1. Podcasts
  2. Bloomberg Tech
  3. Neocloud Concerns Return After Elice Selloff
Bloomberg Tech

20 Oct 2026 · 41m · 21 chapters

Episode artwork

Neocloud Concerns Return After Elice Selloff

In short

AI infrastructure “neocloud” concerns and financing risk; plus major related deals and broader tech/security headlines.

Guests

  1. Peter Elstrom (Bloomberg) — covers markets/tech; explains the Ellis IPO drop and neocloud financing/execution worries.
  2. Bailey Lipscholz (Bloomberg) — reports on Nscale/Anthropic computing deal and Nscale IPO timing/valuation disputes.
  3. Stephanie Aliaga (J.P. Morgan Asset Management) — global market strategist; discusses credit spreads, interest-rate drag, and AI ROI/earnings.
  4. Natasha Mascarenas (Bloomberg) — reports Eric Schmidt’s Anthropic stake and investment history.
  5. Kate O’Keefe (Bloomberg) — national security reporter on China AI ecosystem links.
  6. Philippe Bockendorf (ASI CEO) — discusses AI orchestration for critical infrastructure; FAA/defense contracts.
  7. Bloomberg Intelligence: Mandeep Singh and George Ferguson — preview SpaceX earnings and AI/compute strategy.
  8. Bloomberg’s Felix Gillette and Lauren Grush — Netflix and SpaceX earnings previews.

Key claims + examples

  • Ellis (South Korea data center firm) fell 32% on IPO debut, reviving doubts about neocloud buildouts; similar issues cited: Firmus IPO pullback, CoreWeave troubles, SB Energy valuation pullback, Nscale execution/power financing concerns.
  • Investors question “circular”/creative financing and whether projects will reach online compute; execution (power, GPU financing, construction) is the bottleneck.
  • Anthropic talks with Nscale for Monarch campus: second building deal; right of first refusal for third; potential up to $45B.
  • Nscale IPO valuation dispute: investors push $16–19B vs management ~$30B; SEC refile expected; possible delay to December.
  • J.P. Morgan: higher rates tighten financing; AI ROI hurdle rising (est. $100–$300/month per paying user vs ~$50 earlier); credit spreads reflect repricing risk, not universal collapse; AI enablers expected ~75% of next-year earnings growth.
  • Eric Schmidt owns ~1% of Anthropic; stake could be ~$20B at a $2T IPO valuation.
  • Security: UK MI5 alleges CGTRI ties to China’s Ministry of State Security; Z.ai co-founders worked on a CGTRI-funded project (Z.ai not flagged).
  • ASI raised $325M at $3B valuation; FAA production contract $875M; differentiates via “orchestration” of critical infrastructure systems.
  • SpaceX earnings focus: CapEx level and profitability; AI/compute rental via XAI/AI services alongside Starlink cash flow.

Written by AI. May contain mistakes.

Chapters

1:41

Podcast Introduction

Host Ed Ludlow introduces key topics and the guest for today.

“Bloomberg Audio Studios, podcasts, radio, news.”
2:06

Ellis IPO Analysis

Discussion on the significant drop of Ellis in its market debut and implications.

“Coming up, South Korea data center firm Ellis plunges 32 % in its market debut, soaking fears about the billions being poured into AI infrastructure.”
2:58

Concerns Around Neoclouds

Exploration of broader concerns affecting the neocloud sector and smaller companies.

“the country's worst opening day for a sizable IPO on record, reviving concerns about the billions being poured into AI infrastructure.”
4:00

Nscale Partnership with Anthropic

Details on Nscale's potential partnership with Anthropic for a substantial deal.

“SB Energy had to pull back on their valuation.”
5:10

Investor Sentiment on Nscale IPO

Discussion on the discrepancies in Nscale's valuation and investor sentiments.

“But now investors are questioning whether that's going to get across the line.”
6:47

Financing and Market Dynamics

Analysis of how climate, interest rates, and financing affect the AI market.

“Again, just a few days after Microsoft walked away from a pact on the other side of the pond.”
9:15

AI Build-Out Projections

Examination of projected spending versus expected revenues in AI infrastructure.

“And that matters because the AI build out is increasingly shifted from cash funded CapEx toward debt funded CapEx.”
11:20

Investor Appetite for AI IPOs

Discussion on the current sentiment and outlook for upcoming AI IPOs.

“You know, our estimates put that somewhere between$100 to$300 a month per paying user.”
14:01

Investor Sentiment and AI IPOs

Explore the current state of investor appetite for AI IPOs and market conditions.

“A lot of what's in the news cycle is IPOs around the world going badly, being postponed.”
18:21

Eric Schmidt's Investment in Anthropic

Discuss Eric Schmidt's significant stake in Anthropic and its implications.

“For former Google CEO Eric Schmidt, according to sources, the former Google exec owns around 1 % of Anthropic.”
Show all 21 chapters
20:19

SpaceX's Earnings Preview

Get insights into SpaceX's upcoming earnings report and capital expenditure.

“take a look at SpaceX ahead of earnings, which come after the market close.”
22:55

SpaceX's AI Business and Future Plans

Analyze SpaceX's AI initiatives and their potential future impact.

“And in the near term, they record no revenue for internal launches, which are over two thirds of the launches that they do.”
25:08

U.S. Scrutiny of China's AI Ecosystem

Examine the U.S. government's concerns regarding China's AI companies.

“We have clarity on how they are monetizing the compute rental side, but it's the model part and how much API revenue can they generate from Grok.”
28:06

Regulatory Concerns in AI

Discussing the implications of UK government concerns on AI and its relevance to U.S. policymakers.

“I guess like bigger picture, what's the significance of the report, right?”
29:32

ASI's Funding and Goals

Introduction of ASI and its mission in critical industries after raising $325 million.

