WEBVTT - Equities Mixed After Tech Sell-Off

0:00:02.480 --> 0:00:08.360
<v Speaker 1>Bloomberg Audio Studios, Podcasts, Radio News.

0:00:10.560 --> 0:00:13.840
<v Speaker 2>Welcome to the Daybreak Asia podcast. I'm Doug Prisner. It

0:00:13.880 --> 0:00:16.000
<v Speaker 2>was a mixed day for the US equity market in

0:00:16.120 --> 0:00:19.680
<v Speaker 2>front of the Independence Day holiday. On one hand, we

0:00:19.720 --> 0:00:21.799
<v Speaker 2>had the Dow jumping by more than one percent to

0:00:21.840 --> 0:00:25.600
<v Speaker 2>a record high fifty two nine hundred. However, the semiconductor

0:00:25.600 --> 0:00:29.600
<v Speaker 2>stocks were in retreat. We had the Philadelphia Semiconductor Index

0:00:29.720 --> 0:00:33.360
<v Speaker 2>down for a second straight session, falling on Thursday by

0:00:33.440 --> 0:00:36.000
<v Speaker 2>more than five percent. For a look at some of

0:00:36.040 --> 0:00:38.560
<v Speaker 2>the price section we are seeing in the APAC region,

0:00:38.600 --> 0:00:42.920
<v Speaker 2>I'm joined by Bloomberg Charlotte Young. Charlotte is Asia Equities

0:00:42.960 --> 0:00:45.800
<v Speaker 2>reporter and she joins from our studios in Hong Kong.

0:00:46.120 --> 0:00:48.520
<v Speaker 2>Thank you for being here. Talk to me a little

0:00:48.520 --> 0:00:51.680
<v Speaker 2>bit about what you're seeing on fold today, particularly given

0:00:51.680 --> 0:00:53.720
<v Speaker 2>the weakness in the chip stocks.

0:00:54.680 --> 0:00:57.680
<v Speaker 3>Yeah, so this this morning was seeing a really choppy

0:00:57.800 --> 0:01:01.520
<v Speaker 3>session on Korea and Japan stocks. They open slightly higher,

0:01:01.600 --> 0:01:04.920
<v Speaker 3>but then quickly raised all that gangs. So for careers

0:01:04.959 --> 0:01:08.040
<v Speaker 3>cost being decks, they actually win over two percent at

0:01:08.040 --> 0:01:12.000
<v Speaker 3>one point, but then quickly rebounded again. So I think

0:01:12.080 --> 0:01:14.280
<v Speaker 3>this kind of shows you a little I think there

0:01:14.360 --> 0:01:16.880
<v Speaker 3>is a sense of nervousness with treated that everybody is

0:01:16.880 --> 0:01:21.959
<v Speaker 3>now digesting those headlines, especially the one earlier this week

0:01:22.040 --> 0:01:24.760
<v Speaker 3>about around matter planning a cloud business. You saw a

0:01:24.880 --> 0:01:28.720
<v Speaker 3>computing power that's really causing a lot of concerns over

0:01:28.720 --> 0:01:32.720
<v Speaker 3>this AI over capacity built up. But then there's also

0:01:32.959 --> 0:01:35.959
<v Speaker 3>people are still digesting this, so there is a little

0:01:35.959 --> 0:01:38.399
<v Speaker 3>bit I think lack of conviction right now. But on

0:01:38.440 --> 0:01:40.400
<v Speaker 3>the other hand, it was really interesting that you're seeing

0:01:40.440 --> 0:01:43.360
<v Speaker 3>retail investors again instead of stepping up to buy the

0:01:43.440 --> 0:01:47.520
<v Speaker 3>dips both yesterday and today, while offer investors having selling.

