The Macro Brief by HSBC Global Research

HSBC analysts project AI capex peak and EM tech shift driving 2027 market re-rating

15 min Listen to the original Episode page
AI MonetizationEmerging MarketsSemiconductorsHyperscaler CapexValuationsFederal ReserveCloud RevenueNVIDIA

Summary

HSBC analysts identify post-midterm reassessment of AI monetization, regulatory concerns, and NVIDIA leadership changes as drivers of recent market turbulence. They note hyperscaler free cash flow has swung from positive to negative due to high capex, expecting an inflection point in Q1/Q2 2027 as spending peaks and cloud revenue accelerates. The Emerging Markets index has shifted from under 10% to approximately 40% AI/Tech exposure, driven by semiconductor giants. Despite earnings growth projections, analysts argue valuations are not frothy, citing semiconductor P/E ratios near 10-year lows compared to consumer staples. The outlook for 2027 is bullish, citing AI-driven margin improvements and potential peak bond yields.

Key takeaways

  1. 1
    Hyperscaler aggregate quarterly free cash flow has swung from +$50-60bn to -$50bn, raising sustainability questions about current capex levels.

    Unidentified speaker (speaker_1) · 3:00

  2. 2
    free cash flow starts to improve. Now, not turn positive, but actually starts to improve from current levels. in roughly Q one and Q two of 2027. And that to us is a really important moment because as soon as it starts to improve, then you can start to see this question of like, okay, there's light at the end of the tunnel. Yes, we can start to actually see the monetization pay off. And that's when we think, you know, the market will start to re rate.

    Unidentified speaker (speaker_1) · 3:56

  3. 3
    Totally. I think you know, this year alone it's up around five, six hundred billion dollars. I mean, it's it's wild numbers. I mean, that's for for 2027 estimates. And at the same time, so while that's kind of plateauing, at the same time, the revenue that the you know the hyperscalers are generating um from AI is improving.

    Unidentified speaker (speaker_1) · 4:52

  4. 4
    Cloud revenue is accelerating sharply due to AI compute demand and capacity constraints driving up prices, supporting revenue estimates.

    Unidentified speaker (speaker_1) · 5:12

Episode sections

  1. 0:51 – 3:00

    Market Shakeout Drivers

    Identifies post-midterm reassessment of AI monetization, regulatory concerns, and NVIDIA leadership change as turbulence drivers, compounded by sticky rates.

  2. 3:00 – 6:17

    AI Cash Flow and Monetization

    Discusses hyperscaler FCF swing from +$50-60bn to -$50bn, expecting inflection in Q1/Q2 2027 as capex peaks and cloud revenue accelerates.

  3. 6:17 – 10:02

    Emerging Markets Composition Shift

    Notes EM index shift from <10% to ~40% AI/Tech exposure via TSMC, SK Hynix, Samsung, altering commodity correlations.

  4. 10:02 – 12:01

    Earnings Resilience and Valuations

    Projects EM earnings growth ~70% YoY, argues valuations are not frothy citing semiconductor P/E of 13x vs staples 35-40x.

  5. 12:01 – 14:40

    2027 Outlook Signals

    Bullish 2027 outlook citing AI margin improvements, peak bond yields, and potentially lower oil prices.

Assets discussed

AssetKindStanceContext
NVIDIANVDAequitybearishAnd then in addition to that, we've also had obviously the change in leadership at NVIDIA. NVIDIA has been, you know, it's a huge stock in the SP 500.
GoldcommodityneutralSouth Africa with gold. Interestingly, you did have the GCC, which had a number of the big oil producers in there.

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

Claims worth verifying

Post-US midterms, the market is reassessing AI monetization, data center build constraints, and regulatory concerns from local elections.

Unidentified speaker (speaker_1) ·1:30

How to verify: Check HSBC Global Research equity strategy reports from late 2026 for analysis on post-midterm market drivers and AI capex timelines. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

NVIDIA has experienced a change in leadership, raising concerns about its willingness to financially support other AI ecosystem companies.

