From Meta’s massive offers to Anthropic’s massive valuation, does AI have a ceiling?

Meta is still going all-in on the AI talent war, with Mark Zuckerberg reportedly reaching out to top recruits himself, throwing around jaw-dropping compensation packages that top $1 billion over multiple years. And Meta’s latest target? Mira Murati's new startup, Thinking Machines Lab. It's a bold play in an already overheated market. While Zuck eyes new talent, Anthropic is preparing to raise a massive round of its own at a staggering $170 billion valuation, nearly tripling its worth in just months. On paper, it looks like the AI cash floodgates are wide open. But all this endless money raises some serious questions about sustainability. On today's episode of Equity, Kirsten Korosec, Anthony Ha, and Max Zeff unpack the reality behind these eye-popping figures. With compensation packages skyrocketing and funding rounds swelling, how long can this race actually last? Listen to the full episode to hear more about: Figma’s IPO, which is massively oversubscribed ahead of its NYSE debut Ramp’s rapid rise to a $22.5 billion valuation in just 45 days Why the Pentagon’s Golden Dome defense program may not be the big break startups are hoping for The escalating AI chip race, from Groq’s $600 million raise to Tesla’s $16.5 billion deal with Samsung, all while geopolitical tensions flare over chip exports to China Equity will be back for you next week, so don’t miss it! Equity is TechCrunch’s flagship podcast, produced by Theresa Loconsolo, and posts every Wednesday and Friday. Subscribe to us on Apple Podcasts, Overcast, Spotify and all the casts. You also can follow Equity on X and Threads, at @EquityPod. Learn more about your ad choices. Visit megaphone.fm/adchoices

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Anthony Ha?

  This episode of Equity is brought to you by Schneider Electric. Schneider Electric, your energy technology partner.

• 0:17Kirsten Korosec?

  Hello and welcome back to Equity, TechCrunch's flagship podcast about the business of startups. Today is Friday, August 1st. I'm Kirsten Korisak, Transportation Editor at TechCrunch, and I'm joined by TechCrunch Senior AI Reporter Max Zeph and Weekend Editor Anthony Ha. And Anthony, when you see the news, I'm sure you've seen it, of these compensation packages that are being thrown around in the AI space specifically by Meta, are you regretting your choice to become a journalist?

• 0:51Anthony Ha?

  Always. I'm constantly regretting that choice. I do think that it's worth remembering that for old folks like you and me, Kirsten, probably if we'd gone into a technical field when we were first starting out, that would have meant going into computer science, programming. And that actually seems to be an area where there's a lot of layoffs right now. So like the thing that would have seemed safe at the time is not necessarily what would have paid off in the long term. And I am curious about whether or not, obviously it does seem like there's a little bit of a gold rush mentality around spending so much money on this kind of AI talent. And I'm curious whether or not we're still going to be seeing offers like this five or 10 years from now. Max, what do you think?

• 1:32Speaker 02

  Yeah, people used to tell journalists to learn to code, and now I think they need to learn to research LLMs. It's just that's where the money is. That's the safe career. I really am wondering if colleges are just seeing a crazy influx of people who want to study this kind of stuff. But we are going to get into a lot with the AI race later. On the show today, we have deals of the week from Figma, Anthropic, and Ramp. And then we'll get into two big themes, why the Golden Dome may not be the golden ticket Silicon Valley is hoping for, and how the AI chip race is only getting hotter. Let's get into it with the first deal.

• 2:05Kirsten Korosec?

  Right, so this is Figma. We've been talking about Figma quite a bit. And as we're recording this right now, any moment, this IPO is going to debut. So. Some of these numbers may change, but the TLDR is this, is it's going to do much better than expected. This is a very clearly hot IPO. Shares are indicating that they are going to begin trading at around 105 to 110 each. That is more than 200% above what their IPO price was of $33. So this is going to be a big one. To me, it's really interesting

• 2:44Kirsten Korosec?

  because yes, the company has been incorporating AI tools more recently, but this is not your typical AI company IPO hype situation. This is a design and collaboration software company. They've been adding a ton of tools over the past couple of years, and this is just an IPO that seems really well timed and a company that's doing well.

