Anthropic to Spend $50 Billion Building US Data Centers

12 Nov 2025 · 43 min

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Podcast Summary: Bloomberg Tech - Anthropic to Spend $50 Billion Building US Data Centers

Episode Overview

  • Hosts: Caroline Hyde and Ed Ludlow
  • Focus: Discussion on the data center boom, recent announcements from Anthropic and Meta, and market impacts from companies like AMD and Whoop.

Key Topics Discussed

  1. Anthropic's $50 Billion Investment
  2. Anthropic, an AI startup, announced plans to invest $50 billion in building custom data centers in the U.S., specifically in Texas and New York.
  3. Compared to OpenAI's $1.4 trillion commitment, Anthropic's investment is seen as a more measured approach.
  4. Anthropic's recent success includes raising $13 billion and having a valuation of $183 billion with a strong enterprise customer base (300,000 businesses).
  5. Discussion around their partnership with FluidStack, a less well-known company, to assist in the data center build-out.
  1. Meta's Data Center Plans
  2. Meta has announced a $1 billion data center in Wisconsin, part of its broader strategy to build infrastructure for AI development.
  3. This investment signifies a shift towards controlling their infrastructure, avoiding reliance on third-party providers, and ensuring a competitive edge in AI technologies.
  1. AMD's Market Performance
  2. AMD shares surged nearly 10% following their investor day where they projected a 35% annualized revenue growth over the next three to five years and an 80% compound annual growth rate (CAGR) in the data center sector.
  3. Discussion around whether AMD can close the competitive gap with NVIDIA in the AI accelerator market.
  1. Whoop's Potential IPO
  2. Whoop, a fitness tracker company, is considering an IPO within the next two years.
  3. CEO Will Ahmed discussed the company's evolution and product roadmap, including plans for continuous glucose monitoring technology.
  1. Market Context and Investor Sentiment
  2. The Nasdaq was under pressure, with ongoing concerns about government shutdown impacts and the viability of AI investments.
  3. Discussion with Nicholas Janvier from Columbia Threadneedle on the cautious optimism in tech stocks, the importance of understanding capital expenditures, and the need for tangible returns from investments in AI.
  1. Technological Innovations and Future Outlook
  2. Anticipation surrounding NVIDIA's upcoming earnings report and how their future outlook will impact market sentiment.
  3. Exploration of the AI sector's growth and investor expectations regarding returns on investments.
  1. Defense Technology Funding
  2. Forterra, a defense tech startup, raised $238 million to enhance its autonomous vehicle capabilities for military applications.
  1. General Market Trends and Challenges
  2. Investors expressed skepticism regarding the necessity for the massive build-out of data centers.
  3. Concerns over energy supply, power requirements, and the sustainability of AI growth based on current models.

Key Takeaways

  • Data Center Boom: Significant investments from tech companies highlight a booming demand for AI infrastructure.
  • Strategic Shifts: Companies like Anthropic and Meta are focusing on building their own infrastructure to maintain competitive advantages.
  • Market Performance: AMD's positive outlook reflects strong demand for data center products amid competitive pressures.
  • Skepticism in AI Investment: Investors are questioning the sustainability of current AI growth strategies and the necessity of extensive infrastructure investments.
  • Future of Technology: Innovations in health tech and defense highlight how technological advancements are being pursued across various sectors.

---

This summary encapsulates the major discussions and insights shared during the Bloomberg Tech episode, providing a comprehensive overview of the current technology landscape and market dynamics.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00I'm Barry Ritholtz, inviting you to join me for the Masters in Business Podcast. Every week, we bring you fascinating conversations with the people who shape markets, investing, and business. CEOs, fund managers, billionaires, Nobel laureates, traders, analysts, economists, everybody that affects what's going on in the market. Whether you own stocks, bonds, real estate, commodities, crypto, you really need to hear these conversations. Sometimes it's behaviorists like Dick Thaler or Bob Schiller. Sometimes it's fund managers like Peter Lynch, Bill Miller, Ray Dalio. Sometimes it's authors, Michael Lewis, author of The Big Short and Moneyball.

0:45Regardless of the conversation, these are the folks that move markets each week. That's the Masters in Business podcast with me, Barry Ritholtz. Listen on Apple, Spotify or wherever you get your podcasts.

1:03Bloomberg Audio Studios, podcasts, radio, news.

1:12Bloomberg Tech is live from coast to coast with Caroline Hyde in New York and Ed Ludlow in San Francisco. This is Bloomberg Tech. Coming up, Anthropic plans to spend$50 billion to build AI data centers in the U.S. Plus, data center demand helps AMD shares jump as it predicts accelerating sales growth over the next five years. And Whoop is considering a potential IPO in the next two years. According to its CEO, we'll have the details. OK, let's get on to our top story. Anthropic plans to spend$50 billion to build custom data centers for AI work in several U.S. locations, including Texas and New York.

1:56This is the latest expensive pledge for infrastructure to support the AI boom. Bloomberg's AI editor, Seth Figgerman, joins us for more. There are some details worth understanding and knowing about in what Anthropic plans to do with this$50 billion over which timeline and with who. But there's also some consideration around how big this$50 billion is relative to the rest of the field. That's right. You know, I think in one sense, if you had said a year or two ago that a four-year-old unprofitable startup was going to commit$50 billion to an infrastructure project, you said that's enormous and possibly insane.

