TRANSCRIPT 🔗
00:00:02
Speaker 1: Bloomberg Audio Studios.
00:00:04
Speaker 2: Podcasts. Radio. News.
00:00:13
Speaker 3: Bloomberg Tech is live from the heart of Silicon Valley with Ed Ludlow in San Francisco.
00:00:23
Speaker 4: This is Bloomberg Tech coming up. Oracle kicking off a $ 20 billion capital raise to finance its AI infrastructure spending. Plus, anthropic revenue hits nearly $ 17 billion in Q3 before its blockbuster IPO. We've got details on the amended S1. And Apple is trading at the priciest levels since 2008. Can the $ 1 trillion rally continue? That's an earnings story, too. Let's get to our top story. The race for capital heating up to start the week. Alphabet tapping the European debt markets. Oracle looking to raise $ 20 billion combined through debt and equity. Bloomberg's Brady Ford joins us with more. Let's go to the Oracle piece first because we knew this was coming, right? Investor Day, they outlined their plans, what they'll need on capital. And when cash flow turns positive, what do we need to know here?
00:01:09
Speaker 5: The big question the last week or two has been how deep does the hyperscaler spending go and how much will they need.
00:01:15
Speaker 1: To issue or borrow to make it happen?
00:01:18
Speaker 5: Last week, Oracle said for their current investment cycle for the next couple of years. They will need just 20 billion more. And that's what we're seeing happen today. Stocks are trading down a bit, maybe coming sooner than expected. But the big question is, how far does all this spending go?
00:01:33
Speaker 4: Just real quick, just before we came to air, Bloomberg reporting that Meta, the parent of Facebook, has asked Citigroup and Morgan Stanley to put together some fixed income calls today and tomorrow.
00:01:43
Speaker 1: What may follow? U.S.
00:01:45
Speaker 4: Dollar-denominated offering and also a separate euro-denominated offering today. All according to a Bloomberg source who say Morgan Stanley is coordinating the logistics. So it looks like Meta is getting going on the debt side of things this Monday as well. We'll bring you more when we have it. Then there's Alphabet, Alphabet, the parent of Google. You and I were talking a lot like post earnings, negative free cash flow. We know what the CapEx story is. They, too, are looking to the debt markets, but in Europe.
00:02:09
Speaker 6: Right.
00:02:10
Speaker 5: Raising some money over there in Europe. And I think the backdrop for all of this is that people are getting nervous about all this spending and hyperscalers are needing to offer more, more interest, higher.
00:02:23
Speaker 1: Rates to be able to get this capital.
00:02:25
Speaker 5: I mean, a company like Alphabet or Oracle, they can afford it, but. It's getting to be a more difficult fundraising environment for them.
00:02:32
Speaker 4: Interestingly, they both have $ 125 billion of debt each on the balance sheet, but look at their net debt and their relative cash positions, and they're like two.
00:02:38
Speaker 1: Very different stories.
00:02:39
Speaker 4: Bloomberg's Brady Ford, thank you very much, and happy Monday to you. I want to go to what's happening in markets. Basically, in technology stocks, things are pushing higher. We're kind of rebounding, but remember Friday. The Nasdaq 100 had its biggest drop since June. It was down on the week for the third straight week of weekly declines, the longest streak of weekly drops since March. Yes, there's green on the screen right now. Generally, there's a bit of anxiety, particularly in the AI trade. And on a sort of longer time horizon, technology stocks have been under pressure. I want to dive deeper into that. Joining us now is Celine Wu, Portfolio Manager at Lazard Asset Management. And actually, you know, Alphabet's a name that you're very focused on. This is so interesting. They continue to raise capital through the avenues available to them. Negative free cash flow. We know the CapEx story. How do you feel about all that?
00:03:30
Speaker 7: I mean, obviously, the concerns are mainly about this timing gap, right, between this big CAPEX commitment of fund and the monetization of all the spending that is happening. I think investors need to be a little bit more patient because the underlying demand dynamics that we learned from all the top hyperscalers earning this quarter has been pretty impressive, right? The top three hyperscalers were able to grow 40%, 50%, 100%. on top of the level of $ 100 to $ 200 billion in annual run rate revenue. I think that's pretty impressive. And they're still talking about stretched customer backlog. So, I mean, I think the underlying demand dynamics remain really healthy, and we just have to be a little bit more patient just to understand how the monetization has come through.
00:04:15
Speaker 4: The monetization is coming through. There's lots of competing data points, right? So Google Cloud, 100% year-on-year growth.
00:04:22
Speaker 1: Like, that's pretty amazing. It's impressive. I guess.
00:04:25
Speaker 4: The point of distinction, if you did Microsoft and Alphabet side by side, because they reported on the same day, Microsoft increasingly cash flow positive and Alphabet increasingly pushing into negative free cash flow.
00:04:36
Speaker 1: Is it as simple as that?
