# Chip Selloff Spreads Globally, Apple Makes Smart Home Push

An episode of [Bloomberg Tech](https://invisiblenetwork.ai/podcast/1611767434), 13 Oct 2026, 44 min.

Audio: https://podtrac.com/pts/redirect.mp3/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/41764a4f-fc64-4e11-89ba-ae7c0030ab5e/f41bd3aa-7bc1-475d-b784-b4e20115be0d/audio.mp3?utm_source=Podcast&in_playlist=9caafc41-289c-4115-995d-ae7c0030ab75

Bloomberg’s Ed Ludlow breaks down the accelerating chip selloff as pressure spreads from Asia to the US. Plus, Apple releases a new line of smart home devices including its HomeView Hub, the centerpiece of its push to challenge Amazon and Google in that space. And, Meta releases a new AI model which it says has "frontier capabilities" as it vies to take on OpenAI and Anthropic. See omnystudio.com/listener for privacy information.

## Transcript

[0:02] Speaker 1: Bloomberg Audio Studios.
[0:04] Speaker 2: Podcasts.
[0:05] Speaker 3: Radio.
[0:06] Speaker 2: News.
[0:13] Speaker 4: Bloomberg Tech is live from the heart of Silicon Valley
[0:17] Speaker 4: with Ed Ludlow in San Francisco.
[0:22] Speaker 2: This is Bloomberg Tech.
[0:24] Speaker 3: Coming up, the chip sell-off is ramping up with pressure
[0:26] Speaker 3: spreading fast from.
[0:27] Speaker 2: Asia to the US. We'll discuss what's behind it.
[0:30] Speaker 3: Plus, Apple releases a new line of smart home devices,
[0:33] Speaker 3: including the HomeView Hub, the centerpiece of its push to
[0:37] Speaker 3: challenge Amazon and Google in that space. And Meta releases
[0:41] Speaker 3: a new AI model, which it says has frontier capabilities,
[0:45] Speaker 3: part of its push to take on OpenAI and Anthropic.
[0:50] Speaker 2: A global chip sell-off's accelerating.
[0:52] Speaker 3: The Philadelphia Semiconductor Index is on course for a fifth
[0:55] Speaker 3: straight session of declines, matching the losing streak it posted
[0:58] Speaker 3: in July. Overnight, selling spread across Asia. South Korea's stock
[1:02] Speaker 3: market plunged. Samsung fell 7%. SK Hynix dropping 10%. The drivers,
[1:08] Speaker 3: fresh competition in the memory space from China. Rising concerns
[1:11] Speaker 3: about the cost of AI. Growing doubts about how long
[1:15] Speaker 3: the boom can last. Global tech editor Peter Elstrom joins
[1:18] Speaker 3: us to try and piece together what's going on. The
[1:20] Speaker 3: pressure is in semiconductors right now. Some of what happened
[1:23] Speaker 3: in Asia is specific to the memory space.
[1:26] Speaker 2: So try and summarize what we're seeing.
[1:29] Speaker 5: Yeah, there are quite a few factors going on here.
[1:32] Speaker 5: Of course, the general setting is that there's some nervousness
[1:36] Speaker 5: around this AI trade that's been driving the stock market
[1:38] Speaker 5: for so many months. You know, we saw it go
[1:41] Speaker 5: up for a long period of time. The memory companies
[1:43] Speaker 5: were big beneficiaries of that. We saw SK Hynix, of course,
[1:46] Speaker 5: on a tear, and Samsung to earlier this year, but
[1:50] Speaker 5: that's come back partly because of concerns that memory prices
[1:52] Speaker 5: are getting too high, that there'd be some sort of
[1:55] Speaker 5: pullback in memory. And now one of the more recent
[1:57] Speaker 5: developments is this week, the CXMT, the fourth player in
[2:03] Speaker 5: the memory chip market, the Chinese company, came out and
[2:06] Speaker 5: talked about new innovations that they've got in packing memory
[2:09] Speaker 5: chips together, and it looks like that added to some
[2:12] Speaker 5: of this selling pressure, especially today. What they're talking about
[2:15] Speaker 5: is a technique where they go from something called four
[2:18] Speaker 5: F squared or go from six F squared to four
[2:21] Speaker 5: F squared in particular, the idea is that you can
[2:23] Speaker 5: essentially pack more memory into smaller spaces. It's sort of
[2:26] Speaker 5: like using those vacuum packing techniques in your suitcase. You
[2:30] Speaker 5: can put in the same number of clothes, but it's
[2:32] Speaker 5: going to take up less space. So 4F squared is
[2:34] Speaker 5: something that CXMT said they expect to introduce later on
[2:37] Speaker 5: this year. This is something that Samsung has not talked
[2:39] Speaker 5: about using until 2028. So it looks like they're pushing ahead.
[2:44] Speaker 5: CXMT did not provide a whole lot of evidence that
[2:46] Speaker 5: they are able to do this, but it's a sign
[2:48] Speaker 5: that China is really innovating in some of these key technologies.
[2:51] Speaker 2: What an analogy that was with packing.
[2:56] Speaker 3: The point you ended on there is that Samsung and
[2:58] Speaker 3: SK Hynix haven't got there yet. Is CXMT a real competitor,
[3:04] Speaker 3: a real threat? Is there any evidence that they're kind
[3:06] Speaker 3: of chipping away proportionately in the memory market?
[3:11] Speaker 5: Well, CXMT is certainly investing very aggressively. They've got a
[3:15] Speaker 5: lot of backing domestically. They've got a lot of users
[3:18] Speaker 5: within China, which is the biggest chip market. They do
[3:20] Speaker 5: not have access to the most advanced chip making machines,
[3:23] Speaker 5: of course. They can't go to ASML and get the
[3:25] Speaker 5: EUV machines that we've talked about in the past. So
[3:28] Speaker 5: on that front, there's certainly a couple of steps behind.
[3:30] Speaker 5: But they are innovating with some of these design techniques.
[3:33] Speaker 5: And their goal, of course, is to compete with Samsung
[3:35] Speaker 5: and SK Hynix and Micron and try to be a
[3:38] Speaker 5: real player here. That comes on top of the other
[3:40] Speaker 5: innovations that we've seen come out of China, like Huawei
[3:42] Speaker 5: talking about its own chip innovations. And then you've got
[3:45] Speaker 5: companies like DeepSeek and AI, too. So I think the
[3:47] Speaker 5: broader message here is that China is innovating very aggressively
[3:53] Speaker 5: and making some real progress.
[3:55] Speaker 2: Bloomberg's Peter Elstrom. Thank you very much.
[3:57] Speaker 3: That sell-off in Asian chip stocks underscores a bigger question
[4:01] Speaker 3: facing the AI trade. Are expectations simply getting too high?
[4:05] Speaker 3: I want to bring in Christina Hooper, Chief Market Strategist
[4:08] Speaker 3: at Mangroove. And Christina, I know you're kind of in
[4:09] Speaker 3: the camp that there are several data points that serve
[4:12] Speaker 3: as evidence that we could have some kind of re-rating
[4:16] Speaker 3: in AI-related stocks. In the here and now, that is
[4:19] Speaker 3: specifically in semiconductors is where we see the movement. What
[4:22] Speaker 3: are those data points that you're most focused on?
[4:25] Speaker 6: Well, first and foremost, of course, it's yields, right? What
[4:29] Speaker 6: we've seen is yields on the long end have gone
[4:31] Speaker 6: up quite dramatically And that has taken the wind out
[4:34] Speaker 6: of the sails, in my opinion, of the AI-related trade somewhat.