“Okay, startup ASI, formerly Airspace Intelligence, has raised$325 million, led by Spark Capital, at a$3 billion valuation.”
30:12

Training Costs and Technology Needs

Philippe Bockendorf discusses the importance of talent and technology in scaling ASI.

“So the way you can think about ASI is it's really America's engine.”
31:30

Case Study: FAA's Air Traffic Control

Explaining how ASI's technology improves air traffic management and its significance.

“This is your technology being used for air traffic control, which is serious.”
34:10

Government Agencies and Technology Adoption

Discussion on government agencies' openness to adopting new technology in logistics and military.

“Yeah, this is proven capability, and I think this is a change you see in the government.”
35:36

Global Tech News Roundup

Covering major tech news including China's tech firms and Google's cloud issues.

“And first up, three of China's biggest tech firms, Tencent, Alibaba, and ByteDance, have ordered more than three million new NVIDIA graphics cards.”
39:27

Netflix Earnings Preview

A discussion on Netflix's upcoming earnings and recent developments affecting its growth.

“That's how much Netflix has lost in market value this year.”
42:00

SpaceX Earnings Preview and AI Integration

Discussion focuses on SpaceX's upcoming earnings report and its integration with AI.

“Right, which is something that everyone does.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On Bloomberg Tech, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:00The most effective people at work aren't working harder than everyone else. They're working smarter, inside better systems. Superhuman Go, from the makers of Grammarly, is the AI assistant that works inside every tool you already use, always ready, already aware of what you're working on. It's a teammate whose only job is to help you be better at yours. With Go working with you, you can show off what you do best. See what Superhuman Go can do at superhuman.com. Some people treat ChatGPT like some kind of smart search engine, and some use it to get work done. ChatGPT Work is a new way of working in ChatGPT that can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work.

0:46It's designed to help you move from a chaotic starting point to a reviewable first version. So all the source materials, briefs, and scattered information that you have to grind through to turn into something useful can just become something useful. Put ChatGPT to work on your most ambitious ideas and projects. Get started at ChatGPT.com by selecting Work Mode, available on Plus and Pro plans. For big business, AI opened up a world of promise, but that world of promise turned into a world of pain. Cost spiraling, data trapped, security and governance risks multiplying, and ROI out of reach. That's why enterprises turn to Boomi to connect data, apps, and AI, helping them operate securely, efficiently, and at scale.

1:32Boomi turns a world of AI pain into a world of AI gain. As Boomi says, we got you. Head to boomi.com. That's B-O-O-M-I dot com. Bloomberg Audio Studios, podcasts, radio, news.

1:56Bloomberg Tech is live from the heart of Silicon Valley with Ed Ludlow in San Francisco.

2:06Ed Ludlow: This is Bloomberg Tech. Coming up, South Korea data center firm Ellis plunges 32 % in its market debut, soaking fears about the billions being poured into AI infrastructure. Plus, Anthropik's in talks with Nscale for a massive new computing deal that could be worth as much as$45 billion. And Eric Schmidt's early Anthropik bet could pay off big with a stake worth as much as$20 billion if Anthropik hits a$2 trillion IPO valuation. It is absolutely all changed. The first hour of the session, chip stocks down, tech stocks down, broadly at one point the SOX was down 2.5 % and trading at its lowest level since late July.

2:44Ed Ludlow: In the last hour, complete reversal of that. Tech stocks pushing higher, chip stocks in particular rebounded now up 1%. It was a sell-off in chip stocks and technology that stemmed from a concern around neoclouds. South Korean data center firm Ellis plunged 32 % in its market debut, the country's worst opening day for a sizable IPO on record, reviving concerns about the billions being poured into AI infrastructure. Bloomberg's Peter Elstrom is with us. Everyone is reading about this this morning. What on earth happened with this Ellis IPO? Yeah, it's amazing how many times we've been talking about South Korea in this AI boom and where we're seeing a lot of these issues emanate from it.

3:24This time, it's not Samsung. It's not SK Hynix. It's this company, as you were mentioning, this little data center company called Ellis. So they had their IPO. They actually priced it pretty well. They raised a bit more money. They raised at the top of the range and got about$150 million. But as they began trading, shares traded down very hard. They were down, as you said, more than 30%. They closed down 32%. Now, this is part of this concern about what's going on with the neocloud push at this point and the build-out. This is more than three's a trend. You've got Firmus, of course, the data center company in Australia that had to pull their IPO.

3:59You have these troubles with CoreWeave that we've been detailing in quite a bit. SB Energy had to pull back on their valuation. And also we've heard that Nscale is probably not going to be able to hit the valuation that it wants as it goes public. So they're adding up at this point. But this time was the surprising trigger coming out of South Korea and this company, Elise.

4:18Ed Ludlow: You know, I think you're exactly right. It's actually an aggregate. Everyone's thinking about this sort of change sentiment to Neoclouds. They've lost faith. So park the quite small South Korean company to one side for a minute. There is still in Asia, in Asia trading, this lack of faith. And that's built over into the U.S. What else can you tell us, Peter? I think in particular, there are concerns about the financing behind some of these smaller companies. These are not concerns about OpenAI or Anthropic or Google for that matter. But these are the smaller companies that have gotten pretty creative with the financing.

4:52We've talked about the circular financing. Nscale, of course, is one of those examples where NVIDIA has backed the company. They have a 10 % stake in the company. and N-Scale is a huge customer for NVIDIA too. So there's a lot of financial incentive for companies like NVIDIA to sprinkle their money within the system to be able to increase sales and get these build-outs done. But now investors are questioning whether that's going to get across the line. And as we detailed in the story on N-Scale that we talked about before, they've had a hard time executing here. The issue is not getting contracts like the one that they may be able to get with Anthropic here, but their issue is they have to be able to get power connected to these sites.

