HSBC Global Viewpoint

HSBC analysts discuss AI capex risks, currency outlooks, and US-China tech divergence at Hong Kong event

23 min Listen to the original
AI CapexUS DollarChinese TechGlobal CurrenciesSemiconductersGold

Summary

Fred Neumann, Paul Mackel, Yuqian Ding, and Frank Lee discuss market volatility, noting Nasdaq declines amid rising bond yields and tax concerns. Mackel attributes US dollar strength to Federal Reserve rate hikes and equity inflows, while highlighting RMB outperformance and slow yen recovery. Lee projects US hyperscaler capex could double to $1.5 trillion, raising debt-to-equity ratios but remaining manageable due to high AI profit margins. Ding contrasts US frontier performance spending with China's efficiency-focused approach, noting similar LLM outcomes. Lee identifies optics as a key supply chain overlap where China holds 70% market share, while Van der Linde concludes with optimism on long-term AI investment potential.

Key takeaways

  1. 1
    Well uh absolutely and w what the team has been identifying for for quite a long time, led by Joey Chu, is is that the Raminbi should be an outperformer. I mean the the team has been uh uh having this thinking pretty much since September last year. It's been absolutely right. You know, you you did hit some points where maybe they don't want the currency appreciation to be too strong, that is the the regulators.

    Unidentified speaker (speaker_2) · 6:37

  2. 2
    Gold demand from central banks and retail remains resilient despite price volatility.

    Unidentified speaker (speaker_2) · 8:07

  3. 3
    US hyperscaler capex could double from $750 billion to $1.5 trillion, raising net debt-to-equity to 15-20%.

    Unidentified speaker (speaker_3) · 9:28

Episode sections

  1. 0:09 – 2:57

    Introduction and Market Context

    Fred Neumann introduces the live recording in Hong Kong with co-hosts Paul Mackel, Yuqian Ding, and Frank Lee. They discuss recent market volatility, noting that while the S&P has held up, the Nasdaq has declined, largely attributed to tax stories and rising bond yields.

  2. 2:57 – 7:51

    US Dollar and Currency Outlook

    Paul Mackel argues the US dollar remains strong due to Federal Reserve rate hikes and financial tightening, though equity weakness could challenge this. He highlights the RMB as an outperformer since late 2024 and notes the Japanese yen's recovery is slow due to interest rate differentials.

  3. 7:51 – 8:48

    Gold and De-dollarization

    Mackel describes gold as volatile but long-term constructive, citing resilient demand from central banks and retail investors despite recent price declines. He dismisses immediate fears of de-dollarization, viewing them as recurring cycles that eventually dissipate.

  4. 8:48 – 10:19

    US Tech Capex and Funding Risks

    Frank Lee notes US hyperscalers are turning negative in free cash flow, necessitating debt or equity raises. He projects Capex could double from $750 billion to $1.5 trillion next year, potentially raising net debt-to-equity to 15-20%, which he deems manageable.

  5. 10:19 – 17:57

    NVIDIA Valuation and Vendor Financing

    Lee addresses concerns about vendor financing, stating NVIDIA's P/E multiple is lower than during the crypto era due to financing uncertainties. He asserts underlying compute demand remains positive, evidenced by rising GPU rental prices on Bloomberg.

  6. 17:57 – 20:34

    China vs. US Tech Strategies

    Yuqian Ding explains that US Capex is 7-8 times higher than China's because the US optimizes for frontier performance, while China optimizes for system efficiency under export constraints. She argues both systems achieve similar LLM results despite different spending levels.

  7. 20:34 – 21:57

    Optical Supply Chain and Geopolitics

    Lee identifies optics as the primary area of overlap between US and Chinese tech supply chains, noting China holds 70% of the global market. He suggests the US government is now seeking to reduce dependency on Chinese optical components over time.

  8. 21:57 – 23:01

    Conclusion and Investor Sentiment

    Harold van der Linde concludes the session by expressing optimism, noting that long-term equity investing has historically rewarded optimism. He acknowledges near-term volatility but believes the AI investment story has not yet fully played out.

