SpaceX Kicks Off High-Grade Bond Sale

23:27Ed Ludlow:Thank you very much, Yuhaira. Now, Apple built its reputation on iconic product design, but the influence of the team behind the iPhone and iPad has faded over the past decade. Bloomberg's Mark Gurman writes in this week's Power On newsletter that the incoming CEO knows a major design shakeup is needed and is getting ready to put his firm imprint on the team. Mark joins us now. And this is about the industrial design studio that in the era of first Steve Jobs and then Johnny Ive for a very long time held a lot of power within the organization at Apple. What are you writing about Empower On that's changed and why do you think that the incoming CEO takes us back to that place of power?

24:13Yeah, over the past decade or so, you've seen the industrial design studio go from where exactly the future products are dreamed up, where the whole company would come in to rally around the next big thing, leading to products like the iPhone, the iPad, the iMac, the iPod, the Apple Watch, you name it. And it's now more of a service organization where different teams, operations, engineering, they come in, they get what they need, and they get out. It's no longer the central hub, the influential center of innovation for the company. And John Ternus knows that things need to be shaken up there.

24:48The head of the industrial design team, Molly Anderson, she's known as a great individual designer. She's known as someone who's beloved by her team. But it's not like Apple went out to look for the best industrial design leader in the world to take that position. They made that choice from a place of defense, not wanting more people to leave the team versus the perspective of a company that's worth$3 trillion plus that could go out and get anyone they want to lead design in this post-Johnny Ive era that's already been going on for nearly a decade.

25:22Ed Ludlow:Mark, you've detailed recently the kind of roadmap on the hardware side for not just the rest of this year, but beyond that. Is there a case study of where this kind of return to design focus plays out? We just have about 30 seconds. Yeah, I mean, we would like to see bigger design changes, right? For the first time in 20 years, you're going to have a meaningful shakeup to the iPhone over the next few months with the foldable phone, which I'm calling the iPhone Ultra. But there's a lot more to be done. There's more product categories. Major changes are necessary to freshen up the AirPods and the Apple Watch.

25:58Those are categories that have essentially looked the same since 2016. Right.

26:04Ed Ludlow:Bloomberg's Mark Gurman, who puts power on out on Sunday mornings and then comes on Bloomberg Tech Monday to go through it. Thank you so much. Now, coming up, Micron lands a key deal with Anthropic as AI demand keeps surging. It's so interesting to see Micron and memory name go direct to a frontier lab. We're going to go through that. This is what your markets look like, kind of muted but outperformance in chips, which has been the story of the year.

26:33Ed Ludlow:Welcome back to Bloomberg Tech. Our top stories today, SpaceX's bond sale and Micron striking a partnership with Anthropic. Bloomberg Executive Reporter, Carmen Reiki here with details on both. Carmen. Yeah, thanks, Ed. So we're seeing a lot of red on the screen today here with SpaceX. Shares are down about 8 % off the lows of the day. But if this continues, this will notch a third straight day of selling in the name after its record-breaking IPO. So what we're seeing here this morning is that investors might be reacting to a few different things. One is this investment-grade bond sale. The company looking to raise even more money to spend on AI, fund its sort of future growth after raising a record$75 billion in an IPO.

27:15The other thing we're seeing here is classic volatility that we see from IPOs, especially ones that have low float, like we know is the case here for SpaceX. We also are looking today at shares of Micron. So this stock is looking a little bit better today, up about 4.4%. It's bucking the broader trend down in the market today. What's happening this morning is that the company said that it has a strategic agreement with Anthropic. So it's going to work together on things like memory and storage, architecture for AI, and also gave some money to one of Anthropic's funding rounds. This company has been on an absolutely massive stock rally.

27:56The shares are up more than 260 % from a March 30th trough here. And we're going to be watching this one really closely because it reports earnings on Wednesday after market close.

28:07Ed Ludlow:Bloomberg's Carmen Reineke, thank you very much. Let's stay with Micron. Jake Silverman of Bloomberg Intelligence says, Micron's new supply deal with Anthropic highlights just how tight memory chip capacity has become. He says, major AI data center customers are increasingly locking in long-term agreements to secure that supply. And Jake joins us now. This is the thing, though. It's a deal between Micron and a Frontier Lab Direct, right? Not the middleman that we call hyperscale. How unusual is that? How surprised were you to see it? It's definitely not typical for memory makers themselves to sign these kind of deals, but it's not necessarily uncommon for big customers to look for supply assurances.