5:28They have to be able to get the financing for the GPUs. Then they have to be able to build the buildings. Execution right now is really top of mind for investors.

5:37Ed Ludlow: Bloomberg's Peter Elstrom, thank you very much. And even as investors question this AI infrastructure boom, the deals keep getting bigger. Some new details. Anthropic is in talks with Nscale for a massive new computing deal that could be worth as much as$45 billion, all according to sources. Bloomberg's Bailey Lipscholz broke this story last night. It is new because there's an existing relationship between N-Scale and Anthropic. And it is counter to some of the other reporting we've done earlier in the week. But let's start with what's new. What do we need to know about this deal? The big thing to keep in mind is this is for their Monarch campus.

6:09And people at first, when we broke the story yesterday, were like, wait, wasn't this already announced? No, this is a campus that has three buildings. So the first building was inked in a partnership with Anthropic, a similar nominal announcement. That was back in August. Our reporting is that Anthropic is in talks for a similar deal for the second building in West Virginia. And they also have a right of first refusal for that third building. So to your point, this is a big deal. This is a big potential contract for N-Scale at a time when we've been writing ad nauseum about the neoclouds and kind of whether they deserve to exist or continue to grow.

6:42Well, all signs point towards a potentially big announcement coming from N-Scale that they are leaning further into that partnership with Anthropic. Again, just a few days after Microsoft walked away from a pact on the other side of the pond.

6:55Ed Ludlow: In the background, an N-Scale IPO is in limbo based on our reporting, right? You were on the show 24 hours ago telling us that N-Scale sees itself at one valuation and a group of investors are saying no, significantly lower. Remind us of that and where they stand in this IPO. So our reporting over the weekend was that N-Scale is having some investors push back and say you're worth between 16 and 19 billion dollars that stands opposite of the management kind of vocal push towards a 30 billion dollar valuation so quite the discrepancy when we're talking about a 40 haircut give or take there and what that does or doesn't mean not only for the company but for longer term private investors the interesting thing ed when i talked to folks about our story yesterday is how does this ultimately impact the math for say 2029 or 2030 ebit and how that could kind of potentially be ascribed to a higher valuation, the puts and takes still come back to the need for N-Scale to actually build Monarch 1 in the second building.

7:53That's the big risk. And it seems like Wall Street and management teams are seeing it differently in how you actually can discount whether or not these are really projects that will be fully online in a few years and actually delivering the promised compute that are in some of these deals.

8:09Ed Ludlow: N-Scale's public filing was September 18th. Today is October the 20th. What's going on in this interim period? We're expecting the company to refile updated S1. So that would put the ball back in the SEC's court. Whether that details this contract, whether it details something else, we do not know. But they need to clear those SEC comments before they can begin marketing the deal. Then really, you say about a week from there. The clock is ticking because, again, we sit here on Tuesday. In two weeks, we will have a midterm election. This is a company that does not want to go public the week of midterm, so really needs to get on file potentially this week and start that roadshow.

8:48Otherwise, all signs would point towards potentially being held up until December. I say potentially because, again, our reporting reports, our reporting focuses on Anthropic wanting to go public before Thanksgiving. So a lot of moving parts and a really crowded calendar.

9:02Ed Ludlow: Bloomberg's Bailey Lipschelz, thank you very much. Step back from those headlines. The common thread is financing, huge upfront capital needs, higher borrowing costs, construction delays, revenues that may sit years in the future. And that matters because the AI build out is increasingly shifted from cash funded CapEx toward debt funded CapEx. Our next guest writes that this shift will keep volatility elevated and the AI balance sheets are no longer created equally. Stephanie Aliaga is global market strategist at J.P. Morgan Asset Management. that was dramatic, but across the news headlines of last week, the movement in credit markets, what Bailey just outlined in that report, where do you net out in how deep the concern is in the health of this neocloud sector?

9:49Absolutely. I mean, one thing absolutely top of mind right now is that the AI build-out is not immune to the climate interest rates. And we are seeing that tighter financing conditions are having a toll in some of the more riskier and levered parts of this AI build-out. And so I think a lot of what we're seeing in spreads, yes, there is a degree of this neocloud situation becoming the focal point for AI anxieties that have been bubbling up over the course of this whole year, but it's also a reflection that we should probably reprice the risk around these companies, given that higher financing costs do challenge some of the economics around these players.

10:28That doesn't mean that they're all going to go bust, And it could be that there's maybe more pessimism than optimism priced into some of these CDS spreads, just given some of the inherent uncertainties here. But it is a reflection of just growing drag from higher interest rates on this AI build out.

10:43Ed Ludlow: Some of that pessimism stems from a Gartner report that was published October 14th, where they said expected spending this year, one trillion and then six trillion between 27 and 30, the end of the decade. So seven in aggregate. But the expected revenues generated 3.6 trillion. So they see a mismatch and they expect write downs, particularly in the Neo clouds. Right. Is that something that you're modeling for to JP Morgan Asset Management? I will say there is an inherent uncertainty that as this AI build out grows and grows. Right. The hurdle that you have to reach in terms of ROI, just mathematically, is a bit more difficult.

11:20You know, our estimates put that somewhere between$100 to$300 a month per paying user. And that is up from a year and change ago. That was like$50. So that hurdle has moved higher. Now, the demand environment for AI has also changed enormously and continues to look very constructive. But the reality is the proceeds, the revenue generation, the monetization of AI is not going to be shared evenly. But in the near term, as we do look at an AI build out that is continuing to grow and shows a lot of durability, all of that spend should be directed to various different pieces of this AI infrastructure wave.