Unidentified speaker (speaker_1) ·2:10

How to verify: Verify NVIDIA board changes via SEC Form 8-K filings or NVIDIA investor relations news releases in 2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Primary source

Hyperscaler aggregate quarterly free cash flow has dipped from positive $50-60 billion to roughly negative $50 billion.

Unidentified speaker (speaker_1) ·3:00

How to verify: Calculate aggregate free cash flow for major hyperscalers (e.g., Microsoft, Alphabet, Amazon, Meta) using quarterly earnings reports (10-Q) for 2026-2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Primary source

HSBC expects hyperscaler free cash flow to start improving in Q1 and Q2 of 2027.

Unidentified speaker (speaker_1) ·3:56

How to verify: Locate HSBC Global Research '2027 Outlook' or credit strategy reports published in late 2026 referencing this forecast. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Primary source

AI capex estimates for 2027 have increased by $500-600 billion in the current year (2026).

Unidentified speaker (speaker_1) ·4:52

How to verify: Cross-reference HSBC research notes on AI capex trends with public announcements from hyperscalers and semiconductor equipment makers (e.g., ASML, TSMC) in 2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Cloud revenue from hyperscalers is accelerating sharply due to AI compute demand and capacity constraints.

Unidentified speaker (speaker_1) ·5:12

How to verify: Review earnings call transcripts of major cloud providers (AWS, Azure, Google Cloud) in 2026 for revenue growth metrics and management commentary on AI demand. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

The three largest stocks in the EM index (TSMC, SK Hynix, Samsung) account for roughly 30% of the index.

Unidentified speaker (speaker_1) ·6:37

How to verify: Check the current weighting of the MSCI Emerging Markets Index or FTSE Emerging Markets Index on their respective websites. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Energy and materials weight in the EM index has declined from ~30% post-GFC to ~10% today.

Unidentified speaker (speaker_1) ·7:50

How to verify: Compare historical sector weightings of the MSCI EM Index from 2026-2026 to current data via index provider websites. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

EM earnings are projected to grow close to 70% year-on-year this quarter, with 20-25% growth ex-tech.

Unidentified speaker (speaker_1) ·10:02

How to verify: Check consensus earnings estimates for the MSCI EM Index from Bloomberg or Refinitiv for the current quarter. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Semiconductor stocks are trading at 13 times earnings, near 10-year lows, while consumer staples trade at 35-40 times earnings.

Unidentified speaker (speaker_1) ·11:02

How to verify: Calculate the trailing P/E ratio for the SMH (Semiconductor ETF) and XLP (Consumer Staples ETF) using current market data. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify hyperscaler free cash flow trends using quarterly SEC filings (10-Q) for Microsoft, Alphabet, Amazon, Meta, and NVIDIA.
  • Check HSBC's specific Q1/Q2 2027 free cash flow improvement assumptions against consensus estimates from Bloomberg or Refinitiv.
  • Confirm the current weight of TSMC, SK Hynix, and Samsung in the MSCI Emerging Markets Index using MSCI or FTSE Russell data.
  • Compare current semiconductor sector P/E ratios to 10-year historical averages using S&P Capital IQ or FactSet.
  • Assess the impact of NVIDIA's new leadership on ecosystem financial support by reviewing recent earnings call transcripts and investor presentations.
  • Monitor local election campaign platforms in key markets for specific AI regulation proposals.
  • Track cloud service pricing trends from major providers (AWS, Azure, GCP) to validate revenue acceleration claims.
  • Review HSBC Global Research's full 2027 market watchlist for potential conflicts of interest or bank-specific biases.
Full transcriptexpand

0:09Unidentified speaker (speaker_0): Hello, I'm Aline Van Dyne in New York, and welcome to the Macro Brief from HSBC Global Investment Research, our weekly podcast that looks at the key drivers of financial markets across the globe. Today I'm joined by Alistair Pinder, our global equity strategist. and a regular contributor on global markets to the podcast, to unpack what's been moving markets, whether AI can deliver on its promise, and why emerging markets may tell several different stories in 2027. Alistair, welcome.

0:45Unidentified speaker (UNKNOWN): Thanks

0:45Unidentified speaker (speaker_1): for having me, Ellen.