0:01:47.840 --> 0:01:51.240
<v Speaker 2>So next week we have the preliminary earnings from Samsung,

0:01:51.480 --> 0:01:55.440
<v Speaker 2>and today a very interesting press report in the Information

0:01:55.560 --> 0:01:58.200
<v Speaker 2>I don't know whether you saw this, Anthropic is in

0:01:58.280 --> 0:02:01.200
<v Speaker 2>talks with Samsung to be a man facturer partner for

0:02:01.320 --> 0:02:05.080
<v Speaker 2>a custom AI chip. Now, in the last week, we

0:02:05.160 --> 0:02:07.720
<v Speaker 2>had this announcement in South Korea that Samsung and sk

0:02:07.840 --> 0:02:10.400
<v Speaker 2>Heinix we're going to be spending more than five hundred

0:02:10.440 --> 0:02:14.480
<v Speaker 2>billion US to build up more capacity for these high

0:02:14.480 --> 0:02:17.680
<v Speaker 2>bandwidth memory chips, which as you know well, are in

0:02:17.760 --> 0:02:21.000
<v Speaker 2>high demand given this AI buildout. Talk to me a

0:02:21.080 --> 0:02:24.800
<v Speaker 2>little bit about what we're expecting from Samsung next week,

0:02:25.240 --> 0:02:26.079
<v Speaker 2>so there's.

0:02:25.840 --> 0:02:29.240
<v Speaker 3>A lot of attention on Samsund's pulmonary second quarter results

0:02:29.280 --> 0:02:33.320
<v Speaker 3>next week. It doesn't provide any segment breakdown of business guidance.

0:02:33.360 --> 0:02:35.639
<v Speaker 3>With that said, I think all the tension will be

0:02:35.680 --> 0:02:39.880
<v Speaker 3>focused on second quarter operating profit, so Boombay Intelligence expecting

0:02:39.919 --> 0:02:43.080
<v Speaker 3>it to be slightly exceed on consensus given the strong

0:02:43.120 --> 0:02:46.519
<v Speaker 3>memory chip burnings despite the have that special bonus payments.

0:02:46.880 --> 0:02:49.640
<v Speaker 2>We also have sk Heinix in the news. So I

0:02:49.720 --> 0:02:52.320
<v Speaker 2>mentioned kind of the capex announcement earlier in the week,

0:02:52.360 --> 0:02:54.079
<v Speaker 2>and I also want to point out that the South

0:02:54.160 --> 0:02:57.440
<v Speaker 2>Korean government is very much involved in this initiative. But

0:02:57.560 --> 0:02:59.959
<v Speaker 2>in the coming week, we're going to have ADRs trading

0:03:00.080 --> 0:03:03.119
<v Speaker 2>for sk Heinex for the US market. Is that right?

0:03:03.800 --> 0:03:07.720
<v Speaker 3>Yes, So, esk Heinex ADR listing is expected to happen

0:03:07.800 --> 0:03:11.280
<v Speaker 3>next Friday. The company is looking to raise twenty nine

0:03:11.280 --> 0:03:15.960
<v Speaker 3>billion dollars. This is a really big deal for eskhex

0:03:16.000 --> 0:03:19.360
<v Speaker 3>in that and now that now offers with the ADR listing,

0:03:19.400 --> 0:03:23.239
<v Speaker 3>it now offers global investors easy access to buy the shares.

0:03:23.840 --> 0:03:27.480
<v Speaker 3>So before the ADRs, a lot of the investors, for example,

0:03:27.480 --> 0:03:29.400
<v Speaker 3>if they're based in Hong Kong and they don't have

0:03:29.520 --> 0:03:32.480
<v Speaker 3>direct trading idea with the career stock exchange, they have

0:03:32.600 --> 0:03:35.760
<v Speaker 3>to go through the banks, the trading decks to use

0:03:35.760 --> 0:03:38.600
<v Speaker 3>equity swapsually that's the main way people do it. But

0:03:38.680 --> 0:03:41.760
<v Speaker 3>now with ADRs, people can just easily access it. There

0:03:41.800 --> 0:03:46.880
<v Speaker 3>is some expectation that this will help boost eskahex valuations.