12:02So we do think AI enablers are going to continue to account for the lion's share, roughly 75 percent of earnings growth next year.

12:11Ed Ludlow: The pessimism is also not universal. Goldman upgraded CoreWeave, saying that the debt concern was overdone. And I think at the beginning of this conversation, we pointed out not all balance sheets are the same. CoreWeave is in a very different position than some of the other neoclouds. Bailey just brought you the reporting. Nscale, potentially an additional$45 billion of capacity with Anthropic. How closely do you kind of track the progress of those kinds of deals, those commercial agreements? Yeah. I mean, I think violent moves in some of these, you know, debt and CDS spreads as of late don't necessarily tie cleanly to the fundamentals.

12:49You know, in reality, the interwovenness of this AI ecosystem can actually be a benefit to some of these players, too. If you have some of these players that are a bit weaker, they're having more challenges and raising funds. they can be supported by their much meatier, stronger AI balance sheets across this AI buildout through the strength of these partnerships. So that's one thing I think to keep in mind here. So long as the overall pie of AI revenue growth continues to grow really strongly, it really diminishes the risk of massive upsets in even some of these smaller, more fringe players. Ultimately, leverage in and of itself isn't a bad thing, but it is a riskier thing.

13:31So some of this repricing is rational again. But on the other side, there is probably some room for optimism to reenter the equation once we hear more about the demand picture of AI, once we hear progress around the build out itself and just a clear picture of what the economics of the state center build out are going to look like.

13:50Ed Ludlow: I mean, to draw a line under the credit story, the call we have 2032 notes hit a yield of 15 percent distress. But like in the beginning of October, 20 days ago, they were at 13 percent. So I feel like the credit market's new about that. A lot of what's in the news cycle is IPOs around the world going badly, being postponed. And then Anthropic is the big, like, we're all waiting for that. What do you see in investor appetite for these IPOs? Yeah, these IPOs are very key because sentiment is a very powerful thing in this market. You have a lot running on AI and a lot of high momentum parts of this market.

14:26However, it's also really important to take a step back and to just look at the broader fundamentals of this AI wave, the earnings growth at play, the strength of these players that very much need to go public. So we have clear visibility around the financials of these massively important AI companies and the build out. I think we will get those companies going public. But the details around all of this are absolutely going to be scrutinized. And I think the reality is, you know, there is this air of uncertainty of we don't know what some of these model labs ultimately mean for the broader AI ecosystem until we have that visibility into their earnings.

15:05However, I think on the debt and the equity, markets do seem very well equipped to absorb all of this new AI issuance. And that is something that is important not to lose sight of.

15:17Ed Ludlow: Stephanie, what happens next? Well, literally the midterms happen next. But I mean, what is it that you've got in your market outlook right now? We're not too worried about the midterms, I will say. These are kind of more noise than I think the fundamental drivers of this buildout. Look, this earnings season is going to give us a lot of clarity around how we're tracking around AI revenues, the monetization of the buildout overall. Also, we're going to get visibility into the personal AI assistant craze this week and what that's meant for some certain applications like Muse. and I think all of this is going to give investors a lot to digest, could give us more good news around this AI build-out to calm some of these jitters.

15:56I think that's going to be really key. And then again, just looking at the long-term and that's something that we've been talking about a lot from JP Morgan this week with the launch of our long-term capital market assumptions. We do see AI contributing to this structural climb in productivity which should allow this economy to grow at an above-trend rate of growth. And that's a bit of this rising tide lifting all ships. It's not going to be smooth sailing. There's going to be bumps along the way. But long-term investors should maintain that outlook.

16:23Ed Ludlow: Stephanie Aliago of JPMorgan Asset Management, really grateful for a really broad conversation. A lot of news headlines today. Coming up in the show, we're going to take a look at SpaceX ahead of earnings, which, of course, come after the market close. This is Bloomberg Tech.

16:43The people who seem to get more done than everyone else. They're not working longer hours or running on more caffeine. They've just stopped wasting time on the stuff that doesn't move work forward. Switching apps, re-explaining context, hunting for files. Those aren't small inefficiencies. They're hours wasted every week. Superhuman Go gives you those hours back. From the makers of Grammarly, Go is an AI assistant that sits inside every tab and tool you already use, always available and already aware of what you're working on. Ask it to draft something, summarize a long thread, pull up a file, or prep you for a meeting.

17:23Go handles it without you ever leaving the page you're on. This is what it looks like when AI actually fits into your work, instead of adding to it. It's like having a teammate whose only job is to help you be better at yours. Go keeps up so you can move forward. With Go working with you, you can show off what you do best. Superhuman Go. Find out more at superhuman.com. That's superhuman.com. Your team is on a deadline, but moving all your enterprise data will take time your business simply doesn't have. IBM technology lets AI work with your data where it already lives. With an open hybrid cloud approach designed to build on your existing investments, it's another way IBM helps you put your data to work.

18:11Let's create smarter business. IBM.

18:14Ed Ludlow: Visit ibm.com to learn more.

18:20Ed Ludlow: An early bet on Anthropic could be about to pay off big time. For former Google CEO Eric Schmidt, according to sources, the former Google exec owns around 1 % of Anthropic. from early investments, a state that could be worth as much as$20 billion if Anthropic goes public at a$2 trillion valuation. Bloomberg's Natasha Mascarenas broke the story and is with us. And it kind of reflects some of the early investment history that Anthropic has, the types of people that got in. Give me the story. Yeah. I mean, listen, up until 2023, Anthropic did not have any of those big name institutional investors that you and me are often talking about.

18:58So the first three rounds were actually mostly angels and high net worth individuals. Eric Schmidt, his first investment, I mean, tens of millions, 461 million valuation. And now we're looking at a 1.8 trillion IPO. Pretty good individual return there. Pretty good individual return.

