kind of sounds like you're talking about me and I don't like it.
Right. Yeah, it's the same sort of thing with wars. Some of you are going to have to die. Me and my kids, we're fine. Politicians are never going to send their sons and daughters to fight these wars.
I was recently in Washington and talking to some folks there who are involved with federal agencies and the difference between how Silicon Valley and tech companies and startups think of and use AI. I mean, it's $10,000 token budgets per developer and everybody's using agents to automate workflows and everything. And in Washington, it's like, this went way too far, way too fast. How do we stop it? And meanwhile, now AI is uncovering flaws and vulnerabilities in all of this software and all of these online services and web servers that we've been using for 25 years. So we've got to figure out how to patch all of that. We've got people entering confidential information into Chinese models like DeepSeek. So I think that the tech industry got way ahead of itself and the rest of the country is just starting to catch up. And yeah, maybe in five or ten years we will all be using agents and the infrastructure will have been built out to a point where the costs make sense and the ROI makes sense. But right now there's just this huge disconnect it seems between tech insiders in Silicon Valley and how people actually are using the technology today. So there may be some of that as well.
I think that there is a bubble, an intellectual bubble around Silicon Valley where all of these people are talking to each other and hanging out on X and assuring each other they're the smartest. And they're saying things like, "Oh, this is going to totally change your life." And the normal person who is using ChatGPT to try to replace their search engine is noticing that ChatGPT kind of hallucinates and it's not it's not that good. And there is this impulse on the part of these evangelists to be like, "Well, that's your problem because you don't know how to properly prompt the AI." And it's like, uh-uh, that's not how you get people to adopt technology. If you remember the internet, when the World Wide Web first happened, people didn't start using it until it made sense for them to use it. There were people who liked it because tech was their hobby and they thought it was a cool toy. But it wasn't until there was a reason for ordinary people to start using it that ordinary people started using it. Even online shopping, which I think is one of the very obvious use cases for the internet and was very early on, didn't catch up to in-person shopping for 20 years. So all of this hyperventilating around AI and how it's going to change the entire society and eliminate everybody's job and all of this other stuff, I feel like it's really misplaced. And it's not just misplaced because it underestimates the experience the rest of us are having with AI, which is frankly not that great. You're going to replace my job with this? Good luck.
Also there's just this sort of reality check. If you think about what AI is, it does certain kinds of intellectual work really well. Programming, for instance. Mathematics. But those are not things that require human relationships or resources outside of a computer. Almost all of the rest of us are involved in industries where those things are required. And so, how do you fit AI into that is a really different question. So there are blinders on in Silicon Valley and you can even see it — I live in Oakland and as soon as you leave San Francisco and go to Oakland, you start talking to normal people and they're just like, "I don't see the point of this AI. Why are we doing this?"
Yeah, I mean most people I talk to agree it's kind of useless to a degree, right? The way I think of it is it's almost like it's replaced Wikipedia for me, right? It's replaced — it hasn't even replaced search, it's replaced Wikipedia. It's the place I can go to get a basic overview of something to then do my own research because I can't trust the information that a Google search overview gives me, but at least it's going to maybe surface some resources that I didn't know were there, right?
It's good for summarizing certain kinds of documents. I do hear about it being used in the legal profession pretty extensively. So you don't have to read every line of every briefing. You have to check its work, because sometimes it makes stuff up or invents citations. I think it's gotten better at that.
You know, it's good for things like cleaning up text and it is good for programming. I hear a lot of programmers saying, "Hey, we were using this internal software or we were using a couple of our licenses of a particular enterprise piece of software to accomplish one very specific task and it was too expensive and we vibe coded a replacement for it over the weekend and it actually works. It's good enough. And we don't have to pay enterprise vendor X ten thousand dollars a month or something anymore."
Yeah, just make sure it's not exposed to the internet or next thing you know.
Right. It's not handling any sensitive data that you care about. Right, exactly.
Yeah, I mean I generally think of AI, at least in my experience, because I tend to not use it actually. I have moved to Kagi so that I can just get the old Google search circa 2012, primary source document type deal, because Google search sucks so much.
It really does.