Assets discussed

AssetKindStanceContext
US dollarUSDcurrencymixedThe US dollar relatively resilient this year.
GoldcommoditymixedBut uh y you you mentioned dollar exceptionalism, uh but we also talked about the de-dollarization uh or debasement trade, I don't know what if you want to call it, where people at some point say we're looking for alternatives to the US dollar, we're going to gold.
NVIDIANVDAequitymixedThere's a lot of concern in markets that essentially it's all a shell game. There's companies uh like big uh semiconductor companies, NVIDIA being one, making a lot

“Stance” is how the speakers talked about the asset, not a recommendation.

Claims worth verifying

The Nasdaq is down to a large extent due to a tax story.

Unidentified speaker (speaker_1) ·1:49

How to verify: Check HSBC research reports or market data from 2026-08-13 to 2026-11-13 for explanations of Nasdaq performance linked to tax legislation or policy changes. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

Interest rates are going up in most economies and oil prices are very high.

Fred Neumann ·2:04

How to verify: Verify against Federal Reserve, ECB, and BOJ meeting minutes and oil price data (e.g., Brent Crude) for the period 2026-08-13 to 2026-11-13. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

The US dollar has performed well since largely June 2026.

Unidentified speaker (speaker_2) ·3:15

How to verify: Check US Dollar Index (DXY) data from June 2026 to November 2026 to confirm the outperformance trend. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

Hundreds of billions of dollars have gone into the US equity market this year, more than fixed income.

Unidentified speaker (speaker_2) ·4:41

How to verify: Review U.S. Treasury International Capital (TIC) data for 2026 to verify net flows into US equities versus fixed income. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

Hyperscalers in the US are starting to turn negative free cash flow in 2026.

Unidentified speaker (speaker_3) ·9:08

How to verify: Analyze quarterly earnings reports (10-Q/10-K) from major US hyperscalers (e.g., Microsoft, Amazon, Google) for 2026 free cash flow figures. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

Primary source

US hyperscaler Capex could rise from $750 billion in 2026 to almost $1.5 trillion in 2027.

Unidentified speaker (speaker_3) ·9:28

How to verify: Compare projected Capex guidance from major tech companies in their 2026 earnings calls against 2026 forward estimates. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

NVIDIA earnings are generating nearly 100% growth.

Fred Neumann ·10:19

How to verify: Check NVIDIA's Q3 or Q4 2026 earnings report for year-over-year earnings growth percentage. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

Primary source

GPU rental prices on Bloomberg have continued to go up.

Unidentified speaker (UNKNOWN) ·11:36

How to verify: Access Bloomberg terminal data for GPU rental indices or spot prices for the period 2026-08-13 to 2026-11-13. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

US Capex on AI infrastructure is seven to eight times what China spent.

Unidentified speaker (speaker_1) ·19:19

How to verify: Compare aggregate Capex data from major US tech firms with reported Capex from major Chinese tech firms (e.g., Alibaba, Tencent) for 2026. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

China holds 70% of the overall optical components market.