19:17Ed Ludlow: It's also interesting, right? Because, you know, he continues to have an affiliation with Google. Right. You know, Google is a frontier lab to all intents and purposes. How normal or atypical is it for a 1 % stake in a company of like Anthropic scale? Now, I get it's paper, right? It depends how the IPO goes. Totally. Is he like one of one or are there others in that profile? I would say he's one of few, right? I mean, we know the co-founders of Anthropic are very much expected to also net out billions of billions of returns themselves. But when it comes to the individual, this is a seat that investors usually like to play.

19:55And so for an individual to have this is rare for Eric Schmidt to have this is rare. I mean, he has made many investments before in companies that you and me cover and follow. But Anthropic is by far the most prominent private investment we've seen him make in a very long time.

20:10Ed Ludlow: This is absolutely something everybody is reading about this morning. Catch it on the terminal and on Bloomberg.com. Bloomberg is Natasha Mascarenas. Thank you very much. Coming up, we're going to take a look at SpaceX ahead of earnings, which come after the market close. This is Bloomberg Tech.

20:38Ed Ludlow: SpaceX's third quarter earnings will be out after the bell later today. The report comes after sending a new crew into space, helping Tesla show off a rocket-powered car, all while pouring billions into AI infrastructure. Bloomberg Intelligence's Mandeep Singh and George Ferguson are here for the preview roundtable. Mandeep, I'll start with you. So SpaceX is a quasi-neo cloud, a frontier lab, and it has a chatbot and an AI agent. And I'll go to you first to tell me what we expect to be told about that tonight from SpaceX. Yeah, look, they do have a lot of contracts when it comes to renting the compute capacity, and they have given us guidance for$100 billion in ARR by the end of the year.

21:20So from that perspective, I do think they have visibility into that ARR. But the key question is how much is going to be the 2027 CapEx number? And given all the jitters we have seen with CoreWeave, how will investors look at a 150 billion plus CapEx number for someone like SpaceX, given they don't have the operating cash flow that the other hyperscalers have?

21:46Ed Ludlow: Q2 CapEx was 18.4 billion. The expectation on consensus is 19.6 billion CapEx for this quarter. George, SpaceX, it's a space company, rocket company. That's the bit you're modeling for. And I look at the breakdown of revenues and space is the third bucket way behind AI and connectivity. What are you looking for? Yeah, so like you said, it sounds like a space company, but I'd say it's not a space company. It's all about sort of that AI business and turning that into a very profitable enterprise. In the near term, the connectivity business is generating most of the profits. That's the Starlink satellite system.

22:30And space, the portion that I focus on the most, is really just the enabler of those two businesses. It's going to enable Starlink in the near term as they try to generate as much cash and profit as possible to help offset some of those challenges Matt Deep was just talking about. But in the long term, I guess they're going to put data centers in space, which probably fix a lot of problems on the world. And in the near term, they record no revenue for internal launches, which are over two thirds of the launches that they do.

23:03Ed Ludlow: And so you use the word Georgia enabler that we're talking specifically about Starship, which is the modus operandi by which all of this future vision is achieved. Right. Yes. Yes. Starship really gets them the size they need. i think to put both data centers in space and the larger satellites that will help starlink get better connectivity with the ground right i think we saw spacex make some moves in in the last uh just a couple weeks i think it's actually about a week ago time seems to fly anymore when they bought some uh you know broadband wavelength so they can try to be able to use starlink inside buildings And so, you know, there's lots of developments.

23:48But again, I think the space business right now, it's progressing well, but it's the enabling.

23:56Ed Ludlow: We're also looking at the carriers, right, because Gwynne Shopwell, SpaceX president, could not have been more clear that they want to take customers from those carriers with direct to mobile. Mandeep, when I read your research, we put SpaceX AI, XAI part of it, side by side with the Frontier Labs, Google, et cetera. Where do you look most closely on how they sell AI? You know, the subscriptions for a chatbot, GrokBot agent, and now Cursor bundled in there too. Yes, so Cursor is an asset we do expect to ramp up fairly quickly in terms of the growth rates. And look, we know Coding Agent is one of the most successful use cases of generative AI.

24:38So from that perspective, they should benefit from that trend. But I would point to a comparison with Meta, which has a Muse app now, which is at the frontier. So from that perspective, I feel Grok still hasn't reached that level of capability on its own. And I think that's where investors would want to hear more around the training spend. from XAI and where is Grok when it comes to the frontier capabilities and how they plan to monetize the model part. We have clarity on how they are monetizing the compute rental side, but it's the model part and how much API revenue can they generate from Grok.

25:17I think that would be a fine risk.

25:20Ed Ludlow: Okay. SpaceX third quarter numbers coming after the bell. The big watch is CapEx, maybe for the full year 27, but revenue is expected to get to almost 13 billion driven by AI. Mandeep Singh and George Ferguson of Bloomberg Intelligence, thank you so much for joining us. The U.S. is blacklisting Miracle Plus, a Chinese venture firm founded by former Microsoft AI executive Lu Qi, over its investment in a startup accused of stealing Google AI technology. The sanctions could include freezing U.S. assets and banking restrictions, as Washington intensifies scrutiny of China's AI ecosystem. Four founders of one of China's AI companies, Z.ai, worked on a study with a technology institute that the UK government is accusing of having ties to China's main spy agency.

26:10Ed Ludlow: Joining us now in Washington, D.C., is Bloomberg's Kate O 'Keefe, who covers the US and China tech from a sort of national security perspective. This is a fascinating story. Let's start with the basics of what Z.ai actually is. right ed so zai might not yet be a household name here in the u.s but in china it's it's huge it's one of their most promising artificial intelligence companies and was actually the first of the major ones to list publicly in hong kong where its stock is trading now uh there are other parties involved here and if i don't get this right correct me but there is cgtri What is that and what is going on?