• 3:07Speaker 02

  Yeah.

• 3:08Anthony Ha?

  And it seems like, I think when we talked about it recently, the numbers that Figma wanted to put out and what their goals were. And so it was like, hey, this could go really well, but we don't know yet what Wall Street investors actually think. And now we're seeing that in fact, those investors are even more excited about it than Figma was probably hoping.

• 3:24Kirsten Korosec?

  Yeah, to me, it's like a referendum on like, this is real tools that people use and are popular. And it's kind of nice to see, you know, I like it when companies that have real development tools that are being used are rewarded for that. And that, you know, it's a reminder that not everything is a meme stock and hypey and pre-revenue. There are companies out there that are doing quite well. And, you know, as a reminder, Max, I don't know if you have any thoughts on this, but I'm wondering if they're patting themselves a little bit on the back about getting over that failed merger, potential takeover acquisition by Adobe. Because that was like 20 billion, I think.

• 4:03Speaker 02

  Yeah, I think that that is what everyone's just comparing this IPO to is this almost acquisition for $20 billion. Figma, you know, if it does trade at the indicated range that we're seeing now, could be worth closer to $60 billion. I mean, just on day one. So I mean, this is clear that it was a good decision in hindsight to not take that acquisition. It's interesting because at the time- A lot of people, I think, were throwing their hands in the air upset that the Adobe deal didn't go through because regulatory concerns held it up. Adobe wasn't sure if they could clear the deal with European regulators. But, I mean, look at this, like, we have this, like, thriving company that honestly might become a serious Adobe competitor in the coming years.

• 4:46Speaker 02

  And it already is an Adobe competitor in some ways. So I think this is really fascinating. This is, yeah, like a great software design startup that will have to use AI in the future now.

• 4:57Kirsten Korosec?

  Yeah, I mean, as I mentioned before, they have been incorporating some of the AI tools recently. We'll see how that evolves. And they clearly are trying to appeal to developers as well. And as I mentioned before, they put out this dev mode in 2023. We've seen IPOs that have done quite well and then the business fundamentals aren't so attractive to shareholders later on. And you see the share price drop in the weeks, months and years in the future. I don't want to speculate on that, but clearly this is an indicator that for these types of companies, IPOs make a lot of sense. I don't think IPOs make a lot of sense for every company right now. It's not that solid.

• 5:38Kirsten Korosec?

  It is an indicator too, though, that while we are still seeing some SPACs, people are doing the IPO market. And they're also looking in the private markets for money. And you're seeing companies like Anthropic, which we're going to get to now, having no trouble raising money. And what I'm really curious about is when is it going to shift down to the smaller startups? Because in the case of Anthropic, which Max, I know you're going to talk about, they just raised how much? They're close to raising, close, three to $5 billion. And I'm wondering if other companies, aside from the big, big, big AI companies, are able to even touch this type of capital.

• 6:18Speaker 02

  Yeah, it's a good question. And I mean, just to end on the Figma story, really, we're probably going to see a lot of investors in Figma in the next week post on X and say how they believed in the startup very much right in the beginning and how they knew Dylan, the CEO, right when he was just a young buck. And I can't wait for all those posts. Anthropic is still in the stage of kind of raising those early rounds, although it's much larger than Figma ever raised. This round is three to $5 billion. It's at a $170 billion valuation. This is the kind of capital that AI labs need to raise right now. And it's really just indicative of just the massive capital demands of AI models.

• 7:03Speaker 02

  I will say that Anthropic, for all the money that it has to raise, it is growing a lot. Dario Emery went on a podcast this week, and he said that his company has been essentially 10 X-ing in revenue every year, which really does make Anthropic one of the fastest growing software startups ever. Anthropic is just a few years old. I think they were founded in 2021 or something like that. Like, it's really just a couple of years now. But this is the kind of capital that's reserved for a couple of the biggest AI startups like OpenAI and XAI. And they're even having to turn to Middle Eastern sovereign wealth funds to get some of this.

• 7:39Kirsten Korosec?