0:03:47.200 --> 0:03:50.480
<v Speaker 3>While there is also a key issue that is unresolved

0:03:50.520 --> 0:03:53.840
<v Speaker 3>at the stage, which is people the funngability issue, whether

0:03:53.960 --> 0:03:57.600
<v Speaker 3>that ADRs can be easily transferred into the local shies

0:03:57.640 --> 0:04:00.680
<v Speaker 3>and vice versa. So the fundability issues something people are

0:04:00.680 --> 0:04:02.800
<v Speaker 3>still trying to figure out. But with that said, the

0:04:02.800 --> 0:04:05.280
<v Speaker 3>ADR listing is the one key thing to watch for

0:04:05.600 --> 0:04:08.000
<v Speaker 3>how much I think I s Kahanis valuations can go.

0:04:08.360 --> 0:04:10.360
<v Speaker 2>Let's talk a little bit about what you're seeing right

0:04:10.400 --> 0:04:14.240
<v Speaker 2>now with the smaller suppliers that kind of are a

0:04:14.280 --> 0:04:18.040
<v Speaker 2>part of the supply chain for semiconductor stocks, those suppliers

0:04:18.040 --> 0:04:20.600
<v Speaker 2>that are Chinese based. Talk to me a little bit

0:04:20.600 --> 0:04:22.520
<v Speaker 2>about what you're seeing in the price section there and

0:04:22.560 --> 0:04:24.400
<v Speaker 2>some of the story that's connected with that.

0:04:25.560 --> 0:04:28.120
<v Speaker 3>Yeah, So earlier this year, I would say one of

0:04:28.160 --> 0:04:31.200
<v Speaker 3>the biggest bry spot in the Chinese aial market is

0:04:31.200 --> 0:04:35.159
<v Speaker 3>those hardware names, and usually it's vary in those niche

0:04:35.160 --> 0:04:39.200
<v Speaker 3>corners that nobody has paid attention before, you know, whether

0:04:39.279 --> 0:04:42.479
<v Speaker 3>that is about like optical makers or like making MLCC

0:04:43.160 --> 0:04:46.719
<v Speaker 3>these names that investors usually before wouldn't look at. But

0:04:46.839 --> 0:04:49.919
<v Speaker 3>because the now they offer the most eract exposure to

0:04:50.120 --> 0:04:53.400
<v Speaker 3>the broader chips psycho story and then they are they're

0:04:53.400 --> 0:04:56.640
<v Speaker 3>really seeing real earnings upgrade. So there's been a lot

0:04:56.680 --> 0:05:00.040
<v Speaker 3>of excitement also flows going into there. Uh just this

0:05:00.240 --> 0:05:02.520
<v Speaker 3>week they've also been seeing some of the sell off,

0:05:02.680 --> 0:05:06.080
<v Speaker 3>just as the global uh textile off. Investors are kind

0:05:06.080 --> 0:05:08.159
<v Speaker 3>of partterning a little bit there. But with that set,

0:05:08.200 --> 0:05:11.680
<v Speaker 3>I think so far, if you other strategies about China,

0:05:12.000 --> 0:05:14.440
<v Speaker 3>the key thing they're still stressing about is this AI

0:05:14.560 --> 0:05:18.240
<v Speaker 3>AI semiconductor psycho and how the hardware names still have

0:05:18.360 --> 0:05:21.480
<v Speaker 3>a lot of optimists on that. And also with CXMT,

0:05:21.680 --> 0:05:25.440
<v Speaker 3>which is China's I think memory chip maker Crown DREWNE

0:05:25.640 --> 0:05:29.000
<v Speaker 3>is about is expected to list soon. We don't have

0:05:29.080 --> 0:05:31.080
<v Speaker 3>the fixed day yet, there are some chatters that they

0:05:31.080 --> 0:05:34.159
<v Speaker 3>could becoming, you know, later this month. China's memory chip

0:05:34.200 --> 0:05:38.960
<v Speaker 3>maker Crown CXMT is expected to list soon in on

0:05:39.040 --> 0:05:41.520
<v Speaker 3>the onshore market or we don't have the fixed day yet,

0:05:41.520 --> 0:05:43.839
<v Speaker 3>but some chatters are talking about it to be later

0:05:43.920 --> 0:05:48.640
<v Speaker 3>this month. That's definitely the key, key thing everybody's watching.