Unidentified speaker (speaker_3) ·21:14

How to verify: Review industry reports from firms like LightCounting or Omdia on optical transceiver market share for 2026. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify US Treasury data on equity vs fixed income inflows for the current year.
  • Check Bloomberg for current GPU rental price trends to assess compute demand.
  • Review US hyperscaler financial filings for free cash flow and debt-to-equity projections.
  • Compare US and Chinese LLM performance benchmarks to assess efficiency differences.
  • Monitor Chinese government policy statements on RMB appreciation limits.
  • Analyze HSBC's historical accuracy on currency outperformance claims.
  • Check optical component market share data from industry reports like LightCounting or Omdia.
Full transcriptexpand
0:09
Fred Neumann: Welcome to Under the Banyan Tree, where we put Asian markets and economics in context. My name is Fred Newman, Chief Asian Economist here at HSBC's Asian headquarters in Hong Kong.
0:20
Unidentified speaker (UNKNOWN): Yeah,
0:21
Unidentified speaker (speaker_1): that's right. Um
0:23
Unidentified speaker (UNKNOWN): let
0:23
Unidentified speaker (speaker_1): me introduce uh our um the people that we are
0:28
Unidentified speaker (speaker_1): Uh oh myself. Let me introduce myself. So I'm the co-host uh of Under the Banyan Tree and we have three other people here on the stage today. Yu Chen in the middle, she is our expert on electric vehicles but also on tech hardware. So we're gonna talk about that particular topic today. We have Frank sitting here, the analyst who
0:48
Unidentified speaker (speaker_1): looks at Nvidia and all other matters of tech. in the world. So we're gonna look at US and China tech in this discussion. And of course we got Paul Mackel here who knows everything about the Peruvian what is the Peruvian currency? And uh Seoul. I didn't even know. Okay. And all other currencies in the world as well.
1:05
Unidentified speaker (UNKNOWN): Now
1:05
Fred Neumann: before we get started, we should probably mention to our uh listeners and viewers at home that this is a live recording. And with that, uh Harold, what's uh what's been shaking in the equity world? You've been away a little bit, but I look at those equity you know indices, they don't look that encouraging.
1:23
Unidentified speaker (speaker_1): No, actually I've been on holiday over the last two weeks. And um,
1:28
Unidentified speaker (UNKNOWN): that's maybe
1:29
Unidentified speaker (speaker_1): why. But yeah, I come back and of course these markets uh down. Actually, certain markets have held up, I think, reasonably well. Uh India, for example, is held a up breeze reasonably. well also. Also the S P has actually held up well. The Nasdaq is down, so it's really to a large extent a tax story. But let me quickly go back and maybe you can help me out a little bit here,
1:49
Unidentified speaker (speaker_1): Fred. Um, what has happened? We've seen bond yields move, we can talk with Paul about the currencies. But why has this happened? Has there been In any economic data that's come out that's that's sh that scared markets while I was in uh in in uh seeing Mexican temples?
2:04
Fred Neumann: There are a few themes uh out there in the market. One is that of course there's still worries about inflation and where central banks are taking interest rates. So interest rates are going up in most economies, oil prices very high, that worries investors who invest in bonds, for example. That's one big theme. And then, of course, we also saw a bit of a shake shakeout out
2:24
Fred Neumann: in the AI trade where investors are getting a bit more nervous about this idea about AI valuations, for example. So some volatility in that part is
2:35
Unidentified speaker (UNKNOWN): But that
2:35
Unidentified speaker (speaker_1): had partially also to do, I think, with bond yields rising and people worried that the financing we can talk about this in a second, of course, that maybe financing would be an issue. So there was a bit of an economic sort of component to this as well.
2:47
Unidentified speaker (UNKNOWN): And I love
2:48
Fred Neumann: how the equity strategist always blames the economics when things go slowly.
2:51
Unidentified speaker (UNKNOWN): Yeah, but why not, right? Um but
2:53
Fred Neumann: maybe we should start right here on the more the macro side and bring Paul in here.
2:57
Unidentified speaker (UNKNOWN): Yes.
2:57
Fred Neumann: Paul, um talk to us about the big view on currencies. The US dollar relatively resilient this year. I think uh is that a theme that that you think will remain in play? Uh interest rates in the US high. Does the dollar remain well supported or is there some cracks showing? there?
3:15
Unidentified speaker (speaker_2): Well i i it is very much the baseline scenario that we think that the dollar can perform quite well. I mean that's been the case since largely June this year, maybe a little bit before that. And I think that there are some connection points here about its outperformance, maybe what's happening in in your world too, but what's happening is that the Federal
3:35
Unidentified speaker (speaker_2): Reserve has been telling us that they want to raise rates. And you get some tightening of financial conditions alongside that. That puts the dollar in a stronger orbit. But you're also seeing maybe a longer-term theme, not yet, that could eventually challenge the dollar. I mean, we've had some moments recently where, say, the U.S. uh
3:55
Unidentified speaker (speaker_2): equity market on the technology side is not performing that well, and the dollar weakens. So you're seeing some stretching point. points that can challenge the dollar. If you get enough fear occurring in parts of the equity complex, it's not necessarily good for the currency.