  I have a question for you. So Anthropic is doing really well in terms of growth. Where is it doing particularly well? And why, in your view? And Anthony, I don't know if you have an opinion on this, like, why are they so successful? Because there are other AI labs out there that are raising a lot of money, but I wouldn't call them like, successful on the customer or like, revenue front?

• 8:02Speaker 02

  So the venture firm Menlo VC just came out with this 2025 mid-year report that found that Anthropic's models were the dominant LLMs being used in enterprise settings, more than OpenAI, more than Google. Based on conversations I'm having with people, that sounds roughly right. Like, you honestly don't hear about people using OpenAI or XAI's models that much. It's really Anthropic that's killing it here. Big bet on selling their AI through an API to enterprises that I think a lot of people were skeptical of, but has so far panned out semi well for them. I mean, these companies still burn billions of dollars in cash.

• 8:43Speaker 02

  They're not profitable, but it does seem like they're gaining some traction here, at least in the enterprise.

• 8:49Anthony Ha?

  Is this something where these are still kind of like experimental budgets where it's like, let's try this out? To what extent is paying for Anthropic something that's actually paying off for its customers yet? Or is this something where like they could just stop paying for it in a year or two?

• 9:03Speaker 02

  I mean, they've only been paying for it for a year, like this business is pretty new. So I would not say that like we're still very much in the experimental phase of these tools. That said, I think that it's becoming a habit for people to use Anthropic's AI models, like especially for software engineers who want to code with Anthropic's AI. They have the best in class models. So if companies are committed to supercharging their workforces with AI, for lack of a better term, that's kind of like their corporate marketing phrase. It's just Anthropic really does seem to be the best of the best. Now, we've heard that OpenAI is going to release GPT-5 in, you know, the coming weeks, and that is said to change things, but it's supposed to be better at coding.

• 9:48Speaker 02

  We'll see how much that takes a hit on Anthropic's business. But there is just a lot of excitement around any startup that's using AI, and they seem to be raising these massive rounds.

• 9:59Anthony Ha?

  Max, you also mentioned that they were bringing on some money from these like Middle Eastern sovereign wealth funds. There was some reporting around possible participation from the Qatar Investment Authority. And it seems like that also speaks to how Anthropic in some ways tries to sort of paint itself as different from the other big AI companies, more ethical, more thoughtful. And at the same time, I mean, I think, you know, the CEO even like put out a memo where he basically said, I don't love that we're taking this money, but I think the subtext is sort of like, if we're going to be raising this much money, we're probably going to be taking money from some not 100% ideal investors.

• 10:36Speaker 02

  Yeah, this was clearly a point of tension in Anthropic to try to raise money from, yes, the Qatar Investment Authority. And I think they're also talking with some other investors in the Middle East, which they kind of have to to raise, you know, a billion dollar round at this stage. And it's really fascinating because you hear all of these AI companies kind of bloviate like they are, you know, holier than thou and that they have this grand mission. It's kind of like Google in its early days. And then companies just fall to corporate dynamics eventually, like we need to make money. And suddenly your values aren't that important when you need to raise billions of dollars and generate billions in revenue.

• 11:19Speaker 02

  That's kind of where we're at with them. I think we might see something similar play out with OpenAI as they start to, you know, need to put ads in their products. I think a lot of these very ethical values at AI companies will be tested.

• 11:32Kirsten Korosec?

  It's a slippery slope. I'm wondering if you have a sense of what the ceiling is because we opened up the show kind of making a joke about, and we were, by the way, for the audience, were referring to some of these compensation packages that were by Meta's attempt to poach people from like Thinking Labs and others. And the ranges of these compensation packages were between 100 and 500 million, and in one case, $1 billion over a period of years. And so turning now to Anthropic, when you start thinking about companies like Meta and the amount of money they're throwing around to poach employees, and then you look at what Anthropic is close to raising, three to $5 billion, that sounds like a ton of money. And then if you're like, well, if the talent battle is that brutal

• 12:17Kirsten Korosec?

  and the compensation packages are that high, like how far does this money go? And is there a ceiling to it? Like, there isn't infinite sources of money. At some point, all of these companies will run out. And I guess that's where they go and become publicly traded, right? And it's a whole new pool of capital. But I'm wondering if you have a sense of what the ceiling is and is Anthropic spending its money? differently than the metas of the world that clearly have a lot more to pull from. Are they even participating in this whole talent poaching war?