[4:40] Speaker 6: And what we know historically is that typically when yields
[4:42] Speaker 6: go up on the long end, they do exert downward
[4:45] Speaker 6: pressure on long-duration equities like technology. So that is, to me,
[4:51] Speaker 6: a very powerful force. And then, of course, we've always worried,
[4:56] Speaker 6: there's always been in the back of investors' heads the
[5:00] Speaker 6: idea that we could have another deep-seek moment. In fact,
[5:04] Speaker 6: we could have a number of other deep-seek moments. And
[5:06] Speaker 6: this seems to be what's happening right now in the
[5:08] Speaker 6: chip industry, where we're hearing about perhaps some pretty significant
[5:13] Speaker 6: innovations that could cause pricing pressure. So suddenly the outlook
[5:20] Speaker 6: is not as positive for some of the mainstream chip
[5:26] Speaker 6: makers that have been such a big part of this
[5:29] Speaker 6: AI build-out.
[5:31] Speaker 3: I want to go more broadly than the chips themselves.
[5:34] Speaker 3: You're in the camp of people that believe there are
[5:37] Speaker 3: reasons that we could see AI capex slow. This is
[5:40] Speaker 3: something that's come up a lot actually in the last
[5:42] Speaker 3: seven days. What are your reasons?
[5:45] Speaker 6: Well, of course there is financing. What we're seeing is
[5:48] Speaker 6: the hyperscalers are relying increasingly on debt to finance the
[5:52] Speaker 6: build out. But more than that, I believe it's community opposition, right?
[5:57] Speaker 6: We have seen, and I've heard a number of pollsters
[6:00] Speaker 6: say this, they've never seen an issue where there's been
[6:03] Speaker 6: such a dramatic change in sentiment among the members of
[6:07] Speaker 6: the public in such a short period of time as
[6:09] Speaker 6: they've seen with data centers. So now you have a
[6:12] Speaker 6: pretty significant opposition And of course, there is the potential
[6:17] Speaker 6: for greater regulation, perhaps not at the federal level, but
[6:21] Speaker 6: certainly zoning regulation, all kinds of things that could really
[6:24] Speaker 6: impact the build out, at least in the United States.
[6:28] Speaker 3: I actually am going to reflect on something that Jamie
[6:30] Speaker 3: Dimon said during JP Morgan's quarterly earnings call. And I'm paraphrasing,
[6:35] Speaker 3: but I'll just read it for you. He was asked
[6:37] Speaker 3: about data centers, right? And he basically argues that data
[6:42] Speaker 3: centers are not affected by the cost of money and
[6:45] Speaker 3: that the value of the data centers is high enough
[6:48] Speaker 3: that to pay high interest costs. If JP Morgan and
[6:53] Speaker 3: Jamie Dimon are addressing that in their earnings call, what
[6:55] Speaker 3: does that tell you about the questions we are posing
[6:57] Speaker 3: in the here and now?
[6:58] Speaker 2: And your thoughts on his point of view?
[7:01] Speaker 6: Well, I do worry that there is the potential for
[7:05] Speaker 6: rather swift obsolescence with regards to data centers. that we
[7:10] Speaker 6: could see state-of-the-art technology going in, but that could be
[7:14] Speaker 6: rather old and may depreciate quite significantly in just a
[7:20] Speaker 6: few years. So there could very well be a mismatch
[7:24] Speaker 6: in terms of the lifespan of a data center, or
[7:27] Speaker 6: at least critical components of a data center, relative to
[7:32] Speaker 6: the debt that is being issued for those data centers.
[7:37] Speaker 3: One area that you are looking at closely is cybersecurity.
[7:42] Speaker 3: If you take the trajectory of some of those corners
[7:45] Speaker 3: of the AI trade that you've just outlined, you think
[7:48] Speaker 3: cybersecurity may be an opportunity here.
[7:51] Speaker 6: Well, what we're seeing is that increasingly it looks like
[7:55] Speaker 6: at least some of the cybersecurity names or a complement
[7:58] Speaker 6: to AI as opposed to vulnerable to being replaced by
[8:04] Speaker 6: AI LLMs. So I certainly think that is an evolution
[8:09] Speaker 6: in thinking. Beyond that though, I think cybersecurity as a
[8:14] Speaker 6: spending category for companies is significant and is likely to grow.
[8:19] Speaker 6: When we look at surveys of CEOs in terms of
[8:22] Speaker 6: the risks, the concerns that are top of mind for them, hacking,
[8:27] Speaker 6: the kind of cybersecurity threats that they face are always
[8:33] Speaker 6: top of the list. And so, especially given the reputational
[8:37] Speaker 6: risk that can come with a cyber attack, I think
[8:41] Speaker 6: that we're going to see rather robust budgets for cybersecurity
[8:45] Speaker 6: spending for corporations.
[8:48] Speaker 3: Christina Hooper of Man Group, thank you so much. There's
[8:50] Speaker 3: a lot going on in the world this morning, and
[8:52] Speaker 3: I feel like we've managed to pass through all of it.
[8:55] Speaker 2: Thank you for being with us.
[8:56] Speaker 3: Now, coming up, Apple's released a new line of smart
[8:59] Speaker 3: home devices, which includes the new HomeView Hub. More details
[9:03] Speaker 3: about those new devices when we come back. This is
[9:06] Speaker 3: Bloomberg Tech. Apple released a new line of smart home devices,
[9:18] Speaker 3: including a new gadget called the Home View.
[9:21] Speaker 2: A $ 299.
[9:22] Speaker 3: Display aimed at challenging Amazon and Google in the smart
[9:25] Speaker 3: home market. Joining us, Bloomberg's Mark Gurman, who leads our
[9:28] Speaker 3: coverage of Apple and consumer technology.
[9:30] Speaker 2: Tell me about the device, $ 299. What do I need
[9:32] Speaker 2: to know?
[9:34] Speaker 7: Yeah, it's a pretty good price. This is basically a
[9:37] Speaker 7: HomePod with a built-in display. It has a new operating
[9:40] Speaker 7: system called the HomeView software. And basically, the way it
[9:44] Speaker 7: works is six people in a home could walk up
[9:46] Speaker 7: to it, and everyone will get to experience something different,
[9:49] Speaker 7: something customized just for them. If I walk up to it,
[9:51] Speaker 7: I'll see my music, my messages, my applications. If my
[9:55] Speaker 7: wife walks up to it, she'll see her calendar appointments,
[9:59] Speaker 7: she'll see her clock faces, she'll see her notes, calendars,
[10:02] Speaker 7: and reminders. And so it's a pretty cool device, and
[10:05] Speaker 7: it's going to give the Amazon Echo Show a serious
[10:07] Speaker 7: run for its money, especially at the $ 299 price point.
[10:11] Speaker 3: So it's interesting that there's actually two variants of it, right?
[10:14] Speaker 3: There's the tabletop and then there's a wall mounted one,
[10:17] Speaker 3: both the same price, slightly different form factors. I think
[10:20] Speaker 3: the tabletop just standard USB-C, right? But why did Apple
[10:24] Speaker 3: do that? Why did they want optionality?
[10:28] Speaker 7: Well, they wanted to fit people however they wanted to
[10:30] Speaker 7: use this device. I mean, the Echo show there's a
[10:32] Speaker 7: lot of people who pin theirs on the wall. These
[10:35] Speaker 7: smart home devices, they can serve as a thermostat, can
[10:37] Speaker 7: serve as a control panel. And just like you may
[10:39] Speaker 7: have like an ADT security panel on your wall, you
[10:42] Speaker 7: may want one of these instead or in addition to
[10:44] Speaker 7: to control your thermostat, to control your lights, to see
[10:47] Speaker 7: the weather, to do quick FaceTime calls. So it's pretty
[10:50] Speaker 7: interesting they went in both directions. And this design is
[10:53] Speaker 7: pretty nifty. You have this pretty premium metal base. You
[10:57] Speaker 7: have perforated metal holes. You have a great speaker system.
[11:01] Speaker 7: So I think this is going to be pretty popular,
[11:02] Speaker 7: especially for the holiday season.
[11:04] Speaker 3: There are a range of other smart home devices. We
[11:07] Speaker 3: knew these were coming in large part because of your
[11:09] Speaker 3: reporting in partnership with LG and then with Schneider Electric.
[11:14] Speaker 3: Go through them, but also like the bigger picture of
[11:16] Speaker 3: what Apple's trying to do in the home, moving on
[11:19] Speaker 3: from just the iPhone and the Mac.