2:32But when you compare it to the$1.4 trillion that OpenAI has committed, it feels much more measured. And I do think that is very much the subtext of a lot of this. Anthropic is clearly taking a page from OpenAI in terms of the scale of infrastructure required to build what it wants to build. But I think it also wants to show that it's perhaps a more fiscally responsible company and doing this in order of magnitude less at the moment. They've raised$13 billion recently. They're worth$183 billion. And they've had a really good track record with business and enterprise sales, 300 ,000 business customers.

3:05But what's interesting is the business they're doing with other people. They're using a UK company to build out in the US, Seth. That's right. I mean, FluidStack is really not a household name here. I mean, it's part of that emerging sector of neoclouds, but not nearly as well known as CoreWeave or even Nebius. But, you know, it's an up and coming company that we've previously reported in talks for a large amount of funding. And it's been central to a French effort to build a supercomputer over there. But again, more untested, perhaps, than some other partners you might announce here. It's one of the stories where it's what we do not know that's worth consideration.

3:39So they're not telling us, you know, the total gigawatt figure for the footprint in the United States. And at the same time, they're heavily reliant on these hyperscale partners, Amazon, AWS, and Google, to give them the compute they need away from this plan to build$50 billion capacity. Yeah, and I think, again, if you look at their nearest rival, OpenAI, the answer is sort of it takes everything and everyone to build what they want to build here. And so while they want to potentially have a bit more control and ownership over some of the data centers that they build in the coming years, they're still going to be relying on those big tech partners and the neoclouds and pretty much anyone else in this industry to meet what they imagine to be a pretty intense amount of computing needs here.

4:21I'm going to make Seth Vigerman. It's a great breakdown. Thank you. Look, those intense compute needs, those AI bubble hopes, they're still being questioned by the market, at least on the day. Let's check in on the Nasdaq. It's under pressure by three tenths percent. Even though we've got some significant movers to the upside of the chip sector, Ed, you're going to get to it. But at the moment, we're still trying to digest potential end to the government shutdown. We're still trying to digest really where some of the AI, return on AI is eventually coming from, Ed. There are some outperformers and AMD is probably the starkest example, up almost 10 percent, more than 9 percent.

4:54They had their basic annual capital markets day or investor day. And what they said is that top line revenue growth is going to be annualized 35 % over the next three to five years. But in that key data center category, 80 % CAGR over the same time period. And the question we've asked so many times in this program is, is AMD meaningfully going to close the gap on NVIDIA for AI accelerators or GPUs for data center? The market really likes what it heard. Global Foundry has actually had a really strong earnings beat in the third quarter, particularly on the bottom line. Again, another sort of smaller scale contract manufacturer of chips like TSMC, but the stock down 3%.

5:32All in all, at the index level, the Philadelphia Semiconductor Index, or SOX, up 1.2%. But there is like this weird kind of treading water right now, bigger picture because of the government shutdown. And AMD, as you know, Cara, is probably doing a little bit to boost at the index level. It is. Let's just get the context here for all of this and drill down into what has been some cautious optimism in tech stocks. In an end, of course, to the U.S. government shutdown potentially in sight. Nicolas Janvier is with us. He is the head of North American equities for Columbia Threadneedle. It's got a cool$714 billion in assets under management.

6:04Nick, it's wonderful to have you here in the studio. Thank you for having me, Caroline. And you've got this great take on sort of what the drivers are going forward. You call it tire, but we are still trying to digest trade. We are still trying to think about interest rates. I think you go into what's happening with inflation, but earnings. What are we hearing from earnings to vindicate the AI run-up that we've had of late? Hey, look, thus far, right, I think it's fair to say that earnings have been really, really good. If you look at the performance of the hyperscalers, in particular, this most recent earnings season, you are seeing a growth and, dare I say, return on some of this capex.

6:40All you have to do is look at what's happening with Meta, right? When you look under the surface, you are seeing that the AI is enabling better ad effectiveness. And those are some of the early things that we want to start to see as investors to validate the spending. What's been interesting is investors didn't give much of a longer reign to Meta. It was beaten up in this earnings season because they're taking a lot more debt. They're thinking and committing to that capital expenditure. What do you want to hear from Zuckerberg? Do you want to hear that optimism or do you want to sometimes hear that he's going to be sensible with the money spending too?

7:12Hey, look, investors are doing exactly what we should be doing right now. We're asking questions. And why are we asking questions? Because, Caroline, even in a world where$100 billion is quaint, right, we're talking about really big numbers here. So investors are rightfully saying, hey, help me understand the pace of the spending. Help me understand how do I think about these returns flowing through? What's the time frame on that? And Mark, do you really need to spend all of this money right now or should you take your time? Now, Zuckerberg would say we absolutely have to move forward because this is going to matter for the future of the company.

