# Apple’s Ternus Era Starts the iPhone Duo

An episode of [Bloomberg Tech](https://invisiblenetwork.ai/podcast/1611767434), 10 Sep 2026, 43 min.

Audio: https://podtrac.com/pts/redirect.mp3/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/41764a4f-fc64-4e11-89ba-ae7c0030ab5e/f4fc3deb-5b6b-4bb1-add0-b4c1011915d8/audio.mp3?utm_source=Podcast&in_playlist=9caafc41-289c-4115-995d-ae7c0030ab75

Bloomberg’s Ed Ludlow discusses Apple's iPhone Duo with Bloomberg's Mark Gurman, from its anticipated reception with consumers to what it means for Apple's new era under CEO John Ternus. Plus, the AI chip race is moving to inference, and Elon Musk's Boring Co. reaches a $23 billion valuation. See omnystudio.com/listener for privacy information.

## Transcript

[0:02] Speaker 1: Bloomberg Audio Studios, podcasts, radio, news.
[0:13] Speaker 2: Bloomberg Tech is live from the heart of Silicon Valley
[0:17] Speaker 2: with Ed Ludlow in San Francisco. This is Bloomberg Tech.
[0:22] Speaker 2: Coming up, it's what everyone's talking about. Apple's first foldable smartphone,
[0:27] Speaker 2: the iPhone Duo, will break everything down from yesterday's product launch. Plus,
[0:31] Speaker 2: the AI chip race is moving to inference. D-Matrix CEO
[0:34] Speaker 2: Sid Sheff joins us on its new partnership with NVIDIA
[0:37] Speaker 2: and the battle to make AI cheaper and faster. And
[0:40] Speaker 2: Elon Musk's Boring Co. reaches a $ 23 billion valuation after
[0:44] Speaker 2: its latest funding round. Let's get straight to our top story,
[0:49] Speaker 2: Apple's showstopper.
[0:51] Speaker 3: iPhone Duo brings an entirely new design while preserving everything
[0:55] Speaker 3: you love and expect from iPhone. It gives you the
[0:58] Speaker 3: largest display ever on an iPhone while still fitting easily
[1:02] Speaker 3: in your pocket. It reflects extraordinarily complex engineering yet feels
[1:07] Speaker 3: effortless to use. And it makes new experiences possible.
[1:12] Speaker 2: That was Apple CEO John Turner making his first official
[1:15] Speaker 2: keynote in the top seat, turning heads with the new
[1:18] Speaker 2: foldable iPhone Duo. Shares really accelerating gains 90 minutes into
[1:22] Speaker 2: the session, up 2.5% now. What's interesting is they were
[1:25] Speaker 2: basically flat, bang in line with the average of an
[1:28] Speaker 2: iPhone launch day market performance that we've seen going back
[1:31] Speaker 2: at least a decade. Well, everyone's woken up and said,
[1:34] Speaker 2: this is a big deal. Let's discuss. And who better
[1:36] Speaker 2: to discuss it with than Mark Gurman, who leads our
[1:39] Speaker 2: Apple coverage and basically broke most of the details prior
[1:42] Speaker 2: to yesterday's event anyway. Start with the basics, iPhone Duo.
[1:46] Speaker 2: We were spot on on the price. That's probably the
[1:49] Speaker 2: big news headline, right? Just under $ 2, 000. But the specs
[1:53] Speaker 2: as well. Your first impressions and the most notable points
[1:56] Speaker 2: of the technology.
[1:59] Speaker 4: My first impression is that this device makes every foldable
[2:02] Speaker 4: phone from Samsung, from Google, from Huawei, anyone else before it,
[2:06] Speaker 4: look like a toy, look like a prototype, look like
[2:08] Speaker 4: a beta test, basically, because the fit and finish we're
[2:11] Speaker 4: seeing here is new to the foldable industry. The way
[2:15] Speaker 4: that they are putting, they call it ceramic inserts into
[2:18] Speaker 4: the cellular antenna cutouts, so when you go and try
[2:21] Speaker 4: to bend the phone and break it, people who do that,
[2:24] Speaker 4: it's not gonna bend in half. The way they put
[2:27] Speaker 4: a special cover glass over the flexible display to hide
[2:31] Speaker 4: the crease as much as possible and make it so
[2:33] Speaker 4: your fingernails don't put cuts in the screen. The fact
[2:36] Speaker 4: that they're using titanium around the edges to make the
[2:40] Speaker 4: phone more durable, the fact that it's water and dust resistant, right?
[2:44] Speaker 4: They've hit every checkbox here for it to feel like
[2:47] Speaker 4: a normal iPhone, but to also be foldable. This really
[2:50] Speaker 4: has very few compromises other than the lack of the
[2:54] Speaker 4: telephoto camera on the back and battery life to some extent.
[2:59] Speaker 2: It was amazing to get hands on with it yesterday. And,
[3:01] Speaker 2: you know, everyone's sort of first reaction is to say
[3:05] Speaker 2: there's no crease, there's no margin, there's no seam in
[3:08] Speaker 2: the middle, not one that the naked eye can see.
[3:10] Speaker 2: I do want to talk about the pricing, Mark. You
[3:13] Speaker 2: said this would be $ 2, 000. It was in the end $ 1, 999.
[3:18] Speaker 2: To what extent has Apple eaten some cost in the
[3:21] Speaker 2: environment they're operating in? But also, you know, if you
[3:25] Speaker 2: look at the higher tiers of storage, clearly there are
[3:28] Speaker 2: some handsets out there up beyond $ 3, 000 too.
[3:32] Speaker 4: So they're eating costs for sure because the memory shortage,
[3:36] Speaker 4: you know, really is having an impact on so many companies,
[3:39] Speaker 4: including Apple, delaying products, paying more for memory, needing to
[3:43] Speaker 4: charge customers more. But they've offset it in two key ways.
[3:47] Speaker 4: One is on the storage tiers. So you can get
[3:49] Speaker 4: the 256 gigabytes for that $ 2, 000 price point. You can
[3:53] Speaker 4: get 512 for the $ 2, 200 price point. But if you
[3:56] Speaker 4: want to go to the higher tiers, you're talking 2,600
[3:58] Speaker 4: for a terabyte and then 3,200 for.
[4:01] Speaker 2: For two terabytes.
[4:02] Speaker 4: So they're charging massive premiums on that storage. And their
[4:06] Speaker 4: thinking is that most people probably aren't going to get
[4:09] Speaker 4: the base level version. This is an early adopter product
[4:12] Speaker 4: for all intents and purposes. And that group of people
[4:14] Speaker 4: is going to go for higher storage tiers. So they
[4:17] Speaker 4: are going to get their margin on that. The other
[4:20] Speaker 4: thing they've done is they've really jacked up pricing on
[4:23] Speaker 4: the international customers. So in the US, right, the $ 2, 000
[4:26] Speaker 4: price point globally is going to be the one everyone's
[4:28] Speaker 4: talking about. But you go internationally, the prices in Canada,
[4:32] Speaker 4: the price in Australia, the price in Asia, really sky high.
[4:36] Speaker 4: And if you do some of the conversions, the prices
[4:38] Speaker 4: in those other countries, it's not a one-to-one conversion for
[4:41] Speaker 4: what the U.S.
[4:41] Speaker 1: Price is.
[4:42] Speaker 4: So, for instance, in Australia, there's an $ 800 Australian dollar
[4:47] Speaker 4: premium on top of the converted rate. So that's how
[4:50] Speaker 4: they're making up for that shortfall.
[4:53] Speaker 2: Mark, to me, John Ternus tried to do two things
[4:56] Speaker 2: with his first formal keynote as CEO. Talk about the
[5:00] Speaker 2: smartphone being the device for the AI era, but also
[5:03] Speaker 2: he put a lot of emphasis on your iPhone or
[5:06] Speaker 2: whichever OS you're using knows you. It has your data.