I think of it as being kind of a shitty intern, right? I occasionally use AI to transcribe my interviews and it's good enough. I have to double check it. I've got to listen to the part that I want to quote to make sure that it got it right, because sometimes it really doesn't. But it beats doing the transcription myself. So I have this kind of shitty intern now that I have to keep an eye on. And that's fine because that can be time saving if you need to change a format in an Excel spreadsheet. That feels like it's about the level at which the shitty intern does okay. But anything more serious than that, and I'm like, "I don't know about this. I think I'd better do it if it's going to be done right."
But is it worth $1,000 a month or $10,000 a month?
Right.
With the election and at the same time the finances of Anthropic becoming clearer and clearer and these increasing commitments. They filed their S-1 at precisely the right time when they had this growth spike and they were profitable for one quarter and then you looked at Q3 and lo and behold, they're back in the red again. I just wonder if some of this infrastructure buildout is going to be stalled and how that sort of affects pricing and how that affects competitiveness versus open weights models and those kinds of things. I don't know, I know Liz, you've looked at some of the financing in the industry, particularly around the CoreWeaves of the world, the neo-clouds, what do you think's going on there?
I think that those are only going to live as long as Nvidia wants them to live. You look at CoreWeave and it just doesn't make sense as a business. It's paying too much for debt to make sense as a business. There's just no way to get out from under that debt. And so, as long as Nvidia is willing to continue bailing them out, because it is basically an Nvidia client state – Nvidia is the reason that they were able to have an IPO at all. So as long as Nvidia is willing to bail them out, sure, they're going to continue being a going concern. But if you look at the bond market, people are pretty nervous about those bonds and I think they're right to be nervous. We saw here today – today is Friday – we saw Microsoft pulling out of a deal with Oracle for datacenter capacity for compute. The excuse that they gave was that Oracle had old chips. And maybe that is part of the decision-making, I don't know. I don't live in Microsoft. But it's also possible that Microsoft, who is building its own datacenters, is like, "Well, I don't need to give money to a competitor. We'll figure out something smarter to do in the meantime. And then once our datacenters are built, we'll just use our own."
And that has been the case for all of those neo-clouds for a while now, that their clients are about to become their competitors as the buildout happens. So if the buildout slows down, then every existing datacenter becomes more valuable. And so maybe that gives CoreWeave a lease on life. But maybe not, because who knows what's going to happen with the interest rates.
And how do you feel about Oracle? Their borrowing is getting more and more expensive and it seems like people are sort of walking away. Could Oracle actually be in some kind of real financial trouble?
I think that Oracle has enough going on in its legacy businesses that I doubt bankruptcy is on the table, but I could be wrong. I do think that they have made a really big bet on AI generally and OpenAI specifically, and I'm not sure that OpenAI was a super smart bet. Again, depending on how the buildout goes and what comes online and what doesn't, it might turn out that Oracle can turn around and flip that compute to somebody else if OpenAI isn't good for their money. It might also be the case that Oracle can use that compute itself for its other businesses. I am a little less worried about Oracle than I am about the pure play neo-clouds that basically pivoted from bitcoin mining and have no other business.
We've been talking about that almost weekly, because I feel like so often the Kettle ends up being about AI, right? Because that's the big topic right now. And almost every week we mention the fact that Oracle, Google, Microsoft, these companies are going to be fine if all this AI financing and circular investing starts to shrivel up, right? Who's not going to be fine is OpenAI and Anthropic. And the neo-clouds, right? They're all making these huge bets on AI and they have no other business to fall back on. It's not like someone's going to try to fall back on bitcoin at this point, I don't think.
You never know.
Certainly not at this point. I mean, just looking at looking at what bitcoin did this week, I feel like it's not super valuable to be a bitcoin miner right now.
No, it no longer pays for itself, right?
Let's talk about Elon Musk. Is this sort of the beginning of the bursting of the Musk worship bubble? Are we sort of seeing him lose his influence and power a little bit?
God, I hope so.
I know Liz wrote an article on all of the things he could have spent whatever it was, $500 million on.
Yeah, I saw that. That was good.
Obviously he doesn't really know how to have a lot of fun, but the fact that he lost so completely everywhere that he bet, and the fact that Democrats are starting to whisper about indictments and investigations, is this sort of the end of his reign as the king of tech?