4:10
Fred Neumann: Can I just follow up on that? Is the dollar supported by high interest rates, or because the AI theme is so strong that it sucks in money? What is the bigger driver of the dollar at the moment?
4:21
Unidentified speaker (speaker_2): It's a good question. I think that you know our approach more uh is associated with interest rates and and short term interest rate differentials. But there's no question, if you look at the flow of funds that's been going into the United States, uh some of the data provided by the U.S. Treasury, I mean this year there's been hundreds of billions of dollars going into the U.S. equity market,
4:41
Unidentified speaker (speaker_2): much more than perhaps fixed income. So there uh again, there's definitely a connection between what what's is going on with the equity market.
4:48
Unidentified speaker (speaker_1): Exactly. But also if you believe that you are going to be a winner in the AI sort of new industry that is emerges, if you pay, say, seven or eight or nine percent on interest rates on on a bond, that's not a big thing. You're thinking you're gonna make hundred percent in profits, of five hundred percent in profits, right? So
5:08
Unidentified speaker (speaker_1): one or two percentage point higher interest rates is not really gonna dent your confidence that you can uh that you want to invest. So higher interest rates is probably not gonna slow down these investments that quickly.
5:20
Unidentified speaker (UNKNOWN): Well
5:20
Unidentified speaker (speaker_2): it comes down to time frames of course and uh but ultimately I look I think you're right from a longer term perspective it's feeding into this view of dollar exceptionalism. We've had multiple rounds of this over the last decade and and it's back. But you know it's always in the back of my mind while we've been on more on the right side of the fence over the last six
5:40
Unidentified speaker (speaker_2): months or so, you know, what could go wrong, what could challenge the thinking. And as I said, there's there's been some indications there that uh maybe not everything is great.
5:49
Fred Neumann: But uh y you you mentioned dollar exceptionalism, uh but we also talked about the de-dollarization uh or debasement trade, I don't know what if you want to call it, where people at some point say we're looking for alternatives to the US dollar, we're going to gold. go
6:02
Unidentified speaker (UNKNOWN): from the world.
6:04
Fred Neumann: How
6:04
Unidentified speaker (UNKNOWN): does an
6:04
Fred Neumann: FX guru like you uh respond to those uh worries?
6:08
Unidentified speaker (speaker_2): I I've I've been to that rodeo so many times uh in my career. And
6:12
Fred Neumann: you're Canadian, you've been to a few rodeos.
6:14
Unidentified speaker (speaker_2): Yeah, I've been a few I've been to a few rodeos. Uh but I'm I'm quite patient about it when it comes back and I I do feel a Field lot of questions that is this the moment that it's really gonna hurt? And eventually the storm begins to go away.
6:27
Unidentified speaker (speaker_1): But currencies are always a two-sided game, right? One currency strengthened, the other one is gonna be weakened. So what have we seen in Asia? Which currencies have strengthened, which ones have held up reasonably well?
6:37
Unidentified speaker (speaker_2): Well uh absolutely and w what the team has been identifying for for quite a long time, led by Joey Chu, is is that the Raminbi should be an outperformer. I mean the the team has been uh uh having this thinking pretty much since September last year. It's been absolutely right. You know, you you did hit some points where maybe they don't want the currency appreciation to be too strong, that is
6:57
Unidentified speaker (speaker_2): the the regulators. But on the other hand, the direction of travel has been very, very clear. So there are exceptions. And and that currency is held up very nicely in a strong dollar world. So again, there are some exceptions.
7:08
Unidentified speaker (speaker_1): And a yen?
7:10
Unidentified speaker (UNKNOWN): The
7:10
Unidentified speaker (speaker_2): Japanese yen is is a is a different story. It's probably uh you know it's gone through uh quite significant weakness over the last few years, but that appears to be dissipating. Perhaps uh again the worst has passed. But the recovery will be a lot slower because you're still dealing with quite a a high interest rate differential uh between the U.S.
7:30
Unidentified speaker (UNKNOWN): and Japan.
7:31
Fred Neumann: Um perhaps one last question for you, Paul, before we uh move on to our other guests here. Um obviously gold po falls into your purview. as well. Um we have an environment of high inflation, high geopolitical risk, a lot of uncertainty in the world economy. Is this now the the f era for gold or is that
7:51
Fred Neumann: a bit overdone? Gold has actually declined this year.
7:54
Unidentified speaker (speaker_2): Yeah, it's it's it's a very volatile animal. And when you think it should be holding up in periods of significant risk aversion, again, I I've seen enough moments in my time when that doesn't work. So you have
8:06
Unidentified speaker (UNKNOWN): Another rodeo.
8:07
Unidentified speaker (speaker_2): Another rodeo or series of rodeos. But you have to think about the longer term game and and again the direction of travel is another phrase we're using. And for us we're still long term constructive on it. You know, the the the demand I think is still gonna remain pretty resilient, whether it's from central banks or retail, different types of investors, but again, it's quite volatile.
8:27