[5:09] Speaker 2: It has your context. What did you make of that
[5:11] Speaker 2: messaging and Mr. Ternus's first full iPhone day event?
[5:17] Speaker 4: Yeah, I mean, for the first time, Apple came out
[5:19] Speaker 4: and showcased a real AI strategy. Whether it's an AI
[5:24] Speaker 4: strategy that I agree with, that you agree with, anyone
[5:27] Speaker 4: agrees with, it doesn't matter. At least they have a
[5:30] Speaker 4: strategy now. And it's interesting they haven't really showcased a
[5:34] Speaker 4: coherent strategy until now. He calls it the intelligent personal hub,
[5:38] Speaker 4: the phone being the center of everything for artificial intelligence.
[5:42] Speaker 4: It hearkens back to their previous intelligent hubs That all
[5:46] Speaker 4: started with the Mac. Steve Jobs introduced the Mac as
[5:49] Speaker 4: sort of the center of everything, and you have all
[5:51] Speaker 4: those peripherals around it.
[5:53] Speaker 2: The iPod, you name it. Then came iCloud.
[5:57] Speaker 4: iCloud is the center of everything, and the Mac, the iPad,
[6:00] Speaker 4: the phone, those are all peripherals around the cloud. Now,
[6:03] Speaker 4: the center of everything is sort of that intelligent hub
[6:06] Speaker 4: and the iPhone with everything built around that, whether it's
[6:09] Speaker 4: your iPad, your Mac, your AirPods, your Apple Watch, and
[6:11] Speaker 4: all the AI wearables are going to have in the future.
[6:14] Speaker 4: It's all built around AI in the phone. So it's
[6:16] Speaker 4: a pretty interesting strategy. I do think it's going to
[6:19] Speaker 4: work to some extent, but obviously it's still early innings
[6:23] Speaker 4: for their AI and there's more to come and there's
[6:26] Speaker 4: more to do.
[6:27] Speaker 2: Bloomberg's Mark Gurman, who leads our coverage of Apple, but
[6:30] Speaker 2: also reported so deeply on iPhone Duo, including what its
[6:34] Speaker 2: name would be prior to its announcement yesterday. Let's stick
[6:37] Speaker 2: with Apple, but also we're going to talk more about
[6:39] Speaker 2: technology markets. Nuveen's Chief Investment Officer, Sarah Malik, joins us now. Actually,
[6:43] Speaker 2: the Apple move today is really interesting, up 2.5%. If
[6:46] Speaker 2: you go back and look at sort of the average
[6:48] Speaker 2: performance on the day itself of an iPhone event and
[6:51] Speaker 2: next day, it doesn't often get the momentum we're seeing
[6:54] Speaker 2: right now. I think you probably want to talk to
[6:57] Speaker 2: what Mark was just talking about in terms of Apple
[6:59] Speaker 2: having an AI strategy. That's been a bearing on the stock, Sarah.
[7:02] Speaker 2: Your thoughts, please. Well, definitely.
[7:05] Speaker 5: There's a lot of consumer interest and excitement over the
[7:07] Speaker 5: new foldable phone. But there's two issues that investors are
[7:11] Speaker 5: thinking about around Apple and this new phone. So, first
[7:14] Speaker 5: of all, with the phone, the $ 2, 000 price point or
[7:18] Speaker 5: $ 19. 99 may lead to a more modest addressable market from
[7:20] Speaker 5: the start. And we'll grow from there. But secondarily, is
[7:24] Speaker 5: this question around AI? The new CEO has a hardware
[7:27] Speaker 5: background and Apple's very strong on hardware. But what investors
[7:30] Speaker 5: want to know is, can Apple become more competitive in AI?
[7:33] Speaker 5: They have somewhat lagged their peers in terms of their
[7:36] Speaker 5: push into AI. So really, that is the key question
[7:39] Speaker 5: for investors and Apple going forward. Is this AI strategy
[7:42] Speaker 5: going to turn into something that has monetization for the company?
[7:45] Speaker 5: Of course, this is a wait and see. So happy
[7:47] Speaker 5: to see the stock up based on the excitement over
[7:50] Speaker 5: a new phone. But I think the AI question is
[7:52] Speaker 5: the medium term question for whether the stock.
[7:54] Speaker 6: Outperforms or not.
[7:56] Speaker 2: There is lots happening in financial markets right now. You
[7:58] Speaker 2: look at oil, Brent, $ 105 a barrel. Yields are higher.
[8:01] Speaker 2: Technology stocks are lower. Chip stocks are underperforming. And there
[8:05] Speaker 2: are lots of pieces of economic data and macro events
[8:08] Speaker 2: that are driving this market. For you this morning, what's
[8:11] Speaker 2: going on? What's happening?
[8:13] Speaker 5: It is all about higher rates and higher rates are
[8:16] Speaker 5: tougher for long duration technology stocks. And really, even though
[8:20] Speaker 5: we had a very strong second quarter earnings season driven
[8:22] Speaker 5: by the tech sector, it's higher oil price prices, potentially
[8:27] Speaker 5: higher rates at the FOMC meeting next week, a rate
[8:31] Speaker 5: hike out of the ECB. All of that is causing
[8:33] Speaker 5: this turmoil for the markets. The biggest data point we're
[8:37] Speaker 5: going to see before the FOMC meets next week is CPI,
[8:40] Speaker 5: which comes out tomorrow. If that number comes out hot,
[8:43] Speaker 5: and by that I mean above 0.3% on the plus
[8:47] Speaker 5: side month over month, I think you're going to lock
[8:49] Speaker 5: in a hike for next week. If it comes in soft,
[8:51] Speaker 5: under 0.2% month over month, that might cause the Fed
[8:55] Speaker 5: to pause. But either way, we're seeing long-term treasury yields
[8:59] Speaker 5: at up over 5%, oil prices over 105. That is
[9:02] Speaker 5: very concerning for the markets. And there is no end
[9:04] Speaker 5: in sight at this point for the geopolitical issues and
[9:07] Speaker 5: the war.
[9:07] Speaker 1: In the Middle East.
[9:08] Speaker 5: So I think this turmoil for the markets is going
[9:10] Speaker 5: to continue because it's hard to see the path for
[9:13] Speaker 5: yields to start to decline from here, especially with the
[9:15] Speaker 5: Fed either pausing or hiking rates next week.
[9:19] Speaker 2: For me, Sarah, I'm going to go back to Apple, actually,
[9:21] Speaker 2: because if you think about what the market, at least
[9:23] Speaker 2: on the equity side, is doing this morning, Apple's massively
[9:26] Speaker 2: outperforming from a points perspective at the index level. Would
[9:30] Speaker 2: that indicate to you that investors are genuinely cheering what
[9:33] Speaker 2: was outlined by the company yesterday?
[9:37] Speaker 5: The positive here for Apple and out of the Goldman
[9:39] Speaker 5: Sachs Technology Conference that's happening this week is we are
[9:42] Speaker 5: hearing positive data points on monetization of artificial intelligence, which
[9:46] Speaker 5: is what everybody has been waiting for. Today, what you're
[9:49] Speaker 5: seeing for Apple is great. It's fundamentals outperforming the macro issue.
[9:53] Speaker 5: So overall for the markets, the macro overhang of higher
[9:56] Speaker 5: oil prices and potential spark of inflation because of that
[10:00] Speaker 5: is definitely a cloud over the markets. But Apple's new
[10:03] Speaker 5: product launch, it's exciting. It's a new style of phone.
[10:07] Speaker 5: It is at a higher price point, but that new
[10:09] Speaker 5: style of phone can start a new product cycle for Apple.
[10:12] Speaker 5: And that's what people are excited about. But as I mentioned,
[10:14] Speaker 5: there is a key question for Apple, which is how
[10:16] Speaker 5: are they going to gain more traction in in AI
[10:19] Speaker 5: with their phones. And it's great to see a strategy.