00:04:38
Speaker 7: Well, I think we tend to try to oversimplify certain data points, especially in the near term. But when you actually take a step back and look at what you're trying to do, I think Alphabet really stands out because you're really going... fully vertically integrated, the full stack, the strategic advantage they're trying to build. I mean, in fact, they're the number one company in terms of the custom chips that is also becoming part of their cloud revenue, right? Also, they're trying to sell some of them offerings to the third party. So if you actually look at company specific, directionally, it seems like they're on the same and very similar direction, but what is
Load full transcript
actually happening quarter over quarter can be a little bit different. So we have to take a grain of salt when we actually look at some of the difference in terms of cash flow dynamics.
00:05:22
Speaker 4: We could also bring what we learned from Anthropix updated S1 into this. You have the capital expenditure commitments or the spending commitments across public companies and private companies. The net conclusion is what? That the demand picture is intact given their confidence on spending.
00:05:40
Speaker 7: The net conclusion, I think the most important conclusion is Anthropix, which is one of the biggest customers for the hyperscalers today and then one of the biggest spenders of all these AI computing chips and infrastructure are showcasing pretty impressive growth rate. I mean, the revenue, I think, is a third quarter that came out today, $ 17 billion. But on top of that, more importantly, the unit economics have improved significantly.
00:06:08
Speaker 1: What's the metric you look at there, unit economics?
00:06:10
Speaker 7: So the compute cost to the revenue, right?
00:06:12
Speaker 2: Yes.
00:06:12
Speaker 7: I believe the number used to be, what, 10 times just, what, three years ago? Has rapidly come down to 50% level, right? I mean, that's really impressive. And their cash balance, right, as of third quarter was close to $ 100 billion. So this is actually the business model that is able to generate profits in the middle of this rapid growth when you still need a lot of, you know, capex numbers to continue the growth. And I think that's really key credential in the validation. Also really important to read through for the hyperscalers, again, because this is one of their largest customers and biggest vendors of the competing chips of the world.
00:06:49
Speaker 4: It is so interesting that we still continue to be very fixed on the spending. I'm not an economist, but in my limited study of market economics, demand running far ahead of a market's ability to supply, for those on that side of the table, that's an enviable position to be in, isn't it?
00:07:11
Speaker 7: Definitely. And then also pretty unprecedented, right? No one has seen something like this. This is probably the biggest technological transformation in our lifetime and in history as well. And I would like to quote what Andy Jesse said during the earnings call. He said it would be really convenient if they can spread the CapEx for the next 25 years. But that's not the reality. There has to be the disconnect. There has to be the timing gap. And that's why I like to say the investors need to be a little bit more patient Yes, the quarter over quarter results are very important, guidelines and everything, but I think the longer term, I believe that these companies can actually come through, then the economics and the monetization can improve much materially, and that's what they're investing today, and they're investing further.
00:07:57
Speaker 4: Celine Wu, Portfolio Manager at Lazard Asset Management. It's great to have you back on the show. Thank you very much for being with us. Okay, coming up, we're going to get kind of back to private markets. Matt Withyla from Wellington Management joins us. But we're going to talk about Anthropic's upcoming public debut and the new financial details just out in the amended regulatory filing.
00:08:15
Speaker 1: That's next. This is Bloomberg Tech.
00:08:20
Speaker 2: Bloomberg Tech.
00:08:29
Speaker 4: And for our picks, amended S1 puts third quarter revenue at nearly $ 17 billion. That's up 12-fold from the same period a year earlier. Net loss, $ 23. 3 billion, up from $ 12. 1 billion. Here, back at the desk, Bloomberg's Bailey Lipschultz.
00:08:43
Speaker 1: Let's go through it. I mean, you know, the S1A hits. It was going to hit at some point.
00:08:48
Speaker 4: We go with the third quarter revenue number.
00:08:50
Speaker 1: It's slightly higher than what we reported previously. What else do you see?
00:08:53
Speaker 6: Yeah, modestly higher than what we reported about a month ago, but not all that much, especially when you talk to folks on the buy side. The things that investors really were calling out from this refreshed financial statement was, A, the cost of compute going up pretty markedly from quarter over quarter, or even in particular year over year. And the one big thing that we flagged on the terminal and that investors have been talking about is the commitments. A company committed to spend $ 895 billion in the out years. About $ 448 billion of that through the end of the decade. And Ed, interestingly enough, in the next 12 months, when you include October, about $ 70 billion committed to be spent. So this is a company, as we've reported, looking to raise more than $ 86 billion. Well, they've already got about $ 70 billion they plan on raising. on spending in the next 12 months. So where it goes from there will be interesting. The one thing we did not get is an October revenue run rate update. That is one area that is still blocked out on the S1. So we should expect to see that in a refresh filing. Again, that would be through the end of October. So a bit of tight work, Ed, to turn that around in two days.
00:09:54
Speaker 4: We also did not get a price range, details of the offering in terms of target range, etc.
00:10:03
Speaker 1: I think it is useful.
00:10:04
Speaker 4: Just remind us of where we're at and what our reporting has said about the road to the public markets for Anthropic from this point.