Unidentified speaker (speaker_1): So we've spoken about the dollar, the investment flows coming into the US, and that's to a large extent a US tech story, right? Frank, that's your purview, that's your area. Uh these companies they're still making good profits, right? And they have we seen a slowdown in the investments by these companies uh
8:47
Unidentified speaker (speaker_1): in the US?
8:48
Unidentified speaker (speaker_3): Yeah, I think if you look at you know obviously the market. market's been very volatile with the AI leading the way down. The number one issue, I think, is still something that plagues the market over the last few years, which is funding, right? The amount of capback spending is unprecedented. But this year it's become especially sensitive because this is the first
9:08
Unidentified speaker (speaker_3): year where these hyperscalers in the U.S. are starting to turn negative free cash flow, which means they have to start raising money. either through debt or through equity. And hence, you know, with inflation concerns and bond rates, this has been, you know, a a key issue. Um but I think if you look at the balance sheet of the hyperscalers, generally they're all
9:28
Unidentified speaker (speaker_3): very healthy. Uh we've looked at this earlier this year. Uh and I think next year they could raise their our capex from seven hundred and fifty billion this year to almost one and a half trillion, which is you know. the a hundred percent growth, that would have raised the net to debt to equity for this group probably to about fifteen to twenty percent, which we think is still very manageable.
9:47
Fred Neumann: Well you you you're telling a very bullish picture here. You talk you're saying that um you know US investment could jump to almost one and a half trillion with It would double what it is this year, which would completely blow out all supply chains in Asia. Um
10:06
Fred Neumann: There's a lot of concern in markets that essentially it's all a shell game. There's companies uh like big uh semiconductor companies, NVIDIA being one, making a lot
10:17
Unidentified speaker (UNKNOWN): of money.
10:19
Fred Neumann: They then put money into lend the money to the customers to buy their chips. And so it goes back and forth. And it looks like everybody's making money, but really they're just lending to each other. Is this not ultimately a recipe for kind of a a shot downturn to say the least? Because vendor financing, you know, to finance your own sales, that's a bit of a
10:39
Unidentified speaker (UNKNOWN): uh
10:41
Unidentified speaker (speaker_1): tricky, tricky situation.
10:42
Unidentified speaker (speaker_3): Yeah. Yeah, I I think look that that's ultimately why the valuation for a lot of these companies are not that high. I mean you look at NVIDIA earnings it's generating nearly 100% in earnings, but the multiple is lower now than it's seen in even the crypto area, right? Because of the uncertainty uh because of these financing,
11:02
Unidentified speaker (speaker_3): right? Um but I would say in the last few months though, there are still other issues such as power, right? Um power build out in the US infrastructure. Uh there are concerns that is there a big delay, and this is going to lead to a slowdown in data center builds. There's also talks about, you know, the the political environment in the States, um, whether that's gonna lead to a
11:22
Unidentified speaker (speaker_3): slowdown in data center, which ultimately will hit the hardware guys. But I think when you think about the underlying demand of compute, it still remains quite positive. So you can look at uh for example, um,
11:36
Unidentified speaker (UNKNOWN): you
11:36
Unidentified speaker (speaker_3): know, there are GPU rentals, uh you know, there's people are able to rent out their GPUs, and and you can look on one on Bloomberg, it's it's basically a uh a commodity price. The trend there has continued to go up. So, you know, I I do think that the underlying demand isn't an issue, but the funding is uh is an i is a question mark.
11:54
Unidentified speaker (speaker_1): What does that mean for the Asian producers? Because as Fred said, if we see more investments, we're already running at all sorts of capacity. problems. Certain factories are running at full speed in order to get the DRAM and and any other components out. Can they actually expand faster in the US? Because the components they need to get from the US, and
12:14
Unidentified speaker (speaker_1): we haven't got the additional capacity. And when is that capacity going to come online? For example, in in in memory chips?
12:20
Unidentified speaker (speaker_3): Yeah, I I think when you look at the capacity issue, it will still be a challenge next year. Um but I think at the same time though, you know, the market concern is
12:30
Unidentified speaker (speaker_3): not so much capacity driven but more of sustainability into twenty twenty eight. So it's kind of a little bit of an ironic thing where you have market is extremely tight next year 'cause supply, yet there's fear about twenty twenty eight being sustainable. This is not the first time we've seen this. I mean you in know the last couple of years there's always question marks about sustainability of
12:50
Unidentified speaker (speaker_3): this trend. Uh but so far we've seen after a few months the market basically gained some confidence that the st the uh sustainability will continue. But what's more important though is that I actually think the application of AI is no longer something that needs proof of concept. I think when we started this a few years ago, there was this worry that, well, are we how are we gonna going to use
13:10
Unidentified speaker (speaker_3): AI? But it's pretty clear we are using AI.
13:12