28:51Hyperscalers themselves have often done this in terms of how you think about wafer allocation from someone like TSMC, for example.

28:59Ed Ludlow:In your React, you're talking about it being reflective of the situation, the state of play, and how tight supply is. But this is good news for Micron, right, if you just go on the share reaction alone? Yeah, this is good for Micron in the sense that it gives them a little bit more visibility into the actual downstream demand, at least theoretically based on what we're seeing from the agreement that they actually put out in their press release. So, you know, Micron needs to build out capacity, right? The entire industry is supply constrained, and this allows them to potentially build it out in a responsible way so that when Anthropic is serving their customers, increasing their capacity, not specifically of memory, but the chips to allow people to use their models.

29:43So now, theoretically, Micron can get an idea of exactly how much chip demand there actually is.

29:50Ed Ludlow:you write in the react note that this sets them up well for the next down cycle situation you know historically memory markets have been cyclical right boom and bust we talk about that a lot on the show but there are loads of people that don't think there'll be a down cycle you know in in the traditional sense because of the trajectory of data center building an investment just explain your your point of view on that yeah i think just if you look at historical trends, having covered memory for a while, we've always seen up cycles and down cycles. And I don't think that just because we're seeing very robust demand right now, it doesn't mean that we won't see price declines on, say, a sequential or year-over-year basis in the future.

30:34What these agreements might do, though, on the other hand, is it might reduce some of the volatility. We've seen massive price decreases in the past during down cycles, such as 50 percent or so quarter over quarter. So we might see a reduction in some of that volatility. And at the same time, if you look at past cloud cycles, there are periods where CapEx flattens or declines slightly. And so that might be an opportunity for hyperscalers to get slightly better deals on the actual memory that they're purchasing. So we don't necessarily think that the cycle is not going to happen again or down cycle, but we do think that we might see less volatility, more of a straightforward idea of what the demand levels will be over the next, say, five years.

31:18Ed Ludlow:Jake, as you know, Micron reports fiscal third quarter earnings after market close Wednesday. What are your expectations? Yeah, yeah, yeah. We think they're definitely going to offer guidance results that are above consensus right now. We do think that they'll talk a little bit about some of their contracts, but I think one of the reasons that they make this announcement today with Anthropic is it's something that they can highlight to investors to give everyone an idea of what types of customers that they're signing to these agreements. But we don't necessarily think that they're going to give a lot of details in terms of, you know, we don't think they're going to give anything in terms of pricing.

31:56They might not say anything in terms of the length of these contracts or a lot of the very intricate details. But that said, I think they'll probably give a longer term view of how they see the market given the fact that there are more of these contracts being signed and memory prices continue to be very strong and the environment is very favorable for them at least through this year and we think at least into next year as well.

32:20Ed Ludlow:And to recap, Micron up 4 % on this supply agreement with Anthropic but up more than 300 % year to date. Jake Silverman of Bloomberg Intelligence with the analysis. Now coming up, the CEO of Base 10 and one of its lead investors is coming on the show to talk about the inference startup's latest$1.5 billion fundraise. But this is really a conversation about the economics of AI, and particularly as we transition increasingly to inference, the use of open source, and getting the best cost dollar per token. Really, really looking forward to this one. It's next. It's Bloomberg Tech.

33:05Support for the show comes from Public. Public is an investing platform that offers access to stocks, options, bonds, and crypto. And they've also integrated AI with tools that can assist investors in building customized portfolios. One of these tools is called Generated Assets. It allows you to turn your ideas into investable indexes. So let's say you're interested in something specific like biotech companies with high R &D spend, small cap stocks with improving operating margins, or the S &P 500 minus high debt companies. Chances are there isn't an ETF that fits your exact criteria. But on public, you just type in a prompt and their AI screens thousands of stocks and build a one-of-a-kind index.

33:45You can even backtest it against the S &P 500. Then you can invest in a few clicks. Go to public.com slash market and earn an uncapped 1 % bonus when you transfer your portfolio. That's public.com slash market. Add paid for by Public Holdings. Brokered services by Public Investing, member FINRA SIPC. Advisory services by Public Advisors, SEC Registered Advisor. Crypto services by ZeroHash. Sample prompts are for illustrative purposes only, not investment advice. All investing involves risk of loss. See complete disclosures at public.com slash disclosures. Ask yourself, what are your best people spending their time on right now?