7:54And if we don't do this spending, we're going to fall behind. But investors are going to ask questions. And that's precisely what we should be doing. Nick, it's great to have you on the program. Let me give you some really big numbers. $4.7 trillion market cap for NVIDIA. Top line growth overall, 56%, 57%. Data center, maybe even better. The E in your acronym, it cannot get any bigger than next Wednesday. How should the market brace for that NVIDIA print, either to the upside or to the downside in terms of disappointment? Look, I know that there is a focus, and we like to look at these earnings sort of on a real-time basis.

8:38But what the market is really going to care about, Ed, is how is NVIDIA talking about the future, right? Caroline was just pointing out that we've had lots and lots of CapEx commitments. And what investors want to hear is what is the pace of that CapEx coming through their revenues? And look, I know we're going to focus on a backwards-looking view in terms of the quarter. But what we should be thinking about is, is the pace of their competitive positioning, are they going to be able to maintain that? Are they going to be able to maintain this revenue trajectory that they're on? And at least in the short term, I think they can.

9:19You're the head of North American equities, right? You probably take a much broader macro view. But when you do have a company like this, who's not just technical weighting at the index level is an important consideration, but like the psychological weight that Jensen Wang puts on the market, how do you prepare for that? How do you translate that for your clients? I think what you have as an investor is focused on the things that you can do best, right? And what is our competitive advantage, if you will, from the perspective of Columbia Threat Deal? What we can do is study the market, understand the competitive dynamics, and be able to wade through who the winners and the losers are going to be.

10:02Now, is Jensen's commentary, is that going to matter for investor expectations? expectations it will but to the extent it creates noise I think what we have to do is be ready to take advantage of that on behalf of our clients and so in this moment if there is weakness and we have seen some pockets of weakness look quantum's come off we've had some of the more sort of risk prone maybe the unprofitable side of the tech's been pulling back a little bit nuclear maybe we've got less excited about that are they buying opportunities or are you thinking that you actually want a little bit of a healthy correction as some of these CEOs of financial companies have been talking about?

10:40Hey, look, Caroline, I would segment out the market. I think there are places where there is fundamental strength. And to the extent you see weakness, it's probably a buying opportunity. And I would separate out the places of the market where really what we're talking about is speculative appetite, right? The quantum stocks have melted up thousands of percents. But for most of these companies, they don't even have any revenues for a few years running. So I would not necessarily describe that segment of the market as a buying opportunity. But for the healthier places, the places where you can model out real revenue growth, you can see where the cash flows are going.

11:21There's probably going to be opportunity as the market continues to be weak. Opportunity. And I mean, the opportunity we keep hearing about, the limiting factor we keep hearing about is utilities and is energy. But also the key limiting factor for many is going to be the pace of interest rates and where they go. Yes. Yes. Let's go to interest rates. So in your acronym, we've done a lot on the E, right? What we always say on this program is that higher rates discount the present value of future cash flows. And we try and help the audience understand, like, why do we track the Fed on Bloomberg Tech?

11:53As an investor who's looking at this technology sector, why do you look at the Fed? Do you care? Help us understand. Hey, look, we absolutely care, right? And I'll take you back again to these really big numbers, trillions of dollars that we're talking about. So the Fed matters a lot from the perspective of, is there going to be liquidity in the marketplace? And obviously the Fed, in terms of the short-term interest rates is going to dictate the pricing on that capital. So yes, the Fed matters. I would point to the audience that there's been somewhat of a pivot. Four or five weeks ago, I think it was relatively consensus that interest rates were heading lower.

12:36We're now in somewhat of a data fog, but the little bit of data that we've seen suggests that we might be on hold. So in the short term, I think you're going to see some countervailing winds, if you will. In the long term, you would expect interest rates to start to come down, given what we're seeing in terms of economic performance. Let's end the conversation by you telling us what's going to happen in the balance of 2025 and 2026 for the technology sector. Your crystal ball, please. Look, I think for the next few weeks, it's going to be all about investors trying to figure out where are the returns going to come from.

13:14Let's look through the tea leaves. Let's see the early evidence that this CapEx is going to pay off. So I would expect that that's going to drive a lot of the conversation for the balance of the year. But I'll pull you forward into 2026. I think an important conversation and one that we haven't spent a lot of time on thus far is, are these CapEx numbers one time in nature or are we about to get on a capex treadmill which is akin to telecoms of the 1980s if it's the former i think the market will do well through that if people start to believe the latter i think you you have the real possibility of seeing some nervousness from investors nicholas chambier of colombia thread needle great to have you on bloomberg tech thank you so much Coming up, fitness wearable company Whoop says it's looking to potentially IPO within the next two years.

14:13We're going to get that story next. Cara, what you got? In the here and now, Ed, we're checking out airline stocks because we have heard from Sean Duffy. Here's the transportation secretary speaking to reporters saying the U.S. aims to start lifting flight cuts in a week once the government opens. We're up more than 2 % across the board. This is Bloomberg Tech.

14:38Hello, I'm Stephen Carroll. I'm in Brussels, where many of Europe's biggest decisions get made. And I'm Caroline Hepker in London. We're the hosts of the Bloomberg Daybreak Europe podcast. We're up early every weekday, keeping an eye on what's happening across Europe and around the world. We do it early so the news is fresh, not recycled, and so you know what actually matters as the day gets going. From Brussels, I'm following the politics, policy and the people shaping the European Union right now. And from London, I'm looking at what all that means for markets, money and the wider economy. We've got reporters across Europe and around the globe feeding in as stories break.