Unidentified speaker (speaker_1): Will people pay for it and these sort of things? So that's that's a fair argument. But there are other risks that have come up as well. So we have the funding risk we spoke about. Also Chinese competition is an issue, right? And we'll maybe talk with uh Yu Chen uh about this in in a second. But that's another fear that is hitting the markets as well.
13:30
Unidentified speaker (speaker_3): I I think the Chinese competition I know in this region we like to talk about China versus the US, but I think if you look at where because of geopolitical the aspect of where things are going, I don't see there's a lot of overlap between China and the U.S. I think they will each will kind of go its own path. I give you an example. Early
13:50
Unidentified speaker (speaker_3): in the year there was a lot of concern about Kimi coming out, and that would be very uh that could have a big impact on the US market uh because the open source model was more efficient.
14:01
Unidentified speaker (UNKNOWN): It
14:01
Unidentified speaker (speaker_3): is, it is more efficient. But what it did was that it caused this the US market to look at the open source model approach and create their own open source models. So you're not so basically they looked at what the guys in China did and said that was actually quite interesting. It may be a more efficient way of doing it, but we're not gonna use the Chinese models. We're gonna
14:21
Unidentified speaker (speaker_3): basically continue to to to have the US model, but we'll will change the approach. So I I think you're not really gonna going see to you know a direct head-to-head competition.
14:29
Fred Neumann: Frank, when you I mean everybody talks about AI. Everybody's now suddenly an expert. I know Harold is suddenly an expert. I'm now suddenly an expert just as much as I was a virus expert a few years ago.
14:42
Unidentified speaker (UNKNOWN): Just as much as I was a terror expert a few years ago.
14:45
Fred Neumann: Um But you're the real expert. You follow the semiconductor companies, you you're in the weeds there. What is the one thing you think all of us get wrong? What is the one thing where you say, guys, you just not getting it? You just not getting it. Is it is it that there is much more investment coming around the pike, uh down the pike on on
15:04
Fred Neumann: data centers? Is it that the rest of the world hasn't even started to invest? Is it the bullish case that we don't get? Is it the bearish case? Tell
15:13
Unidentified speaker (UNKNOWN): me.
15:14
Unidentified speaker (speaker_3): I I think the bullish case is really
15:18
Unidentified speaker (speaker_3): the application of AI, right? And what that how big that could be. So for example, I think one of the things we saw this year was the to talk about agentech AI, right? Um you know that's that wasn't even in discussion up until the beginning of this year. And we're just hitting the surface of what agentec a Gentech AI means, what we could
15:38
Unidentified speaker (speaker_3): do, what with a Gentech AI. So that's driving another part of the compute demand on CPUs, right? So I think this is the part that it's hard to know, is because the applications of it continue to change and continue to adapt. And I think that's where you know the uh the perhaps the more cautious view is because we don't see how what the next
15:58
Unidentified speaker (speaker_3): application layer is gonna look like.
15:59
Fred Neumann: Is there something when Frank Lee goes to bed at night? and anything that worries you keeps you up at night? Anything that could go wrong?
16:07
Unidentified speaker (speaker_3): I I think well it is ever it's very interesting because we did a tech tour recently in in in the in the Bay Area and you go there and it it's like not a care in the world. These guys you know are super confident. They think 2027 twenty seven, there's no cloud demand's gonna be great. But Wall Street doesn't feel that way, right? Wall Street has it
16:27
Unidentified speaker (speaker_3): a different perception because they think there will be a slowdown. So there's this constant debate between the two sides.
16:33
Fred Neumann: Um we could talk for hours on AI. Uh we're gonna grill you a bit more maybe later, but I wanna uh turn to Yu Chin, who is our expert on China technology, but particularly automotive technology. But I have one question I want to start off with. When, when finally can I drive in Hong Kong be driven around in an autonomous
16:53
Fred Neumann: vehicle? When is it gonna happen? Because we promised that.
16:59
Unidentified speaker (speaker_1): Yeah, I remember this question last year um when we do the session. Again, it's a geofence, um uh situation. So you specifically define in Hong Kong, so it's up for the Hong Kong government. I would say the technology is now viable. So Hera and I have been
17:13
Fred Neumann: discussing the trip, like
17:14
Unidentified speaker (speaker_1): we want
17:15
Fred Neumann: to take a Tim's driving car to drove drove it from Shenzhen to Beijing and then see uh how much of the disengagement. I think uh the the major obstacles uh preventing us from doing the trip is we don't have time. We haven't synced the schedule.
17:29
Unidentified speaker (speaker_1): It's like no doubt on
17:30
Fred Neumann: the commercial uh the technology and the product.
17:34
Unidentified speaker (speaker_1): That's right. But so China has done remarkably good on the EV side, right? How much is that mirrored on the tech hardware side? We just talked about uh what happens in the US. Uh is is is we seeing something very similar on the tech hardware side or is that a completely different sort of story?
17:54
Fred Neumann: The the tech hardware, broadly speaking, right?
17:57
Unidentified speaker (speaker_1): So
17:57