00:10:11
Speaker 6: So what we had expected, as you mentioned, third quarter numbers were going to be dropped in.
00:10:14
Speaker 1: That's what we got today.
00:10:15
Speaker 6: We will now, based on the filing, expect to get some type of new run rate update, again, including the end of October. That likely will take place over the coming days. And then, as you mentioned, next week is when we should get the targeted range and offer size. Then we will
know the valuation the company is targeting, the expected total raise, and that'll lay the groundwork formally for a formal marketing period for the company to price its IPO in the days or week ahead of that Thanksgiving holiday here in the U.S. So it'll be an interesting week next week, and then obviously all eyes will be on the company and its pricing ambitions in the week that follows.
00:10:51
Speaker 1: Okay, Bloomberg's Bailey Lipschultz, thank you very much.
00:10:52
Speaker 4: I want to go a bit deeper on a metric that Bailey just went through, and that's today's big number, $ 895. 4 billion. That is how much Anthropic says it has in committed spending in this latest S1A. I want to go into it a little bit more. So of that $ 895.
00:11:08
Speaker 7: 4 billion, $ 447.
00:11:08
Speaker 4: 5 is by the end of 2030. Almost $ 70 billion is due over the next 12 months. The balance is due right up until 2045. It's anticipated in large part. What do anthropic investors make of all of those updates? Matt Withyla is head of late-stage growth at Wellington Management, which has been an investor in this company since 2025. It's great to have you back on the show, Matt. I mean, sometimes you just want to put the numbers away. What's the story here with Anthropic?
00:11:39
Speaker 1: What are we learning?
00:11:40
Speaker 8: I mean, I think the incremental that we've learned over the past day with the new Q3 filings is the fact that, one, growth remains incredibly high for this business.
00:11:48
Speaker 2: It's growing 1,200% year over year. Two, the company remains profitable.
00:11:53
Speaker 8: So they're doing this and generating all this revenue, not to no end, but to real operating profit. And three, as you highlighted, I think that the company is setting out to commit to large commits of spend over the coming decade. And that's all in support of the demand that they're seeing.
00:12:10
Speaker 4: Q3, adjusted operating income, $ 2. 5 billion with a 15% margin. Gap operating loss, $ 567 million with a negative 3% margin. everyone's trying to understand the unit economics, right? You might have heard that earlier in the show. What a dollar of spend on compute gets you in terms of revenue. How do you digest that?
00:12:31
Speaker 8: I think that the company has repeatedly shown that they can squeeze more and more margin out of that dollar of revenue by being incredibly efficient on the compute side. And if I think back to how Anthropic has evolved, I think one of the key things that they've done ever since inception is ensure that their models can run on a variety of hardware types. They're not beholden to one particular type of silicon. They can run on NVIDIA. They can run on Google TPUs. They can run on the Amazon hardware. And so it's that plus the recursive learning that the company is doing with its inference is what is driving more and more margin dollars out of each dollar of revenue generated.
00:13:10
Speaker 4: I think one other really interesting metric to me was how much is still being spent on training relative to running inference, right? Clearly, training is still a major focus for Anthropic. But for investors, how do you go with that? You want to see more commercialization, right?
00:13:29
Speaker 8: You do want to see commercialization, but what is at the tip of the spear of commercialization is what the models are delivering. And the only way you stay at the frontier, which I think anybody would argue is an important characteristic of Anthropic, in order to stay at the frontier, you really have to do a lot of training in order to develop the best models and therefore have the.
00:13:50
Speaker 2: Most inference revenue.
00:13:52
Speaker 8: And so I think that that is a really important justification for the spend they do on the training side.
00:13:58
Speaker 1: I find this so fascinating.
00:13:59
Speaker 4: It's like the fixation that we have with the iPhone upgrade cycle. I know it's different, but one of the things that jumped out at me, model versions released in the six months ended September 30th, 2026, accounted for over 80% of our revenue for the three months ended September 30th. In other words, more than 80% of that Q3 revenue came from models that released in the preceding six months, like being at the frontier.
00:14:22
Speaker 1: Yeah, sorry, Matt.
00:14:23
Speaker 2: I was going to say, I think that's exactly right.
00:14:25
Speaker 8: And I think that if you look at how the model cadence has gone, Anthropic had a big model drop in September. And I think the belief is that some of this October ARR number that is blank in the S1 on page 198 will be filled in and that I think there's general consensus that that number will be strong because of the model releases that were released in September. And so I think being at the frontier does provide and generate outside return for the business.
00:14:54
Speaker 4: You know, there's market penetration and then there's also the competitive environment, right? So I'm just hitting you with stats, but price cuts from new models and contractual discounts shaped about $ 600 million off Q3 revenue. That's in page 161 of the S1A. What would that be evidence of? That they, you know, they're still in a race here.
00:15:16
Speaker 2: They're still in a race. They're in a competitive environment.