• 12:52Speaker 02

  Yeah, it's interesting. I referenced earlier an interview that Dario Amadei gave this week, and he actually said that they haven't materially increased compensation for employees in light of the poaching. They said that they think it's unfair to just treat employees a certain way because Mark Zuckerberg picked their name out of a hat. Anthropic has consistently done a lot with far less than OpenAI, than Meta, than Google. It's actually pretty impressive. You know, there's a lot of criticisms to be made about Anthropic, but they have been able to punch above their weight for a long time now. They continue to do that. I don't think that this capital is endless. I think these companies have to go public in the next few years because I just don't think that there's enough money in the private markets

• 13:35Speaker 02

  to sustain them for much longer. So we will see.

• 13:38Kirsten Korosec?

  Yeah, it reminds me, not every big massive raise comes from these AI labs. And Ramp, which is the expense management startup, is a good reminder of that. Anthony, this company just raised, what, 500 million with pretty eye popping valuation?

• 13:56Anthony Ha?

  That's right. So the valuation is $22.5 billion. I mean, that's already, those are very impressive numbers, but it's only been 45 days since I think they announced raising 200 million at a $16 billion valuation. So essentially, the valuation grew by more than $6 billion in a month and a half, which I don't have access to their internal numbers or anything like that. I think it's probably fair to say that's not necessarily a reflection of they had just an amazing 45 days and suddenly they're just making, you know, that much more money. I think it's a sign of just investor excitement around this company, around the sector. I mean, what was striking to me when I was reading the CEO's blog post about this

• 14:39Anthony Ha?

  funding was, again, even though we talk about Ramp as being, this is a distinct category, and it is, it's not Anthropic, but the number one word that just came up over and over again was AI. And it was about presenting this vision of what he's calling autonomous finance. And presumably, that is what the investors, including Iconic Growth, which is apparently likely to be the lead investor in the Anthropic round. And in my head, I'm just seeing sort of like dollar signs over their eyes and just be like, oh, did you say AI and finance? I'm in.

• 15:08Kirsten Korosec?

  Well, to me, what's interesting is that if you were to go back to, I want to say 2023, definitely 2024, and listen to an episode of Equity, there were a lot of fintechs that weren't doing well. I don't know if this is just a one-off outlier, or if the sector as a whole seems to be doing better. My gut, and based on also not just vibes, but actual information, I think a lot of the companies that were, in a way, either zombie companies, or just weren't able to get over like some of the economic humps, particularly around the fact that we didn't have access to free cash anymore with interest rates. Those kind of failed, and now, most of the ones that have survived are like pretty strong. And then now you're

• 15:53Kirsten Korosec?

  seeing greater amounts of money just going into fewer fintech startups. Basically, it seems like things have cycled through. I don't know if that's your sense of things, but it is interesting to me that there were so many fintech failures just a couple of years ago.

• 16:08Anthony Ha?

  Ramp does seem to have like a genuinely like very successful core business. It says it has 40,000 customers, $700 million in annualized revenue, and that was earlier this year. So I would imagine all those numbers are higher now. So I think it's true that it has a successful core business in a way that maybe some of those other fintechs did not. And also, I think a lot of these current investments are sort of highly speculative in the idea that this is just the beginning. Again, it's this blog post that's like full of all these charts. This is what the finance team of the future looks like. I think both are true. Is that like Ramp weathered a tough period for fintech, has done well, has built a real business. They've done a very good job now of convincing investors that they're not just a finance company, they're not just a fintech, they're an AI company.

• 16:49Kirsten Korosec?

  They've survived just long enough to then also jump on the AI hype wagon.

• 16:54Anthony Ha?

  Yeah, the next hype cycle.

• 16:55Kirsten Korosec?