Fred Neumann: basically uh we talked about like a one point five trillion um um uh cap capacks pax spending in the US US. So basically that's gonna benefit like China hardware supply chain. So basically agentic AI, more tokens, like uh more CapEx from the US C SPs, meaning they're gonna they're gonna build more data center. They're gonna they're gonna procure more um
18:17
Fred Neumann: servers and uh they need a faster connection, they need a higher bandwidth, they need uh more reliable power delivery. So these are the components China has been doing. So the optics, uh these uh tech
18:30
Unidentified speaker (speaker_1): items yet
18:31
Fred Neumann: these bomb items are basically benefiting uh the China supply chain. So
18:35
Unidentified speaker (speaker_1): we've seen the huge rally.
18:37
Fred Neumann: Can I can I just ask you actually on that and maybe also a Frank. We're talking about the US investing in data centers and we're providing all the components. China is Korea is etcet. et cetera. Why not China itself? Why are these data centers not built at a grander scale in China as well? Wouldn't China is China falling behind in the data center race? Why why not
18:57
Fred Neumann: China?
18:59
Unidentified speaker (speaker_1): Yeah, I'm glad you asked. So let's say if you're looking at the end um end game, let's say the large language model, so China US similar. But if you compare the CapExten spanning, spending US is seven to eight times versus what China spent. So I think there is a fundamental difference between China and US because they optimize for different
19:19
Unidentified speaker (speaker_1): objective functions. US more like um um optimized on the frontier. So China is more like optimized the system efficiency under constraint.
19:30
Unidentified speaker (UNKNOWN): Because the
19:30
Unidentified speaker (speaker_1): China
19:31
Unidentified speaker (speaker_1): sorry, what do you mean? S system under on the constraint?
19:34
Unidentified speaker (UNKNOWN): Yeah system on the c constraint. Let's say China CSP um better than Alibaba they wanted to procure the most advanced NVIDIA chips, but they couldn't
19:42
Unidentified speaker (speaker_1): because of the
19:43
Unidentified speaker (UNKNOWN): export control. And uh the the China um semi um industry is uh constrained to purchase the most advanced equipment, say EUV. So this is the system has to live under constraint.
19:56
Fred Neumann: But that does not stop the system to evolve, to to update. But the objective function is different. We can't do this, we can't do that, but we need to deliver the results. So when we talk to the China tech guys, they are mostly talking about the system efficiency. So um if we don't have strong hardware, maybe we do more in terms of the software
20:16
Fred Neumann: efficiency so that my end product, let's say, when we use the Chinese AI, is still very like good and advanced. So uh I
20:23
Unidentified speaker (speaker_1): would like to stress this fundamental difference, even though like the money spent has been very different. But if you look at the results, it's actually quite similar.
20:32
Fred Neumann: Frank, do you do you have a view on this?
20:34
Unidentified speaker (speaker_3): I I think you're talking again like two different markets. Um I think people like to talk about China and the US, but there's only one area in the tech hardware where there's a huge overlap, and that's on the optical side. Um you know, if you look at people talk about China. Chinese developing their own CPUs and GPUs, it will be for the China market,
20:54
Unidentified speaker (speaker_3): but it won't really be Apple to the US. But what you are seeing and and I think the question is, is there gonna be even greater um uh diversification moving away? Because uh on the optical side, now the US government has realized that China's seventy percent of the overall market. So I think they're now looking
21:14
Unidentified speaker (speaker_3): at over time do we find a way to reduce the dependency on China.
21:20
Unidentified speaker (speaker_1): So we have a new technology. It's it's it's had massive macroeconomic consequences. Bond yields currencies have moved on that. We're still in a world whereby we're not quite sure how much demand will there be, but the investments are pretty significant. Now we're seeing that the market over the last couple of weeks,
21:40
Unidentified speaker (speaker_1): while I was on holiday, uh have started to get seriously worried about this. But the the story that I get here is that that's not really been played out yet. It might work out that actually there is demand and that things we're gonna see a lot of volatility in this story.
21:57
Fred Neumann: So are you saying you're coming away bullish from this conversation, uh Harold? Because as an equity strategist that really surprises me.
22:06
Unidentified speaker (speaker_1): Yeah, you're always joking that we're positive, but I'd like to highlight that if you do invest in equities, you would have actually done pretty well over the last one or two decades. It pays to be an optimist, Fred. It's good.
22:17
Fred Neumann: Yes, yeah. Um no but I think this was a fantastic conversation. Um thank you very much for all of your contributions and uh um thank you very much for joining us here for our live video recording of Under the Banyan Tree where we put Asian markets in economics in context. And we want you to
22:37
Fred Neumann: not miss an episode. Therefore you have to uh like and subscribe wherever you get your podcast to Under the Banyan Tree. Is that right?
22:46
Unidentified speaker (speaker_1): That is absolutely correct. And you remembered that very well, that sentence. Thank
22:50
Unidentified speaker (UNKNOWN): you. Thank you
22:50
Fred Neumann: very much for joining
22:52
Unidentified speaker (UNKNOWN): us.
This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.