[10:22] Speaker 5: And now we'd like to see the traction and the
[10:24] Speaker 5: monetization from that.
[10:26] Speaker 2: We are kind of late in a technology earnings cycle
[10:29] Speaker 2: where they keep coming. Oracle and Adobe's after the bell.
[10:33] Speaker 2: One software name, one infrastructure name with a software element.
[10:37] Speaker 2: How much of importance do investors continue to assign the
[10:41] Speaker 2: data points coming out of earnings in terms of where
[10:43] Speaker 2: we are in the cycle right now?
[10:45] Speaker 5: Well, every technology data point counts because technology earnings are
[10:49] Speaker 5: what are leading overall earnings growth, especially for the S &
[10:52] Speaker 5: P 500 and in the U.S. So Oracle and Adobe
[10:55] Speaker 5: will be important. We may get some commentary around hyperscalers
[10:59] Speaker 5: and CapEx and whether we're seeing any monetization. from AI investments,
[11:03] Speaker 5: all of that's going to be important. But I think
[11:05] Speaker 5: the issue is with an FOMC meeting next week and
[11:08] Speaker 5: CPI data coming out tomorrow, people are more focused right
[11:11] Speaker 5: now on what do these higher oil prices mean for inflation?
[11:15] Speaker 5: What does it mean for the US economy? As an example,
[11:18] Speaker 5: we did get very strong payrolls last week, but if
[11:22] Speaker 5: inflation stays high and we become worried that oil prices
[11:24] Speaker 5: will cause the economy to slow, then we could start
[11:27] Speaker 5: talking about stagflation. I think all of that for now
[11:30] Speaker 5: is going to be a bigger issue for the markets
[11:32] Speaker 5: than a couple of more incremental technology earnings data points.
[11:36] Speaker 2: We've covered an iPhone day. We covered technology earnings. We
[11:38] Speaker 2: covered macros. We covered higher rates. We covered oil. We
[11:42] Speaker 2: covered inflation and then the Fed meeting next week. That's
[11:44] Speaker 2: pretty good going. Nuveen Chief Investments Officer Sarah Malik back
[11:47] Speaker 2: on Bloomberg Tech. Thank you very much. Coming up, D-Matrix
[11:51] Speaker 2: is plugging into NVIDIA's AI ecosystem. CEO Sid Sheff joins
[11:55] Speaker 2: us on the new partnership and why the next big
[11:58] Speaker 2: battle in AI chips is inference. This is Bloomberg Tech.
[12:09] Speaker 2: TSMC reported better than expected sales in August, but its
[12:12] Speaker 2: US-listed shares are under pressure. As investors and analysts start
[12:16] Speaker 2: to worry the company doesn't have enough supply to meet
[12:18] Speaker 2: robust demand, let's discuss with Bloomberg's Peter Elstrom. I mean,
[12:21] Speaker 2: this is the dynamic of industry right now. In the
[12:24] Speaker 2: AI context, demand for lead-edge chips is running ahead of
[12:27] Speaker 2: the industry's ability to supply. What did TSMC say?
[12:32] Speaker 7: Yeah, the numbers are pretty staggering. I think we're familiar
[12:34] Speaker 7: with the trend, but the numbers are quite staggering here.
[12:37] Speaker 7: So First of all, TSMC reported a 53% rise in
[12:41] Speaker 7: revenue for August. Taiwan does this thing where they have
[12:44] Speaker 7: companies report every month. So we get a sense of
[12:47] Speaker 7: how the quarter is going for them. That's above the
[12:50] Speaker 7: expectations that analysts have for the company to boost revenues
[12:53] Speaker 7: about 47% for the quarter. We had an interview with
[12:56] Speaker 7: the COO though, Cliff Ho, and he talked to us
[12:58] Speaker 7: a bit about the constraints that they've got on supply
[13:01] Speaker 7: right now. They say that typically in a typical year,
[13:04] Speaker 7: they'd be building four or five fabs someplace. Right now
[13:07] Speaker 7: they're trying to build 20 fabs. They're so far behind
[13:10] Speaker 7: on demand here that they just can't keep up. So
[13:12] Speaker 7: they're trying to build fabs as quickly as they can.
[13:14] Speaker 7: Of course, fabs takes years to build. They're very complicated.
[13:18] Speaker 7: They're very expensive. TSMC is the best in the world
[13:21] Speaker 7: at building them, but they're having a hard time getting
[13:23] Speaker 7: them built in as quick of a manner as they
[13:26] Speaker 7: would like right now. And partly it's what you were
[13:28] Speaker 7: talking about before. Apple certainly wants more advanced chips, especially
[13:32] Speaker 7: Nvidia in the AI ecosystem wants to be able to
[13:35] Speaker 7: get more capacity online as soon as they possibly.
[13:37] Speaker 2: Can get it. That's on the manufacturing side. On the
[13:40] Speaker 2: fabless chip maker side, MediaTek, a company we've been covering
[13:43] Speaker 2: very closely on the show recently, out with very strong numbers,
[13:46] Speaker 2: a lot of momentum for their chips.
[13:49] Speaker 7: Yeah, I mean, you alluded to this earlier that we've
[13:52] Speaker 7: had MediaTek CEO on the show talking about this new
[13:55] Speaker 7: joint venture that they've got with NVIDIA in particular. MediaTek
[14:00] Speaker 7: is also growing very quickly. They reported strong revenue growth
[14:03] Speaker 7: of 44%. They historically have done very well on the
[14:06] Speaker 7: smartphone market. Now they're moving into these adjacent markets, especially
[14:10] Speaker 7: in AI services. They've struck this deal with NVIDIA where
[14:13] Speaker 7: they want to be able to incorporate their technology in
[14:17] Speaker 7: with NVIDIA's technology. They want to also move into being
[14:19] Speaker 7: able to make chips that will work in laptops and
[14:22] Speaker 7: computers and bring more AI capabilities there too. So the
[14:26] Speaker 7: company is growing quite strong at this point. They also
[14:28] Speaker 7: are trying to keep up with demand.
[14:31] Speaker 2: Bloomberg's Peter Elstrom, who leads our coverage of Asian technology companies.
[14:34] Speaker 2: Thank you very much indeed. NVIDIA is confronting a world
[14:37] Speaker 2: of more custom chips for more players and growth in
[14:40] Speaker 2: specialized inference systems. Its response has been embrace it.
[14:45] Speaker 8: There's no question NVIDIA's footprint is the largest. There's no
[14:47] Speaker 8: question we're growing share in closed models and open models.
[14:53] Speaker 8: We're growing share across the entire AI opportunity. And so
[14:58] Speaker 8: we're not intimidated by this. We're, in fact, welcoming it
[15:04] Speaker 8: and opening our platform so that they could connect XPUs
[15:07] Speaker 8: into it.
[15:08] Speaker 2: So that was NVIDIA CEO Jensen Wang speaking to me
[15:11] Speaker 2: after the deal with MediaTek on XPUs last month. Today,
[15:14] Speaker 2: startup D-Matrix is the latest to partner with NVIDIA to
[15:18] Speaker 2: plug its specialized inference XPUs into NVIDIA's platforms, targeting ultra-low
[15:22] Speaker 2: latency AI workloads, where its tech works alongside NVIDIA GPUs
[15:27] Speaker 2: using NVLink fusions. Dematrix CEO Sid Shash joins us now.
[15:32] Speaker 2: Very interesting. Why is it a benefit to you to
[15:36] Speaker 2: be able to plug into a largely NVIDIA-based architecture? Yes.
[15:40] Speaker 9: First of all, Ed, I'm a fan of the show,
[15:42] Speaker 9: so thanks for having me on it. This is a
[15:45] Speaker 9: landmark announcement for Dematrix, obviously, and the biggest reason we
[15:51] Speaker 9: embarked on this partnership with NVIDIA, and we've been working
[15:54] Speaker 9: on this for six-plus months with them, is exactly what
[15:58] Speaker 9: Jensen said in his comments earlier. This is probably, most certainly,
[16:03] Speaker 9: the largest deployed infrastructure base for AI in the world.