15:15So whether it's geopolitics, energy, tech or markets, you're hearing it while it happens. It's smart, calm and to the point. And it fits into your morning. You can find new episodes of the Bloomberg Daybreak Europe podcast by 7am in Dublin or 8am in Brussels, Berlin and Paris. on Apple, Spotify, YouTube, or wherever you get your podcasts.

15:43Waymo has begun offering rides that include highways in San Francisco, Los Angeles, and Phoenix. The company is also widening its service area in the Bay Area in the coming weeks. It's all offering rides across Silicon Valley more broadly. Bluebird's consumer app reporter, Natalie Lung, joins us now. And why is rides on highways such a big step? It is a big deal because it can cut down trip time for a lot of customers. Right now, before freeway access was available, Waymo had to plan routes, avoiding the highways and navigating the narrow city streets and lengthening trip times a lot. Natalie, we were talking with the team this morning about is this the first time that a genuine robo-taxi service, no safety driver, is going on a freeway or highway and charging a fare?

16:32The answer is probably yes. Just explain that footprint in those markets, SFLA and Phoenix area, against what others have or have not yet done. So opening the freeway access would connect some of Waymo's existing markets like SF down south, down all the way to San Jose and even to the airport. And so this would make them very competitive against traditional taxis and Ubers and Lyfts and make it, you know, not different from regular ride hailing. And compared with Tesla, which currently still have safety monitors in the cars, whether it's within city trips or freeway trips. So Waymo is really the first one that's offering like a fully driverless experience for customers.

17:17So I'm flying into Phoenix. I've got a nice little Airbnb that I've booked and I'm taking my Waymo to my Airbnb. And now I'm going to be able to get instacart ordered there as well you've been a busy reporter what's happening over at airbnb with instacart yes so um airbnb is piloting this kitchen stocking service with instacart where hosts can receive your instacart order and then pre-stock um your airbnb and so you can make use of that kitchen while you're in a short-term rental and so this is a three-month pilot that will start next year um and airbnb is incentivizing hosts to provide this service by giving them some cash incentives.

17:51Bloomberg's Natalie Lung. Thank you very much. Let's get to another story. Fitness wearables company Whoop says they're considering a potential IPO in the next two years. That's according to CEO Will Ahmed, who spoke with Bloomberg's Samantha Kelly. Sam joins us now. We've been tracking Whoop closely with you in particular on this program. It's not, how should I put this, unusual for a CEO of a closely followed startup to say, we're thinking about an IPO or a listing in the not too distant future. But give us some of the background context of why that conversation came up. Sure. So this is something that he's kind of hinted at before, but this is the first time he's really outlined perhaps in the next two years.

18:35He pointed to the fact that the company is pretty much a standalone business at this point, has hardware, software, has accessories, apparel, its own proprietary technology. and he said that it's really gotten to a point where they can move forward. And, you know, he raised the question, if you think about a big company that really owns the idea of personalized health at the moment in the public space, what comes to mind? And he said Whoop is the company that he wants people to think of when the time comes. I do feel like the health and gadgets in health offering has just got ever more competitive, though, throughout Will Ahmed's journey.

19:14Talk us through what the product roadmap looks like for that couple of years timeframe. Absolutely. Yeah. To your point, a lot of companies now are really stepping it up, doing a lot of the same type of features and type of results and data. But to your point, as far as product roadmap goes, one of the areas that a lot of companies have yet to master is the idea of continuous glucose monitoring. A lot of these devices, they monitor your glucose levels. A lot of people with diabetes, this is something that they need to do. And the technology behind it is quite clunky. It's very invasive. It usually involves needles and the puncturing of the skin.

19:56A lot of companies want to perhaps change the form factor. Whoop is one of these companies. They say that they are interested in not only taking that information and using it as part of its overall dashboard, but also to figure out a new form factor that makes it easier, more seamless, and is non-invasive for users. Another area is what he calls a health operating system. So basically something that you can come to and do a little bit of everything. And similar to how an AI model perhaps might predict the next word if you're using a chat GPT or something like that, OOP has its own AI models, and what it wants to do is to look at predictive health.

20:40So perhaps taking all the information and being preventative when it comes to heart attacks, strokes, and other things down the line. Also, some ongoing conversations, therefore, with the US FDA. Sam Kelly, you always bring us the greatest reporting. Thank you very much indeed.

21:02real quick i want to go back to the story of softbank selling out of its invidia stake entirely 5.6 billion dollar stake now gone this is softbank's u.s listed shares down almost six percent a somewhat delayed reaction to that news of 24 hours ago but overnight in japan in tokyo those shares also dropping and again we're going back to a story we did well yesterday But this isn't concern about NVIDIA. It's just more about SoftBank and what it's up to. Karen. Good reminder. And meanwhile, it's talking tech now. And first up, NVIDIA partner Honhai was offering a rosy outlook for 2026, citing AI as its key growth driver.