0:05:48.680 --> 0:05:50.960
<v Speaker 3>And also how they could read a cross for the

0:05:51.040 --> 0:05:53.960
<v Speaker 3>career memory chip makers as well. So the Chinese name,

0:05:54.000 --> 0:05:57.320
<v Speaker 3>even though their technology is not as advanced as as

0:05:57.400 --> 0:06:01.200
<v Speaker 3>Kinex or Samsung. But the now right now is that

0:06:01.240 --> 0:06:04.119
<v Speaker 3>if they get a lot of if they were able

0:06:04.160 --> 0:06:07.040
<v Speaker 3>to raise a lot of money from the equity market,

0:06:07.080 --> 0:06:10.440
<v Speaker 3>they can instead put it back into the R and

0:06:10.520 --> 0:06:13.679
<v Speaker 3>D and all those big capital expenditure and that could

0:06:13.680 --> 0:06:16.479
<v Speaker 3>help them catch up as a career makers.

0:06:16.760 --> 0:06:19.040
<v Speaker 2>Okay, Charlotte, thank you so very much. We'll leave it there,

0:06:19.120 --> 0:06:22.720
<v Speaker 2>Bloomberg Charlotte Young. She joins from Hong Kong. She covers

0:06:22.720 --> 0:06:25.560
<v Speaker 2>the Asia Equity Markets Force. Joining us here on the

0:06:25.600 --> 0:06:36.600
<v Speaker 2>Daybreak Asia podcast. Welcome back to the Daybreak Asia Podcast.

0:06:36.640 --> 0:06:40.839
<v Speaker 2>I'm deg Krisner. Samsung will be reporting preliminary earnings next

0:06:40.839 --> 0:06:45.120
<v Speaker 2>week on Thursday. The tech news outlet the Information Report

0:06:45.120 --> 0:06:48.919
<v Speaker 2>it Anthropic is talking with Samsung about being a manufacturing

0:06:49.000 --> 0:06:51.839
<v Speaker 2>partner for a custom AI chip, and that's where we

0:06:51.880 --> 0:06:55.240
<v Speaker 2>begin our conversation with Bob O'Donnell Bob as president and

0:06:55.320 --> 0:06:59.159
<v Speaker 2>chief analyst at TECHnalysis Research. He spoke with Bloomberg TV

0:06:59.240 --> 0:07:00.839
<v Speaker 2>host sharing how.

0:07:00.680 --> 0:07:04.800
<v Speaker 1>Does this meaningfully change the market if this more custom

0:07:04.880 --> 0:07:07.520
<v Speaker 1>chips come to play, are they a meaningful threat to

0:07:07.680 --> 0:07:10.040
<v Speaker 1>Nvidia and the likes of chip makers or are we

0:07:10.160 --> 0:07:12.040
<v Speaker 1>just going to see more specialized chips.

0:07:14.000 --> 0:07:15.960
<v Speaker 4>I think we're just going to see more specialized chips,

0:07:15.960 --> 0:07:18.360
<v Speaker 4>to be honest with you, Sherry. I mean, look, there

0:07:18.440 --> 0:07:22.320
<v Speaker 4>is such demand for different types of architectures. And one

0:07:22.320 --> 0:07:25.040
<v Speaker 4>of the interesting things that we're seeing that's driving this

0:07:25.280 --> 0:07:30.240
<v Speaker 4>movement towards specialization is the fact that we're seeing companies

0:07:30.320 --> 0:07:33.400
<v Speaker 4>trying to figure out how can they reduce the cost

0:07:33.600 --> 0:07:37.200
<v Speaker 4>and increase the efficiency of how they generate tokens for

0:07:37.280 --> 0:07:39.480
<v Speaker 4>all the queries that we're doing in all the agents

0:07:39.480 --> 0:07:42.640
<v Speaker 4>that people are running, and so what they've realized is

0:07:42.680 --> 0:07:45.720
<v Speaker 4>they can break it up into different chunks or disaggregate.