26:51There are other parties involved. Right. So CGTRI is an institute that until very recently, probably no one had ever really heard of. But that changed on September 30th when Britain, the UK's MI5 put out an espionage alert warning people that CGTRI has very close ties to China's Ministry of State Security, which is its main spy agency.

27:18Ed Ludlow: So that's the UK government's concern. It's around CGTRI and its links to that agency. The news or the development is this body of work that CGTRI and ZAI are doing together. Why are they doing that? What are they looking at? Right. So the story is specifically about the ZAI founders. Four of them, it turns out, participated in a research project that was co-funded by CGTRI. And that project was four years long and it overlapped with the founding of the company, ZAI. To be clear, ZAI is not actually flagged in this espionage alert. But what we're reporting on is the connection here between these four founders who worked on a project that CGTRI was involved in.

28:05Ed Ludlow: A lot of people are reading about and clicking on the story this morning. I guess like bigger picture, what's the significance of the report, right? Like how this is about a UK government concern in particular, but like what is the relevance to what's happening in Washington? Like why might there be an interest in this? Right. I mean, I guess when I when I think about the story, I think about how much the U.S. is right now struggling to figure out how to regulate U.S. AI companies. Right. I mean, every day we're sort of hearing news about that. And now I don't think people, policymakers have really even started thinking that deeply about how to consider China's AI champions.

28:45And so I think anything that sort of sheds light on the people involved in China's AI ecosystem will be very relevant to U.S. policymakers. Because, of course, AI is global, right? It's not like these models are just locked up in China. So it's something that everyone really needs to think about. And so when you think about the fact that ZAI's co-founders had a relationship through this research project with an institute that we now know is an alleged front for Chinese intelligence, I think that is definitely something that people will take notice of.

29:23Ed Ludlow: Right. It's one of the things people are reading about the most on Bloomberg this morning. Kate O 'Keefe, Bloomberg News, thank you for bringing us the story. Okay, startup ASI, formerly Airspace Intelligence, has raised$325 million, led by Spark Capital, at a$3 billion valuation. The company builds AI software for operating critical industries, including national security, transportation, and energy. ASI founder and CEO, Philippe Bockendorf, is with us. It's great to have you on the program. Thank you for being with us. People know you because of the work you've done with the FAA, and we'll get into how that works in practice.

30:00Ed Ludlow: but this is a pretty big vote in confidence from your investors. What are you going to do in the first instance with this capital? Because you're already out there in the world. You're already doing real deals. Well, absolutely, Ed. So the way you can think about ASI is it's really America's engine. We built the AI orchestration systems for the critical industries, from logistics to defense to energy. So, yeah, we're really excited to announce this round today. It's at a$3 billion valuation,$325 million. to scale our work across America's most critical industries, no matter if that's energy, defense, or transportation.

30:37For you, the technology, your core competence,

30:41Ed Ludlow: this is an issue of training costs and running the platform costs. It's talent. What do you need most right now to continue growing? Yeah, we're seeing an enormous amount of demand out there. I think it's really a function of the world being increasingly more dynamic, but then also if you look at the critical industries that really run on a mix of like 1980s technology and software great technology 30 years ago but not necessarily in 2026 and then a lot of consumer ai or dashboards rearview mirrors and we're providing the technology to the operators where they have foresight okay so that's interesting even if it's an agency or a commercial partner that is using a chatbot right the whole point is that the capabilities of that technology are not as predictive as ASI.

31:30Ed Ludlow: Explain the FAA case study to me. This is in the real world. This is your technology being used for air traffic control, which is serious. That's consequential for Americans flying around this country. Yeah, it's a great example on how orchestration AI can work, right? So just to be clear, we're not, like, tactically separating aircraft. Those are separate, different systems. But we are orchestrating the system as a whole from the airport to the airspace to the traffic flows. So just a few months ago, we were awarded an$875 million production contract. And just to put that in perspective, over the last 30 years, only four technology companies have been founded that were awarded a contract of this size.

32:10So it's pretty meaningful.

32:11Ed Ludlow: Palantir was in the running for that contract, and you were able to get there in the end. What would you say is the point of differentiation, the defining feature of your technology that it won out in the end? Yeah, I would say ASI is really a company one-off-one. The way we look at the AI landscape, I would say you have three categories. So one category is probably the one that's most talked about. It's really the generation of text, images, and videos. And then you have AI for autonomy, no matter if that's self-driving cars or the autonomy that goes on a drone for the battlefield or the humanoid that eventually might do your laundry.

32:51And we think there's a third category, which is really the orchestration of the most critical infrastructure, the most complex systems, no matter if that's powering the energy grid, air traffic systems, transportation, logistics, military operations. And that's really where we specialize as a company.

33:09Ed Ludlow: Spark led the round, but Kraft Ventures, Andreessen Horowitz, Affiliates of Citadel Securities all participated. Are you in this place in the company's growth where you just need the capital or you need more sort of operational help from those investors that are on the cap table? Yeah, we're scaling an enduring company here. We've always been a very robust and enduring business. And I mean, a lot of the investors that you see in this round know the company for a long time. Spark led our Series A back in the day. Now they're doubling down in this round. I think it's one of the largest checks their growth team has ever written.

33:46And recent Horowitz, who led our last round, is a returning investor in this round. And I think it's really a testament to the robustness of this company.

33:56Ed Ludlow: So, look, we're showing some footage here which speaks to the other area you're working on, which is with the U.S. military. There are also some recent case studies where this is real. Your technology is being used. Just explain that. Yeah, this is proven capability, and I think this is a change you see in the government. Instead of relying on research work that might take five or ten years, the government is acquiring proven capability from the private sector. We are deeply dual-use, dual-purpose. I mean, if you just look at the sectors— You mean civilian and military. Exactly. I mean, just if you look at the sectors where we operate, transportation, there's an interplay between the government and the airlines, right, through air traffic control.