21:40But the company reported a better than expected 17 % jump in net income for the September quarter. Chairman Young Liu struck a tone of cautious optimism, saying the company remains very positive about AI. Plus, JD.com says it's notched record sales during China's Singles Day event, with a reported 40 % jump in shoppers. Orders, they surged nearly 60 % during the sale, defying concerns over weak consumer spending amid persistent deflation in the country. And shares of Infineon, well, they are soaring after the company forecast a return to growth for revenue in 2026, driven by global AI data center boom, of course.

22:15Bloomberg's Guy Johnson spoke with Infineon CFO Sven Schneider. We just raised the guidance for our AI revenue for this year from$1 billion to$1.5 billion. So we are more than doubling in this fiscal year and it will not end in 26. We expect the addressable market for Infineon for powering the AI data centers in the magnitude of$8 to$12 billion. So it's a significant idiosyncratic growth driver for the company going ahead. the ai boom has also led to rising demand for electricity and plans to build out nuclear power facilities to meet that demand but despite billions in investment hardly any of that power will be ready for a decade our nuclear energy reporter will wade here with more that's the story that's the reporting lots of investment up front and you've got to wait a long time to get more yeah the thing to keep in mind about nuclear is that everything moves really slowly you know i I mean, any of these power plants, they're major construction projects.

23:14So already that's a slow thing. And then it's nuclear. So you know that they are going to take their time because it has to be safe. So it's a very slow moving industry. There's huge demand for it now. There's, as you know, there's amazing demand for electricity now. All these big tech companies, they want electricity yesterday. But what we're going to see in the near term is a lot more gas power. That's so interesting. And just briefly, Will, we hear talk about SMRs. We hear talk about new nuclear, but it's actually retrofitting or it's old nuclear plants that are coming back online. Yeah. So there's a push to reopen some of the nuclear plants that closed earlier this century.

23:53You know, I was tracking them for years and more than a dozen reactors closed in the decade through about 2022. And we're seeing several of them getting restarted. They were way too expensive to operate when people didn't really value it. But now they are willing to spend a lot of money, especially the tech companies, because they've got these big, deep pockets. They can spend money for it. We may see the first one of those come online January, February, early next year in Michigan, the Palisades plant. All eyes peeled on your reporting. Will Wade, it's always great to have you on Around New Zealand.

24:33Welcome back to Bloomberg Tech. AMD is probably our big story and mover of the day, best performer, both in terms of percentage gain and points on NASDAQ 100 and the Philadelphia Semiconductor Index. The story is really clear. At its investors' day, it says top line overall, 35 % annualized growth over the next three to five years. But in that key AI data center category, 80, 80 % CAGR over the same time period. And clearly, investors like it, that 10 % gain. It's basically biggest jump in about a month. Global Foundries is a US-based contract manufacturer of chips. It reported pretty strong earnings, actually, in the quarter just gone.

Read the full transcript

25:12But the stock down 6%, I guess mixed feelings, a high bar quarter, that kind of thing. But right now, with the effort of this country to re-onshore manufacturing assemblies, we tend to take a close look at it. And then at the index level, we were talking about this earlier in the program, Caro, where AMD is probably artificially boosting a little bit the picture on the socks right now, which is up 1.6%. But as we discussed with some brilliant guests throughout the show, government shutdown, the big E for earnings next Wednesday of NVIDIA, those are probably things that we're bracing for. Meanwhile, though, the AMD story is all about data centers, isn't it?

25:47We've got yet more breaking news on that front. Metashares are currently off by 2.7%. But this is as the social media giant is also planning to spend yet more money, more than a billion dollars on a new AI data center in Wisconsin. Now, the announcement actually confirms Bloomberg's reporting from April about the plan. Let's get a reminder on it. Kurt Wagner joins us, who covers the company and almost feels a drop in the bucket of what the ultimate capex that Mark Zuckerberg is currently saying is going to spend. But this is big for Wisconsin. Yeah, it's big for Wisconsin. And it's a sign, Caroline, that this is just a snowball that continues to roll downhill in terms of AI infrastructure for Meta.

26:26We're now talking about data centers in Ohio, in Texas, in Louisiana, this one in Wisconsin that they just announced this morning. And so these are projects that aren't even going to be really online and operational for several years. It just gives you a real sense of how much this company is betting on the future, betting on AI, and wanting to control this stuff themselves, right? Not necessarily going out and paying another tech company to use their data center, but really trying to own that infrastructure themselves. That's where I wanted to go. So the headline right now is a$1 billion data center in Wisconsin.

27:02But the bigger figure is$600 billion, as Caro pointed out, over a period of years. Meta is not a hyperscaler. It's a technology company, social media focus that also has an AI lab within it. Just explain why Meta thinks the best approach is to build its own infrastructure and have its own computer available to its teams internally as much as anything else. I mean, my read on this, Ed, is to go back more than well over a decade ago when Meta really missed out on the mobile phone. And Mark Zuckerberg has spent now years and years and years reliant on Apple and Google primarily to distribute his products to his users.

27:45And we see this with, I think he regrets that. He's talked about regretting that. He's talked about the challenges that come with being reliant on another major tech company. And so we see what they're building around their own hardware in terms of AR and VR with glasses. We see what they do around their own data centers. I really think this is very much a control thing. This is something that the company has been burned on in the past, if you ask Mark. And, you know, they have a chance to control their own destiny a little bit with this AI infrastructure. And so I think they're leaning into it because of that.