0:07:45.720 --> 0:07:48.800
<v Speaker 4>That's what everybody calls it these days, is disaggregation. And

0:07:48.920 --> 0:07:53.280
<v Speaker 4>what they're figuring out is different aspects of those workloads

0:07:53.720 --> 0:07:56.280
<v Speaker 4>require different types of computing. So you're going to see

0:07:56.320 --> 0:07:59.280
<v Speaker 4>companies say, aha, I'm going to work on prefill or

0:07:59.320 --> 0:08:01.720
<v Speaker 4>I'm going to do the decode portion. I'm going to

0:08:01.800 --> 0:08:04.600
<v Speaker 4>do this part of the inference workload. There's still the

0:08:04.640 --> 0:08:07.880
<v Speaker 4>training workloads as well. So all of what is to say,

0:08:08.440 --> 0:08:11.800
<v Speaker 4>a lot of different ways to look at this problem,

0:08:11.960 --> 0:08:14.040
<v Speaker 4>but a lot of different ways for companies to come

0:08:14.120 --> 0:08:17.840
<v Speaker 4>up with creative and differentiated ways to solve this because

0:08:17.880 --> 0:08:19.760
<v Speaker 4>at the end of the day, look, people want to

0:08:19.760 --> 0:08:23.480
<v Speaker 4>be able to generate more AI tokens to fulfill like

0:08:23.520 --> 0:08:26.440
<v Speaker 4>I said, all these inquiries to let all the agents

0:08:26.440 --> 0:08:28.200
<v Speaker 4>do the work they want to do, and they want

0:08:28.240 --> 0:08:31.040
<v Speaker 4>to do it cost efficiently, so they can actually start making.

0:08:30.800 --> 0:08:35.360
<v Speaker 1>Certain when they partner with these facts are making ready

0:08:35.400 --> 0:08:38.760
<v Speaker 1>to be clear A yeah, to be fair, right, I

0:08:38.760 --> 0:08:40.960
<v Speaker 1>mean they are helping in the manufacturing process. I mean,

0:08:41.040 --> 0:08:44.040
<v Speaker 1>Samson could partner up with Anthropic. But when it comes

0:08:44.040 --> 0:08:46.680
<v Speaker 1>to choosing those partners, does it make a difference who

0:08:46.679 --> 0:08:48.800
<v Speaker 1>they are and what sort of facilities they have like

0:08:48.880 --> 0:08:52.679
<v Speaker 1>Samson for example, full Stock memory founder, advanced packaging.

0:08:55.280 --> 0:08:56.720
<v Speaker 5>Yeah, I mean all of those things matter.

0:08:56.760 --> 0:08:58.440
<v Speaker 4>I mean, look, we all know there's only a couple

0:08:58.480 --> 0:09:01.160
<v Speaker 4>of places they can go, Right, You're gonna go to TSMC,

0:09:01.559 --> 0:09:02.320
<v Speaker 4>You're gonna go to.

0:09:02.240 --> 0:09:05.040
<v Speaker 5>Samsung, and maybe you'll go to Intel.

0:09:05.080 --> 0:09:06.520
<v Speaker 4>But really, I mean, there's not a whole lot of

0:09:06.600 --> 0:09:09.240
<v Speaker 4>choices when we're talking, you know, the most advanced silicon.

0:09:09.400 --> 0:09:11.679
<v Speaker 4>So look, it's a great win if Samsung gets this.

0:09:11.920 --> 0:09:16.280
<v Speaker 4>We've seen some other companies showing interest in getting access

0:09:16.320 --> 0:09:19.080
<v Speaker 4>to Samsung. Part of it is because TSMC is completely

0:09:19.080 --> 0:09:21.160
<v Speaker 4>sold out, right, I mean, so look, you've got to

0:09:21.200 --> 0:09:23.160
<v Speaker 4>go where there's availability.