  To like propel that growth even faster, even further. Yeah, so I'm really interested to see if we see that same pace. So are we gonna in another 45 to 60 to 90 days see another? My suspicions are no. And to your point earlier, I think that maybe some of that deal flow was already in the works, and then it was more of a timing thing as opposed to Something magical happened in the 45 days to get all these other VCs to suddenly sign checks. Is that kind of your take?

• 17:25Anthony Ha?

  I think so. I think it feels like there's just not that many people who can write checks that are much bigger than this. If the CEO of Ramp is getting on a plane to the Middle East, then maybe we'll get another round like this in 45 or 90 days. But otherwise, it seems unlikely.

• 17:39Kirsten Korosec?

  All right. Well, we have to stop for a break now, but when we come back, we're going to get into the Golden Dome. I have some thoughts about that one. And then we're going to get a little bit of an AI at Chips race update.

• 17:57Anthony Ha?

  What moves our world forward? Energy. History has been shaped by it, revolutions won with it, and progress dependent on it. For 190 years, Schneider Electric has been a leading force in this story, creating the technologies that harness energy and transform the competitive landscape. And today, in the dawning era of energy and industrial intelligence, Schneider is helping customers build more efficient, resilient, and sustainable operations across every sector, from data centers to homes, from buildings to industry. Schneider Electric, your energy technology partner. Discover more at se.com.

• 18:35Kirsten Korosec?

  One of our reporters, our space reporter, actually, Aria, wrote a story recently that truly caught my eye. And it's about building the Golden Dome, which is this next gen missile defense system, sort of a kin to Israel's Iron Dome. I'm wondering, Anthony and Max, if you've heard any startups that you've spoken to, any founders or VCs that have talked about the Golden Dome strategy? Because this is actually a thing that's happening in and around Silicon Valley and other areas.

• 19:07Speaker 02

  I've certainly heard about a lot of people planning their Golden Dome strategy, a kind of a way to buddy up with the Trump administration to, you know, sell these defense tech systems. But it sounds like Aria really went deep here. And here's, I'm curious what you found from her story and what investors were really seeing here.

• 19:28Kirsten Korosec?

  Yeah. So not to explode anyone's dreams out there.

• 19:33Kirsten Korosec?

  But really, what this is about is that there is absolutely, in certain sectors, if you're talking to VCs, or if you're talking to startup founders, everyone has a golden dome strategy. And you see this a lot with dual-use companies, in particular, defense space, things like that. The reality is that even if they have the tech that might actually be quite innovative for the golden dome, for this next-gen missile defense system, they're probably not going to get access to this contract. And so the contract we're talking about, it's a multi-year, it's about a decade long. Think of it as an umbrella vehicle, or an umbrella contract. It's $151 billion. You have to be able to qualify to access that umbrella, or that

• 20:16Kirsten Korosec?

  pool of money. And from there, you compete for specific contracts. The legacy companies, the big five defense manufacturers out there, so L3 Harris, Lockheed Martin, Raytheon, these types of companies are sort of like, a poor analogy, but sort of the, in the David Goliath fight, they're the Goliath. The sense in the startup world was that all these startups were going to have this chance to hit this pool of money. And there will be some startups that do that, but they will be in the very mature SpaceX, Palantir, Anduril category. The smaller startups, they're simply not going to be able to access this, specifically

• 20:57Kirsten Korosec?

  because it's very difficult to comply with the government contracts. So, security compliance, filing paperwork, these things that are actually quite difficult hoops to jump through, that a small startup with amazing tech just doesn't have either the cash or people, resources, to comply with. But it's not all lost. Do you know what the solution is? What's the solution?

• 21:24Kirsten Korosec?

  Tell me you didn't read the story without telling me you didn't read the story.

• 21:26Anthony Ha?

  Actually, the solution is subcontracting, but I wanted to set you up.

• 21:30Kirsten Korosec?

  Okay, yes, you win a prize.

• 21:32Anthony Ha?

  But now you're trying to call me out, geez.

• 21:35Kirsten Korosec?

  You win a prize, you win a prize. Yeah, so subcontractors. Now, what I'm curious about is, does that suddenly change the strategy of how startups and the VCs that might be advising and backing them, how they access this pool of cash?