28:18Kurt, what happens next with Meta, right? They have these big projects like there's this map that people show on the social medias of Meta's data centers occupying New York City in Central Park. Just give the sort of headline summary of their total ambitions. Yeah, so you mentioned that's the data center in Louisiana, this one that could be as large as five gigawatts of power and compute. Ultimately, I think this is a company that, as we've talked about, is building several different data centers in order to spread this out. But I think the big thing for them, quite frankly, data centers aside, because those are still a few years away, they need to show that all of this talent that they spent billions of dollars on over the summer, these AI researchers and engineers, are going to deliver a world-class model.

29:06And so we will see that at some point, I imagine, in the next six months or so where they will come out with a new model that they need to be on par with, you know, OpenAI and Google and other top class models to sort of show that it's not just building the infrastructure, but it's actually capitalizing on that talent and having a product to deploy to people that sort of reflects this investment on infrastructure. So the data centers are years off. I think more urgently, you know, the pressure is on them to deliver a real world class model here in the next six months. Bluebird's Kurt Wagner. Thank you very much.

29:41Let's get back to that AMD story. Shares absolutely surging up almost 10 percent over the course of the day. That shows the trading over a couple of days. It sees accelerating sales growth over the next five years, driven by strong demand for its data center products. CEO Lisa Su updated investors on AMD's long-term outlook amid rising concerns of massive spending, some of which Kurt just outlined. Let's get to Bloomberg's Ian King, who leads our coverage of semiconductors. Clearly, this is a positive market reaction to AMD saying, here's what we see for the next five years. But that 80 % CAGR, compound annualized growth, for the data center business, that's the headline, really, isn't it?

30:20Yeah, I mean, there were a lot of headlines, a lot of numbers. And if you actually look at what happened as they were speaking, the stock really didn't do very much. Then after hours, when the management were on stage and were being questioned, they put in a strong performance. And that's really when the stock started to pick up. And that's what the analysts and investors have reacted to, really. Was enough articulation given by Lisa Su as to if OpenAI is good for the money longer term? She was asked a lot about some of these nervousness about these circular deals, about an AI bubble in other kind of words.

30:55Yeah, I mean, you're right. She was asked about that, sort of talked around it. Then the analysts were like, no, come on, you've got to help us out here. And what she said really was, look, if everything happens as though it is laid out, as we get the users, the amount of models deployed and everything, the money is going to be there. Real quick, the money is going to be there. I believe Lisa Sue was asked by some of the analyst community about the relationship with open AI. And they posed the question that we tried to some weeks ago when she was in the seat you're in. Is open AI good for it? And she said, yes, don't rule them out was her strong sort of retort to that push.

31:31Ian King, we love it. Thank you very much. AMD on an absolute tear.

31:42Startup Forterra has just closed a$238 million funding round that raises its valuation to more than a billion dollars. Forterra's autonomous technology is used for battlefield vehicles that can operate individually or as a swarm to conduct surveillance or to carry items such as missiles. Here to discuss Forterra's CEO. Josh Araujo. Josh, you join us from the Web Summit in Lisbon. Thank you for being here on Bloomberg Tech. I want to bring back some of that video of the technology and we'll get to the raise. But just explain the autonomous system, the payload and the swarm effect that we were just discussing.

32:23Yeah. So what we're providing or really enabling is the use of any vehicle to carry any mission capability into combat in support of our men and women who are serving our country. So you think about, you know, a lot of investment in defense tech has happened in many other domains, space, air, maritime. But really where the wars are fought and where wars are won are by the men and women on the ground doing the hard, dangerous work of our ground combat forces. So this technology is really, you know, in support of them. It's what gets us excited, gets us up in the morning is to go deliver, you know, really mission critical, life saving technologies that takes them out of harm's way and allows them to accomplish those missions.

33:02really from anywhere in the world and particularly from a much safer position than riding in that vehicle. The$238 million Series C is split between equity and debt. What is it that you need the funds for in scaling and deploying your technology to the battlefield? Yeah. So if you look at the priorities for the Department of War and Secretary Hegseth had a really great speech last week, I've had the pleasure to be there. Really what they're indexed is speed, scale, and capability. So how do you get more capability into the hands of warfighters faster and at scale? So what this funding is going to enable us is really to do just that.

33:42Everything from acquisitions and M &A, we recently acquired Gotenna, which really brings more capability tightly integrated to the warfighter. We're using to invest in additional capabilities. If you look at these vehicles on the screen, those are now mobile data centers with comms, compute, sensing, that are now fully distributed on the battlefield will deliver more than 200 systems into operational and combat environments this year. So it's how do we take more capability and leverage that compute comms and sensing at the edge to bring additional payloads, additional sensing, additional firepower, and really help those missions close for those men and women that are out there defending our country.

34:20Josh, let's just talk about the underlying technology, the innovation that you've brought here, Because you're kind of outfitting these existing vehicles with your own sensors. How are you working with partners and what is it that you brought that no one else really had? Yeah, I think it's our ability to work with best in class capability, best in class partners. So traditional defense primes, they are excellent at building vehicles at scale. So we can take their capability, made it with our technology, which is really the compute sensors, the autonomy software, leveraging the latest and greatest in AI.

