Anthropic’s AI Spending Surges as Microsoft Rethinks Copilot

Bloomberg Tech

The episode explores the massive surge in AI infrastructure spending, highlighted by Anthropic's $11.6 billion contract with Akamai for compu

Key takeaways

• Anthropic secured a $11.6 billion seven-year deal with Akamai to meet growing demand for Claude.
• Microsoft is refocusing Copilot as a productivity tool by merging consumer and workplace versions.

Main topics

• AI Infrastructure and Computing Power
• US-China Geopolitics and Trade

Notable quotes

The purpose of AI is to serve humanity and accelerate human flourishing.

Conclusion

The episode demonstrates that while AI innovation is driving massive c

Transcript preview

Speaker 3 (0:00) Here are five reasons to subscribe to Bloomberg News Now. Number five, you get the latest news from around the world instantly. Number four, it saves you time. Your day's busy. Catching up takes just a few minutes. Three, it's available 24 hours a day whenever you need it. Two, it's convenient. Listen on your smartphone or smart speaker. And number one, it's easy to find. Subscribe on Apple Podcasts, Spotify, or anywhere you listen. Five reasons, one Speaker 7 (0:26) simple way to stay on top of the news. Bloomberg News Now. Speaker 3 (0:31) Bloomberg Audio Studios. Podcasts, radio, news. Bloomberg Tech is live from the heart of Silicon Valley with Ed Ludlow in San Francisco. Speaker 7 (0:53) Coming up, Anthropic signs an $11.6 billion seven-year contract with Akamai for computing power. It's the largest deal in Akamai's history. Plus, President Trump calls on his meeting with Xi Jinping, quote, very productive. We'll discuss if there's anything to show for it. And Microsoft steps back from the race to build a personal chatbot and merging the consumer and workplace versions of its co-pilot AI assistant into one. I'm Tim Stenevek. In for Ed Ludlow, it is Bloomberg Technology. First, though, a check on markets. Bloomberg's Nora Melinda joins us now for more. Hey, Nora. Speaker 1 (1:29) Hey, Tim. We are seeing a bounce back of major indexes from early session lows. Those session lows, of course, coming off of that consumer sentiment data that we got earlier this morning, showing that consumer sentiment fell to a four-month low. We've also got cooling oil really aiding sentiment here, and a NASDAQ 100 that's actually heading for its biggest weekly gain in seven weeks. But one of the laggards that I want to look at right now, let's look at shares of Tesla currently pointing lower by about 1.7 percent. Now, this coming after the fact that the company's ramping up production of its Optimus humanoid robot. But the concern here is about the production specifically as it pertains to the robot's arms. Now, we know that this robot is something specifically that the company plans to lease to other commercial companies. clients but not necessarily sell because they would like to have the opportunity to be able to actually revamp and refurbish some of these robots. And lastly, looking at shares of Meta, they are softer today but the stock actually heading for its biggest monthly gain in since 2013 here. And this all really of course coming from the optimism that we've been seeing on Wall Street about that Muse AI Assistant. Speaker 7 (2:34) All right, Norm Malinda, thank you. We'll be checking in with Nora throughout the hour here on Bloomberg Tech. Meantime, Anthropic locking up even more computing power to keep pace with demand for Claude. The AI startup has struck an $11.6 billion seven-year deal with Akamai. It comes, though, with an unusual twist, a warrant that could give Anthropic a stake in the company. Bloomberg's Seth Figerman joins us now with more. Hey, Seth, I was saying, I mean, when this crossed yesterday, I'm thinking, and Akamai, like... You know, CDN, like Content Delivery Network, Netflix, like a dozen years ago. Are they going all in on AI? Speaker 8 (3:07) Yeah, well, it feels like at the moment, Anthropic and OpenAI and others are turning over every stone they can right now to get the chips and data centers and computer power they need to meet demands. So certainly for Akamai, you're right. I think they're taking a firm that focused on cybersecurity and making their biggest bet yet to really claim a piece of the AI pie. But for Anthropic, these deals, which can be sort of unprecedented for Akamai, are becoming routine. They're tapping Bitcoin miners and Neo clouds, as well as rivals like Google and SpaceX. to meet their demands. Speaker 7 (3:36) So it's part of the demand story here, like going anywhere to look for this compute at this point. Yeah, I mean, Speaker 8 (3:42) at one point, Anthropic said that their demand was growing 80x from the prior year. All of a sudden, they had to kind of plan for a rapid acceleration in computing capacity far beyond what I think they had anticipated earlier this year. And while you can do that by trying to launch these kind of Stargate initiatives like Opening Eye has done, I think you're seeing Anthropic move more in the direction of trying to get a lot of these kind of bespoke partnerships. off the ground and as quickly as they can. On Speaker 7 (4:05) this idea of warrants, I mean, I think increasingly we're seeing this, not just from companies like Anthropic, but also companies like NVIDIA that are in this space, this idea that they can become owners or investors in other publicly traded companies. How do you look at this in terms of the cycle? Speaker 8 (4:19) Yeah, it certainly feels very much a piece of the circular financing criticism that we've all heard. Speaker 1 (4:23) You Speaker 8 (4:23) know, AMD did stock warrants as part of a larger computing deal with OpenAI. So it's certainly not unprecedented here. But I think it adds to that concern or anxiety that these suppliers Buyers and developers are taking stakes in one another and potentially inflating the demand and also raising questions about the financial viability of these deals. So it's Speaker 7 (4:41) not, Speaker 8 (4:42) this Speaker 7 (4:42) demand isn't coming from everywhere. Let's take the news from Oracle yesterday, for example, in this idea of a force majeure on a data center in the Southwest. How can we look at what we heard from Anthropic and Akamai yesterday and also what we heard