PIMCO analysts discuss AI CapEx timing, European energy constraints, and credit risk in a high-yield environment
Summary
Speakers argue that US permitting frictions will slow the pace of AI infrastructure build-out rather than reduce its total size, shifting value accrual across the investment cycle. They note that Europe's status as a net energy importer limits its ability to absorb AI data centers compared to energy-sufficient regions like the Nordics. The rise in US yields is attributed to real yield resets driven by growth expectations rather than inflation, creating a favorable risk-reward profile for high-quality fixed income. Rising funding costs are pressuring low-quality borrowers, particularly in the CCC high-yield segment and commercial real estate. AI debt issuance has not crowded out broader credit sectors, as spreads remain stable and issuance is absorbed within the AI ecosystem. High-quality balance sheets are positioned to benefit from the slowdown as they can wait out market imbalances.
Key takeaways
- 1 US permitting friction is likely to slow the pace of AI CapEx rather than reduce the total size, shifting value accrual across the investment cycle.
Unidentified speaker (speaker_2) · 1:53
- 2 four of the five hyperscalers are essentially negative free cash flows. And so the reliance on external funding to continue to sustain the current pace of CapEx is really critical.
Unidentified speaker (speaker_2) · 2:53
- 3 Europe's status as a net energy importer limits its ability to absorb AI infrastructure build-out compared to energy-sufficient regions.
Unidentified speaker (speaker_2) · 4:32
- 4 The recent rise in US yields was driven primarily by real yields (growth expectations), not inflation, distinguishing it from typical inflationary periods.
Unidentified speaker (speaker_2) · 6:55
- 5 AI debt supply has been absorbed by wider spreads within the AI ecosystem itself, rather than crowding out broader corporate credit.
Unidentified speaker (speaker_2) · 15:27
- 6 To me, the clear winner or the clear beneficiary, at least in relative terms, is probably high quality balance sheets in general because they can afford sort of waiting.
Unidentified speaker (speaker_2) · 25:49
Episode sections
- 0:37 – 3:59
Introduction and US Political Impact on AI CapEx
The speakers discuss how the US midterm election results introduce permitting frictions that may slow the pace of AI infrastructure build-out without necessarily reducing its ultimate size. This shift in timing affects relative value performance and the predictability of the credit supply pipeline.
- 3:59 – 6:31
European AI Infrastructure and Energy Constraints
Europe may benefit from a rerouting of AI data center construction due to US regulatory hurdles, though energy availability remains a primary constraint. Planning is currently concentrated in five major cities, but opportunities exist in energy-sufficient regions like the Nordics and parts of France.
- 6:31 – 10:35
Yield Drivers and Portfolio Implications
Higher global yields are driven by real yield resets rather than inflation expectations, improving the risk-reward profile of high-quality fixed income. This environment restores bonds' role as portfolio protection, as evidenced by the October sell-off where bonds rallied while equities fell.
- 10:35 – 13:54
Borrower Stress and Credit Selection
Rising funding costs pressure low-quality borrowers, particularly in the high-yield CCC segment and commercial real estate, testing their ability to sustain leverage. High base yields reduce the incentive to take excessive credit risk, making active selection and dispersion management critical for alpha generation.
- 13:54 – 17:35
AI Debt Supply and Crowding Out Debate
The speakers argue that AI-related debt issuance has not crowded out other credit sectors or government bonds, as non-AI spreads remain stable. They attribute yield movements to macroeconomic resource mobilization rather than portfolio rebalancing, noting that AI issuance is absorbed within the AI ecosystem's own spreads.
- 17:35 – 21:50
Monetization and Future Supply Predictability
While AI hyperscalers show revenue growth, spending continues to outpace it, creating negative market reactions. The speakers expect AI debt supply to become more predictable in 2028 as CapEx growth slows, allowing investors to better assess terminal values and manage issuance pipelines.
- 21:50 – 24:21
European Fiscal Capacity and Diversification
Germany offers fiscal space to stimulate its economy, and Europe's lagging position in AI infrastructure provides diversification benefits for investors. Regulatory changes in Spain and energy self-sufficiency in the Nordics create specific opportunities, though structural capital market fragmentation remains a hurdle.