[16:08] Speaker 1: It is in every cloud.
[16:10] Speaker 9: NVIDIA hardware is in every cloud, every country, pretty much
[16:13] Speaker 9: every data center in the world. At D-Matrix, we have
[16:16] Speaker 9: a very unique memory-centric inference computing technology. We have been
[16:21] Speaker 9: working on this for seven plus years, really targeted for
[16:24] Speaker 9: ultra-low latency inference. And ultra low latency inference computing just
[16:30] Speaker 9: took off in a massive way in the last 12 months,
[16:33] Speaker 9: post GPT Codecs and Cloud Code and with the arrival
[16:36] Speaker 9: of agentic coding, everyone is looking for how they can
[16:40] Speaker 9: quickly get access to low latency compute so that they
[16:43] Speaker 9: can really interact with these agentic coding tools.
[16:46] Speaker 1: And that's where D-Matrix comes in.
[16:49] Speaker 9: And this partnership will really allow us to take our
[16:52] Speaker 9: products to market very rapidly with the NVIDIA partnership.
[16:57] Speaker 2: Sid, in Mediatek's case, NVIDIA made a very significant investment
[17:02] Speaker 2: through the purchase of convertible notes. Under your arrangement with NVIDIA,
[17:06] Speaker 2: are they making an equity or any other kind of
[17:09] Speaker 2: investment into D-Matrix to help you get moving?
[17:13] Speaker 9: At this point, Ted, I couldn't comment on the investment
[17:15] Speaker 9: piece specifically, but look, this is a partnership that is
[17:20] Speaker 9: very strategic to NVIDIA. And they are putting a lot
[17:24] Speaker 9: of resources, engineering resources. There is obviously a lot of
[17:30] Speaker 9: work the two teams have been doing for six plus months.
[17:34] Speaker 9: And our first goal is to really make sure that
[17:37] Speaker 9: the Raptor XPU that is going to be integrated into
[17:40] Speaker 9: the NVIDIA ecosystem is going to be ready to come
[17:43] Speaker 9: to market in about 12 months. So the biggest opportunity
[17:47] Speaker 9: here is really about getting that product in customers' hands
[17:50] Speaker 9: because we've got a lot of customer pull for that
[17:53] Speaker 9: specific product in the NVIDIA ecosystem.
[17:55] Speaker 2: So on your newest chip or your newest platform, I
[17:57] Speaker 2: should say, Is there any data point you can give
[18:00] Speaker 2: me on how the partnership with NVIDIA tangibly allows you
[18:04] Speaker 2: to move faster? In other words, will this deploy into
[18:06] Speaker 2: the real world quicker than it otherwise would had you
[18:09] Speaker 2: not struck a deal with NVIDIA for NVLink Fusion?
[18:12] Speaker 9: Yeah, 100%. Look, the Raptor XPU, which is the specific
[18:18] Speaker 9: XPU that's in partnership with NVIDIA and will be launched
[18:22] Speaker 9: as part of this partnership first, is going to be
[18:24] Speaker 9: the world's first 3D-stacked DRAM XPU, right? And just let
[18:30] Speaker 9: me touch upon why that is so important. It is
[18:33] Speaker 9: important because memory is really the biggest bottleneck for AI
[18:37] Speaker 9: right now.
[18:38] Speaker 1: And we all know that.
[18:39] Speaker 9: And what we have done is instead of using HBM,
[18:41] Speaker 9: we don't use any HBM at D-Matrix. What we have
[18:45] Speaker 9: done is we have taken DRAM and found a way
[18:49] Speaker 9: to package it directly.
[18:51] Speaker 1: With the compute.
[18:53] Speaker 9: And it is 3D stacked exactly like you just saw
[18:56] Speaker 9: in the picture there. And that allows for the architecture
[19:00] Speaker 9: to punch through the memory wall, which is really the
[19:03] Speaker 9: biggest bottleneck in AI today in terms of addressing latency,
[19:07] Speaker 9: addressing energy efficiency, addressing cost. And we have found a recipe.
[19:11] Speaker 9: We have cracked the code on how to do it.
[19:13] Speaker 9: And we'll be the first to bring that solution to market.
[19:16] Speaker 9: So there is tremendous customer pull for that product right
[19:19] Speaker 9: now because no other company is going to be within
[19:22] Speaker 9: a two-year window of getting access to that technology. And
[19:26] Speaker 9: this NVIDIA partnership will truly allow us to take that
[19:28] Speaker 9: to market the fastest because there is already a deployed
[19:31] Speaker 9: base of all that infrastructure and all the data centers.
[19:33] Speaker 9: And we don't want to reinvent that part of the wheel.
[19:35] Speaker 9: We'd much rather just ride on the NVIDIA ecosystem and
[19:37] Speaker 9: get the product to our customers as soon as possible.
[19:40] Speaker 2: So, Sid, we're short of time here. Who are those
[19:42] Speaker 2: new customers? If I look at your cap table, there's
[19:45] Speaker 2: lots of sovereign wealth funds on there who have lots
[19:47] Speaker 2: of national AI projects. We just have 30 seconds.
[19:50] Speaker 6: Yes, yes.
[19:51] Speaker 9: It's a combination, Ed, and we'll be announcing these customers.
[19:54] Speaker 9: Stay tuned. It is a combination of hyperscalers, frontier labs, sovereigns,
[19:58] Speaker 9: inference clouds, high-frequency traders.
[20:01] Speaker 1: Stay tuned.
[20:02] Speaker 9: We're going to be announcing these customers in pretty short order.
[20:05] Speaker 2: Sid Sheff of D-Matrix, thank you so much for your time.
[20:08] Speaker 2: There's another important story. As NVIDIA expands its ties with
[20:11] Speaker 2: specialized AI chip makers, one of those deals is drawing scrutiny.
[20:15] Speaker 2: The Justice Department is investigating NVIDIA's $ 20 billion tie-up with
[20:20] Speaker 2: AI inference startup GROK and whether it was structured to
[20:24] Speaker 2: avoid an antitrust review, according to sources. Grok remained independent,
[20:29] Speaker 2: but its CEO and COO moved to NVIDIA as a
[20:32] Speaker 2: part of the deal. US lawmakers have called the arrangement
[20:35] Speaker 2: a takeover that stifled competition. NVIDIA representatives and a DOJ
[20:40] Speaker 2: spokesperson declined to comment. A representative for Grok did not
[20:43] Speaker 2: respond to requests for comment.
[20:47] Speaker 10: It's time now for Talking Tech. I'm Yahaira Ana. And
[20:49] Speaker 10: first up, OpenAI is ending its deal for government agencies
[20:53] Speaker 10: to use its models for $ 1 per year. That's according
[20:57] Speaker 10: to the General Services Administration. Starting October 1st, agencies will
[21:01] Speaker 10: have to pay based on how much they use them
[21:03] Speaker 10: with a 50% discount off standard pricing. Plus, Anthropic is
[21:08] Speaker 10: giving the EU's cybersecurity agency access to its most powerful
[21:12] Speaker 10: mythos AI model after months of back and forth over
[21:16] Speaker 10: who should be allowed to use it. Of course, Anthropic
[21:18] Speaker 10: has tightly restricted access because it says the model is
[21:22] Speaker 10: so powerful it can uncover cybersecurity flaws on its own.
[21:27] Speaker 10: And DeepSeek unveiled its latest AI model that charges as
[21:31] Speaker 10: little as a fraction of a cent per million tokens,
[21:35] Speaker 10: putting even more pressure on rivals to bring down prices.
[21:39] Speaker 10: DeepSeek even claims the new model outperforms Moonshot's Kimi K3.
[21:43] Speaker 11: Ed.