[11:22] Speaker 1: Yeah, so they announced a thermostat.
[11:24] Speaker 7: They announced several cameras, one of the floodlight camera, an
[11:26] Speaker 7: indoor camera, an outside camera. And they basically have created
[11:30] Speaker 7: a smart home starter kit. So now if you want
[11:32] Speaker 7: a smart home, you can go to Apple.
[11:34] Speaker 1: You can get everything in one full swoop.
[11:37] Speaker 7: Before, everything was very bunched, not as bunched up. You
[11:41] Speaker 7: had to go to some places to buy your camera.
[11:44] Speaker 7: You had to go to some places to buy your thermostat.
[11:46] Speaker 7: You didn't know what was compatible. But now there's one
[11:48] Speaker 7: place where you can get everything. And if you want
[11:50] Speaker 7: an Apple smart home experience, now you're able to get
[11:53] Speaker 7: that really for the first time.
[11:55] Speaker 1: Before, it was all discombobulated.
[12:00] Speaker 3: So those products coming in December, but I'm looking forward
[12:02] Speaker 3: to the Power On once you get to put them
[12:04] Speaker 3: in your home, Mark, and get some hands-on time with them.
[12:08] Speaker 3: I remember tuning for more with Mark's Power On podcast,
[12:10] Speaker 3: which drops on Thursday. That's where you can find it.
[12:13] Speaker 3: Apple isn't the only tech giant rolling out new gadgets today.
[12:16] Speaker 3: Alibaba unveiled its latest Qen AI smart glasses. I want
[12:20] Speaker 3: to take a look at shares of Meta over the
[12:21] Speaker 3: last two days. You can see that kind of big
[12:22] Speaker 3: spike that came in the session yesterday. And that is
[12:24] Speaker 3: after the company released a new AI model, Muse Wonder,
[12:27] Speaker 3: that it says has frontier capabilities.
[12:28] Speaker 2: It's all part of a push.
[12:48] Speaker 3: To take on the likes of open AI and improper.
[12:52] Speaker 3: It's also the most capable the meta has released since
[12:55] Speaker 3: CEO Mark Zuckerberg kind of overhauled its AI operation through
[12:59] Speaker 3: aggressive hiring that included naming Alexander Wang as the company's
[13:03] Speaker 3: AI chief. Bloomberg's Riley Griffin has been speaking with Alexander Wang.
[13:07] Speaker 3: So let's start with the model itself, what it is,
[13:10] Speaker 3: what it represents, but also what Alexander Wang had to
[13:12] Speaker 3: say about it.
[13:13] Speaker 4: Yeah, I mean, this is what he's been working towards
[13:15] Speaker 4: since he joined the company about a year and a
[13:17] Speaker 4: half ago. This is the moment we have.
[13:20] Speaker 8: Been waiting for.
[13:21] Speaker 4: As Meta tries to get back to the frontier, that
[13:24] Speaker 4: cutting edge of its competitors. Those who have been following
[13:27] Speaker 4: this the most closely may know this as Watermelon. It
[13:30] Speaker 4: was expected to be released in October, and here it is.
[13:33] Speaker 4: And they are coming in confident. I mean, the claims
[13:37] Speaker 4: from Alexander Wang and the rest of the Meta executive
[13:40] Speaker 4: team are strong. They feel like they have reached capabilities
[13:45] Speaker 4: that match their rivals in terms of coding, in terms
[13:48] Speaker 4: of finance and legal expertise.
[13:50] Speaker 3: So time will tell but if there is a point
[13:52] Speaker 3: of distinction it's like how meta thinks about consumers so
[13:55] Speaker 3: when i was reading your report i really appreciated this
[13:58] Speaker 3: detail like more big models are coming yes but focused
[14:02] Speaker 3: on how they empower consumer applications and muse the the
[14:06] Speaker 3: assistant version of muse yeah.
[14:07] Speaker 4: Alex told me yesterday morning that they really purpose built
[14:10] Speaker 4: muse wonder to power this consumer application they are seeking
[14:15] Speaker 4: to make All the fine little details appropriate for remembering
[14:19] Speaker 4: what a user is like over time, right? And Muse,
[14:24] Speaker 4: the app, it's a shame they all have the same
[14:26] Speaker 4: name because it makes it.
[14:27] Speaker 2: Hard to talk about. Right, it does, yeah.
[14:28] Speaker 3: But we are talking about the personal AI agent or assistant.
[14:31] Speaker 2: Totally.
[14:32] Speaker 4: The personal AI agent or assistant, an app that you
[14:34] Speaker 4: can download on the App Store, will be powered by
[14:37] Speaker 4: this app. Highly capable model, but for the subscription user.
[14:42] Speaker 2: So those who pay a.
[14:43] Speaker 4: Little bit more of a premium, $.
[14:45] Speaker 1: 20 or $ 100.
[14:45] Speaker 3: So I've been using Base Muse, the AI agent or assistant,
[14:49] Speaker 3: a lot recently. And I use it a lot for
[14:52] Speaker 3: social media posts to Instagram, for example, to help with composition. Or,
[14:56] Speaker 3: for example, if I'm planning a trip to the UK
[14:58] Speaker 3: next year to monitor price flights, I just seem to
[15:01] Speaker 3: get a lot of bad news from my Muse. Neither
[15:04] Speaker 3: here nor there. What's coming next? What's Alexander Wang and
[15:07] Speaker 3: the team working on? What do they think they want
[15:09] Speaker 3: to focus on from this point?
[15:10] Speaker 4: Yeah, one particular thing they're working on is actually making
[15:13] Speaker 4: Muse Wonder, that big, highly capable model, more efficient. Because
[15:17] Speaker 4: this comes down to compute. It's pretty expensive to scale
[15:21] Speaker 4: out a model like this to billions of users. And
[15:24] Speaker 4: right now they're saying our most capable model, it's only
[15:27] Speaker 4: going to power Muse for those who are paying for it.
[15:30] Speaker 4: But they'd like free users actually to have some of
[15:32] Speaker 4: those capabilities too. So they're looking to make the model
[15:35] Speaker 4: more efficient so that they don't have to have those
[15:38] Speaker 4: steep costs at the end of the day.
[15:40] Speaker 3: The stock is pushing higher again in today's session, 3%.
[15:44] Speaker 1: Good reporting.
[15:45] Speaker 3: It's really worth reading the detail in Riley's piece about
[15:48] Speaker 3: MuseWonder and kind of the bigger picture for Meta. Thank
[15:50] Speaker 3: you very much indeed. AI requires strong regulation to ensure
[15:54] Speaker 3: that it's safe. That's according to Lila Ibrahim, Google DeepMind's
[15:57] Speaker 3: Chief AI Readiness Officer. She made those comments as part
[16:00] Speaker 3: of a panel discussion moderated by Bloomberg's Francine Lacroix at
[16:04] Speaker 3: the Bloomberg New Economy Forum in New Delhi.
[16:07] Speaker 9: We've been long proponents of AI is too important not
[16:10] Speaker 9: to regulate and not to regulate well. So we have
[16:13] Speaker 9: to think If this is one of the most transformational
[16:16] Speaker 9: technologies of our time, how do we work to make
[16:20] Speaker 9: sure that it is safe for humanity? We've done a
[16:24] Speaker 9: lot of work specifically around agentics, agentic workflows as well,
[16:30] Speaker 9: in terms of ethics research, trying to provide the community
[16:34] Speaker 9: an opportunity to talk about the risks and the opportunities together.
[16:38] Speaker 3: And we'll have more from the Bloomberg New Economy Forum
[16:40] Speaker 3: after the break and hear from Infosys chairman Nandan Nilekani.
[16:44] Speaker 3: Also coming up, the space race is turning into a
[16:46] Speaker 3: money race with annual investments set to top $ 100 billion
[16:49] Speaker 3: for the first time.
[16:51] Speaker 2: More on that next. This is Bloomberg Tech.