0:46Unidentified speaker (speaker_0): So, Alistair, resilient earnings, restless markets, and a big question. what happens when the AI boom has to pay its bills? Over the past few weeks, of course, equities have been pushed around by shifting rate expectations, post-US election policy uncertainty, and questions about the scale and payoff of AI spending. Emerging markets have also felt that turbulence, but there the story has not been. uniform. So as we look towards 2027, we're gonna find out if this is a shake out in positioning or a change in fundamentals. So, Alistair, let's kick off with the recent shakeout. What's been driving it?

1:30Unidentified speaker (speaker_1): Well, I think post the US midterms, you've really had probably two or three things that the market is trying to reassess right now. Uh the first is uh, you know, the AI trade itself and particularly the monetization of this and how fast we can get these data centers uh being built. You know, that has been it is a constraint for a lot of these hyperscalers. And a lot of the Democrats who campaigned and and won uh their local elections, you know, campaigned on basically more regulation. And so that has, you know, brought up a lot of concerns. for the market. And then in addition to that, we've also had obviously the change in leadership at NVIDIA. NVIDIA has been, you know, it's a huge stock in the SP 500. It has basically backstop and financed a number of uh the AI ecosystem or companies in the AI ecosystem. And there is worries that with the a change in leadership, maybe NVIDIA might not be willing to support. some of these companies like they have done before. And so again, that I think that has also um reinforced some of this concern um around the AI monetization debate. And then on top of all of that, we've had sticky rates, potentially a more hawkish Fed. It feels like it's just a number of crosswinds which have really been challenging the market at this point.

2:45Unidentified speaker (speaker_0): Zooming in on on AI and I guess AI's next test, what about the cash flow and translating that into sustainable returns? Are the doubts around that justified or what is the data actually telling us?

3:00Unidentified speaker (speaker_1): Well, I mean, if you take the big hyperscalers, uh, you know, their free cash flow, which was, you know, running positive on a quarterly basis of roughly $50 to $60 billion dollars, you know, in 24-25 24, 25 per quarter, uh, is now basically dipped quite sharply. negative. Actually, their quarterly free cash flow in aggregate is roughly minus $50 billion. And so you've gone from generating lots of cash to now spending lots of cash. And that again has raised this question of like, well, how sustainable can you continue to pay for all of this capex if the revenues from AI haven't really come through yet?

3:34Unidentified speaker (speaker_0): So it's a legitimate question. It

3:36Unidentified speaker (speaker_1): is a legitimate question. It's one of the reasons why the hyperscale hyperscalers have been issuing huge amounts of their, you know, the credit strategy team have, you know, done great reports on this, particularly in their 2027 outlook. And they still need to finance more of this because they just don't have that revenue coming in. You know, our kind of expectation is that the free cash flow starts to improve. Now, not turn positive, but actually starts to improve from current levels. in roughly Q one and Q two of 2027. And that to us is a really important moment because as soon as it starts to improve, then you can start to see this question of like, okay, there's light at the end of the tunnel. Yes, we can start to actually see the monetization pay off. And that's when we think, you know, the market will start to re rate. I guess the question is like, well, why do we think uh the free cash flow will improve? Two reasons. One is because the capex estimates have probably peaked. You know, we're we're probably kind of close to the top in the amount of capex spending that we're gonna get. It's gonna be sustained all of this capex for a number of years, but the estimates aren't probably gonna be upgraded any higher.

4:41Unidentified speaker (speaker_0): And of course, as regular listeners of the podcast will know, we've discussed many times over this year just how much the estimates have grown. Totally. Just in this year alone.

4:52Unidentified speaker (speaker_1): Totally. I think you know, this year alone it's up around five, six hundred billion dollars. I mean, it's it's wild numbers. I mean, that's for for 2027 estimates. And at the same time, so while that's kind of plateauing, at the same time, the revenue that the you know the hyperscalers are generating um from AI is improving. And and importantly, what we're talking about here is the cloud revenue. So the revenue. that you get from your cloud businesses, that's actually starting to accelerate really, really sharply. And again, that's because as we use AI, we have to use the cloud infrastructure. Because there's a lack of compute or capacity, prices to use the cloud are increasing. And again, that's driving this increase in revenue estimates. And so hopefully by Q1, Q2, To we'll start to see that inflection point. And that should start to soothe some of the concerns around the AI monetizing. monetization debate.