- 24:21 – 28:59
Scarcity Value and Beneficiaries
A slowdown in AI infrastructure build-out creates scarcity value that benefits hyperscaler credits more than equities due to supply dynamics. High-quality balance sheets are positioned to benefit from the ability to wait out market imbalances, while 'picks and shovels' sectors face demand destruction risks.
Claims worth verifying
Four of the five hyperscalers are essentially negative free cash flows.
Unidentified speaker (speaker_2) ·2:53
How to verify: Review quarterly financial statements (10-Q/10-K) for the five largest cloud providers (e.g., Microsoft, Amazon, Alphabet, Meta, NVIDIA) on SEC EDGAR to calculate Free Cash Flow (Operating Cash Flow minus Capital Expenditures). For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Primary sourceUS yields are driven by real yields, not inflation expectations.
Unidentified speaker (speaker_2) ·6:55
How to verify: Compare the change in nominal 10-year Treasury yields with the change in 10-year TIPS yields and breakeven inflation rates over the relevant period using Federal Reserve H.15 data. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Base yield in the dollar market is around five percent.
Unidentified speaker (speaker_2) ·9:06
How to verify: Check the yield on the US 10-Year Treasury Note or the ICE BofA US Corporate Bond Index yield at the time of the podcast (November 2026). For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
CCC high-yield bonds have been under pressure since late April.
Unidentified speaker (speaker_2) ·10:54
How to verify: Analyze the ICE BofA US High Yield CCC Index total return or spread data from April to November 2026. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Primary sourceAI debt issuance has not crowded out other credit sectors.
Unidentified speaker (speaker_2) ·18:57
How to verify: Compare the spread performance of non-AI Investment Grade Corporate Bonds against AI-related bonds using Bloomberg or ICE BofA indices. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Europe is a net importer of energy.
Unidentified speaker (UNKNOWN) ·4:31
How to verify: Review Eurostat or IEA energy balance data for the European Union showing net energy imports. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Spain relaxed requirements for non-renewables out to 2029.
Unidentified speaker (speaker_1) ·23:58
How to verify: Search Spanish government regulatory filings or energy ministry announcements regarding data center energy sourcing rules effective post-2026. For recent events, verify 2026-08-17 through 2026-11-17. Use another period only when the speaker explicitly names it.
Primary sourceDo your own research
- Verify US permitting timelines and regulatory hurdles for AI data centers via US federal and state agency filings.
- Check hyperscaler free cash flow statements and debt issuance schedules in SEC 10-Q and 10-K filings.
- Assess European energy grid capacity and import dependency data from Eurostat and national energy regulators.
- Compare current US Treasury real yields against historical 25-year averages using FRED or Treasury.gov data.
- Monitor high-yield CCC bond default rates and spread data from S&P Global or Bloomberg.
- Review AI hyperscaler revenue growth versus CapEx spending trends in quarterly earnings calls.
- Evaluate commercial real estate loan maturity walls and refinancing risks via mortgage bank reports or Fed data.
- Analyze European fiscal stimulus plans and regulatory changes in Spain and the Nordics via government press releases.
Full transcriptexpand
- 0:00
- Unidentified speaker (speaker_0): Stay tuned after the conclusion of the podcast for additional important information. Subscribe for more episodes connecting macro trends to portfolio strategy and visit Pimco.com for extensive research and resources.
- 0:17
- Unidentified speaker (speaker_0): Welcome to Fixing Your Interest.
- 0:20
- Unidentified speaker (UNKNOWN): In
- 0:21
- Unidentified speaker (speaker_0): today's episode, Lot Varui, Viruy, multi-asset strategist and head of client Quiet solutions. and Analytics, and Nidhi Nakra, Credit Portfolio Manager, explore how the AI investment boom is reshaping global credit markets.
- 0:37
- Unidentified speaker (UNKNOWN): From
- 0:37
- Unidentified speaker (speaker_0): the financing demands of the data center build-out to higher funding costs and growing corporate issuance. The discussion looks at where risks may be emerging, where opportunities are developing across the US and Europe. and why active credit selection could become increasingly important as this investment cycle
- 0:56
- Unidentified speaker (speaker_0): unfolds.
- 0:58
- Unidentified speaker (speaker_1): Lockby, welcome once again to London.
