00:00Our Bloomberg Intelligence team, Tom, noted or out with a note today that says investors sold
00:05a record $30.5 billion of Korean equities in June with July selling already nearing $8 billion and
00:11showing little signs of stopping. To you, what does this show to you? Or what, is that an important
00:16sign? Yeah, I think it's, I mean, it's obvious. I mean, when I came on here late May, I said,
00:23you know, the positioning with leverage and semiconductors and the Korean
00:27markets and investors were cashing out of their life insurance policies to put money into SK
00:34Hynex. The SK Hynex two-time levered ETF is down 73% from the high. So that's pretty ugly out
00:44there.
00:44And look, the fundamentals can be great, but positioning on one side of the boat,
00:49the boat can tip over. And that's basically what's happened.
00:52So stock, not so bad, but down. But to be levered against it, that gets a little crazy.
00:59Is that your point, kind of-ish? Or a little bit of both?
01:01I think you've had a gambler's mentality in the market everywhere, and especially in Korea. But
01:07in the US, you know, the short-term options markets, levered ETFs, it's just getting a little
01:15nuts. The cost be down about 30% just in the last month. We have a chart that shows what
01:21it's done
01:21over the last few months. I'm curious about not looking at that in a vacuum and looking at that
01:27in the context of the wider global market and what you see as challenges that you think will come when
01:32it comes to AI CapEx. Well, I think it's not a matter of if, but a matter of when there's
01:38going
01:38to be CapEx cuts or perhaps just no growth given on the guidance.
01:44Well, I think, you know, of course we could say that because this can't go on forever.
01:48It could be this quarter. It could be, I think it'll be next quarter. And we're starting to
01:52already see some cracks when you see, you know, Meta, SpaceX leasing out some of their compute.
01:59And that puts into question some Frontier models in the US. If they have to lease out compute,
02:06which I will admit, it brings in a lot of money at which they need.
02:11But you're saying they're leasing it out because they don't need it?
02:15Possibly. Possibly their models aren't ready.
02:17And what is that? What story? Okay. So that tells a story about models not being ready.
02:20Well, I think that Meta, for example, they've had some really good models recently.
02:26But I also think that they're still in the formation age of where they are, what their
02:33models look like. They brought on a lot of new people. It takes time and a lot of discovery
02:39on what works and what doesn't work. And so just if they're giving, you know, leasing out
02:44to Anthropic or others, I think it's a nice way to just bring in some revenue. And cash flows
02:51are down with all the hyperscalers. So there is going to be a wet blanket thrown on this market.
02:59I'm not sure which company it'll be. I mean, Oracle has some very high lofty goals of going into 2030.
03:08I think that's coming into question right now. I think that's going to be the issue in the next,
03:15this quarter, perhaps next quarter. Goldman Sachs said that the first one that cuts CapEx is going
03:22to be the winner. And I don't know about that. But I think I think Alphabet will probably keep
03:27CapEx up. They raised money. They have other businesses that generate income. So yeah,
03:33you're bullish on Alphabet. Yeah, I like I like the company a lot. I think that they're going to
03:37succeed with their Gemini. The Gemini is going to work with the consumer because they're going to be
03:43in Samsung phones with Android and iPhones. So I think that's really a big deal. Anthropic probably
03:51for enterprise right now where we stand. But still, that's where we are in terms of AI. So we see
03:58with
03:59some of the big hyperscalers, less CapEx spending. That's kind of a canary in the coal mine, or at least
04:07a red flag. Well, back in the, you know, the tech bubble, it was when the, the Selex and the
04:13telco companies said, Hey, I think we've got enough Cisco equipment, they still were buying,
04:19but it just leveled off. And then everything started to, you know, domino effect down the road.
04:26And that was Q3 of 2000. And it could be again this year. I think, though, that AI has a
04:35longer
04:35path. And I think that there's also going to be more disruption in other, in other places,
04:44particularly China. Let's talk China. Okay, because you're, you're long China and Hong Kong
04:49technology. You sent some notes to our producer, Talia that said you're, you're buying Alibaba and
04:54others, your crane shares, CSI, China internet fund, the iShares, China large cap ETF. What,
05:00what attracts you to these stocks into this market? The China ETFs and Alibaba were down about 50%
05:10from their highs. The businesses are good. You have Alibaba, for example, they're taking a different
05:17approach than some of the other US companies. They're, they're doing chips, they're doing AI models,
05:23they're doing the cloud, consumer AI, other apps that they're going to come out with. Plus,
05:30their, their core business, which is okay, not great. But still, it's there. China's going to do
05:36a stimulus, it sounds like. Nobody wanted these a few weeks ago. Nobody, I told everybody, buy these
05:44with two hands. And basically, it was crickets. Nobody wanted them. And now they're up 20%. And you see
05:51Moonshot AI, which Alibaba has a 35% stake in that Alibaba has stakes in every big AI startup,
06:01as do Tencent and, and others. So I think that there's an open source roadmap that's going to be
06:10very popular for the next six months. And it's going to be disruptive for US AI companies,
06:16especially if the models are more efficient. And cheaper, cheaper, still, still do what needs to
06:25be done. Yeah. And you have risk of the US putting the brakes on people in the US using the
06:32Chinese
06:33models for whatever reason. I mean, I can name a couple reasons. But the fact is, the AI models in
06:40the US are expensive, the token prices are still high for the major ones, anthropic, open AI. And so
06:47if you have an alternative, and it does the job, I think that's going to work.
06:53Before Oh, go ahead, Carol.
06:54No, it's something we talked about on Friday, all of these well known companies that are tapping into,
07:00I believe, the Chinese AI model. That's something we talked about.
07:03So before we let you go, you're long China and Hong Kong tech, you made your case, you're also
07:08short a couple names, or sectors. What are you short right now, very briefly?
07:14Well, I'm short Tesla going into the earnings, it's a smaller position. But I think they have,
07:20you know, it's moving on three years of negative growth. And they have two cars that they basically
07:25sell that are getting stale. I think there's a long list of promises that haven't been kept
07:31from the CEO. I think that's going to be a question that people want answered. When is the
07:38robo taxi coming? When is the optimist coming? When are they going to scale all these things? And
07:44they really haven't been able to do that.
07:48Interesting take. Really, and I think, an important perspective in terms of the overall
07:54conversation that we're having.
07:55You didn't take a robo taxi here, did you?
07:57No, I didn't take a Waymo. I flew.
08:01But we've talked to you. You like Waymo's though, right?
08:04I like Waymo's, yes. They work.
08:07They work.
08:07I live.
Comments