34:37energy same between the utilities and the local governments and then look at military logistics a lot of the logistics capability and and capacity actually comes out of the private sector and is

34:48Ed Ludlow: then used by the government you also work with the private sector right some of the commercial airlines but what is your sort of assessment about present day government agencies willingness to commit to a new technology platform you talked about how in the faa context right some of those systems are 30 years old, but you're able to do business with different arms of government and actually get things done? Yeah, I think this is really what we've seen over the last year. I mean, in the beginning of the year, we scaled our work with the Department of War very significantly through Operation Epic Fury.

35:20Then we were awarded the FAA contract you just mentioned. And then just a couple of weeks ago, we were awarded a$99 million contract with the U.S. Army.

35:27Ed Ludlow: Philip Buckendorf, founder and CEO of ASI, really grateful to have you with us here today. in San Francisco. Thank you. There's lots of other news out there around the world. It's time for talking tech. And first up, three of China's biggest tech firms, Tencent, Alibaba, and ByteDance, have ordered more than three million new NVIDIA graphics cards. Reuters is reporting, citing people familiar with the matter. The new RTX Pro 5500s are not the highest end chips Washington bars from exports to China, though they can support AI services. NVIDIA said it cannot guarantee that it can fill the orders.

36:00Ed Ludlow: Reuters put the sticker price of that deal at over$35 billion. Plus, a power outage affecting Google Cloud knocked out some services for banking firms in the UK. Digital Bank Revolut, HSBC and Lloyds were impacted. On its status page, Google customers in the London region could try to migrate their services to other cloud regions. And Australia is set to sue five platforms, including TikTok and Facebook, for violating the country's landmark social media ban for children. The law requires companies to take steps to stop underage users from holding social media accounts. But a report from Australia's online safety regulator earlier this year found underage users were able to keep or create new accounts.

36:43Ed Ludlow: Australia's communications minister, Anika Wells, spoke with Bloomberg. We do seek to enforce our Australian laws. They say that they wish to uphold them. The safety commissioner isn't convinced. There are five investigations on foot. and now with the laws that we recently strengthened in the Parliament a few weeks ago now, she has more powers to be able to gather the evidence that she needs to take those five companies to the federal court. So can we expect a court case against those five companies? Yes. It's not my decision, but I would expect it. Representatives for the companies either declined or did not respond to requests for comment.

37:24Ed Ludlow: Okay, coming up, lots of folks going to be watching Netflix this afternoon, third quarter earnings, that is. We're going to preview that next. There's a lot to talk about. This is Spoonbook Tech.

37:43The people who seem to get more done than everyone else. They're not working longer hours or running on more caffeine. They've just stopped wasting time on the stuff that doesn't move work forward. Switching apps, re-explaining context, hunting for files. Those aren't small inefficiencies. They're hours wasted every week. Superhuman Go gives you those hours back. From the makers of Grammarly, Go is an AI assistant that sits inside every tab and tool you already use. Always available and already aware of what you're working on. Ask it to draft something, summarize a long thread, pull up a file, or prep you for a meeting.

38:24Go handles it without you ever leaving the page you're on. This is what it looks like when AI actually fits into your work, instead of adding to it. It's like having a teammate whose only job is to help you be better at yours. Go keeps up so you can move forward. With Go working with you, you can show off what you do best. Superhuman Go. Find out more at superhuman.com. That's superhuman.com. Your business operates under pressure and you need control over your data. But IBM research shows only 9 % of companies have real oversight of their technology. IBM's open hybrid cloud approach is designed to give you visibility with built-in governance tools, data sovereignty, and compliance reporting.

39:08It's another way IBM helps you stay in control of your data. Let's create smarter business. IBM. Visit IBM.com to learn more.

39:26Ed Ludlow: Let's take a look at today's big number,$135 billion. That's how much Netflix has lost in market value this year. The stock's down 25 % in 2026, putting Netflix among the 10 worst performers in the tech-heavy Nasdaq 100 index. ahead of its earnings later today, shares completely flat and treading water. But there's been so much this week in the Netflix news cycle. Let's talk about what to expect and some of those news items with Bloomberg's Felix Gillette. There's a lot of concern about Netflix and slowing sales, sales as a measure of growth, and it all stems back to content. What do we need to know?

40:05Well, obviously, they made big headlines this week by dismissing their chief creative officer, the head of their film division. But the realities of TV and movie development is they're going to be stuck with the previous leadership's film slate and TV slate for the next year. So I think the question is, what else are you going to do strategically to try and boost engagement on the platform, particularly in the US and Canada where it's been lagging?

40:28Ed Ludlow: Right. The other headline was a week ago when Netflix announced it would fire 6 % of global staff. That's about a thousand people. So like senior departures is quite normal. Netflix happens time to time, but like that is a quite wide layoff. What is that signal about the action they're taking? I mean, I think part of it is that the question is, you know, they've proven they can boost engagement with live sports, but that's incredibly expensive. And how are you going to keep the margins up over 30 % while also boosting engagement without breaking the bank? I think, you know, they've tried these experiments with, you know, putting on podcasting and creative creators content, cheaper programming.

41:07That doesn't seem like it's moved the needle. You know, 2022, when they went through their last slump, they made some big changes. They introduced advertising to the platform, which is something they said they'd never do. They did the big password sharing crackdown. So would they be willing to make other big changes this time around? I think that's what people are waiting to see.