[21:45] Speaker 2: Thank you. You're high, Rick. Welcome back to Bloomberg Tech.
[21:53] Speaker 2: Let's get back to Apple and the iPhone duo. Bloomberg
[21:56] Speaker 2: Intelligence Senior Analyst Anurag Rana writing today that, quote, we
[22:00] Speaker 2: anticipate a much better consumer reception than recent models. Anurag
[22:05] Speaker 2: joins us now. I knew this was a big deal
[22:08] Speaker 2: because while I was at Apple Park running around the
[22:11] Speaker 2: Steve Jobs Theater, I got an email from you. And
[22:14] Speaker 2: I thought, OK, Anurag has something important to say here.
[22:16] Speaker 2: You're really bullish on how many of these you think
[22:18] Speaker 2: that Apple will ship in the first year.
[22:21] Speaker 12: Yeah, I think the rumor was that they had asked
[22:23] Speaker 12: suppliers to do about 10 million units or so. But
[22:27] Speaker 12: I think for the first time in years, I've seen
[22:29] Speaker 12: a product that is designed far better than the earlier
[22:32] Speaker 12: generation of products. The UI was completely mind-blowing in terms
[22:36] Speaker 12: of how the software was infused. I mean, I could
[22:40] Speaker 12: see so many new use cases in terms of consumers
[22:43] Speaker 12: having things side by side. So again, you have to remember,
[22:46] Speaker 12: the ultra side of the Apple consumer is fairly expensive
[22:49] Speaker 12: and rich, I would say. So if you are selling about,
[22:52] Speaker 12: let's say, 240, 250 million units in a year, you know,
[22:56] Speaker 12: 14 million is not that big of a deal in
[22:58] Speaker 12: terms of like that's not even, you know, 6% of
[23:00] Speaker 12: the total units you're going to sell in a year.
[23:03] Speaker 12: And there are massive amount of affluent consumers who would
[23:06] Speaker 12: do that. In fact, in our own circle, I asked
[23:09] Speaker 12: a bunch of people and I got an overwhelming response
[23:11] Speaker 12: that they actually may be thinking of upgrading for the
[23:14] Speaker 12: first time in years.
[23:17] Speaker 2: The price point, Anurag, what was notable about the psychology
[23:23] Speaker 2: of coming just under $ 2, 000 at the entry level?
[23:28] Speaker 12: See, I think 2000 is exactly what Mark Gurman had reported.
[23:31] Speaker 12: So we were looking into that. But what I was
[23:34] Speaker 12: surprised about is when you look at the upgrade plan,
[23:37] Speaker 12: it was priced at about $ 57. 99 a month for 24 months,
[23:41] Speaker 12: which means the residual value after two years is also
[23:44] Speaker 12: still over $ 600. So that number was a bit surprising
[23:47] Speaker 12: to me because I would have thought that would be
[23:48] Speaker 12: a little higher. I think that's a very attractive price
[23:51] Speaker 12: point for people who want to finance it. I mean,
[23:54] Speaker 12: $ 58 a month is not that much if you are
[23:56] Speaker 12: thinking about it from that point of view.
[23:59] Speaker 2: That's interesting because this is, I imagine, just first generation, right?
[24:02] Speaker 2: They'll iterate on top of this. Also, Anurag, if you would,
[24:07] Speaker 2: your assessment of John Ternus as CEO, the Apple iPhone
[24:12] Speaker 2: Day keynote for the first time in that role, what
[24:14] Speaker 2: did you make of all of it?
[24:16] Speaker 12: I think one of the things that really caught my
[24:18] Speaker 12: attention was that the center of discussion was around the
[24:22] Speaker 12: product quality, design, and the integration of the whole ecosystem.
[24:26] Speaker 12: He talked about it a lot, and I think that's where,
[24:29] Speaker 12: going back to the roots of Steve Jobs, you know,
[24:32] Speaker 12: the focus on that. And again, you know, Mark has
[24:35] Speaker 12: said that, you know, he has been working on this
[24:37] Speaker 12: project for a very long period of time. I'm really
[24:40] Speaker 12: curious to see if he's able to spend some more money,
[24:44] Speaker 12: I would say a little more capex into new product developments.
[24:47] Speaker 12: Even if it's a dud, it's okay. But, you know,
[24:49] Speaker 12: more experimentation would be really nice. And I have a
[24:52] Speaker 12: feeling that we're going to see that from him.
[24:55] Speaker 2: Anurag Rana of Bloomberg Intelligence with the React. Thank you
[24:58] Speaker 2: so much. I want to get more reaction from the
[25:00] Speaker 2: sell side. Wamsi Mohan, who leads research covering IT hardware
[25:03] Speaker 2: and tech supply chain at Bank of America, reiterating a
[25:06] Speaker 2: buy rating on the stock following the launch of iPhone
[25:10] Speaker 2: Duo and saying that the iPhone Duo, quote, impresses. Wamsi
[25:14] Speaker 2: Mohan of Bank of America joins us now. What about
[25:17] Speaker 2: it specifically impresses you?
[25:20] Speaker 1: Yeah, good morning, Ed.
[25:22] Speaker 13: Thanks for having me. Look, I think that this is
[25:25] Speaker 13: where Apple excels. It's in bringing the hardware and the
[25:28] Speaker 13: software and the services, all of those, and tightly integrating it.
[25:33] Speaker 13: I was at Apple Park yesterday. I had an opportunity
[25:36] Speaker 13: to feel the device. The quality of it really feels excellent.
[25:41] Speaker 13: You can play with it, open it, close it. And
[25:44] Speaker 13: the way that the software morphs around that, the way
[25:47] Speaker 13: that the apps rearrange themselves around that, that's the magic
[25:50] Speaker 13: that Apple brings. And it was truly an excellent product.
[25:53] Speaker 13: It was definitely a home run from a new keynote
[25:58] Speaker 13: from a new CEO standpoint. I think one of the
[26:01] Speaker 13: things that's also very impressive is that obviously Apple has
[26:05] Speaker 13: worked on this for a long time, but we're seeing
[26:07] Speaker 13: the same ethos of product design, the importance of this
[26:10] Speaker 13: whole stack You've heard of a lot of new chips
[26:14] Speaker 13: that are being designed into this product specifically for this product.
[26:18] Speaker 13: I think all of these are things that are very
[26:20] Speaker 13: core to the Apple ecosystem, and we're seeing a reiteration
[26:22] Speaker 13: of that again.
[26:24] Speaker 2: Wamsi, how many models or do you have a model
[26:26] Speaker 2: for the number of units you think Apple can do
[26:28] Speaker 2: on Duo in the December quarter and its first 12-month period?
[26:34] Speaker 13: Yeah, so actually when we were in Asia back in March,
[26:38] Speaker 13: we picked up that the foldable was somewhere between 10
[26:41] Speaker 13: to 20 million units is what Apple was asking for production.
[26:44] Speaker 13: And so as we look through what can happen in
[26:48] Speaker 13: the first year, it's going to depend a little bit
[26:50] Speaker 13: on whether or not some of these installment plans are
[26:53] Speaker 13: taken advantage of. There's also the handoff feature, which I
[26:55] Speaker 13: think is very interesting and coming at a time when
[26:59] Speaker 13: people are also facing pressure from a spending perspective. And
[27:03] Speaker 13: so I would say somewhere between that 10 to 20
[27:06] Speaker 13: million units for the year is pretty realistic.
[27:08] Speaker 1: In year one.
[27:09] Speaker 13: I think about 17 million units for the year or
[27:11] Speaker 13: so is where we shake out. Now, as we look
[27:14] Speaker 13: into the following year, I think that TAM really starts
[27:16] Speaker 13: to expand. You can also bring pricing down a little
[27:19] Speaker 13: bit on some of the models and have a broader
[27:20] Speaker 13: skew and a broader price spectrum around this product as well.