00:15:19
Speaker 8: And I think that Q3 number reflects the fact that In Q3, they did not have a large model release until towards the end of that quarter. And they faced some competition from OpenAI releasing their frontier model. And so I think that maybe put some pressure on the pricing capability of the business in Q3. And I think what is most exciting to me is the fact that they were able to navigate through that pricing environment and still generate operating income.
00:15:48
Speaker 1: Matt, I appreciate you coming on the show.
00:15:50
Speaker 9: Right.
00:15:50
Speaker 4: I get that there is a limit to what we can discuss
in practice with the S1A now out. But like Anthropic going public is a big moment, you know, for industry, for private markets, for your portfolio.
00:16:05
Speaker 1: Would you just speak a little bit about that?
00:16:07
Speaker 4: Summarize what you think this this sort of foreshadows in terms of everyone else going to the market?
00:16:12
Speaker 8: Yeah, I think it's a really important milestone because if you look back at how this year has proceeded, the market has been waiting for kind of the big three to come out. We had SpaceX go out first, Anthropic, I believe, will be public in the coming weeks here, and then OpenAI. And I think what's unique about this is these companies are not coming public with $ 100 million, $ 100 billion.
00:16:34
Speaker 2: Of market cap.
00:16:36
Speaker 8: They're coming public at north of a trillion dollars of market cap. And so I think this is just validation of the fact that the public market isn't dead. IPOs are not a thing of the past. The public market has just been sleeping. And these IPOs are starting to wake up and reinvigorate the.
00:16:55
Speaker 2: Activity in the public side.
00:16:57
Speaker 8: And I think that's one of the benefits that we have as investors here at Wellington is the fact that we can see that whole picture. We can see the private markets. We can see the public markets. And we've always been of the belief that These late-stage private companies are growing into exciting stories for the public market and are not doing it and staying trying to be perpetually private, if you will.
00:17:20
Speaker 4: Matt Withile, head of late-stage growth at Wellington Management. Thank you so much for being back with us on Bloomberg Tech. Now, coming up, the next test for the tech trade on deck, Apple results. Out after the closing bell, stock down about a percentage point, treading water ahead of that. This is Bloomberg Tech. Shares of Apple falling ahead of its earnings report today after the closing bell. Early order trends for its newest product slate and profitability measures top of mind for investors. Bloomberg's Ryan Vestelica. joins us with the preview. There's kind of like, you know, there's the corporate earnings story and then there's the stock story slightly separately. So let's start with expectations for earnings. What are the metrics that the street's looking for with Apple?
00:18:11
Speaker 3: Well, I think the major thing is that people really want to get a sense of the demand for the iPhone Duo, so the foldable iPhone. Now, I think there's some positive early indications that there is a lot of consumer interest, but how much of that reflects supply availability? How much of that reflects, you know, demand being taken from other forms of the iPhone, including the Pro and Pro Max models? If we are seeing a lot of higher-end customers, paying up for the foldable phone, I think that's going to be a very positive signal for their ability to continue growing average revenue per device
and so forth. That's going to be a major thing that people are going to be watching for in order to justify how much the stock has been moving up this year and as another potential indication for growth accelerating.
00:18:58
Speaker 4: Okay, so let's go there. You got this piece out on the Bloomberg today. Apple's $ 1 trillion rally leaves no room for error on earnings.
00:19:05
Speaker 1: There's been this run up. Tell me the stock story.
00:19:10
Speaker 8: Sure.
00:19:11
Speaker 3: So a major factor behind Apple's gain this year, it is actually the top performing member of the Mag 7. It's almost, I think, 30% or so. And a lot of that is because it is seen as an alternative to AI. So whenever there is some kind of consternation or fear about the state of the AI trade, Apple is a very common place that investors are rotating into. Obviously, it isn't spending money on CapEx with AI in the way that the hyperscalers are. So it doesn't have questions about negative cash burn. It doesn't have questions about the ROI side of things. Now, there are plenty of people who say Apple needs to be more aggressive in AI. But at the moment, the fact that it doesn't have those spending risks is something that people view really in its favor. So that's really helped the stock outperform, especially over the past month, past six weeks or so, when we started to see other AI names really taking a step back. However, this has really resulted in an elevated valuation for Apple. I believe it's trading around its most expensive... since 2007. So like I said, there are a lot of high expectations going into this report, especially since a lot of it hasn't really been directly fueled by the company's own fundamentals.
00:20:28
Speaker 4: So the point on it being at a priceless level since 2007, 2008 is that the stock's vulnerable, right?
00:20:34
Speaker 1: Just explain that logic.
00:20:37
Speaker 3: It's come up so much and the valuation is so elevated that people really think we need to see a strong report here in order to justify how much the stock has moved up. Again, a lot of the stocks advanced this year hasn't really been on signs of improving fundamentals at Apple. There's been some of that. But it is largely in response to sort of the state of the AI trade that's really been helping the stock. But someone I was speaking to for the story really said, you know, this has been an environment where Apple's fundamentals have sort of taken a seat back. There's obviously a focus on its quality characteristics, the cash flow, the sustainability and consistency of earnings. But now we're getting to a point where we really need to see what is growth going to look like from here. That's going to be the real metric that determines whether the stock can keep moving up at the way it has been. Or if people think this is so expensive, the growth isn't there or what we're looking for. And that's an opportunity to start stepping back.