[17:06] Speaker 3: Infosys chairman Nandan Nilekani sees some job losses from AI,
[17:10] Speaker 3: but expects, quote, millions of small businesses to be the.
[17:13] Speaker 2: Future of job creation.
[17:15] Speaker 3: He spoke with Bloomberg's Sherry Ahn on the sidelines of
[17:17] Speaker 3: the New Economy Forum in New Delhi.
[17:20] Speaker 10: Job loss could happen for larger organizations, but India is
[17:25] Speaker 10: a whole dense network of small companies, and if we
[17:28] Speaker 10: make them successful, they will hire people. So I visualize
[17:32] Speaker 10: millions of small businesses each hiring two or three people,
[17:35] Speaker 10: and that's how jobs will get created. Because in a
[17:38] Speaker 10: small company, there are no set roles. You can't make
[17:41] Speaker 10: those jobs.
[17:41] Speaker 11: That are AI-approved.
[17:42] Speaker 12: Yeah, you basically do everybody's job.
[17:44] Speaker 10: Exactly, so that's where I think the future is.
[17:47] Speaker 12: At the same time, SMEs don't necessarily have that much
[17:49] Speaker 12: capital to be putting into innovation and transformation.
[17:52] Speaker 10: That's why many things we are doing in India, SMEs
[17:56] Speaker 10: get access to capital, access to markets, access to know-how,
[18:00] Speaker 10: access to compliance rules. So creating a whole infrastructure to
[18:03] Speaker 10: make small companies more effective.
[18:05] Speaker 12: This is my first time in India, and I've got
[18:07] Speaker 12: to say, I'm loving New Delhi, but at the same time,
[18:10] Speaker 12: there have been a lot of security concerns around protests,
[18:13] Speaker 12: a lot of discontent among the youth. How do you view,
[18:16] Speaker 12: right now, the social change that's happening?
[18:19] Speaker 10: Well, India is facing an issue which is there in
[18:20] Speaker 10: the rest of the world, which is, how do you
[18:22] Speaker 10: create jobs for millions of young people? And I think
[18:26] Speaker 10: you have to, and of course, give them the proper education,
[18:28] Speaker 10: which is why I think we have to really boost entrepreneurship,
[18:32] Speaker 10: Because only by creating millions of small companies and startups
[18:36] Speaker 10: and so on, they will create the jobs, which India
[18:39] Speaker 10: is doing, by the way. You know, 10 years back,
[18:41] Speaker 10: there were only 1,000 small companies, startups. Today, we have
[18:45] Speaker 10: 150,000 startups. In 2035, we'll have a million startups. So
[18:49] Speaker 10: these startups, in turn, will create jobs.
[18:52] Speaker 12: It's all about meritocracy, democratization of education. I know that
[18:56] Speaker 12: you're also working on a report when it comes to
[18:59] Speaker 12: the exams.
[18:59] Speaker 10: That's right.
[19:00] Speaker 12: What can you tell us and where are you looking
[19:03] Speaker 12: for sort of guidelines and lessons on how you form
[19:06] Speaker 12: that report?
[19:07] Speaker 10: Well, it's basically, the report is not out, but we
[19:10] Speaker 10: are trying to provide a roadmap on how.
[19:13] Speaker 1: To make exam papers.
[19:16] Speaker 10: Dynamically generated with AI just before the exam.
[19:20] Speaker 12: Are you incorporating AI into the examination system?
[19:22] Speaker 10: Yeah, to create questions and so on, but of course,
[19:25] Speaker 10: checked by human beings. And also, how do you have
[19:28] Speaker 10: a network of locations where people can do exams safely
[19:31] Speaker 10: and securely if you want to do them online? So
[19:33] Speaker 10: there's a number of things we.
[19:34] Speaker 11: Have to do.
[19:35] Speaker 3: That was Infosys chairman Nandan Nilekani along with her own
[19:38] Speaker 3: Sherry Ahn. Let's take a look at today's big number,
[19:40] Speaker 3: $ 100 billion. That's how much investment could pour into the
[19:44] Speaker 3: space economy this year, a first for the industry. SpaceX's
[19:48] Speaker 3: blockbuster IPO is helping fuel the boom. alongside rising defense
[19:51] Speaker 3: spending demand for new space infrastructure. Bloomberg Space Reporter Lauren
[19:55] Speaker 3: Grush is with us. This is data from Space Capital,
[19:58] Speaker 3: the investment firm. But it's a big number. And it's interesting,
[20:02] Speaker 3: you know, this kind of milestone for the space economy.
[20:05] Speaker 8: Yeah, absolutely.
[20:06] Speaker 13: I was really interested when I saw this report come
[20:09] Speaker 13: out just because I have been tracking anecdotally what seems
[20:12] Speaker 13: like really big numbers and big money floating around the
[20:16] Speaker 13: space industry, at least in the last few months. As
[20:19] Speaker 13: we know, for instance, Stokes Space, an aerospace startup building
[20:23] Speaker 13: a reusable launch rocket, had announced a $ 1 billion raise
[20:27] Speaker 13: a few months ago.
[20:28] Speaker 1: And then on top of that, Blue.
[20:31] Speaker 13: Origin has taken outside investment for the first time reaching
[20:34] Speaker 13: $ 10 billion in its round, possibly going to have more soon.
[20:39] Speaker 13: So I had noticed that there was a lot of
[20:41] Speaker 13: money floating around. And this just confirmed that, you know,
[20:43] Speaker 13: the space economy is on track to exceed this $ 100
[20:46] Speaker 13: billion number, which is fueled by a lot of things,
[20:50] Speaker 13: but the SpaceX IPO being one of them.
[20:53] Speaker 3: Yeah, the SpaceX IPO is a big chunk of it.
[20:55] Speaker 3: And we should, you know, we should just acknowledge that
[20:57] Speaker 3: out there. But it's kind of like the bigger picture
[20:59] Speaker 3: of what's happening. The US defense apparatus has demand for
[21:04] Speaker 3: space assets. Sovereigns have demand for space assets. Give us
[21:07] Speaker 3: some of those case studies.
[21:09] Speaker 13: Yeah, I would absolutely say that all roads lead back
[21:12] Speaker 13: to SpaceX in this scenario. It's the $ 2 trillion elephant
[21:17] Speaker 13: in the room. So obviously the IPO has re-rated what
[21:22] Speaker 13: Chad Anderson says at Space Capital, the space sector, so
[21:26] Speaker 13: becoming a little bit more essential than before. But also
[21:29] Speaker 13: SpaceX's plans for its business are having ripple effects across
[21:33] Speaker 13: the industry. So as we've reported, the Falcon 9 rocket
[21:37] Speaker 13: that SpaceX relies on, they're planning to sunset it. And
[21:40] Speaker 13: so there's creating this massive launch bottleneck that is providing
[21:44] Speaker 13: an opportunity for other launchers and other startups to come
[21:48] Speaker 13: in and fill that gap. And so that is another
[21:51] Speaker 13: reason that we're seeing a lot of money being invested
[21:54] Speaker 13: in the sector right now is because people are understanding
[21:57] Speaker 13: we actually need more hardware. And SpaceX is part of
[22:01] Speaker 13: those decision makings.
[22:02] Speaker 3: And so Lauren, super, super quick, you've just published a
[22:05] Speaker 3: report about SpaceX's plans for more Starship launches.
[22:08] Speaker 2: Just what's the number?
[22:10] Speaker 13: Yes, SpaceX is proposing to launch up to 456 Starship
[22:12] Speaker 13: launches a year. And on top of that, 912 landings
[22:19] Speaker 13: of Starship out of its Starbase launch facility in South Texas.
[22:24] Speaker 14: That's a massive number.
[22:25] Speaker 13: Though it isn't really a surprise. SpaceX CEO Elon Musk
[22:29] Speaker 13: has talked extensively about how Starship will be launching once
[22:34] Speaker 13: an hour when it gets up to full speed.
[22:37] Speaker 3: Check out the report from Lauren that published just before
[22:39] Speaker 3: we came on air. Lauren Grush, thank you very much.