[27:24] Speaker 13: But I think to start off, this is going to
[27:26] Speaker 13: be a very, very well-coveted and desired product.
[27:30] Speaker 2: Well, let's talk a little bit more about the handoff.
[27:33] Speaker 2: It's something we discuss less with Bloomberg's Mark Gurman. But
[27:35] Speaker 2: the idea is that that could drive an upgrade cycle
[27:39] Speaker 2: in and of itself because you make it easy for
[27:42] Speaker 2: a consumer to go from a this shape phone candy
[27:46] Speaker 2: bar shape to to the foldable. Do you really see
[27:50] Speaker 2: that as the catalyst?
[27:53] Speaker 13: Yeah, I wouldn't say that's the only catalyst. I think
[27:55] Speaker 13: the feature functionality, it's a very good standalone device by itself.
[27:59] Speaker 13: I just think that for pro users, the ability to
[28:03] Speaker 13: have handoff, for instance, if I'm carrying two separate phones,
[28:06] Speaker 13: one for my work, one for my personal use, and
[28:08] Speaker 13: then I'm able to hand off any of those at
[28:11] Speaker 13: any given point in time for productivity reasons, that can
[28:14] Speaker 13: be really appealing. So I think it does sort of
[28:17] Speaker 13: substitute almost as a productivity enhancer, which I would say...
[28:22] Speaker 13: is more of me not using an iPad mini almost
[28:25] Speaker 13: instead of using and instead use the Duo is the
[28:28] Speaker 13: way that I would.
[28:28] Speaker 1: Think about it.
[28:29] Speaker 13: It's complimentary more so to the iPad from a productivity
[28:33] Speaker 13: standpoint rather than an incremental yet another device that you're
[28:37] Speaker 13: going to have to carry.
[28:39] Speaker 2: I'd love to get into the specifics of the A20
[28:41] Speaker 2: Pro chip. You mentioned the chip innovation, two nanometer process,
[28:45] Speaker 2: six core CPU, seven core GPU, 40% faster graphics, but
[28:49] Speaker 2: basically much better in the AI context. Was that a
[28:53] Speaker 2: real move forward from Apple, you think?
[28:56] Speaker 13: Yeah, look, I mean, I think it's very clear that
[28:59] Speaker 13: right at the beginning, right, John Ternes' opening statements and
[29:03] Speaker 13: in the keynote, it was very obvious that AI is
[29:06] Speaker 13: going to be very, very central to the strategy. We've
[29:09] Speaker 13: written extensively about the fact that Apple is going to
[29:11] Speaker 13: be a winner at AI.
[29:12] Speaker 2: At the Edge.
[29:13] Speaker 13: And we think that this whole... notion around privacy, being
[29:17] Speaker 13: able to do a lot of things on device is
[29:20] Speaker 13: where Apple's going to excel. And for that, you're going
[29:22] Speaker 13: to need more memory. You're going to need faster processors.
[29:25] Speaker 13: You're seeing innovation over there with the A20, as you
[29:27] Speaker 13: just mentioned that. And I think that the key here
[29:30] Speaker 13: is to be able to do more by loading sparse models,
[29:35] Speaker 13: by being able to take advantage of the architecture, which
[29:38] Speaker 13: is so tightly integrated with the hardware design overall. And
[29:42] Speaker 13: I think that this is where Apple, again, excels, and
[29:44] Speaker 13: we're going to see a lot more innovation in AI
[29:46] Speaker 13: across all sorts of devices from Apple. And this is
[29:50] Speaker 13: just the start of that. I think AI is going
[29:52] Speaker 13: to be very dominant because this is a modernization vehicle,
[29:55] Speaker 13: not just for devices, but also for first-party services and
[29:58] Speaker 13: eventually for third-party services. And we think the time for
[30:01] Speaker 13: this is much, much larger than what people expect. So
[30:03] Speaker 13: we're very excited about what's yet to come from an
[30:05] Speaker 13: AI standpoint from Apple.
[30:08] Speaker 2: Wamsi, I heard many Apple executives say, and indeed post,
[30:11] Speaker 2: this is a new chapter for Apple. What do you
[30:14] Speaker 2: think changes from this point on under John Ternus?
[30:18] Speaker 13: Yeah, I think the most notable change here is what
[30:21] Speaker 13: was announced as a departure from net cash neutral. And
[30:25] Speaker 13: what that really means from an Apple standpoint, in our opinion,
[30:28] Speaker 13: is that we're giving John Ternus the ability to go
[30:33] Speaker 13: and invest.
[30:34] Speaker 1: As he would like, right?
[30:35] Speaker 13: So we've taken away some of the barriers around maintaining
[30:38] Speaker 13: net cash neutral, which means you opened up the aperture
[30:41] Speaker 13: for more M & A, you opened up the aperture
[30:44] Speaker 13: for spending more CapEx. And we've done some analysis around
[30:47] Speaker 13: this and think that if Apple were to spend $ 100
[30:50] Speaker 13: billion cumulatively through 2030, so $ 20 billion a year, you
[30:55] Speaker 13: can get a sizable amount of infrastructure set in place
[30:59] Speaker 13: for a lot of things from an AI workload perspective
[31:03] Speaker 13: to be done on device and Apple private cloud compute.
[31:05] Speaker 13: And so you don't really have to go off to
[31:07] Speaker 13: other cloud.
[31:09] Speaker 1: Providers to really be able to.
[31:13] Speaker 13: Do all the AI, to take all the AI workloads
[31:18] Speaker 13: and be able to satisfy those requests that are coming
[31:20] Speaker 13: to them and instead be able to do it either
[31:23] Speaker 13: locally on device or do it on private cloud compute.
[31:26] Speaker 13: We think this is a big deal, and we think
[31:28] Speaker 13: that this is where John Ternus will be able to
[31:31] Speaker 13: differentiate on what we can do on device and on
[31:33] Speaker 13: Apple private cloud compute in the age of AI.
[31:35] Speaker 2: I'm Zimon of Bank of America from Apple Park. Next
[31:39] Speaker 2: day on the show. Appreciate it a lot. Thank you.
[31:41] Speaker 2: Two other companies that we're watching today, Oracle and Adobe.
[31:44] Speaker 2: By the way, both have earnings due out after the
[31:46] Speaker 2: closing bell. One more of an infrastructure hyperscaler story. The
[31:50] Speaker 2: other is software. Both of them an AI story. Both
[31:53] Speaker 2: down pretty significantly ahead of those prints. Coming up on
[31:56] Speaker 2: the show, Positron CEO, Mitek Agrawal, joins us to discuss
[31:59] Speaker 2: the chip startup's latest funding round at a $ 5 billion
[32:03] Speaker 2: valuation in the private market. Now, if you're a Bloomberg
[32:06] Speaker 2: Terminal user, check out Live Go. Bloomberg Power Players is
[32:10] Speaker 2: live in New York. Some of the leading voices from
[32:13] Speaker 2: the business of sport join us to discuss the multi-trillion
[32:16] Speaker 2: dollar global industry.
[32:19] Speaker 9: Al, Silas, a reporter here at Bloomberg News, we broke
[32:22] Speaker 9: a deal that you're involved in, which was the Yankees
[32:24] Speaker 9: deal with Apollo.
[32:26] Speaker 1: And then, of course, are they announcing?
[32:36] Speaker 2: Okay, across private markets, investors continue to show an appetite
[32:39] Speaker 2: for backing startups aiming to take on NVIDIA's lead, but
[32:43] Speaker 2: focused on inference. The latest example, Positron, has raised $ 875
[32:48] Speaker 2: million at a $ 5 billion valuation to bring its inference-focused
[32:52] Speaker 2: silicon to market. Positron CEO, Mitesh Agrawal, is with us. Okay,
[32:56] Speaker 2: let's just do the round. That was a big round,
[32:58] Speaker 2: split into two tranches, but also the participants, super interesting.
[33:02] Speaker 2: There must be a need for capital. What is it?