from Oracle? Speaker 8 (4:53) Yeah, I mean, I think two things are happening at the same time, that we are seeing a surge in business and consumer adoption of these products as they get better. And we're also seeing a lot of groundwork groundswell of criticism and pushback on the ground from lawmakers, public officials, and communities. And so I think these firms have to figure out how to navigate doing these massive data center build-outs that raise the concerns with other projects that might come to Speaker 7 (5:14) fruition more quickly. Well, Seth, you and your team help us navigate it all, so we appreciate you joining us on Bloomberg Tech. Bloomberg's Seth Figgerman joining us here in New York. Shares of NVIDIA are higher today as Elon Musk races to scale up XAI's computing power. Musk says its Colossus 2 cluster in Memphis now has $500 50,000 Nvidia GB200 and GB300 chips, with hundreds of thousands more expected online in the coming months. Now, if all goes according to plan, the cluster could more than double its current chip count by the end of the year. The AI boom may be showing up in the bond market, too. The 10-year Treasury yield jumped again yesterday. This is investors' way stronger growth, stubborn inflation, and a more hawkish Federal Reserve. At the same time, the massive AI build-out is flooding the market with new corporate debt. Ana Rathbun, Grenadilla Advisory Founder, joins us now for more. Ana, good to have you on the program. I want to talk about the relationship between tech stocks and the 10-year, because yesterday's increase in the 10-year, I think, was surprising to a lot of people, you know, with a yield higher than, you know, 5.2%. Yet here we are in a week where the NASDAQ 100 hit a new all-time high. Typically, tech stocks don't really like higher yields. Speaker 5 (6:28) Yes. No, they don't because they're long duration assets, right? So first... Talking bond speak, the duration here matters. But, you know, actually, it doesn't always move with a 10 year. And I think that's what we're seeing today. And if we go back a little bit, you know, tech stocks have been pressured for many, many weeks. And so there's some a little bit of a selling fatigue that I think we've seen. So we see that bouncing. But over the long term, there has been repricing of AI stocks and tech stocks that will match what we're expecting to see in the longer term, especially in the 10 year year. for the foreseeable future. And I do think that we are in a higher for longer environment. Speaker 7 (7:08) Ana, I'm so sorry to do this. I want to get to President Trump now. He's with the First Lady. They're now doing a site visit. This is the National Archives with Chinese President Xi Jinping and Madam Pang. You can see the live pictures coming from Washington, D.C. right now. This after the U.S. and China extended a trade truce for another two months following President Trump and Xi Jinping's meeting yesterday. We're going to continue to monitor the situation. You can also see it there at LiveGo on the Bloomberg terminal. We'll bring you any headlines as we continue to get them. Ana Rothman joins us from Grenadilla Advisory. I want to get right back to you. Ana, I want to bring you in here. Thanks for your patience with that, by the way. I appreciate it. We're watching everything happening in Washington at this, I think it's fair to say, historic visit. I do want to bring in geopolitics, though, to this conversation because this is really important, Ana. And it's the question of, you know, the relationship between the U.S. and China and really what it means for not just tech companies. I mean, look at who you had there last night. It was absolutely a who's who of... I think it's fair to say the most important people in Silicon Valley, like not just from the venture capital community and the technology community, but from business leaders all over the country. How did you look at that? Speaker 5 (8:27) Well, that tells you where the priorities are, right? Not only from the United States standpoint, but also in our relationship with China. And I think AI is at the center of a lot of these talks, even though we are talking about soybeans and we are talking about tariffs. AI is front and center, and everyone's been talking about this potential global governing body so that we can put some safety rules around AI development. But the truth is... no one's going to slow it down. Even between competitors in the United States, there is no slowing down. I mean, we just talked about Anthrop and getting more compute. And definitely between U.S. and China, I mean, we can't expect China to slow down and they're not expecting us to slow down. So all of these talks, we're saying that these things are very important to the two parties. But at the end of the day, the reality is that it's going to be more of the same. Speaker 7 (9:18) You know, I'm wondering about... the different messages that we're getting from the president and his view of a so-called slowdown in this technology versus, you know, Dario Amadei addressing the UN just a couple days ago. Sam Altman of OpenAI addressing the UN, talking about the dangers and the importance of sort of international cooperation with this technology. I'm curious about your view on the different messages that we're getting from the president versus the people who are actually developing the frontier models. Speaker 5 (9:52) Well, I think that there are reasons to be concerned. The reality is, I mean, if we have an international governing body, who's going to enforce it if somebody breaks the rules, right? And I think President Trump, what he is saying, it is different, but I think it's a little bit more realistic from my perspective. Nobody's slowing down. I think this is just talk at this point in time because there's no way to enforce it. Speaker 7 (10:15) But do you think there needs to be? some sort of international standard. I'm going to try not to talk to you about the AI apocalypse because we've been talking about that so much. But the fact of the matter is there's a growing chorus of people who do think that this technology could be dangerous to humankind. I mean, we're talking about something existential here. Speaker 5 (10:37) Yeah, and I think humans have the key to... producing guardrails as well