[22:40] Speaker 3: Coming up, we go live to the CES Summit in
[22:43] Speaker 3: Beverly Hills for a conversation with the Samba Nova CEO,
[22:46] Speaker 3: Rodrigo Liang.
[22:48] Speaker 2: Halftime, it's Bloomberg Tech. Welcome back to Bloomberg Tech.
[22:59] Speaker 3: Let's get back to today's top story, the chip sell-off
[23:01] Speaker 3: that's been accelerating worldwide. Bloomberg Tech's equities reporter, Ryan Vostelica,
[23:05] Speaker 3: joins us now. That's where the pressure is, is in chips.
[23:09] Speaker 15: Yes, this is a real reversal of the momentum that
[23:11] Speaker 15: we've seen throughout this year. Now, I would say most
[23:14] Speaker 15: people aren't saying that the story has really broken or
[23:17] Speaker 15: that the narrative has shifted. There's continues to be a
[23:20] Speaker 15: lot of demand for AI infrastructure and a lot of
[23:23] Speaker 15: optimism about the long term growth prospects. But at least
[23:26] Speaker 15: at the moment, a lot of the big high flyers
[23:28] Speaker 15: that we've seen throughout this year, especially the memory related
[23:32] Speaker 15: chip makers, those are the ones that are kind of
[23:34] Speaker 15: bearing the brunt of all of the selling pressure.
[23:38] Speaker 3: Okay, bloomers round Vestelica with the summary of what's happening
[23:40] Speaker 3: in chip stocks and where the pressure is being felt.
[23:42] Speaker 3: I want to get out to Beverly Hills where the
[23:44] Speaker 3: annual CIS Alternative Investment Summit is underway. Our very own
[23:47] Speaker 3: Danny Berger is standing by. Danny.
[23:53] Speaker 8: Ed, thank you so much.
[23:54] Speaker 14: I am so pleased to say, and it's a perfect
[23:56] Speaker 14: time for this considering the conversation you were just having,
[23:58] Speaker 14: Rodrigo Leong, the CEO and co-founder of Samba Nova joins now. Rodrigo,
[24:03] Speaker 14: great to see you.
[24:04] Speaker 8: Yeah, great to see you.
[24:05] Speaker 14: And so this week you are closing your Series F
[24:08] Speaker 14: fundraising round. For viewers who are paying attention, you already
[24:11] Speaker 14: had kind of an announcement of a closing in July.
[24:14] Speaker 14: This one is different though. We were just talking about,
[24:16] Speaker 14: you don't just have financial backers, you have strategic corporate backers.
[24:20] Speaker 14: Talk to me about the significance of bringing in companies
[24:23] Speaker 14: into the fold beyond just financial firms.
[24:25] Speaker 8: Yeah, no, it was so exciting to see.
[24:27] Speaker 11: We've got so much interest and we did that first
[24:29] Speaker 11: close back in July when you and I talked. And
[24:32] Speaker 11: so all this interest and then we had to extend
[24:34] Speaker 11: the close just because the strategic started coming in. These
[24:37] Speaker 11: are huge tech companies who are coming in and partnering
[24:40] Speaker 11: with us to actually help SambaNova grow in the market.
[24:44] Speaker 2: So they're partnering with you.
[24:45] Speaker 14: Does that also mean we're going to start to see
[24:46] Speaker 14: large tech companies double down and use SambaNova technology too?
[24:50] Speaker 13: Yes.
[24:51] Speaker 11: I mean, you're seeing this incredible interest in the market
[24:54] Speaker 11: with inference growing and people are looking at heterogeneous environments
[24:58] Speaker 11: using chips from Nvidia and AMD and someone over to
[25:01] Speaker 11: intermix into in order to drive the best economics And
[25:04] Speaker 11: so we're seeing these these data centers come up They're
[25:07] Speaker 11: all gonna have all these hybrid hybrid environments and we're
[25:10] Speaker 11: super excited to be part of that.
[25:11] Speaker 14: Another reason this conversation is so apt for today is
[25:13] Speaker 14: we had JP Morgan earnings and on the earnings call
[25:16] Speaker 14: Jamie diamond was talking about this idea We are using
[25:18] Speaker 14: a wide array of chips. They are one of your
[25:21] Speaker 14: customers to what is JP Morgan using the Samba Nova
[25:24] Speaker 14: ecosystem for exactly.
[25:26] Speaker 8: We're really excited with the partnership.
[25:28] Speaker 11: We have our racks at JPMorgan behind their firewalls. And
[25:31] Speaker 11: this is one of the things that's really important for
[25:33] Speaker 11: enterprises where security, safety is important. People are bringing their
[25:37] Speaker 11: technology back on-prem, bringing it behind their own firewalls, protecting
[25:41] Speaker 11: their customers' data behind those firewalls. And because Summable is
[25:44] Speaker 11: so power efficient, so low cost, we're able to deliver
[25:47] Speaker 11: that on their own premise inside their own data centers.
[25:50] Speaker 14: Does that also necessarily mean if they're using your technology on-prem,
[25:53] Speaker 14: maybe using their own model or open source, that companies
[25:57] Speaker 14: like JP Morgan are going to start drifting away from
[25:59] Speaker 14: using open AI, from using Cloud, from using the large
[26:03] Speaker 14: language models?
[26:04] Speaker 11: Well, I think it's going to be hybrid. I think
[26:05] Speaker 11: these frontier models are just tremendous, and they do all
[26:07] Speaker 11: sorts of great things. And actually, Salmanova will work closely
[26:10] Speaker 11: with those frontier labs too so that you can intermix.
[26:13] Speaker 11: There's some things where those models are just really, really
[26:15] Speaker 11: good at, and some things where companies want to keep private,
[26:18] Speaker 11: companies want to keep secure. They have their own IP,
[26:21] Speaker 11: and so they create their own models. They run it
[26:23] Speaker 11: privately and securely on Salmanova on-prem, and they don't have
[26:26] Speaker 11: to worry about that information getting out.
[26:28] Speaker 14: Well, part of the reason I ask is the Broadcom
[26:30] Speaker 14: CEO is also speaking on Monday at a conference. He
[26:33] Speaker 14: had a lot of questions about the ecosystem, but about
[26:36] Speaker 14: these large language models, saying they're losing money and questioned
[26:40] Speaker 14: investors into them. And I wonder if you also question
[26:43] Speaker 14: the economics.
[26:43] Speaker 2: Of it, because.
[26:44] Speaker 14: your very existence shows that they're not the only game
[26:47] Speaker 14: in town.
[26:47] Speaker 11: Yeah, well, look, you know, Broadcom's a great partner of ours,
[26:50] Speaker 11: and we've been talking about it. In the world of inference,
[26:53] Speaker 11: economics is paramount, right? That as you scale, it's about
[26:57] Speaker 11: how quickly can you actually deliver returns on that capital
[27:00] Speaker 11: investment you're making. And so with Salmanova, we're able to
[27:02] Speaker 11: give you six-month returns, where otherwise it might take you 18,
[27:06] Speaker 11: 24 months to do it. And so as you're growing,
[27:08] Speaker 11: I think you're going to see people be much, much
[27:10] Speaker 11: more conscious about dollars, Return on investment, how quickly can
[27:14] Speaker 11: you get your payback on the investments you're making?
[27:16] Speaker 14: Rodrigo, again, all of that kind of feels like the
[27:19] Speaker 14: future that we see for these large language models and
[27:21] Speaker 14: thus the build out of things like NVIDIA and things
[27:25] Speaker 14: that are training models. Do we go too far then?
[27:28] Speaker 14: Is there a risk that we are going too far
[27:30] Speaker 14: if the world is waking up to things that Samba
[27:33] Speaker 14: Nova offers that are cheaper and are more secure in
[27:35] Speaker 14: a closed ecosystem?