[33:05] Speaker 14: Yeah, no, thanks, Ed, for having me. And like, look,
[33:07] Speaker 14: the need for capital is growth. I mean, we've seen
[33:09] Speaker 14: the last, literally the last three days of news, you know, Navier-Stokes, Muse, Astra.
[33:15] Speaker 14: Look at the amount of compute increases that's going. So
[33:18] Speaker 14: the scale is very important. And supply chain, obviously, is
[33:21] Speaker 14: a big kind of constraint. So Anyone that can scale
[33:24] Speaker 14: out in the current supply chain regime and can use
[33:26] Speaker 14: capital to do that, that's kind of where the need
[33:28] Speaker 14: comes from. And that's kind of what this is really
[33:30] Speaker 14: driving is the scale out of our next generation silicon
[33:32] Speaker 14: and systems.
[33:33] Speaker 2: So let's talk about that next generation of system. Where
[33:36] Speaker 2: is Positron at with it? And just explain the basics
[33:39] Speaker 2: of what you offer.
[33:40] Speaker 1: Yeah, yeah.
[33:41] Speaker 14: Look, we're really focused on inference silicon and systems. And
[33:44] Speaker 14: we deploy it at a rack scale. We have our
[33:46] Speaker 14: first product out recently. First generation called Atlas already out.
[33:49] Speaker 14: 50 racks at Oracle getting deployed. And then second generation,
[33:52] Speaker 14: the silicon is called Asimov. The systems are called Titan.
[33:55] Speaker 14: and that scales out of the rack scale setup. Basically,
[33:58] Speaker 14: what we are really, really focused on is memory-first architecture,
[34:02] Speaker 14: so we drive massive amounts of memory capacity and very
[34:06] Speaker 14: high realized memory bandwidth to really get fast inference, but
[34:09] Speaker 14: at really good TCO. That's kind of the, look, making
[34:12] Speaker 14: inference cheaper so that you can have more access to compute,
[34:15] Speaker 14: making all these kind of applications, code generation, research, all
[34:18] Speaker 14: those faster, is very, very critical, right? So we are
[34:21] Speaker 14: taping out end of this year with our next generation silicon, Asimov,
[34:25] Speaker 14: and then Titan will be in production sometime second half
[34:28] Speaker 14: of next year.
[34:28] Speaker 2: Taping out at the end of this year, take us
[34:31] Speaker 2: from there throughout the process. How does that work out?
[34:33] Speaker 14: Yeah, so taping out and then you get your first
[34:36] Speaker 14: silicon back in about anywhere between 8 to 12 weeks.
[34:40] Speaker 14: And we're taping out with TSMC and 3P. They're close
[34:43] Speaker 14: partners for us. And then N3P is the node. And
[34:47] Speaker 14: then you get your first silicon back.
[34:49] Speaker 2: And then validation?
[34:50] Speaker 1: Yeah.
[34:52] Speaker 14: As expected with most kind of companies nowadays for silicon companies,
[34:55] Speaker 14: you want your first silicon to be a success, as
[34:57] Speaker 14: in it should work.
[34:58] Speaker 1: We expect that.
[34:59] Speaker 2: Authorizing it's expensive.
[35:00] Speaker 6: Yes, yes.
[35:01] Speaker 14: And look, we spend on emulation and other kind of
[35:04] Speaker 14: techniques to make sure that there are as less bugs
[35:07] Speaker 14: as possible as we move forward. But yeah, validation, bring up.
[35:11] Speaker 14: You get your first in production systems. You put your
[35:14] Speaker 14: silicon into your systems. And then you're really testing out
[35:17] Speaker 14: all the cutting-edge models of whatever the models will be
[35:21] Speaker 14: in June and seeing how that works.
[35:22] Speaker 2: You talked about the existing generation 50 racks of Oracle.
[35:25] Speaker 1: Yeah.
[35:26] Speaker 2: But is Oracle the customer? How does your go-to-market work?
[35:29] Speaker 1: Yeah.
[35:29] Speaker 14: So for our first generation, the way we have done
[35:32] Speaker 14: generally for go-to-market is kind of typical silicon semiconductor company.
[35:36] Speaker 1: So it's a CapEx sell to Oracle.
[35:38] Speaker 14: So Oracle buys it as a CapEx purchase on their
[35:40] Speaker 14: balance sheet and they pay us as a PO.
[35:43] Speaker 1: So it's a very, very standard purchase.
[35:45] Speaker 14: Obviously, as we grow, you know, there is like more
[35:48] Speaker 14: motions around how do you sell, you know, obviously NeoClouds
[35:51] Speaker 14: and Hyperscale at a dollar per GPU or dollar per
[35:55] Speaker 14: systems per hour. And then there's even further optimization into,
[35:58] Speaker 14: you know, dollar per million tokens and then all of
[36:00] Speaker 14: those things, right? So as a company, we have to
[36:02] Speaker 14: evolve through partnerships and working with all these players. But
[36:05] Speaker 14: the first generation is very simple.
[36:07] Speaker 2: Do you have some customers lined up for these next generations,
[36:10] Speaker 2: even though given the stages they're at? And what kinds
[36:12] Speaker 2: of companies are those? Or name them. It's up to you.
[36:15] Speaker 1: Won't name them yet.
[36:17] Speaker 2: Suspected.
[36:18] Speaker 1: Yeah, yeah.
[36:19] Speaker 14: But really, you have to think about the very, very
[36:22] Speaker 14: typical kind of space, you know, the kind of AI
[36:25] Speaker 14: labs or model labs companies. Neo clouds that are out there,
[36:28] Speaker 14: hyperscalers that, you know, and one already, Oracle, our customer.
[36:33] Speaker 14: You know, you obviously have a little bit of look
[36:35] Speaker 14: into the sovereign play as well, because they are deploying
[36:37] Speaker 14: and building out a lot of their own inference data
[36:39] Speaker 14: center capacities. And then really, anyone that has to do
[36:43] Speaker 14: token as a service or inference as a service, So
[36:45] Speaker 14: either they can be their net consumers of themselves of
[36:47] Speaker 14: tokens or video generations or someone who's packaging them and
[36:51] Speaker 14: selling them into the market, right? So that's kind of
[36:53] Speaker 14: your broad ecosystem of the customer base for us where
[36:57] Speaker 14: we are really focused on. Look, it is a go
[36:59] Speaker 14: big or go home market kind of thing in the
[37:01] Speaker 14: sense that you have to focus on these very large
[37:03] Speaker 14: application companies to really get a large footing in the market.
[37:07] Speaker 2: Let's finish by talking about Jevons Paradox and also NVIDIA.
[37:10] Speaker 2: So on Jevons Paradox, the idea that a resource gets
[37:13] Speaker 2: more efficient, the economics improve, and then demand goes up.
[37:15] Speaker 5: Yeah.
[37:16] Speaker 2: Is that what you see? And then earlier you would
[37:18] Speaker 2: have heard D-Matrix on the show, you know, partnering with
[37:21] Speaker 2: NVIDIA on MV-Link so that their custom platform slots into
[37:24] Speaker 2: existing GPU-based architecture.
[37:26] Speaker 3: Yeah.
[37:27] Speaker 2: Both of those things are true at the same time? Yeah.
[37:30] Speaker 14: So on the Germans' paradox, I mean, like, look, I
[37:32] Speaker 14: think it's, you know, NVIDIA, sorry, OpenAI recently said that
[37:36] Speaker 14: when they slashed the prices by half, they saw the
[37:38] Speaker 14: utilization go up by 10x.
[37:39] Speaker 1: Today, DeepSeek, we 4.1. Amazing, amazing data. launch, I think.
[37:44] Speaker 14: Very cheap, sole-level performance based on early results.
[37:47] Speaker 1: I haven't looked into the architecture yet. But look, again, really.
[37:51] Speaker 14: Great for use cases and utilization.