00:21:34
Speaker 1: Bloomberg's Ryan Plastelica, thank you very much.
00:21:36
Speaker 4: Coming up, a $ 2 billion valuation and a mission to better connect Earth to space. Northwood CEO and founder, Bridget Mendler, on the latest funding round from that company. Markets up. This is Bloomberg Tech. Welcome back to Bloomberg Tech. Marvell has agreed to buy Lamentum at an enterprise value of $ 115 billion, one of the biggest ever tech tie-ups. Both stocks sinking on the deal, which is expected to close in the first half of 2028. Joining us, Bloomberg Deals reporter Ryan Gould, who broke the story prior to its formal announcement.
00:22:21
Speaker 1: $ 115 billion. Go through the terms. Help me understand this.
00:22:25
Speaker 8: Yes.
00:22:25
Speaker 10: So on the headline, they're saying it's $ 115 billion enterprise value deal for Lumentum. We've broken that down for an equity value of about $ 102 billion. So comfortably up there with the biggest ever, as you.
00:22:39
Speaker 1: Put it at the top.
00:22:40
Speaker 10: And I think just in and of itself, you're looking at a deal that essentially based on the premium for the unaffected, it's about 24%. So quite a big deal for Marvell, definitely among the biggest ever across M & A, just on an equity value perspective. And so this is quite a gutsy bet by Matt Murphy, the CEO of Marvell, to go out here and say, listen, we're making a big push here in optical services, the types of things that Lumentum supply, and we think it's going to be a tailwind for us in the next five to ten years as AI is built out in these data centers.
00:23:17
Speaker 1: All stock transaction.
00:23:19
Speaker 4: But when you broke the reporting that the talks were happening, it was close. A lot of people asking the question, how on earth is this going to work and be financed because of the relative market caps and the cash on each balance sheet?
00:23:35
Speaker 10: Yeah, I mean, there was no way that Marvell could have done this deal had it not been a stock deal. They only had about $ 4 or $ 5 billion on their balance sheet in cash, as you say. I think some of the sell-off you're seeing as you look at the share price there, Marvell down, Lumentum also down. Some of this took us by surprise last week when we broke the initial story because Marvell sold off by 12%. Lumentum was up and then gave away some of those gains. Now, both stocks are firmly down today on a day that's green for the NASDAQ. I think the big concern here for investors, I think, well, there are two things probably. There's one, which is just the dilution. If you're a Marvell shareholder, if you're a Lumentum shareholder, you're looking at what that means for you. There's a lot of crossover in the shareholder registers. I think that People are also starting to get their heads around. And I think the other lingering question is, you know, this is a deal that they're saying is going to close in the first half of 2028. So quite a long, long stop here. The big question you have to ask, is it
China? As far as the regulatory goes, I mean, you know, there were, there have been a couple of chips deals that have been cleared by China in the past year or so, but none as far as, you know, none as big as this, either from a sort of industry significance perspective or indeed.
00:24:44
Speaker 1: A, you know, pure dollar value.
00:24:47
Speaker 4: On an equity basis, top M & A deal of the year, largest ever in global tech, excluding the SpaceX-XAI merger, according to our data and reporting. Ryan Gould. Thank you very much indeed. I want to go to private markets, startups, venture capital. The next big opportunity in space could actually be here on Earth. Aerospace startup Northwood builds the ground networks that connect satellites and just raised $ 375 million at a $ 2 billion valuation. Delighted to welcome back to the show Northwood co-founder and CEO Bridget Mendler.
00:25:18
Speaker 1: Really rapid growth.
00:25:20
Speaker 4: In many different metrics but you are trying to get that ground network at the same pace that the satellite network's being built i guess to start how does this capital help you do that Yeah.
00:25:32
Speaker 9: Great to be back on the show, Ed. Great to see you.
00:25:35
Speaker 11: We're super excited to announce the fundraise and for all that it's going to support with Northwood going forward. We have really great partners on the equity side. We have A16Z and DFJ co-leading the $ 300 million equity portion of the deal and HSBC covering the $ 75 million debt financing. And yeah, really, really excited for their partnership. And what does it go towards? It goes towards building the network. It goes towards expanding our performance and our functionality into broader and broader pieces of the networking stack. There's a lot to solve if we want to build and bring a ton of industry on earth into space.
00:26:12
Speaker 9: And so that's what we're going forward with.
00:26:16
Speaker 4: The story, the backstory is really important to talk about, right? We've for a long time been talking about all of these satellite deployments, the launch cadence has increased. And simply put, the pace of modernization of the ground network and literally building it out has not kept pace.
00:26:34
Speaker 10: Why?
00:26:34
Speaker 1: Why was that the case?