41:25Ed Ludlow: I just got through The Gentleman, season two. Like, that was good. That was good TV. But my understanding is the problem they've got is carrying the audience over from season one to season two or something. Yeah, they're having a hard time bringing back the audiences. One thing they've always said they wouldn't change was how they roll out new episodes, right? They have this binge model. They put all the episodes up at once. Is that something they're willing to consider changing, going back to more of a traditional model, rolling out one episode per week to kind of build up engagement and interest and momentum around a series?

41:58I think that's something they might be considering.

42:00Ed Ludlow: Right, which is something that everyone does. You know, Apple TV, find out on HBO Max, Paramount, etc. Bloomberg's Felix Gillette, busy evening for you guys. Thank you very much. The other one many will be watching is SpaceX's third quarter earnings. The upcoming report follows an increase in AI spending. Another crew sent into space. A lot of activity on the AI side. His talk in preview is Bloomberg's Lauren Grush, who leads our coverage of the company. Again, it's been difficult for you and I, right? This is really becoming an AI company. What's your main focus for when the earnings hit this evening?

42:32Well, I think the big question is, you know, will this actually be a profitable quarter for them? I think one of the most staggering things we found when they did go public this summer is just how much money the company was losing. This supposedly very successful launch in satellite company, the numbers just weren't adding up. So it will be really interesting, especially since a lot of folks, including myself, maybe, questioned the merger about why they would merge with XAI, I guess, SI now. But now we might actually see that merger pay off as a lot of that revenue that we're expecting is coming from that side of the business.

43:09Ed Ludlow: Actually, thank you for that. We're 54 minutes into the show. And I think I've said SpaceX AI quite a few times. They went with SI after the president. I mean, you're right. The expectation is 15 cents for EPS and a positive net income. But that comes from Starlink. So like Even if most of revenue is AI, the connectivity piece is the cash cow. Is that right? It is the cash cow. But I think we might actually see revenue from the AI or SI side. I'm sorry. I'll stop. It's OK. Actually surpass the connectivity business. That's not to say that the connectivity business isn't a heavy hitter. It's still expected to bring in, I think, over$4 billion.

43:46It'll increase upon last quarter. but it's possible we might actually see the AI business surpass that this time around.

43:55Ed Ludlow: So there is a connect here, which is SpaceX wants to boost the constellation of Starlink, which is constellation-based internet, do direct to cell, have space-based data centers, and Starship, the next generation launch platform, is the enabler. So we're still tracking progress there, right? Yeah, I feel like I'm constantly beating the drum on Starship. It's the thing that will unlock all of these big ambitious projects of the future. And so it has to work. SpaceX has put it in the critical path for all of the things that it wants to do with its future. And so I really would love to see more of an in-depth update.

44:29We got something from Musk a couple of weeks ago about what happened during that last launch that actually went to orbit for the first time and put upgraded Starlink satellites in orbit for the first time. It wasn't a slam dunk of a mission. There were some issues with an engine out, so I'd like to hear a little bit more about that. But definitely want to know about what the plan is for going forward, how they're going to keep developing the vehicle. And then, of course, you know, there's so much more that Starship has to do for NASA and for SpaceX's long-term ambitions of going into deep space.

45:01How are they going to do that test regime while also launching this massive Starlink constellation and StarMine constellation for their AI Orbital Data Center project? It's a lot that Starship has to do.

45:14Ed Ludlow: It's a lot. And we're going to be doing the live blog on the terminal online. So join Lauren, me and the team for that. Lauren Grush of Bloomberg News. Thank you very much. How does it for this edition of Bloomberg Tech? Again, this is what markets look like. In the first hour of trading this Tuesday, everything was down and quite a lot. The stocks was down to an half percent. Then there was a change in. The world of business is constantly evolving and Comcast business keeps you totally in step with secure AI back networking in more than 100 countries. They're powering over 90 % of the Fortune 500 and millions of small businesses.

45:46That's a lot of muscle. And behind it all, thousands of experts answering your call at 2 a.m. like it's 2 p.m. One partner powering how business gets done for companies around the globe. When you add it all up, no one does business like Comcast Business. Most meeting note takers get the easy words right and the important ones wrong. The client's name, the acronym, Whisperflow Notetaker is built to get the hard words right. It takes notes on any meeting platform, Zoom, Google Meet, Teams, even WhatsApp calls, keeps every meeting searchable, and connects to Claude or Chat GPT so you can ask, what did we agree to last Tuesday?

46:24Free to try on Mac and Windows at whisperflow.ai. That's W-I-S-P-R flow dot A-I. This message comes from Edward Jones, where they believe rich is about more than saving and investing. It's also taking care of what gives your life meaning. That's why your dedicated financial advisor meets you where you are with personalized financial strategies that help protect what matters so you can preserve your progress while creating a path forward. The key to being rich is knowing what counts. Edward Jones, member SIPC. If you're a business owner, you need a card that can keep up. With the Delta SkyMiles Reserve Business American Express card, everyday purchases transform into miles you can use towards future travel.

47:07So business as usual can lead to a well-deserved getaway. And for a limited time, card members can earn 200 ,000 bonus miles, the highest miles offered this year. Whether you're traveling for business or planning your next adventure, now's a great time to make every purchase go further. Apply today at www.delta.com slash podcast. Offer ends November 4th. Minimum spending requirements and terms apply.

From the publisher

Bloomberg’s Ed Ludlow breaks down why fears about the AI trade are resurfacing after South Korean data center firm Elice plunged 32% in its market debut. Plus, Anthropic is in talks with Nscale for a massive new computing deal that could be worth as much as $45 billion; and Eric Schmidt's early Anthropic bet could pay off big, with a stake worth as much as $20 billion if Anthropic hits a $2 trillion IPO valuation.

See omnystudio.com/listener for privacy information.

More from Bloomberg Tech

All episodes