0:09:22.520 --> 0:09:24.080
<v Speaker 5>To manufacture these chips.

0:09:24.200 --> 0:09:26.480
<v Speaker 4>So that's all part of the game as well, is

0:09:26.559 --> 0:09:29.079
<v Speaker 4>just figuring out where that capacity is available.

0:09:30.280 --> 0:09:32.280
<v Speaker 1>But then again, you can make your own, right, You

0:09:32.360 --> 0:09:34.640
<v Speaker 1>have the likes of these hyperscalers now coming up with

0:09:34.720 --> 0:09:38.360
<v Speaker 1>custom chips. So is this eventually a threat down the

0:09:38.400 --> 0:09:41.719
<v Speaker 1>line when it comes to a few years from now.

0:09:44.200 --> 0:09:46.400
<v Speaker 4>Well no, but I mean again, all of those guys,

0:09:46.440 --> 0:09:49.640
<v Speaker 4>all the hyperscalers, again, they're doing the same thing that

0:09:49.960 --> 0:09:52.720
<v Speaker 4>anthropics trying to do that met has talked about doing.

0:09:53.080 --> 0:09:55.480
<v Speaker 4>Everybody is again looking at how can I do my

0:09:55.600 --> 0:10:00.000
<v Speaker 4>own custom silicon to allow me to do these different

0:10:00.080 --> 0:10:03.600
<v Speaker 4>types of workloads more efficiently, and everybody has different ideas

0:10:03.600 --> 0:10:05.319
<v Speaker 4>on the way to approach it. But again, when it

0:10:05.360 --> 0:10:07.839
<v Speaker 4>comes to the actual manufacturing, they don't have a lot

0:10:07.880 --> 0:10:08.960
<v Speaker 4>of choice.

0:10:09.720 --> 0:10:12.720
<v Speaker 1>Okay, so you mentioned Meta, we are also hearing that

0:10:12.760 --> 0:10:15.560
<v Speaker 1>they want to sell AI computing capacity, the same thing

0:10:15.880 --> 0:10:19.240
<v Speaker 1>that Softbak seems to be doing or Xai. Why is

0:10:19.280 --> 0:10:21.200
<v Speaker 1>this important for them and does that give you sort

0:10:21.240 --> 0:10:25.239
<v Speaker 1>of a picture of the capacity and the need to

0:10:25.240 --> 0:10:26.720
<v Speaker 1>start monetizing at this point?

0:10:29.240 --> 0:10:31.880
<v Speaker 4>Yeah, I mean what's interesting with the Meta news is

0:10:32.000 --> 0:10:35.800
<v Speaker 4>it shows the fact that AI compute supply and demand

0:10:35.880 --> 0:10:38.640
<v Speaker 4>are completely out of whack. Some people have more than

0:10:38.640 --> 0:10:40.640
<v Speaker 4>they need and a lot of people have less than

0:10:40.679 --> 0:10:42.800
<v Speaker 4>they need, and so it's all about kind of trying

0:10:42.800 --> 0:10:46.040
<v Speaker 4>to balance that out. Meta made a bunch of early

0:10:46.120 --> 0:10:49.160
<v Speaker 4>bets to get as much capacity as they could, and

0:10:49.200 --> 0:10:52.240
<v Speaker 4>it turns out that they're not using all of that capacity.

0:10:52.320 --> 0:10:54.320
<v Speaker 4>So they think, all right, we can be kind of

0:10:54.360 --> 0:10:57.000
<v Speaker 4>like a neocloud, just like Xai did last week, and

0:10:57.040 --> 0:10:59.640
<v Speaker 4>say we're going to sell some of that excess capacity.

0:10:59.640 --> 0:11:01.319
<v Speaker 4>And there are plenty of people willing to do it,

0:11:01.440 --> 0:11:03.360
<v Speaker 4>and in point of fact, it's proving to be.