34:54to enable those vehicles to do really complex missions. And it's not just, hey, I have a vehicle that can drive point A to point B safely. It's how do I have that vehicle or system really close the gap for an entire mission, whether it be a mine clearing mission, a ship interdiction mission, a resupplied mission, and have multiple vehicles coordinate together where there would otherwise be a human doing that. So that requires a pretty unique insight into how military operations work on the technology that's needed to enable that. And so it's really working with best in class partners. We want to bring the best in class of defense tech, the defense primes, and then help integrate those into a cohesive system that really serves that, again, that 18 to 25 year old who's out there, you know, on the front line, standing there for a nation's defense.

35:39You've been building this business since 2002, Josh. And I just want to get your context on how much has changed just in the last few months, say a year, because we're on track for a record amount of defense tech investing. We've got, as you talk about, Department of War really thinking about bringing on startups. Does that feel seismically different? It seems a lot has changed. But if you look back when the company was founded in 2002, the threat that was really facing pre-9-11 was near-peer competitive threats, tank-on-tank battles. And autonomy and robotics were looked at even back then on how do you deter a near-peer conflict?

36:18How do you prevent the next large-scale warfare from happening? And that's really through robotics and autonomy. So you fast forward to today, we're now several years into the Ukraine-Russia conflict. We're looking at tensions in the Pacific theater. And again, we don't have the size of the military, both in terms of distance and space, to go occupy and project the firepower to deter the next conflict. And so it's just, you know, what's old is new again. It's how do you leverage capability and technology to do that? I think we have the benefit of today that we didn't have in 2002 was very low cost, very capable compute, orders of magnitude more compute that's available to do this technology.

36:58sensors, an ecosystem of technology providers that are focused on the defense industry. And if you look at how the defense industry and Secretary Hegseth said it himself last week is we have to get back to the rapid innovation, rapid deployment of capability at scale. We used to do that back in the World War II and beyond. We've lost that along the way. And that's really that emergence is coming back, which is exciting. Josh Raho, Fortara CEO, joining from Lisbon Web Summit. We appreciate it. Thank you. And coming up, investors, IAI, to help Cisco reach dot-com levels. We're close, about 13 % off of those numbers hit back at the beginning of the century.

37:38This is Bloomberg Tech.

37:46This is Caroline Hyde. And I'm Ed Ludlow, inviting you to join us for Bloomberg Tech, a daily podcast focusing exclusively on technology, innovation and the future of business. Every weekday, we bring you the top headlines from the world's biggest tech companies. From finance to defence, AI to entertainment and from startups to the magnificent seven. We highlight the latest stories of the people and companies pushing the tech sector to new frontiers and the politics that shape global tech markets. We do this all every weekday, then bring you the most important conversations and analysis in our podcast.

38:19Search for Bloomberg Tech on YouTube, Apple, Spotify or anywhere else you listen. Join us every afternoon on your commute home and stay ahead of the tech news cycle. That's the Bloomberg Tech Podcast. I'm Caroline Hyde in New York. And I'm Ed Ludlow in San Francisco. Subscribe today wherever you get your podcasts.

38:45Cisco reports earnings after the closing bell. Investors are watching for signs of gains from its AI investments, though some on Wall Street remain wary it will ever reach its dot-com era peak. Despite being pretty close, Bloomberg's Dina Bass joins us. That's kind of been the story with Cisco for a little while. You know, back in the day, it was the big technology company. And present day, we're kind of waiting for it to get involved and see benefit from all the things that others are pulling ahead in. Is that a fair preview? Absolutely. Look, when I started as a reporter more than 25 years ago, it was right around where Business Week dubbed Cisco one of the four horsemen of the internet era.

39:25And, you know, those horsemen all to varying degrees ran out of steam. Cisco's trying to use AI to relitigate that and to get back into a leadership role. But when you compare it to some of the other AI bets, the other AI companies, it's, you know, their forecast increase in revenue and profit is quite a bit more modest. You go to August when it had its fiscal fourth quarter report and they talked about the one billion that they saw in the AI sales coming from fiscal 2025. How better can they articulate what they're providing in the AI ecosystem right now, Dina? Investors probably want to see just something that's a little bit more sustained and just, you know, continued winning deals, continued development of the products to capitalize.

40:15So what they've basically been doing is, you know, redoing their networking products, a bunch of their software to really take advantage, you know, particularly on the networking side of the opportunity to connect server racks, to connect, you know, data centers in between different data centers to better position those products for enabling AI workloads. But, you know, that's a competitive space. That's a space where Broadcom plays as well and where Broadcom is, you know, definitely better known. So it's really kind of a, you know, show me consistent performance, consistent deal winning, and a consistent roadmap of products that are going to go where the AI market, you know, isn't not just now, but will be in the future.

40:59And show me a return to the highs in the$80 price tag that we saw back in the old noughties. Dina Bass, we thank you so much for things. Cisco, and look, Cisco is aiming at AI data centers. We're just talking about it. Kim Forrest, Spoker Capital Partner CIO, says actually the build out of a data center will take a lot longer than Wall Street thinks. You're joining us now on the theme of the day, I think. It has been all about data centers, whether it's 50 billion coming from Anthropic, whether we're thinking about Meta continuing its focus in Wisconsin. But why do you think it's too much too soon from the Wall Street's perspective?