00:26:37
Speaker 11: I mean, it's due to a lot of disruption and innovation in certain parts of the industry that has not translated to other parts of the industry.
00:26:45
Speaker 9: The ground segment, it's pretty fragmented.
00:26:47
Speaker 11: It's pretty divided between a lot of different vendors and a divided value chain that just makes it hard to disrupt. And so for Northwood, we've taken on that end-to-end value chain all the way from the hardware to putting sites and locations around the globe. We are currently deploying and building sites across nine countries right now. And then software that stitches it all together. It is a rather sizable scope. It's different than launching and making a rocket, but it's also very hard. And so that has been the big challenge that Northwood has taken on as a business. And we are really grateful for the support and partnership of both the folks that have joined the team and been building the product as well as those on the financing side.
00:27:29
Speaker 4: You and I have been talking about this for a couple of years now, and the audience will be familiar with the very big mechanically rotating satellite dishes here on Earth. They'll know exactly what I mean. You do phased array antenna. Why is that technology better? Why is it quicker to deploy, to get going on?
00:27:48
Speaker 11: Yeah, so we do phased array technology for our portal product line. We do PRISM, which is a parabolic technology for our PRISM high throughput product line. And each of those technologies have really transformed what it means to build and deploy antenna capacity.
00:28:04
Speaker 9: So we kind of clean slated it.
00:28:06
Speaker 11: We've been able to build these modular units that you can recombine together and you can ship and deploy in a really rapid fashion. So recently, we've actually been able to deploy demonstrate this deploying a equivalent of a 20 foot dish. So think of, you know, nearly two stories high, really, really large units. kind of footprint that you use to communicate with satellites that are far away, we've been able to actually deploy that capacity in a matter of days. And being able to do that with a traditional parabolic solution for that would have required months in order to resolve the construction project that would be attached to that. So really tight compression of the timeline. And we've been excited to be able to flex those capabilities with real customers on real missions They really matter.
00:28:55
Speaker 9: So it's exciting to be at this stage of the business.
00:28:59
Speaker 4: Bridget, who are those customers? Is this private commercial space? Is this the government?
00:29:04
Speaker 9: It's dual use. So the technology is used in both instances.
00:29:08
Speaker 11: And so while the exact customer that we've demonstrated those capabilities with, we can't publicly state at this time, we'll be able to share more soon and excited to release that.
00:29:20
Speaker 1: Bridget, where are you right now?
00:29:21
Speaker 4: Talk a little bit about what's over your shoulder because part of the story of you is like you've grown the space that you need a lot. You plan to add several hundred staff.
00:29:31
Speaker 1: What's going on there?
00:29:32
Speaker 9: Yeah. So behind me, we're now in the new facility. We're at our 185,000 foot facility, which.
00:29:40
Speaker 11: Allows us to scale up production and produce a lot more volume of antennas so that we can ship those out around the world and and get more capacity live and online. What it represents is kind of an evolution of our manufacturing approach as well.
00:29:55
Speaker 9: As an early stage company, we
approach things clean state.
00:29:58
Speaker 11: We totally revisited the supply chain for the space industry and got really creative about which kind of vendors we could partner with.
00:30:05
Speaker 9: And how we would build the product. And now at this level of.
00:30:07
Speaker 11: Maturity of the company, we're at the stage where we can deepen those partnerships with the vendors that we've established really good relationships with, but we can also bring more things in-house.
00:30:17
Speaker 9: So the facility does both of those things.
00:30:19
Speaker 11: We have large areas for inventory as we bring our supply in, we have machining, we have board production and assembly, we have many other in-house components. And then we also have the testing apparatuses so that we can send our systems out into the field completely ready to go and have that really aggressive deploy time that we've advertised.
00:30:42
Speaker 4: Bridget Mendler, CEO and co-founder of Northwood, $ 2 billion valuation, equity raise, line of credit. Really great to have you back on the show. Okay, coming up, we're going to speak.
00:30:54
Speaker 1: With Will Kemp.
00:30:55
Speaker 4: Meridis, co-author of the new essay. You're all talking about it on the socials. The commanding heights. Who the economic winners from AI could be.
00:31:04
Speaker 1: Spoiler alert.
00:31:06
Speaker 4: Doesn't think they'll be today's AI leaders. Really big one coming up next. This is Bloomberg Tech. We spend a lot of time talking about who's winning the AI race. Open AI, Anthropic, NVIDIA, the hyperscalers. But what if we're looking in the wrong place? Will Menides is the co-author of a new essay called The Commanding Heights, which argues that as artificial intelligence becomes cheaper and more widely available, the biggest economic gains could ultimately flow elsewhere. Delighted to say that Wilma knows joins us now. I'll be able to start with just the thesis in simple terms, right? What are the commanding heights of the AI economy?
00:31:53
Speaker 1: Absolutely. And it's good to be here.