0:11:03.320 --> 0:11:05.120
<v Speaker 5>A very profitable business model.

0:11:05.320 --> 0:11:08.120
<v Speaker 4>Now there are questions longer term, like is this something

0:11:08.160 --> 0:11:10.400
<v Speaker 4>they're just doing to kind of spice up the stock

0:11:10.480 --> 0:11:13.000
<v Speaker 4>for a year or two That could be the case,

0:11:13.480 --> 0:11:17.120
<v Speaker 4>but Meta does have a lot of capacity also being built,

0:11:17.440 --> 0:11:19.720
<v Speaker 4>so they're looking ahead and saying, Okay, we think we're

0:11:19.720 --> 0:11:21.839
<v Speaker 4>going to need this, and we know we're going to

0:11:21.880 --> 0:11:24.080
<v Speaker 4>have this much capacity, So it looks like we've got

0:11:24.080 --> 0:11:26.440
<v Speaker 4>a little bit left and the market seems to react

0:11:26.559 --> 0:11:29.000
<v Speaker 4>very favorably to selling this stuff. We can make really

0:11:29.000 --> 0:11:31.080
<v Speaker 4>great margins with it, so why.

0:11:30.880 --> 0:11:35.040
<v Speaker 1>Not I mean great for the big guys. What happens

0:11:35.040 --> 0:11:37.320
<v Speaker 1>to the smaller guys? Right, we're talking about core wave

0:11:37.440 --> 0:11:40.120
<v Speaker 1>anebias as well. I mean, are we going to see

0:11:40.320 --> 0:11:43.480
<v Speaker 1>this stack in divergence of winners and losers?

0:11:46.160 --> 0:11:48.800
<v Speaker 4>Well, it's a great question, but I look, I still

0:11:48.840 --> 0:11:52.040
<v Speaker 4>see think overall, I said, the supply and demand rod

0:11:52.080 --> 0:11:54.480
<v Speaker 4>of wack. But overall, if you look at the picture,

0:11:54.559 --> 0:11:57.520
<v Speaker 4>based on all the data that I can see, demand

0:11:57.559 --> 0:12:00.559
<v Speaker 4>is still significantly higher than supply. So it's just a

0:12:00.640 --> 0:12:02.160
<v Speaker 4>question of how people are.

0:12:02.280 --> 0:12:03.280
<v Speaker 5>Moving these things around.

0:12:03.480 --> 0:12:06.280
<v Speaker 4>I think certainly in the near term there's still plenty

0:12:06.320 --> 0:12:09.160
<v Speaker 4>of opportunity for the core weaves and the other neoclouds

0:12:09.200 --> 0:12:12.840
<v Speaker 4>that are out there, because again people are looking for capacity. Also,

0:12:12.920 --> 0:12:16.800
<v Speaker 4>remember there are hundreds thousands of different companies looking for

0:12:16.840 --> 0:12:20.880
<v Speaker 4>compute resources it may be that you see the core

0:12:20.920 --> 0:12:24.800
<v Speaker 4>weaves and the other neoclouds focus on say smaller vendors

0:12:24.880 --> 0:12:28.439
<v Speaker 4>or more specialized types of applications, and the metas only

0:12:28.480 --> 0:12:30.520
<v Speaker 4>sell to one or two people a couple of the

0:12:30.559 --> 0:12:33.800
<v Speaker 4>big guys right some of their coopetitions. So I think

0:12:33.840 --> 0:12:36.040
<v Speaker 4>you're going to see a lot of diversity in who

0:12:36.160 --> 0:12:39.640
<v Speaker 4>sells to whom, and that still gives plenty of opportunity

0:12:39.679 --> 0:12:43.079
<v Speaker 4>because again, fundamentally, there's still plenty of room for people

0:12:43.080 --> 0:12:45.000
<v Speaker 4>to grow, which is why, by the way, I didn't

0:12:45.000 --> 0:12:47.200
<v Speaker 4>get to it, but Nvidia, I think still has plenty

0:12:47.200 --> 0:12:50.000
<v Speaker 4>of room to grow even with all this competition because

0:12:50.040 --> 0:12:52.640
<v Speaker 4>of the overall demand just raises all boats.