• 21:50Speaker 02

  I think this is really fascinating. We were talking last week on the show about Trump's AI action plan, and I think that this was kind of a theme there, too, where the headline of Trump's plans for the tech space are, we're going to increase business for you guys, we're going to work with you guys, we're going to help startups out, we're going to boost up this ecosystem. And then the reality seems to be that the plans are actually the most beneficial for the space's biggest players, for the things that investors have put the most money in. This is a really good bet for the people at the top of the ecosystem. But it's a little bit more unclear for how smaller startups really win here. I'm really

• 22:30Speaker 02

  fascinated about this. Kirsten, do you think that subcontracting is a legitimate way for startups to grow and get access to this? Or do you feel like this is just kind of a farce?

• 22:42Kirsten Korosec?

  No, I mean, I think subcontracting is a normal state of business in multiple industries, and this is especially with large government contracts. In a way, it could be a blessing for a lot of these startups, because as long as they find a good partner, and that's going to be, to me, the important strategy, which is finding the right partner, really winning over one of the big five, or the mature startups like a SpaceX or Palantir, to bring them on as a subcontractor, that, to me, is going to be the crux. But it does eliminate some of the sort of arduous bureaucracy tasks that they just don't have the ability to complete. And I'm not saying that, and the story doesn't say this either, that no, there won't be exceptions to the rule. But generally speaking, based on the people that

• 23:27Kirsten Korosec?

  Ariya spoke to on the record, and, you know, just folks in the industry, who weren't quoted in the story, was the sense that it really is quite a high bar. And so, most startups just aren't going to see that magical windfall, and that they're going to have to be strategic about it, and really partner up. With either the big legacy or primes, as they call them, or you know, these up and coming mature startups that have a lot of money.

• 23:52Anthony Ha?

  Something else I'm curious about, and this is even more speculative, but obviously we don't know who's going to be in charge of Washington DC in two years, in four years. I mean, probably who's going to be president. But you know, that there are always changes in administration. In the same way, Trump's AI plan is in some ways sort of a repudiation of the Biden AI plan. I could imagine that a Democratic administration coming in and not necessarily want to continue this. So I'm curious to what extent, when you put deregulation on the AI side or major funding into the defense side, is that something that a bipartisan consensus can get behind? That like this will continue regardless of who's in charge? Or is something where just all the funding could get pulled in four years from now?

• 24:29Kirsten Korosec?

  Yeah, I think that's really good question because the golden dome is a very specific Trump administration idea. Now, could it just be called something different and essentially achieving the same thing? And all the fundings exist? That's politics, right? And I think that could absolutely play out. But it seems early days to be talking about the next administration, but it's not because these projects are so long. There's a real long tail and it really will come down to what the priorities of the DoD are and Space Force and others. Is it about missile defense? Or is it about something else more exploratory, like space exploration? What are the defense priorities? If it can continue to another administration,

• 25:14Kirsten Korosec?

  regardless of whether it's GOP or Democratic, I think that that obviously would be the best case scenario for these startups. But they should be developing their tech certainly for other use cases and probably not like just making this one bet. But for now, there's obviously some extreme interest in this. And there's another thing that has extreme interest in this world, which is AI chips.

• 25:38Kirsten Korosec?

  And this is your area, Max, for sure. But it's been really interesting to me to see how many companies with the AI chips race. And there's just so much money going behind this.

• 25:48Speaker 02

  Yeah, we've seen a lot of interest in developing more AI chips from startups. There was even more interest in this a few years ago, but it's since proven really difficult. And I'll explain why. But the news this week is that Groq, which is a startup building AI chips, take on Nvidia. They're raising $600 million at a $6 billion valuation. That's more than double its valuation since last year. And the really interesting thing about Groq is that their chips claim to be really fast. They say that their chips are faster than Nvidia, so they're really good for inference. But the problem is that Nvidia just has such a strong moat here. Their chips are best in class. And a big part of it is because of not just the chips, but the software around Nvidia's chips,

• 26:33Speaker 02

  which is CUDA. And I think they have a really strong ecosystem to support people using their chips to train AI models, not just the hardware itself. And that is a really hard thing for Groq to build up. I think that a large part of this round is probably going to not only building out more chips, but also building up the software ecosystem around it. There was reports that Groq was struggling also, like at the same time as they're raising this big round. They told investors this year they would make $2 billion and they ended up making less than half of that. And they said that maybe in 2026, we'll make that much. I think that this is a really tough space to be in right now, competing with Nvidia, Google.