00:31:55
Speaker 12: The commanding heights we think of as the basket of businesses that gain durable economic advantage as a result of continued scaling of artificial intelligence. We're really interested in businesses that have an economic position that becomes more advantaged, not only by dint of being able to train frontier intelligence, but being able to capture the networks and rails that that intelligence and the corresponding commerce will flow across.
00:32:19
Speaker 4: This is all predicated on AI itself becoming abundant and cheap. And so you must have conviction, Will, that that is what is happening and will happen. What are some of the data points that you look at most closely to kind of model for that?
00:32:31
Speaker 7: Absolutely.
00:32:32
Speaker 12: We believe in scaling laws constitutionally. Leopold was correct to say that intelligence will continue to scale, that the build-out will continue, and these models will get better. We believe that very, very strongly. However, the cost of an individual unit of frontier intelligence, that is a token that is at the frontier being used to power a workflow, has fallen about 13 times over the last 12 months. It's our belief that as intelligence gets better, it will become cheaper and become more ubiquitous, and that will lead to weird knock-on effects across the economy. And Commanding Heights is an essay that ultimately concerns itself about where those economic benefits will accrue, and it's our strong conviction that they may be far away from the foggy hills of San Francisco.
00:33:13
Speaker 2: Right.
00:33:13
Speaker 4: So this is the bit that I struggle to reach a conclusion of my own on for what it's worth. So it seems like the model, the underlying model, is not where the lasting value sits. So where does the money go if that's the case, if I've interpreted that correctly?
00:33:30
Speaker 1: Yeah, absolutely.
00:33:31
Speaker 12: And this is not a short frontier lab essay, right? We believe Anthropic and OpenAI are great businesses. What we're really interested in is a narrow set of businesses that we call the commanding heights. which we think about as the rate of commerce in the world increases, as the number of firms that are distilling intelligence increase, we're really in the first inning here. The number of firms that have deployed these things well is vanishingly small.
00:33:55
Speaker 1: And the argument we make in the piece is the.
00:33:57
Speaker 12: Companies that control the rails, the companies that control the distribution, to take the intelligence that's developed inside of firms like OpenAI and Anthropic and diffuse it into the real world will benefit disproportionately. We categorize this in two different categories. First is firms that just benefit by dint of having more intelligence, right? Businesses that use lots of human cognitive labor today are obviously able to see margin expansion. But with a second interesting category that we declare as the neo-dollar, the neo-dollar essentially, you know, intelligence is dollar denominated, right? AI is an American business. It's the American economy. Demand for U.S. dollar flows and demands for U.S. commerce is going to increase. So we believe there's a whole new set of businesses and the market has not adequately priced that are going to facilitate commerce in this new intelligent economy that have not adequately been identified.
00:34:45
Speaker 4: It's like this era of American industrialization, re-industrialization. The reason I find that part interesting is like, you know, you look back at history, the case study of the railroads in this country, which is fascinating. What does that period of history teach us about what is now contemporaneously sort of technological revolution?
00:35:09
Speaker 12: Yeah, very rarely is the place where technology invented also the place where that economic benefit is
captured.
00:35:15
Speaker 1: Right.
00:35:16
Speaker 12: Gutenberg invented the printing press and still went bankrupt. The railroad fortunes never built a single locomotive.
00:35:21
Speaker 1: Right.
00:35:22
Speaker 12: If intelligence is going to work, which we believe it will, it needs to be dragged kind of kicking and screaming from the share houses in San Francisco to New York and put to work in the real economy. We believe this diffusion is the real story of AI and the companies that power that diffusion will become the overwhelming chunk of the economy over the next decade.
00:35:41
Speaker 4: Will, you've said we quite a few times. I think explain what you mean by that, where the school of thought comes from. But I think a lot of people are sharing this essay on socials and, you know, people are engaged.
00:35:52
Speaker 1: They're talking about it. What prompted you to write it?
00:35:55
Speaker 4: I mean, what was the catalyst that had you sit down and go, you know, I need to look into this and I believe this?
00:36:01
Speaker 12: Yeah, I grew up inside the AI economy, right? I was 19 when the attention is all you need paper came out. I built and sold a foundation model company called Science.io. I wrote this with a good friend of mine, Joe Roberts, who is a special situations investor. And we've been working on this piece for the past year because we didn't feel like there was adequate discussion about what the ultimate economic structure of intelligence would look like. It's our view that the market is not adequately pricing these things in totality, right? I think everyone's aware that equities are mispriced as a result of this technology. But right now, we've just been bottleneck trading from trade to trade to trade, right? We'll run memory up, then we'll run storage up, then we'll run RAM up. But we have not adequately understood that the fundamental structure that underlies the U.S. economy has changed. All equities will reprice as a result of this. And it's worth suggesting a scenario, which we do in Commanding Heights, of where those economic benefits could ultimately end up.
00:36:55
Speaker 4: Do you present a range of scenarios or you just see one end destination based on where we're currently at in the development of artificial intelligence?