[27:36] Speaker 11: Well, I think we're at early stages still, despite all
[27:38] Speaker 11: this build out. You look at most enterprises are just
[27:41] Speaker 11: getting started with delivering applications on AI. And so I
[27:44] Speaker 11: think we're just at the early days. Now, we're entering
[27:47] Speaker 11: the phase where it is about efficiency. It is about,
[27:50] Speaker 11: you know, you think about data centers and some of
[27:52] Speaker 11: them will focus so much on energy efficiency so that
[27:54] Speaker 11: we can actually reduce the number of data centers required.
[27:57] Speaker 11: And so we want to focus on making sure that
[27:59] Speaker 11: we deliver technology that can scale and allow people to
[28:02] Speaker 11: use it without all this extra cost that today we're facing.
[28:06] Speaker 14: I was also thinking about Hock Tan's comments, sort of
[28:09] Speaker 14: ironically coinciding with this close you're doing, because one of
[28:12] Speaker 14: the things he said was one of the biggest bottlenecks
[28:14] Speaker 14: was financing. And I wonder what you think of that comment,
[28:17] Speaker 14: considering you've been able to bring funds in. And I
[28:19] Speaker 14: wonder if you think perhaps other parts of the ecosystem
[28:22] Speaker 14: are a little bit harder to finance now.
[28:24] Speaker 8: Well, look, I mean, it's a capital intensive business.
[28:26] Speaker 11: And as you scale, you're having to put all these
[28:29] Speaker 11: chips into the factory and taking nine to 12 months
[28:31] Speaker 11: in order for you to actually stay in the data centers.
[28:33] Speaker 11: And so I think all of that's going to be
[28:35] Speaker 11: basically capital that's.
[28:37] Speaker 8: In process of delivering value.
[28:39] Speaker 11: Right. And so I think we're going to need to
[28:40] Speaker 11: figure out how to actually accelerate that because the growth
[28:43] Speaker 11: is so extensive. And so we need to actually make
[28:46] Speaker 11: sure the capital is there to support it.
[28:47] Speaker 14: I would love to talk about your growth because, again,
[28:50] Speaker 14: the number that you had in July was $ 11 billion.
[28:52] Speaker 14: We'll wait for a final announcement to see where that goes.
[28:55] Speaker 14: So at that size, you're about 4% of ARM's market cap, 1%
[28:58] Speaker 14: of AMD, just a rounding error to Nvidia. Do you
[29:01] Speaker 14: see a future where your size rivals those?
[29:04] Speaker 11: Well, look, I think it's an exciting time to be.
[29:06] Speaker 11: You see these companies where a year before that, it
[29:09] Speaker 11: was a quarter the size maybe.
[29:11] Speaker 8: And so the growth is incredible.
[29:12] Speaker 11: What we have to do as a company, and look,
[29:14] Speaker 11: we're a startup still, and we're focusing on growing, but
[29:17] Speaker 11: we're actually seeing this market turn right into what we're good,
[29:21] Speaker 11: which is inference.
[29:22] Speaker 8: It's about energy efficiency. It's about cost efficiency.
[29:25] Speaker 11: And so if you are thinking about that, some of
[29:27] Speaker 11: it becomes a very, very good option as you scale
[29:30] Speaker 11: your inference workloads.
[29:31] Speaker 14: Well, and presumably you have less of the whole data
[29:33] Speaker 14: center backlash. How are you viewing that and how you
[29:36] Speaker 14: position the company considering some of these other large data
[29:38] Speaker 14: center projects have stalled with the regulation and the local pushback?
[29:43] Speaker 8: Well, we're partnering with Vista on this.
[29:45] Speaker 11: We're super excited about the new cloud that Vista is building, VC2,
[29:48] Speaker 11: and what their strategy is and some of us powering
[29:51] Speaker 11: it is using existing data centers, reusing it, taking what
[29:55] Speaker 11: we call brownfield data centers that already have the space,
[29:59] Speaker 11: already have the power, already have the infrastructure developed, and
[30:02] Speaker 11: actually reusing it and reenergizing that. It's all over the
[30:06] Speaker 11: world in terms of how many of those exist. We
[30:08] Speaker 11: just gotta fit it into that power footprint. And so
[30:10] Speaker 11: with Salmanova at 10 kilowatts versus 100 kilowatts per rack,
[30:15] Speaker 11: we're able to actually reuse all the data center, which
[30:17] Speaker 11: is actually gonna help not just businesses, but also the environment.
[30:22] Speaker 14: How receptive are people to that message? And part of
[30:24] Speaker 14: the reason I ask that is a lot has been
[30:27] Speaker 14: said that part of the issue with AI and the
[30:29] Speaker 14: data center pushback has just been, they haven't done a
[30:32] Speaker 14: good job marketing themselves. It's been doom and gloom and
[30:36] Speaker 14: the safety is a concern and not things like, hey,
[30:38] Speaker 14: maybe this stuff can cure cancer. And I wonder as
[30:41] Speaker 14: a business leader, in this space, what you see as
[30:43] Speaker 14: your responsibility?
[30:44] Speaker 11: Well, incredibly important for these things to come out. And
[30:47] Speaker 11: if you kind of look back 20 years ago at
[30:49] Speaker 11: the internet, all these things were coming up around security
[30:51] Speaker 11: and privacy. And then the entire industry around cybersecurity was
[30:55] Speaker 11: born out of that. And so it's great that these
[30:57] Speaker 11: conversations are happening now because safety is important, right?
[31:00] Speaker 8: Efficiency is important.
[31:01] Speaker 11: It's about making sure that we actually have sustainable technologies
[31:05] Speaker 11: and so we're all part of this we're doing our
[31:08] Speaker 11: part we're driving the power down we're driving the cost
[31:10] Speaker 11: down making sure that we're reusing data centers centers that
[31:13] Speaker 11: already exist right and so we all have to be
[31:16] Speaker 11: more conscious about how we actually scale this because.
[31:19] Speaker 14: It's going to be important do you think that that's
[31:20] Speaker 14: the ultimate bottleneck maybe not capital but the pushback you're
[31:23] Speaker 14: getting in local communities for others Well.
[31:25] Speaker 8: I think it's important, right?
[31:26] Speaker 11: Sustainability is important, and I think you're gonna have to
[31:28] Speaker 11: figure out where these things go. And so we start
[31:31] Speaker 11: by saying, look, can we deliver the value and the
[31:33] Speaker 11: technology you need by reducing the power that's required?
[31:36] Speaker 8: And so if you start.
[31:38] Speaker 11: There, it actually opens up opportunities in a lot of
[31:41] Speaker 11: different ways.
[31:42] Speaker 14: Rodrigo, tracking Samba Nova's journey, just how it exists as
[31:45] Speaker 14: a company, has been really interesting. There were rumors that
[31:48] Speaker 14: you and I talked about months ago about Intel partnership.
[31:52] Speaker 14: You've addressed whether you would go public before. You've said
[31:55] Speaker 14: maybe if it makes sense for you, but now you
[31:57] Speaker 14: have the capital, there's no need for that.
[31:59] Speaker 8: When would that change?
[32:00] Speaker 14: What would need to go right, or maybe even wrong,
[32:03] Speaker 14: but what would need to go right in order for
[32:04] Speaker 14: you to consider becoming a public company?
[32:06] Speaker 11: Well, look, I think we're in a high growth mode
[32:09] Speaker 11: right now. And so we are partnering with Intel. We're
[32:12] Speaker 11: partnering with several others because, you know, as a company
[32:14] Speaker 11: in this space, partnership is really important. And you want
[32:17] Speaker 11: to continue to actually leverage that into the business because
[32:21] Speaker 11: it's not just one piece of technology.
[32:23] Speaker 8: That you need.
[32:24] Speaker 11: You need all the CPUs and the memory and the storage.
[32:27] Speaker 11: And so you need to partner with everybody. And so
[32:30] Speaker 11: as we do that and we grow to the next level,
[32:32] Speaker 11: we can see what the capital needs are. And you
[32:34] Speaker 11: see even the largest companies, you'll be looking for additional capital.