0:12:54.880 --> 0:12:57.000
<v Speaker 1>Well, then before we let you go, I mean, how

0:12:57.040 --> 0:12:59.120
<v Speaker 1>bad is the memory shortage? Right now, we're hearing that

0:12:59.120 --> 0:13:02.880
<v Speaker 1>Apple is considered in Chinese suppliers that are on a

0:13:02.920 --> 0:13:04.439
<v Speaker 1>black list from the Pentagon.

0:13:07.000 --> 0:13:07.280
<v Speaker 5>Yeah.

0:13:07.320 --> 0:13:10.840
<v Speaker 4>Well, the Apple thing is interesting because what Apple said is, look,

0:13:10.880 --> 0:13:14.640
<v Speaker 4>we want to use the Chinese memory suppliers for all

0:13:14.679 --> 0:13:17.600
<v Speaker 4>the products we ship into China, which you know makes

0:13:17.679 --> 0:13:20.320
<v Speaker 4>logical sense. I mean, Apple needs a ton of memory,

0:13:20.760 --> 0:13:23.440
<v Speaker 4>and if they can use all the existing memory they

0:13:23.480 --> 0:13:25.680
<v Speaker 4>have for the US and Europe and other markets, and

0:13:25.679 --> 0:13:29.439
<v Speaker 4>then get access to Chinese memory for the China market,

0:13:29.720 --> 0:13:33.280
<v Speaker 4>that tremendously relieves you know, the challenges that they're facing.

0:13:33.320 --> 0:13:35.839
<v Speaker 4>I mean, we already saw the big price increases. I

0:13:35.840 --> 0:13:37.480
<v Speaker 4>don't think those things are going to go away. I

0:13:37.480 --> 0:13:39.800
<v Speaker 4>think that's basically where the world is going to be.

0:13:39.880 --> 0:13:42.720
<v Speaker 4>And my expectation is when the new iPhones come out,

0:13:42.920 --> 0:13:45.000
<v Speaker 4>they may not raise the current iPhone price. When the

0:13:45.000 --> 0:13:46.199
<v Speaker 4>new ones come out, they're going to be at a

0:13:46.240 --> 0:13:49.120
<v Speaker 4>higher price point because of how bad the memory crunch

0:13:49.200 --> 0:13:50.920
<v Speaker 4>is for everybody, including Apple.

0:13:51.840 --> 0:13:54.600
<v Speaker 2>That was Bob o'donald, President and chief analyst at Tech

0:13:54.600 --> 0:13:58.920
<v Speaker 2>Analysis Research, speaking with Bloomberg TV host Sherry on bringing

0:13:58.960 --> 0:14:05.079
<v Speaker 2>you their conversation here on the Daybreak Asia Podcast. Thanks

0:14:05.080 --> 0:14:08.680
<v Speaker 2>for listening to today's episode of the Bloomberg Daybreak Asia

0:14:08.840 --> 0:14:13.319
<v Speaker 2>Edition podcast. Each weekday, we look at the story shaping markets, finance,

0:14:13.640 --> 0:14:16.720
<v Speaker 2>and geopolitics in the Asia Pacific. You can find us

0:14:16.760 --> 0:14:20.960
<v Speaker 2>on Apple, Spotify, the Bloomberg Podcast YouTube channel, or anywhere

0:14:21.000 --> 0:14:24.080
<v Speaker 2>else you listen. Join us again tomorrow for insight on

0:14:24.120 --> 0:14:28.280
<v Speaker 2>the market moves from Hong Kong to Singapore and Australia.

0:14:28.680 --> 0:14:31.160
<v Speaker 2>I'm Doug Chrisner, and this is Bloomberg.