5:39Unidentified speaker (UNKNOWN): So,

5:40Unidentified speaker (speaker_0): Alistair, let's take a quick break. And when we get back, we'll talk about emerging markets and the 2027 market watch list.

5:55Unidentified speaker (UNKNOWN): So,

5:56Unidentified speaker (speaker_0): welcome back, everybody. We've covered the drivers behind the market shakeout. We've talked about AI's next. test. So let's shift the focus somewhat to emerging markets. Now, one label, different stories. I think that sums up your view and what you've been writing about emerging markets. Just tell us a bit more about what the stories are.

6:17Unidentified speaker (speaker_1): The emerging market index equity index as a whole has just completely and fundamentally changed over the last few years. I mean, actually, at the moment, the three lines largest stocks in that EM index, which is uh TSMC, SK Heinnix, and Samsung Electronics, all vital AI plays account for roughly 30% of the index. So you could say, how much does AI account for EM equities? A decade ago, probably tech and AI related stuff accounted for less than 10%. Today it's around 40% of the index. So huge, huge increase. So that's one component of EM is your AI area. And that really is Taiwan and Korea. But outside of that, you have essentially 24 other countries which touch on a lot of other things beyond AI. In in Asia, you have ASEAN, you have India, you obviously have mainland China. Then you have areas of EMIA where it's the GCC, which is contending with the tensions in the region. You have Turkey, South Africa, and of course Latin America. where, you know, not just in Brazil, but you know, across all of the a lot of the markets have seen vital elections this year as well. So there's been a whole, you know, a whole host of different dynamics, which means that, you know, yes, the EM index in aggregate is very much driven by AI, but there are many pockets of opportunities which are non-AI

7:35related.

7:36Unidentified speaker (speaker_0): So where does, for example, oil sense? sensitivity or commodities more broadly, which is often associated with many of the emerging market indices, where does that fit in these days given that shift in the index composition?

7:50Unidentified speaker (speaker_1): Well the weight of energy and and materials in the EM index has just, you know, decline significantly. Probably, you know, post uh global financial crisis it was around 30%. Today it's more in the in the 10% area. So you can see there's been this big shift, but there's still ways to play it. I mean, Brazil, obviously uh a key one, Chile, which has exposure to copper, South Africa with gold. Interestingly, you did have the GCC, which had a number of the big oil producers in there. But the relationship with oil has become very complicated Located in that region because higher oil prices at the moment just reflects higher geopolitical risk. And that means that the normal relationship of higher oil prices benefiting these GCC equities really has broken down quite a lot recently.

8:34Unidentified speaker (speaker_0): And is the trade theme or trade tensions or trade shifts a topic that you can express via emerging marketplace?

8:45Unidentified speaker (speaker_1): I mean, I I think a lot of the debate or a lot of the issues around tariffs um have have played throughout emerging markets already. And when we look at things like, you know, the earnings call transcripts, actually surprisingly, the discussion of tariffs, the you know, the sentiment around it, the potential impact for corporates, doesn't seem as large as as what it has been for a while. And so I don't think it's it's a it's a topic which is, you know, a main concern for people. I think, you know, the big issue or the big thing that people are thinking about now in terms of where to reallocate an emergency. in emerging markets is basically what anti AI trades do we have? And there has been a number of markets, be it India, uh, be it places in uh ASEAN, also in Latin America, which have been incredibly beaten up, you know, this year and have all of a sudden seen a really sharp rotation as people try to reduce their exposure to AI and get to some sort of more grounded structural stories within that EM complex.

9:42Unidentified speaker (speaker_0): Really interesting. And one thing you said about trade, just in some ways, the surprising resilience. We should say that the fundamentals globally have been surprisingly resilient. And of course, earnings, corporate earnings in the US, but also elsewhere, have as well. So that of course continues to be a backdrop. Yeah.