[32:38] Speaker 11: So there are moments where we're gonna see, do we
[32:40] Speaker 11: need more capital in order to actually take us to
[32:42] Speaker 11: the next level?
[32:42] Speaker 14: I see, so it basically becomes a matter of you
[32:45] Speaker 14: need that extra juice that public markets would give you.
[32:47] Speaker 14: But for now, in your growth stage, you're not there yet.
[32:49] Speaker 14: Is that kind of the right way on that? Yeah,
[32:51] Speaker 14: we're just focusing on execution right now.
[32:52] Speaker 11: I mean, right now it's about delivering value, getting the
[32:55] Speaker 11: customer wins, getting some of the technology deployed. Supply chain's
[32:59] Speaker 11: a big issue, and so we're always kind of making
[33:01] Speaker 11: sure that our customers get the supplies they need. And
[33:03] Speaker 11: so we've got to continue to execute and create that flywheel,
[33:05] Speaker 11: and at some point, if it makes sense, we discuss it.
[33:08] Speaker 14: Rodrigo, thank you so much for joining. Great to see
[33:10] Speaker 14: you out here in L.A.
[33:10] Speaker 8: Yeah, great to see you as well.
[33:11] Speaker 14: And Ed, with that, I'll send it back to you.
[33:13] Speaker 14: That is Rodrigo Leyang of Samba Nova.
[33:16] Speaker 2: All right.
[33:17] Speaker 3: Big thank you to Bloomberg's Danny Berger out there in
[33:19] Speaker 3: Beverly Hills. Coming up in the show, we're going to
[33:21] Speaker 3: speak with Mainspring Energy CEO Tom Leinberger about the startup's
[33:24] Speaker 3: deal with American Electric Power that could hit $ 4 billion.
[33:30] Speaker 2: This is Bloomberg Tech.
[33:44] Speaker 3: Startup Mainspring Energy has signed a deal to sell up
[33:47] Speaker 3: to 1.6 gigawatts of small generators to American Electric Power.
[33:51] Speaker 3: The deal could be worth up to $ 4 billion over
[33:54] Speaker 3: the next eight years and signals how the data center
[33:57] Speaker 3: build out is pushing providers to find faster ways to
[34:00] Speaker 3: To get new power online, mainstream CEO Tom Leimberg is
[34:03] Speaker 3: with us. The structure of this, Tom, and welcome to
[34:06] Speaker 3: the program, really interesting. Because initially it's quite small, right?
[34:10] Speaker 3: 190 megawatts. And people would say, oh, it's over eight years.
[34:13] Speaker 3: That seems like a really long time to kind of
[34:15] Speaker 3: ramp up in little increments. Still faster than a nuclear
[34:19] Speaker 3: reactor can come online. Just explain how you're going to
[34:21] Speaker 3: deliver over time.
[34:23] Speaker 16: And thanks so much for having me, by the way.
[34:25] Speaker 16: It's a pleasure to be here. And it's a super
[34:27] Speaker 16: exciting deal for us. AEP, as you may know, one
[34:31] Speaker 16: of the nation's largest utilities, already has a significant investment
[34:36] Speaker 16: in new technology. It's got the largest transmission system today,
[34:40] Speaker 16: and it's investing for its customers to make sure service
[34:43] Speaker 16: is great. That's not just data centers, that's all their customers.
[34:46] Speaker 16: And we're just really excited to be part of that
[34:48] Speaker 16: nearly $ 80 billion build-out.
[34:50] Speaker 1: Over the next five years.
[34:52] Speaker 16: Our structure is really to make sure that they have
[34:55] Speaker 16: the ability to get power fast. So we've committed to
[34:58] Speaker 16: delivering power within 12 months of their order. They put
[35:02] Speaker 16: the first order in, as you can see, of 190 megawatts.
[35:05] Speaker 16: And then the rest of the deal is structured to
[35:08] Speaker 16: give them an option each year to purchase around the
[35:11] Speaker 16: same 200 megawatts a year with a 12-month lead time
[35:14] Speaker 16: so they can make quick decisions immediately. about how to
[35:17] Speaker 16: use our modular reactors across the U.S. wherever they want
[35:21] Speaker 16: to put them to serve their customers better.
[35:24] Speaker 2: What AEP is talking about this morning is speed, right?
[35:26] Speaker 3: This is about getting power where it's needed and online quickly.
[35:30] Speaker 3: So let's talk a bit about the technology, the module.
[35:34] Speaker 3: You know, it's about the size of a shipping container,
[35:37] Speaker 3: natural gas. Why is this a good solution?
[35:41] Speaker 16: Well, the exciting thing about mainspring generators is they do
[35:44] Speaker 16: provide significant flexibility to AP. They can use our generators
[35:50] Speaker 16: and dispatch them. So use power when they want to.
[35:53] Speaker 16: They can firm solar power. They can run any fuel. Yes,
[35:57] Speaker 16: they run on natural gas. They also run on propane.
[35:59] Speaker 16: They run on hydrogen. So any fuel they have available.
[36:03] Speaker 16: And so having the ability to both react to the
[36:07] Speaker 16: customer's needs and the grid's needs at the same time
[36:10] Speaker 16: gives them incredible flexibility. And of course, the speed to
[36:13] Speaker 16: power means they can deploy them quickly where they need
[36:16] Speaker 16: them to serve customers better.
[36:19] Speaker 3: Let me ask you something real quick about competing goals.
[36:22] Speaker 3: I had this conversation with Microsoft very recently. There's the
[36:26] Speaker 3: power requirements. You talked about, okay, multiple fuel sources, but
[36:29] Speaker 3: natural gas. A lot of these companies, the hyperscalers in particular,
[36:33] Speaker 3: also have climate goals. How do you meet both of
[36:37] Speaker 3: those obligations?
[36:39] Speaker 16: Well, Mainspring's mission really is to make sure that we
[36:43] Speaker 16: have a sustainable, abundant, and affordable power grid for everybody.
[36:47] Speaker 16: It's really important to us that the grid is able
[36:50] Speaker 16: to take in renewable energy use it where necessary and
[36:54] Speaker 16: where possible, excuse me, and then also to have a
[36:57] Speaker 16: firming source to make sure it's available at all times
[37:00] Speaker 16: and affordable.
[37:01] Speaker 2: That's our goal as a company.
[37:03] Speaker 16: And we think having a dispatchable source, whether it runs
[37:07] Speaker 16: on natural gas or hydrogen, allows you to use more
[37:10] Speaker 16: renewables on the grid and for behind the meter sources.
[37:13] Speaker 16: So our view is that we will help sustain a
[37:17] Speaker 16: low-carbon goal. The fact that we can run on any
[37:20] Speaker 16: fuel means that as new fuels come in, hydrogen and
[37:23] Speaker 16: other low-carbon fuels, we can also react with those, and
[37:27] Speaker 16: that means that our energy becomes completely carbon-free. So our
[37:31] Speaker 16: goal is that sustainable grid, no question about it, while
[37:35] Speaker 16: doing it in an affordable and abundant way.
[37:38] Speaker 3: You've also raised money in the private markets. Is that
[37:42] Speaker 3: kind of a stop gap to help you scale? Why
[37:45] Speaker 3: would you do that if you're getting into these sizable
[37:48] Speaker 3: commercial arrangements?
[37:49] Speaker 1: As you said, we are we're a growing company.
[37:51] Speaker 16: So we have used we've had private investors for many
[37:54] Speaker 16: years now that have supported the company for many years
[37:58] Speaker 16: as we've developed this remarkable technology. And that's been those
[38:02] Speaker 16: investors have seen us through from the very beginning. They
[38:07] Speaker 16: continue to support us in our last private round. And
[38:10] Speaker 16: we've brought in some new investors, investors that can make
[38:14] Speaker 16: larger investments because we're a growing company. And then we'll
[38:17] Speaker 16: continue to find new sources of capital as we need it.
[38:20] Speaker 16: But I do want to thank these investors. They have
[38:22] Speaker 16: supported us for many years. And the company would not
[38:25] Speaker 16: be where it is without them.