00:37:04
Speaker 12: I mean, we'd be crazy people to come here and say we have all the answers. The main thing we wanted to develop was this idea of commanding heights, right? These businesses that have... unbelievable positioning. An example we call out in the piece as Meta, right? Meta is an incredibly interesting business because not only can it do frontier training itself, but it also owns the rails to handle the diffusion of that intelligence to the global economy and also handle the commerce on the back end of it. We think this is massive tailwinds for businesses that have that legacy distribution and are able to capture the benefits of frontier intelligence and bring it to the
global economy.
00:37:40
Speaker 4: Will Menoudis, SVP at Veridigm, co-author of the essay, The Commanding Heights, that so many of you are talking about in real time. Thank you so much for being with us. Coming up on the rest of the show, the tech industry is preparing for a midterm election that could have an impact on AI and tech leaders' own wallets. We're going to have those details next. This is Bloomberg Tech. Midterms are just a day away, but it's not only candidates that Americans will be casting their votes for. Ballot measures across the country could have a big impact on tech, from data center moratoriums to California's Proposition 40, the so-called billionaire's tax that could hit ultra-wealthy tech leaders in this state. I want to get more from Bloomberg's Felipe Marquez. Let's start with Prop 40. Explain the basics of it, what the proposal is, and then we can get into some very recent case studies.
00:38:41
Speaker 6: Of course.
00:38:41
Speaker 13: So it's been the better part of my year covering Prop 40 already. So it's actually glad seeing the endgame on Tuesday. Prop 40 is a one-time tax, a one-time 5% tax on the net worth of the state's billionaires. It affects billionaires that have lived in California since January 1st, 2026. So moving now doesn't actually change.
00:39:00
Speaker 1: The fact that you have to pay the tax.
00:39:02
Speaker 13: And it basically encompasses all your wealth with some exceptions, especially in real estate, but it would include artwork or whatever illiquid stocks you have. It would impact your illiquid positions, and it would obviously then directly impact some of the wealthiest people in the world, including Zuckerberg, including Sergey.
00:39:22
Speaker 2: Brin, Larry Page.
00:39:23
Speaker 13: A lot of those have taken steps to distance themselves from California and actually don't.
00:39:27
Speaker 4: You and I were at the desk together looking at this, class uh c of meta shares and the idea that it was put there to give mr zuckerberg liquidity to cover a potential tax bill actually even this morning the team has broken some news on another billionaire that might be thinking about what to do.
00:39:44
Speaker 13: Here So yeah, Josh Sisko and Paige Smith broke news on Max Levchin, a firm CEO and also a billionaire, who is contemplating moving out of California if the tax passes.
00:39:55
Speaker 1: Wow.
00:39:56
Speaker 13: One of his arguments, and it's an argument that we're seeing reflected in the Zuckerberg debate you and I follow very closely, is this idea of having to divest yourself from the company you helped found.
00:40:07
Speaker 7: Yes.
00:40:08
Speaker 13: Like Max said, he would have to sell some of his founder shares to cover the tax. Yes. And he was also pointing out that especially for illiquid instruments, it's really hard to actually put a price and sell them without creating a cascading of effects in how founders are attached to the company or in the incentives package, right?
00:40:29
Speaker 1: But it's also about voting control, right?
00:40:30
Speaker 4: You were saying I have to sell a certain class of shares to get liquidity and an indirect result of that.
00:40:36
Speaker 13: How do I do that without losing voting control of the company, right?
00:40:40
Speaker 4: It's not the only tech consideration in this cycle. We just got about a minute left, Felipe. But what else is the team so focused on that's on the ballot?
00:40:47
Speaker 7: Yeah.
00:40:48
Speaker 13: So we, I mean, we are obviously looking at Proposition 40, but as you pointed out before, we're also looking at data center restrictions. We are also looking here specifically in California at the governor's race. There's not a lot of, let's just say the polls show a clear front runner on that, but there's been a lot of tech money pouring into that front runner. So we're probably going to see a very symbiotic relationship between tech and whoever gets elected. We're also seeing some tech money pouring into like local races. I mean, it's Even today, we had a story about how Anthropic is also helping fund BART, which is our public transportation system.
00:41:24
Speaker 1: Here in the city, yeah.
00:41:25
Speaker 13: Yes, who has a big budget hole that needs filling. And we're looking to raise sales taxes in California, which you already know are very, very high.
00:41:33
Speaker 1: In this state, sure.
00:41:35
Speaker 13: And the idea is to fund a campaign to raise sales taxes to... protect BART so workers can.
00:41:42
Speaker 1: Actually get to the office.
00:41:43
Speaker 4: Bloomberg's VP Marques will be here over the next days and week covering catch up Bloomberg specials coverage of the midterm elections tomorrow. That starts at 7 p.m.
00:41:52
Speaker 1: New York time.
00:41:53
Speaker 4: That does it for this edition of Bloomberg Tech. We are rebounding in equity markets up modestly on the Nasdaq 100 semiconductors under pressure. Apple earnings after the bell are a big factor here. Recap the show on the podcast. Such wide ranging conversations. So much news from Bloomberg. This is Bloomberg Tech.