41:33Well, for a lot of reasons. Living in a virtual world, as most of Silicon Valley does, it's really hard whenever they have to, you know, go out into the real world and do physical things. And that's what we're talking about with data centers. We need power. We need really kind of exotic. They're kind of boring buildings, but they need to be very uniform, and there's only certain kind of people that can build them. And then we have the demand. So I've seen some reports, maybe on Bloomberg, about how some data centers that are already established are having problems selling their capacity. So we have a lot of physical world problems.

42:19But I think my biggest thing is I'm not a believer that we really need all these data centers. I think human beings are incredibly smart and are going to work around what we're doing now, which is the brute force method of training large language models. I believe in AI, just not AI as it's presented right now. So, Kim, there's the rub, right? Because you're not just saying that we've got utilities issues, power issues, supply chain issues. You're saying we're going to innovate ourselves out of that. Are we going to the United States or is it going to be Asia? Well, right now it looks like Asia was first on track with the whole DeepSeq product that they came out with, right?

42:58We can quibble about if they unfairly used OpenAI to leverage, but they did leverage. And that's the most important part. They thought out of the box as opposed to saying, you know, we need to have the state of Texas be one big data center. I'm using hyperbole here. But, you know, they thought about how can we do this differently? And I think ultimately it's a good race between the U.S. and the West in general and China about who is going to be able to come up with usable AI that doesn't break the bank and cost every last dollar that Microsoft, Meta and the rest of them have. Kim, you've presented your thesis.

43:42So as an investor, custodian, fiduciary of your clients, how do you invest to support the thesis you've just outlined? Sure. Well, I think we have to have many options, right? We do own some NVIDIA, but we also own AMD, Micron, and believe it or not, Intel. AI is real, and AI is going to have very different needs. Sorry, may I just interrupt you? Like the AMD story today, the whole thing is predicated on the idea that they do participate in some massive data center build out. And you just said we don't need the data centers. Well, there's going to be some data centers, but not as many as are planned currently.

44:28And I think that's the big thing that even, you know, last week's big cha-cha-cha in the world of AI was the Brad Gessner and Sam Altman interview where, you know, they're saying, where are you getting the money, Sam? So, I mean, there's a lot of concern about the whole theory of we need all of these data centers. But I think over time we need them. But isn't AMD currently trading an evaluation that is pricing in an enormous data center boob that will come to light? Or are you thinking actually AMD at this price is reflective of perhaps not such a significant build out? Well, I am. And the other thing is, is whenever we discover that we don't need the brute force method, these chips will be instrumental in that.

45:17right? So I think we're going to deploy money. It's just differently than the way we're currently thinking. How about that? Kim Forrest of Boca Capital Partners. As Caroline made the point of, this is the key topic of conversation today, and we're grateful for the robust conversation. Good conversation. That does it for this edition of Bloomberg Tech. It does. You should go back and listen to the conversation. Listen to all of it. Don't forget to check out the podcast. You can find it on the terminal as well as online on Apple, Spotify, by Anne iHeart. We're going to keep on banging that drum about the AI infrastructure build-out.

45:49We gave our cynics, such as what we've just heard from Kim Forrest, and we have those that think it is significant. Look, AMD, it's a significant move today, Ed. Yeah, look, it's actually only its biggest jump in a month, but that was the whole point. It's them outlining growth tied to the build-out of data centers, and what we're trying to understand is, will in the end we need all of them? That's what your markets look like. This is Bloomberg Tech.

46:18This is Scarlett Fu. And I'm Paul Sweeney, inviting you to join us for the Bloomberg Intelligence Podcast. Every day, we harness the power of Bloomberg Intelligence to bring you deep dives into the companies that are moving markets from publicly traded companies like Apple to those that are privately owned but known by everyone on Earth like OpenAI. Now, I helped to build Bloomberg Intelligence to what it is today, Scarlett. And now, our analysts are the best in the world, covering more than 2 ,000 global companies. That is your legacy, Paul. And we speak to those in-house experts every day. They are Bloomberg's go-to authorities on sectors, companies, and legal processes.

46:50And we do it all live each weekday, then bring you the best conversations in our daily podcast. So be sure to search for Bloomberg Intelligence on YouTube, Apple, Spotify, or anywhere else you listen. Listen in the afternoons on your way home from work to catch up on the market news you missed during the business day. That is the Bloomberg Intelligence Podcast. I'm Scarlett Fu. And I'm Paul Sweeney. Subscribe today, wherever you get your podcasts. you

From the publisher

Bloomberg’s Caroline Hyde and Ed Ludlow discuss the continued data center boom, as Anthropic pledges to spend $50 billion building data centers in the US, and Meta announces plans for a $1 billion data center in Wisconsin. Plus, shares of AMD spike after the Nvidia rival projected accelerating growth for its chips. And fitness tracker Whoop eyes an IPO.

See omnystudio.com/listener for privacy information.

More from Bloomberg Tech

All 343 episodes
Anthropic to Spend $50 Billion Building US Data CentersBloomberg Tech · 43 min
Listen in VO