• 27:15Speaker 02

  And at the same time, there's a lot of opportunity because companies like Meta and Microsoft, as we saw this week in their earnings, they just can't get enough of these AI chips. It's unsatiable.

• 27:26Anthony Ha?

  One thing I would add for listeners maybe who aren't familiar with Groq is that they may be like, wait, this sounds like a different Groq than what I'm used to hearing about. So this is Groq spelled G R O Q. And as far as I know, not affiliated with X or XAI in any way.

• 27:41Speaker 02

  Not at all. They are both based on the same nerdy sci fi concept of this reference from like, I think it's a

• 27:48Anthony Ha?

  Stranger in a strange land.

• 27:49Speaker 02

  Yes, it's like to try to understand something, to grok something, but spelled differently, just a lot of people in tech really love sci fi, I guess, is the message here. But yeah, I wanted to just talk for a second about Meta and Microsoft's earnings that we saw this week. Microsoft said that Azure, its cloud platform, which largely supplies AI chips for training, brought in $75 billion. They've never reported Azure's revenue before, but it's doing a lot. Meta, on the other hand, they increased their capital expenditures. The low end is now $66 billion for this year that they're going to spend largely on AI infrastructure. Meta spent like $30 billion on capital expenditures last year.

• 28:33Speaker 02

  I mean, they are just literally spending twice as much in this extreme, just outrageous amount of money on these AI chips. So you can see that even if Grok can grab a slice of that, they are going to be a multi-billion dollar business, and I guess they already are.

• 28:50Kirsten Korosec?

  I did have a question about the AI infrastructure that you mentioned that they're spending it on. And when you say that, just for the audience, are you talking about data centers? Are you talking about buying literal chips? Like, how is this capex money being dispersed under that umbrella of AI infrastructure?

• 29:06Speaker 02

  Yeah, so for Meta, it's definitely they're buying chips from Nvidia. They're working with data center operators to build out these massive data centers. Meta said that they're trying to bring a five gigawatt data center online. And for people who don't know what, you know, that means, one gigawatt can power about 800,000 homes. So five of them is a lot of energy that that's going to suck up from the grid just to basically, I don't even know, power Meta's AI models. They're like, you know, influencer chatbots or whatever. So this is a lot of hardware and building that they're doing. For Azure, it's a little bit different because they make their own chips, but they also buy a ton from Nvidia. That's still where they get most of them from.

• 29:50Kirsten Korosec?

  I have one last question for you before we wrap. And we've talked a lot about Nvidia being sort of the gold standard and the clear winner. But is there like a chance for any company to really truly compete with them? Because the demand seems to be there.

• 30:03Speaker 02

  I think there is a lot of people betting that AI startups that can make faster chips, that can make more efficient chips can compete here. And as AI inference, people using AI chatbots and AI agents becomes more widespread, there will be a lot of interest in making these chips as fast as possible, which just hasn't been Nvidia's focus historically. That's where Grok comes in. That's where Cerebrus and a lot of these other chip startups kind of play here. And then I think you can't rule out Google, who is working on TPUs, and Amazon, which has their Tranium chips, and Microsoft has their own chips, too. They're all trying to build out their own infrastructure here. I think that this is something over the next 10 years, really, we're probably going

• 30:47Speaker 02

  to see a lot of battling here, just because it's just such a massive market.

• 30:51Kirsten Korosec?

  And so much more to talk about, but not today. Equity will be back next week. Until then, you can find Equity under the handle at EquityPod on X and Brite.

• 31:03Rebecca Bellan?

  Equity is produced by Teresa Locan Solo with editing by Cal. We'd also like to thank TechCrunch's audience development team. Thank you so much for listening and we'll talk to you next time.