More from HSBC Global Viewpoint

HSBC Global Viewpoint

Asia beyond the headlines

Episode discussion: HSBC analysts challenge Western narratives on China's growth and Asia's structural shifts

Speakers’ claims: Piers Butler hosts Jing Liu and Ines Lam to discuss Asia's economic trajectory beyond mainstream headlines. Jing Liu argues that China's 4-5% growth remains significant for its size, driven by a transition to high-quality growth where high-tech sectors now exceed the property sector's peak GDP contribution. Ines Lam highlights regional challenges including demographic aging and AI-driven labor displacement risks. The speakers emphasize interconnected drivers across Asia, such as fiscal health and industrial restructuring, while noting China's strategic avoidance of large-scale stimulus due to local debt concerns.

China EconomyAsia DemographicsAI and Labor
15 min0 assets10 claims to verify
HSBC Global Viewpoint

Perspectives: Transforming treasury through connectivity, data and AI

Episode discussion: HSBC and SAP discuss embedding treasury services in ERP and AI prerequisites

Speakers’ claims: Amber Henderson-Smart and Falk Rieker discuss the shift from manual file-based treasury operations to embedded connectivity within ERP systems. They highlight HSBC's 'Payments for HSBC' app in SAP and the standardization of multi-bank connectivity. Rieker outlines how real-time data and AI improve cash management and fraud detection. He identifies four prerequisites for AI adoption: a single finance database, robust connectivity, documented policies, and an AI governance model. The speakers conclude that AI shifts the treasurer's role from execution to strategic monitoring and policy definition.

Treasury ManagementERP IntegrationArtificial Intelligence
21 min0 assets7 claims to verify
All insights