34:22Expense reports, receipt chasing, month-end close that takes weeks. You become what you spend on, and that's not what you're building toward. Brex is the intelligent finance platform that eliminates that work before it starts. AI agents that handle the manual stuff automatically, so your team can spend their time on what actually compounds. It's time to get Brex AF. Learn more at brex.com slash AF. Ryan Reynolds here from Mint Mobile, with a message for everyone paying big wireless way too much. please for the love of everything good in this world stop with mint you can get premium wireless for just 15 a month of course if you enjoy overpaying no judgments but that's weird okay one judgment anyway give it a try at mintmobile.com slash switch up front payment of 45 for three-month plan equivalent to 15 per month required intro rate first three months only then full price plan options available taxes and fees extra see full terms at mintmobile.com

35:24Ed Ludlow:NVIDIA is developing chips and software that will make humanoid robots safer around people. Humanoid robotics are expected to generate$200 billion in revenues by 2035, according to Barclays estimates. But current safety systems can make the machines less productive. NVIDIA's Halos software will give robots a much better sense of their surroundings so they can interact and even make contact with people. We're going to talk a little bit more about that later in the week. AI inference startup Base10 has raised$1.5 billion across two tranches, one at an$11 billion valuation, the other at$13 billion.

35:59Ed Ludlow:Base10 specializes in delivering software and computing capacity to companies tapping open source, typically lower cost AI models. The firm's co-founder and CEO, Tihim Srivavastra, joins along with one of its investors, Altimeter Capital partner, Apoor of Agrawal, whose firm co-led the round. And actually, it was really interesting, the list of participants that backed you. Let's start with the basic questions that your audience always has, which is, wow,$1.5 billion round, it's a lot of funds. But the stories seem to not change, right? Compute is a constraint. Talent is a competitive marketplace.

36:36Ed Ludlow:I'm assuming that's where those proceeds go. Yeah, absolutely. Look, with inference right now, what we are seeing is there's a lot of demand for inserting intelligence everywhere we possibly can. Because open source models are getting very good. Right. And post-training techniques to specialize those models for specific tasks are getting very good. And as the app layer comes online, there's an enormous surge of demand. For us, what that means is, look, we need to go procure a lot of compute to be able to fulfill that demand and also hire amazing infrastructure engineers and research engineers to be able to build the software layer on top of that.

37:14Ed Ludlow:Just real quick, what's that like trying to acquire the compute? Do you have any leverage in that market? If everyone is supply constrained, how is the experience of being a company of base 10's age and size going to NVIDIA, as an example, saying we kind of need this? Yeah, look, we have a very strong relationship with NVIDIA and they've been very supportive of us over the years. Look, one thing that we have realized is that you cannot be reliant on one compute source. You have to diversify, not necessarily from a chip perspective, from a cloud perspective. So we acquire compute from 18 different clouds now.

37:54We sit in around 90 different clusters. And that's the flexibility we need to be able to acquire compute and fulfill the demand.

38:01Ed Ludlow:I posted on X that you were both coming on the show, and Paul put this question out there, that if you could only invest in one part of the five-layer cake, as Jensen calls it, what would it be? But Altimus has been so busy. you know, with the frontier labs at one end and other parts of the five layer cake at the other end. But what was the thesis around base 10 and why you wanted to get in here to co-lead the round? Ed, it's good to be here. Some truths cannot be said enough times. And one of those is that inference is going to be one of the largest, if not the largest markets, not in AI in the world.

38:39And why is that the case? You know, the early innings of AI were Q &A. Tell me about my trip. We've come a long way since then. You know, it's multi-agent. They've gotten longer context. You go retrieve, analyze, synthesize, verify, and then you go again. And each request is kicking off hundreds, if not thousands, of inference requests. The second thing is, you know, I might correct something you said, Ed. It's not just cost. It's capability, control, and cost. Frontier models, the post-trained open source models comes very close to frontier model performance. sometimes even better for specific workflows.

39:16Control. As Satya Nadella said last weekend, you've got to compound your unique advantages as an enterprise, your data, your knowledge, your know-how, in a way that you're not giving away your intelligence on rent. And that's what Base10 allows. And so to your question from this morning, is like where does value accrue? Is it at the app layer? Is it at the model layer? The way I like to frame it is, look, the model layer is a phenomenal business for very few people. It's a game of emperors. You've got to be at the frontier to accrue a lot of value. And it's a knife fight, as you know. I've started joking there's four seasons to the year.

39:49It's Anthropic, OpenAI, SpaceX, Google with an evergreen of open source. And so that's...