10:02Unidentified speaker (speaker_1): And I mean, again, you go on the EM as a whole. I think this quarter, earnings will be close to 70% year-on-year growth. Now, a huge portion of that is tech, but strip out tech, uh earnings in EM this quarter will be around 20-25%. That is actually a really healthy backdrop and will be one of the strongest, you know, periods of growth really since uh the post-COVID post COVID bounce. So again, it shows you that despite a lot of these perceived headwinds that we see in the news, actually on the ground, companies are adapting, they're resilient, and they're delivering earnings growth. And that I think is you know not just a story in EM, it's a story in the US and and also Europe. Uh in the US, uh this quarter again, we saw almost another 30% earnings growth. And what's kind of very interesting is that you look at the performance of equity markets, and and really they've been quite choppy the last you know three to four months. But what that also implies is that valuations have also derated significantly. And I always try to push back a little bit on this because I think particularly in the world of AI, people say, oh, AI bubble, valuations are really expensive. And it's like when you look at the data, it's just not true. Actually, semis are basically trading uh close to this, you know, the lowest levels in 10 years. I think semis actually in particular trade on 13 times earnings. If you take consumer staples, which is like your

11:20Unidentified speaker (speaker_0): This is back in the US market.

11:21Unidentified speaker (UNKNOWN): Back in the

11:22Unidentified speaker (speaker_1): US market, yeah. You know, you look at your consumer staples, which is more your kind of traditional retailers or whatever else, they're trading at some of them trading 35 to 40 earnings. So actually, you really don't see, I would say, a huge amount of froth, particularly amongst the the big, most liquid areas. AI names. And that's another reason why we kind of feel quite confident going into next year.

11:43Unidentified speaker (speaker_0): Well, perfect segue to the last segment. What to look out for in 2027? So just recap as you just started doing your broad view on the markets. And just what are the two or three signals that you'll be watching most closely? Because obviously there is a chance that things could be changed.

12:01Unidentified speaker (speaker_1): Completely. First things first. is that we are bullish going into next year and almost feel like there's this kind of potential for a Goldilocks environment for equities where you have earnings uh still growing very strong as a result of AI monetization. I think the next big step for the earnings story is AI adoption. Our data science team, uh, led by Mark McDonald, does a fantastic amount of work in terms of drilling into the earnings and discussing or seeing what earnings are saying. And it's one of my favorite things to do is look at these and look at the examples of companies giving. And it's clear that they're seeing, you know, margin improvements. That I think is going to be a really big story in 2027. That helps the earnings front and

12:44Unidentified speaker (UNKNOWN): the

12:44Unidentified speaker (speaker_1): backdrop of, you know, higher bond deals, which has been a really big story this year, uh, a more hawkish fed, that we probably feel like we've reached the peak of. And that starts to help valuations a bit more. And potentially, depending on what the geopolitical backdrop is going into next year, we could also see lower oil prices. And actually, we're already seeing lower oil prices, and that can provide a bit of a relief to uh corporate margins as well. So they're kind of the three things that we're, you know, most closely following going into twenty 2027. twenty seven.

13:12Unidentified speaker (speaker_0): Alistair, we've covered a lot of ground. Um thank you so much. So uh we'll keep looking out for updates. You obviously publish on all of these developments on a regular basis. Thank you so much for joining us. Thank

13:25Unidentified speaker (speaker_1): you for having

13:26Unidentified speaker (UNKNOWN): me. So

13:31Unidentified speaker (speaker_0): that was Alistair Pinder on the Outlook for Global Equity Markets. Now, three quick things before we go. First, please like and subscribe us wherever you get your podcasts. Second, don't forget to check out our sister podcast under the Banyan Tree, where hosts Fred Newman and Harold van der Linden put Asia's economies and markets into context. And actually, this week they're talking about India as an alternative to AI exposure. So very, very topical. Thirdly, a reminder. to HSBC clients that voting is open for the Xtel Asia Survey 2027. So if you valued our insights this year, then please consider voting. Head to extelinsights.com/slash voting to cast your ballot. But that's it from us. This episode was hosted by me, Aline Van Dyne in New York. I was joined by Alice. Alistair Pinder, and it was produced by Tom Barton in London. We'll be back again next week, so until then, thanks very much for joining

14:35Unidentified 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.
All insights