[38:27] Speaker 3: I think it's a reasonable question with this AEP deal.
[38:30] Speaker 3: How solid is it, right? How confident are you that
[38:33] Speaker 3: they will continue investing? those next increments of orders. You
[38:37] Speaker 3: have parallel examples. You did a similar deal in 2024
[38:40] Speaker 3: with Bloom Energy, which was for initial 100 megawatts. Have
[38:45] Speaker 3: they moved beyond that 100 megawatts and are they going
[38:48] Speaker 3: into more installments?
[38:50] Speaker 1: They have indeed.
[38:51] Speaker 16: It's a good point that they have done investments like this.
[38:54] Speaker 16: One of the things I love about AEP, they are
[38:56] Speaker 16: not only a large and highly capable utility but they
[39:00] Speaker 16: have actually taken chances on developing new technology to make
[39:05] Speaker 16: sure they can serve their customers with lower cost and
[39:08] Speaker 16: more sustainable energy over time.
[39:10] Speaker 2: I give them a lot of credit for that. And
[39:13] Speaker 2: they've been quite successful doing it.
[39:14] Speaker 16: And their Bloom contract is an example of that.
[39:16] Speaker 1: They were able to do an initial deal with Bloom and.
[39:19] Speaker 2: Then they've been able to purchase more.
[39:21] Speaker 16: And so that's what gives us confidence that if we
[39:23] Speaker 16: demonstrate the capability, I think we will, that we'll have
[39:27] Speaker 16: more opportunity with AEP. And they're one of the the
[39:30] Speaker 16: markets uh people who assess new technologies and give credibility
[39:34] Speaker 16: new technologies so being part of their their growth is
[39:38] Speaker 16: really really it's a great vote of confidence for all
[39:40] Speaker 16: of us at mainspring.
[39:42] Speaker 2: Tom Lineberger, Chairman and CEO of Mainspring. I really appreciate
[39:44] Speaker 2: your time.
[39:45] Speaker 3: Two pieces of news, the deal with AEP and some
[39:47] Speaker 3: funding in the private markets too.
[39:49] Speaker 2: Thank you very much.
[39:50] Speaker 3: Now coming up, Epic Games has delayed the release of
[39:53] Speaker 3: an upcoming video game collaboration with Disney. There's a Bloomberg
[39:57] Speaker 3: report about a big pivot going on at Epic.
[40:00] Speaker 2: We have the details. This is Bloomberg Tech.
[40:12] Speaker 3: Epic Games has delayed a new video game they've partnered
[40:15] Speaker 3: with Disney to make, saying they want their developers to
[40:17] Speaker 3: start using more AI tools. The game was supposed to
[40:20] Speaker 3: be released in March, but now there's no definite release date,
[40:23] Speaker 3: all according to Bloomberg sources and the reporting of Bloomberg
[40:27] Speaker 3: News video game reporter Jason Schreier. Let's start with this
[40:30] Speaker 3: shooter that was supposed to be based on Marvel and
[40:33] Speaker 3: Star Wars franchises.
[40:35] Speaker 2: Tell me your reporting.
[40:37] Speaker 17: Yeah, so this shooter is part of this huge multi-game
[40:40] Speaker 17: collaboration between Epic Games and Disney, which invested $ 1. 5 billion
[40:46] Speaker 17: in Epic back in 2024 in exchange for an equity
[40:49] Speaker 17: stake and this multi-game collaboration. And this game, it was
[40:53] Speaker 17: not going super well, been delayed a couple times, and
[40:57] Speaker 17: now it's delayed again from March. And Epic's solution to
[41:00] Speaker 17: fix things is AI.
[41:04] Speaker 3: So you have a lot of detail in this report,
[41:06] Speaker 3: but basically the instruction to developers on the game is
[41:11] Speaker 3: take an AI-first approach or production model, use things like Claude,
[41:16] Speaker 3: but what is it that they need to do?
[41:18] Speaker 2: Is there something that needs fixing? Is it too slow?
[41:22] Speaker 1: Yeah, it's interesting.
[41:23] Speaker 17: Epic has been investing and kind of doubling down on
[41:27] Speaker 17: generative AI in game development for a while now. A
[41:30] Speaker 17: lot of game companies are, but Epic is really ahead
[41:33] Speaker 17: of this charge. But this is the first time we've
[41:36] Speaker 17: seen them really say, hey, we are going to make
[41:39] Speaker 17: this game AI first. It is going to be built
[41:42] Speaker 17: from now on largely with AI, with human supervision. which
[41:46] Speaker 17: I suppose is them hoping they can get things back
[41:50] Speaker 17: on track, make a more efficient production, make it more quickly,
[41:53] Speaker 17: make it with fewer people.
[41:55] Speaker 8: I'm not sure.
[41:56] Speaker 17: Tim Sweeney, the CEO of Epic, has said that the
[41:58] Speaker 17: goal of this AI push is not to replace people,
[42:02] Speaker 17: but of course some important context here is that Back
[42:05] Speaker 17: in March, Epic Games laid off 1,000 employees, which I
[42:10] Speaker 17: imagine had a big impact on their game development.
[42:15] Speaker 3: Tim Sweeney, the CEO and founder of Epic, he just mentioned,
[42:18] Speaker 3: he posted on X, the social media platform, Just before
[42:21] Speaker 3: our story went out, actually, like literally just, this is
[42:23] Speaker 3: what he said. We have a huge pivot underway here
[42:26] Speaker 3: at Epic, moving developers around to retool for maximum AI acceleration.
[42:32] Speaker 3: You get to this kind of a little bit in
[42:34] Speaker 3: the story, but what happens next?
[42:38] Speaker 17: Yeah, so the big question is just how efficient this
[42:41] Speaker 17: is going to be. So over the next few weeks,
[42:43] Speaker 17: the developers on this project have been told essentially, hey,
[42:47] Speaker 17: forget about the milestones, forget about our upcoming deadlines. Instead,
[42:51] Speaker 17: spend the next few weeks playing around with Claude. see
[42:53] Speaker 17: what you can do with it, see how efficient it
[42:56] Speaker 17: makes you, see how creative it makes you, see if
[42:57] Speaker 17: it can lead to better results than we've been having,
[43:00] Speaker 17: and see what happens. Needless to say, there were some
[43:05] Speaker 17: mixed reactions from people, to say the least. Some consternation,
[43:08] Speaker 17: some stunned reactions of people saying, is this kind of
[43:11] Speaker 17: a joke?
[43:11] Speaker 8: Like, what do we do here? But yes, Epic has.
[43:14] Speaker 17: Said, quote, no handwritten code, and that all of their
[43:18] Speaker 17: art assets need to be, all of the art they're
[43:20] Speaker 17: making for the game needs to be part of an
[43:21] Speaker 17: AI pipeline.
[43:23] Speaker 1: Moving forward.
[43:25] Speaker 3: Jason Schreier of Bloomberg. I really go and read Jason's
[43:27] Speaker 3: report on Bloomberg.com. Thank you very much indeed. One quick
[43:31] Speaker 3: look at markets, particularly the chip sector. It's in chips
[43:34] Speaker 3: where we've seen pain. It started overnight in Korea. particular
[43:37] Speaker 3: selling in Samsung and SK Hynix, the memory names. Now
[43:40] Speaker 3: in the US session, the stock's down for a.
[43:41] Speaker 2: Fifth straight day.
[43:42] Speaker 3: That does it for this edition of Bloomberg Tech. Tune
[43:45] Speaker 3: in later today for an interview with Applied Materials CEO
[43:48] Speaker 3: Gary Dickson on the heels of the company's investor day.
[43:50] Speaker 3: We're going to talk about how the company expected to
[43:52] Speaker 3: double its revenue over the next few years.
[43:54] Speaker 12: Why?
[43:54] Speaker 3: Because of AI. 3.30pm Eastern. Don't forget also to check
[43:58] Speaker 3: out the podcast. Thanks for being with us. This is
[44:00] Speaker 3: Bloomberg Tech.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.