39:56Ed Ludlow:That part is so fascinating, right? If we could just go to open source. When this story started, for us, let's say 2022, but for you guys probably before that, it was this idea that it was only the largest models with the hundreds of billions of parameters that mattered. And in the world of open source, the issue was that it just didn't work. People couldn't make any real use of them. I go back to what Apov just said, you are making utility out of open source models, but probably also at a slightly smaller scale of model. Is that fair? Yes. Look, I think if you look at, you know, when you think about open source, it's kind of gone through a bit of a journey.

40:38Right. You know, there was Llama 3 maybe two and a half years ago, and then there was a bit of a winter. And then I think the deep seek moment last year really put it back on the map, and there was definitely the capability gap shrunk. I think this is happening again just over the last couple of weeks with GLM 5.2, which is, I'd say, a frontier-level model that is actually usable and it isn't just doing well on the benchmarks. When people are using it, they are feeling that, you know, at the frontier capability. For us, it's just like, how do we make these models very, very easy to use? And it doesn't really matter what size they are.

41:13Like we work with small models, big models, massive models. What we are trying to give our customers the ability to run these models when they don't necessarily have access to the infrastructure teams of the frontier labs doing it for them.

41:26Ed Ludlow:So let's go back to what you were talking about, Paul. cost but also utility and capability. You know, the other case study in this open source debate was Meta kind of changing its narrative saying, well, actually, we are committed to open source, but at some point we need to be realistic about the benefit of closed models. What is the point on the capability part? And actually, take a moment, explain what Base 10 does, why it's such an important player in that respect. Sure. Look, yesterday it was DeepSeek. Today it's GLM. It'll be something tomorrow. And as I said, it's seasonal. And that battle at the model layer is not – we're nowhere near the endgame.

42:11What Base 10 does is it helps customers like Cursor, like Abridge, like Open Evidence harness the power of the model underneath, combine that with your unique advantage as an enterprise, your workflow, and deliver something that is best for that specific workflow. Now, it's not just one model. It's typically a collection, a portfolio of models that are fine-tuned to deliver that, and that compounds over time. And so, Base 10 is successful as these models get better. It benefits us. And ultimately, in the service of the app layer, which is the customers that we discussed, who are delivering AI to the customer.

42:47Ed Ludlow:Jensen's been repeating a lot recently. Jensen Wang, the CEO of NVIDIA. Sometimes you have to remind the audience just in case. you know he wants to see his teams burning tokens um how does that apply to your customers right so so paul just explained really clearly the value that you're adding but um the economics are really really important yeah to some of these companies yeah look i i think using tokens is obviously just what's anonymous with hey i want us to be using ai yeah everywhere. Yeah, it's in their interest for that to be the story right now. And I think for us, what we see is customers kind of go through the same curve over and over again is that they go very aggressively, they start using AI everywhere, they start to see gains but they don't necessarily do it profitably and then they need to figure out how to do it profitably.

43:44And that's when they come to open source models, that's when they come to base 10, that's when they come to post-trained models on base 10 to be able to do it better, faster, and cheaper. And that's when you decide to get both intelligence everywhere but also unit economics that make sense for your business.

44:01Ed Ludlow:What I've been trying to reconcile is Altimeter is invested in the big frontier labs as well. And those big frontier labs plus SpaceX, depending on how you define what they are, would also say that they have a role to play in that market. Where are we in this cycle? of were there just in the NBA very big players? Where does Face 10 fit in that? Look, a lot of ink has been spilled on this debate of open versus closed. Let me just take that head on. The closed models, the frontier models, are very, very good. And they're very good for the use case that require highest intelligence, highest reasoning, new use case discovery.

44:41If you didn't want to think about just one model as a service, they're very good, and they will always be very good for that. for a class of customers, that's not a long list. You know, as Jonathan Ross said, there's about 35 companies that drive 99 % of all inference. The top five are the labs that you just listed, top four, top five. The next 30, you know, that's people like Cursor, like Open Evidence, like Abridge, like Harvey. And for them, if you're not playing the game of the emperors and you're thinking about how do I deliver something that is not rented from somebody else, but mine that I can compound with, that's where post-trained open source can help.

45:19And so we posted this chart this morning about how Harvey, for example, achieved frontier capabilities post-training an open source model, but obviously with a much better cost and control frontier as an example of what the next 30 are thinking about and then the next 1 ,000 are thinking about.

45:38Ed Ludlow:To him, Shrifastava, back on Bloomberg Tech, founder, CEO of Base Center, Paul Bagawal, of Altimeter back again on Bloomberg Tech. It's really great to have you here in person. Thank you very much indeed. Now, coming up on the show, Amazon is pairing Prime Day with pizza delivery. We're going to talk about what's behind the e-commerce giant's fast food pivot next. This is Bloomberg Tech.