[37:53] Speaker 1: I think that will go up.
[37:54] Speaker 14: It's a good story for us as well, especially around memory.
[37:57] Speaker 14: They're using LPDDR5X. We are also doing that. So that's
[38:00] Speaker 14: kind of really interesting. On NVIDIA's kind of working, look,
[38:04] Speaker 14: NVIDIA is absolutely the main player in the market, right?
[38:07] Speaker 14: Whether it's training, absolutely main one, and then even in inference,
[38:10] Speaker 14: they are the majority.
[38:11] Speaker 1: Kind of there.
[38:12] Speaker 14: So for all of us, including Positron AI, we really
[38:14] Speaker 14: focus on how can we work with them, right? You know,
[38:16] Speaker 14: we have to be in the same data centers, in fact,
[38:18] Speaker 14: even same racks as them to work with them, right?
[38:20] Speaker 1: So we also plan on connecting.
[38:22] Speaker 2: Will you do an MVLink Fusion agreement?
[38:23] Speaker 1: We will definitely, you know.
[38:25] Speaker 11: Try?
[38:25] Speaker 1: Yeah, we will definitely, like, talk to them and try it.
[38:27] Speaker 14: But, like, without even, like, just wanted to make clear
[38:30] Speaker 14: also on technical side of things, you can directly connect
[38:32] Speaker 14: with NVIDIA GPUs even today through standard, you know, Ethernet
[38:35] Speaker 14: and InfiniBand connections, right? So this is not, like, this
[38:37] Speaker 14: doesn't stop you from doing it, but obviously you want
[38:39] Speaker 14: to get closer. And then, like, look, you mentioned the
[38:41] Speaker 14: people on our board, right?
[38:42] Speaker 11: I know.
[38:43] Speaker 1: Gavin is on our board. He's very, very clear.
[38:46] Speaker 2: Gavin Baker at Trades.
[38:47] Speaker 1: Correct. Gavin Baker at Trades.
[38:48] Speaker 14: He's very clear about, you know, like, look, we have to,
[38:51] Speaker 14: you know, this is a heterogeneous computer market.
[38:53] Speaker 1: This is how it's going to grow.
[38:54] Speaker 2: Post John C. Mitesh Agrawal.
[38:56] Speaker 1: Thank you.
[38:56] Speaker 2: Really appreciate it. Some breaking news crossing the Bloomberg. Saudi
[38:59] Speaker 2: Arabia's wealth fund is considering combining electronic arts. with Savvy Games,
[39:05] Speaker 2: sources say, a deal that would bring blockbuster titles like
[39:08] Speaker 2: Madden NFL and Pokemon Go under one roof. The sources
[39:12] Speaker 2: point out no final decisions have been made and the
[39:14] Speaker 2: deal is unlikely to happen until Savvy completes its $ 6
[39:18] Speaker 2: billion acquisition of Chinese mobile gaming business Moonton. Coinbase is
[39:24] Speaker 2: betting stablecoins could become the next big growth engine. CEO
[39:27] Speaker 2: Brian Armstrong says the roughly $ 300 billion market could grow
[39:31] Speaker 2: tenfold by 2030 as more businesses use digital dollars for
[39:35] Speaker 2: payments and moving money across borders. He says improving regulation
[39:38] Speaker 2: in the U.S. could accelerate that adoption with a key
[39:41] Speaker 2: piece of crypto legislation facing a Senate vote next week.
[39:45] Speaker 2: He spoke with Bloomberg's Haslinda Arman.
[39:47] Speaker 11: Well, regulatory clarity has been getting a lot better.
[39:49] Speaker 2: You know, from a U.S.
[39:50] Speaker 11: Lens, last year, the Genius Act was passed into law
[39:53] Speaker 11: for stablecoins, and we saw 150 large companies adopt stablecoins
[39:57] Speaker 11: in the wake of that. And now the Clarity Act,
[39:59] Speaker 11: which is the second piece, which deals with crypto market structure,
[40:03] Speaker 11: it's right on the finish line. We actually have a
[40:05] Speaker 11: big vote next week in the Senate. to see if
[40:07] Speaker 11: that legislation is ready to become law. We've been meeting
[40:10] Speaker 11: with both sides of the aisle for many years on this.
[40:12] Speaker 11: There's hundreds of pages of compromises. Many banks have come
[40:16] Speaker 11: out and endorsed it, law enforcement groups, the crypto companies.
[40:19] Speaker 11: And it's good for America, good for the consumers. Today's
[40:24] Speaker 11: big number, $ 23 billion.
[40:26] Speaker 2: That's how much Elon Musk boring companies valued at after
[40:29] Speaker 2: raising $ 3 billion in his latest funding round from the
[40:32] Speaker 2: UAE and affiliated investment entities. That's after slow progress over
[40:36] Speaker 2: the last decade in realizing Musk's vision for a series
[40:40] Speaker 2: of ultra-fast underground hyperloops. Bloomberg's Kyle Porter joins us with
[40:43] Speaker 2: the details. Let's go into the round. Quite a big raise.
[40:46] Speaker 2: Interesting valuation.
[40:49] Speaker 1: Yeah, on both counts.
[40:50] Speaker 6: It's quadruple what they got valued at in their last
[40:53] Speaker 6: fundraising round. And it comes along with a new anchor
[40:56] Speaker 6: investor and an incredibly ambitious plan.
[40:59] Speaker 2: You've done a big take on Boring this year. Where
[41:02] Speaker 2: are they at? The idea was, by the way, Boring Co.
[41:05] Speaker 2: bores tunnels and those tunnels get used for things. But
[41:08] Speaker 2: are there actually any of them out there in the world?
[41:11] Speaker 6: There is an active unit you can go on in
[41:13] Speaker 6: Las Vegas. They started in 2019. The eventual plan is
[41:17] Speaker 6: for a, I believe, 60 mile network with a series
[41:22] Speaker 6: of interconnected tunnels. They currently have 10 tunnels operational. I
[41:25] Speaker 6: believe it's about 11 miles. And the original completion date
[41:28] Speaker 6: of 2027 for a complete network certainly seems very unlikely.
[41:32] Speaker 6: They're also boring down in Nashville. They agreed that last year.
[41:36] Speaker 6: The original plan was that that 10 mile music city
[41:39] Speaker 6: loop would be completed by 2027. I'm being told by
[41:43] Speaker 6: sources that is a little bit unrealistic. And then they've
[41:46] Speaker 6: struck a memorandum of understanding with the Emirate of Dubai.
[41:50] Speaker 6: And they also offered a few free projects in two
[41:54] Speaker 6: US cities, one in Louisiana and the.
[42:00] Speaker 8: Other in Texas.
[42:01] Speaker 2: Very quickly, do we have any sense of like how
[42:03] Speaker 2: involved Elon Musk is in the company?
[42:06] Speaker 6: He defers on boring a lot more than he does
[42:09] Speaker 6: on his other companies. He has a very well-established team
[42:12] Speaker 6: in place there. He's talked about it a lot less
[42:16] Speaker 6: recently compared to a lot of his other endeavors. He
[42:18] Speaker 6: was certainly very, very involved at the start when he
[42:20] Speaker 6: was living in Los Angeles and frustrated with the traffic.
[42:23] Speaker 2: Bloomberg's Kyle Porter on the big round that Boring Co.
[42:26] Speaker 2: just did. Thank you very much indeed. That does it
[42:28] Speaker 2: for this edition of Bloomberg Tech. A lot going on
[42:31] Speaker 2: in markets. There's a lot of news in the world
[42:32] Speaker 2: of technology as well. Recap it on the podcast, which
[42:35] Speaker 2: you can find on the Bloomberg Terminal, as well as
[42:37] Speaker 2: online on Apple, Spotify, and iHeart. It's Thursday, one more
[42:41] Speaker 2: day to go in the week. Thanks for being with us.
[42:43] Speaker 2: This is 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.