It's a good question. One of the things that's been interesting to me is that the Twitter acquisition really made the public turn on Elon Musk, right? That was when people were like, "Is he a genius? Because that doesn't seem like the move of a — why is he doing this?" And then of course, following that, the stuff with the Trump administration, people started protesting Tesla, like he kind of skunked his brand. So there is the popular view of Elon Musk, which is that this guy sucks. And then there is Silicon Valley, where for some reason he is still held in really high esteem. And I don't know why that is, except that maybe he's very good at raising money.
But that's real and that's still a thing. And so I am interested to see what happens here because if these sweetheart deals do start getting unearthed and we start seeing some of the private correspondence and like, "Oops, there might have been a little bit of bribery here and there," or "Oops, there might have been something else that was quasi-legal," those are things that are significant that become risks to his companies because they are so closely tied to him, right? For a long time, Tesla didn't have to buy advertising because Elon Musk's Twitter account was good enough. You know, and so you live with Elon Musk, you die with Elon Musk. That's just what it is. He can delete all the tweets he wants, but it doesn't change what we all saw over the last two years. And so I do think that there is a very real chance that he's got Texas where he's got an attorney general who entered basically with a mandate to investigate him. He has a senator who is like, "Hey, I'm proof that Texas can't be bought."
He's got a governor who isn't that fond of him either. And there's that. And then broadly across the country, again, this man is unpopular, so he's a great target for investigations.
Right.
I think we are looking at a real change of his fortunes. And whether that will stick or not is anybody's guess. Because one of the things that has certainly been true for the entire time that I have covered Elon Musk, which at this point is more than a decade, is that you discover which laws have teeth and which ones do not. Turns out copyright law really matters. Turns out the Chancery Court of Delaware counts. That one matters too.
The SEC not so much.
I don't know about the SEC. I don't know that that's real.
Or like when he was smoking weed on camera and then got a slap on the wrist. I guess our drug laws around contractors, if you're rich enough, they don't matter. So there's a real sense in which we're going to find out exactly which laws apply to Elon Musk and which ones do not.
I also think it was interesting his purchase of Twitter and the way he sort of tilted the platform or the way the platform got tilted, so it kind of became this center-right, right-wing, Silicon Valley echo chamber. And I think a lot of people were surprised by the election results on Tuesday. As a result of that, if you're only on Twitter and talking to your very rich friends in VC, you might not actually know what real people think.
Yeah, there was a whole story I read about Bluesky basically saying, "Hey, we were right the whole time. Twitter's become this echo chamber of right-wing thought that doesn't realize that the rest of the world is a little pissed off at the tech industry." It's been kind of—
I think there isn't what Twitter used to be, which was where everyone was, right? After Musk bought Twitter, I think the social networks really fragmented. So I'm not going to sit here and — I like Bluesky, I use Bluesky. I don't think it's necessarily representative of the entire country. I think it is representative of a lot of people who are left-wing. And I think it can be useful in that way, in the same way that the left-wing blogosphere was useful in the early 2000s. That's a great way to organize, it's a great way to strategize. Fantastic. But similarly, I'm watching with a lot of schadenfreude the right-wing infighting on Twitter right now. They did not see this coming. And to be honest, neither did I. I knew there was going to be a blue wave; I did not think that my home state of Iowa was going to elect a bunch of Democrats. I didn't see Texas coming.
Yeah, no. I mean I figured [] Talarico was maybe a toss-up, right? But he beat [] Paxton handily. You know, it was insane to see the degree to which people really blew back on Trump and, like we were saying, on the tech industry too, to a degree.
And if you think about the differences between Bluesky and X, there's another significant one, which is that X has an algorithm that is being tweaked by Elon Musk to show Elon Musk things he likes. And Bluesky, there are a bunch of algorithms that you can adopt, but I'm just using the same reverse chronology timeline that I have used forever on social media. And that again It's not going to show me just things I want to see. And so like again, like there are uses for algorithms, there are uses for filters, but you have to keep in mind that if you're being shown things you like, you're being shown things you like.
Echo chambers, for sure.
I think it's going to be interesting to see kind of what happens in the next — well, it's going to be at least a few months before things start happening, right? Because the new politicians who've been elected don't really take office until the beginning of the year. But once that happens, will the Democrats actually manage to do anything with this mandate they've been given? Will Elon Musk face some degree of accountability and will datacenter projects be stalled? Don't know.
Could happen, might not. Either way, we're gonna be here on the kettle to talk about it and we hope that you'll join us. Thanks for coming on, guys, and we'll talk soon. ®