46:06Ed Ludlow:Today's big number,$26.3 billion. That's how much US consumers are expected to spend on Amazon and other retailers during the four-day sale that we all know as Prime Day. That figure comes from Adobe. Amazon's facing headwinds, though. Consumer confidence to competition. But the e-commerce giant has a few tricks up its sleeve to encourage shoppers to spend, including pizza delivery. Bloomberg Spencer Soper here with the latest. OK, we're getting ready for Prime Day. Adobe Analytics always puts out numbers like this. Put that$26.3 billion into context for us. That's big. It's small. It's growth.

46:42Ed Ludlow:It's not. It's big. I mean, Adobe's predicting 9 % growth from the four-day Prime Day period last year, which was July. You know, it's always tricky to compare Prime Day to the year earlier because the date moves around, and this is going to be the second year that they do a four-day sale. But Adobe is predicting a 9 % lift, which is pretty sizable. And a lot of that can be, you know, it's counterintuitive, even though there's a lot of things weighing on consumers. Sometimes that compels them to kind of save up money and hold off a big ticket purchase for a big deal, big deal period, which can which can raise spending.

47:17And Walmart and Target both have overlapping sales. So there's going to be just a lot of a lot of deals going on those four days.

47:25Ed Ludlow:OK, the latest tech in depth. Amazon offers pizza in prime day experiment. Explain. So this is the first year I've seen this. Amazon's offering a$5 pizza from Little Caesars. So basically, if you go to the site, you'll see this pretty prominently, a Little Caesars emblem on the site. And you can click over there to redeem a$5 pizza. And they're running that every day of Prime Day, even a couple days before it. And it seems to be something to get people to come to the site. Just another sweetener to get people to come to the site and check out the deals. because Amazon really doesn't have anything in hot food delivery.

48:07You know, they had an Amazon restaurants thing years ago that they shuttered. They now have kind of a tie in with Grubhub. But, you know, when you think about grocery and Amazon really is eye on the grocery business, a lot of people couple grocery and dinner, you know, especially when it's hot in the summer, you don't get the kitchen hot. So this might be a way to just, you know, get people to check out their deals and also cross something off the list that they got to get done that night.

48:34Ed Ludlow:Bloomberg Spencer Soper with a look ahead to Prime Day, but go and read the tech in depth that went out this morning as well. Thank you very much. Before you go, Disney's Toy Story 5 has gone to infinity and beyond on its opening weekend. Woody, Buzz and their friends brought in$160 million in U.S. box office sales, marking the best launch in that franchise's history, and lassoed the title of best 2026 debut from the Super Mario Galaxy movie. The hall was at the low end of Bloomberg Intelligence's lofty expectations for the film. Enthusiasm for blockbusters, including Toy Story, Spider-Man, The Odyssey, is anticipated to fuel a 10 % year-over-year growth rate and the best summer for cinemas since the pandemic.

49:20Ed Ludlow:All right, that does it for this edition of Bloomberg Tech from here in San Francisco. This is what markets look like. A little softer on the NASDAQ 100, down a percentage point, half a percentage point, but still outperformance in semiconductors. The SOX up almost a percent. A lot of that being the news flow around Micron and Anthropics deals. SpaceX down for a third straight session, but well above its IPO price as it goes to the bond market for the first time. Check out the pod. You know where to find it. This is Bloomberg Tech.

49:57Ask yourself, what are your best people spending their time on right now? Expense reports, receipt chasing, month-end close that takes weeks. You become what you spend on, and that's not what you're building toward. Brex is the intelligent finance platform that eliminates that work before it starts. AI agents that handle the manual stuff automatically, so your team can spend their time on what actually compounds. It's time to get Brex AF. Learn more at brex.com slash AF.

50:53Then full price plan options available. Taxes and fees extra. See full terms at mintmobile.com. Aging doesn't stop, and neither should you, with Vital Proteins Collagen and Protein Shakes. Because around the age of 30, your body needs more support for movement and recovery. On workout and rest days, reach for a 30-gram total protein shake, or go with our classic collagen peptides. Help support healthy hair, skin, nails, bones, and joints, so you can stay vital, stay you. Visit vitalproteins.com to learn more and where to buy. These statements have not been evaluated by the Food and Drug Administration.

51:24This product is not intended to diagnose, treat, cure, or prevent any disease.