- 1:01
- Unidentified speaker (speaker_1): It's been an unusually foggy two days.
- 1:04
- Unidentified speaker (speaker_2): Yes. Thank you for having me.
- 1:06
- Unidentified speaker (speaker_1): Yeah, it's a warm welcome to you. So as a credit portfolio manager, uh the AI CapEx theme team has really dominated spreads.
- 1:14
- Unidentified speaker (UNKNOWN): Um,
- 1:15
- Unidentified speaker (speaker_1): so you know, as such, I think it would be nice to um clear the fog
- 1:19
- Unidentified speaker (UNKNOWN): on
- 1:19
- Unidentified speaker (speaker_1): a few of the AI CapEx questions. Um, let's start with the US midterms. Um, what are the implications for the large Democratic congressional majority for the AI Capex bailout?
- 1:34
- Unidentified speaker (speaker_2): Well, I guess prediction markets didn't wait for the election to happen to kind of start pricing. in some of that. But the most immediate implication is that it probably introduces a serious friction in this AI infrastructure build-out. It does make permitting somewhat more complicated than initially, you know, anticipated. Um, the
- 1:53
- Unidentified speaker (speaker_2): question is, is that friction going to slow down the pace of this CapEx cycle without necessarily impacting its ultimate size? or will it kind of change both? My view would be that it's probably more likely that this is just a change in the pace of the build-out without necessarily impacting the
- 2:13
- Unidentified speaker (speaker_2): ultimate size. And so, in other words, the destination is the same. The journey might take a little longer. And that has plenty of implications in terms of relative value performance, in terms of where actually value in that AI infrastructure build. build-out is accrued. But size, I think, is unlikely to change. You know, the pace of the build out will likely
- 2:33
- Unidentified speaker (speaker_2): slow down a little bit. And if you go back, you're right, AI as a theme has been a major driver of relative performance in credit markets. And if you go back to the last 12 months, I think the reason why AI has been such a dominant driver of performance in corporate credit is because of the interaction between capital. CapEx and supply. Today,
- 2:53
- Unidentified speaker (speaker_2): four of the five hyperscalers are essentially negative free cash flows. And so the reliance on external funding to continue to sustain the current pace of CapEx is really critical. And I think credit has been kind of caught, or at least the AI ecosystem in credit has been caught in this negative interaction
- 3:13
- Unidentified speaker (speaker_2): between CapEx and
- 3:15
- Unidentified speaker (speaker_2): supply because every incremental new dollar of CapEx, there's probably a decent portion of that that is funded in debt. If you slow down CapEx, you also slow down supply, and you make that pipeline on the margin a little more predictable for investors. And so for some parts of the
- 3:35
- Unidentified speaker (speaker_2): AI ecosystem, this is probably good news. And I'm talking
- 3:38
- Unidentified speaker (UNKNOWN): about the
- 3:39
- Unidentified speaker (speaker_2): from a credit perspective here, not necessarily the equity perspective.
- 3:43
- Unidentified speaker (speaker_1): Interesting. So your um expectation would be more rather s you know slower build out rather than less CapEx. And specifically for the US, right? So what does this mean for the rest of the world? Is that greater opposition?
- 3:59
- Unidentified speaker (speaker_2): I mean if you have to reroute some of the building you know, outside of jurisdictions that are somewhat less friendly towards AI, then I think the rest of the world might on the margin benefit from that. I'm thinking Europe, Canada is another example, you know, abundant land, uh great, you know, great system, etc. I would keep
- 4:19
- Unidentified speaker (speaker_2): space. Uh but I would just keep in mind one thing, which is permitting is one friction, but there are other constraints. I think the other major constraint to the build-out is is the grid system, power, energy, et cetera.
- 4:31
- Unidentified speaker (UNKNOWN): And
- 4:32
- Unidentified speaker (speaker_2): so I would keep that in mind in those conversations about potential rerouting, because you know, it's not like energy is abundant in the rest of the world. I mean, Europe is a good example. Obviously, Europe is a net importer of energy. Uh, and so that ultimately puts the brakes on how much you can relocate, which is why I said initially. that this
- 4:51
- Unidentified speaker (speaker_2): is probably uh a catalyst for a slower pace without necessarily changing, you know, the outcome. And by the way, rerouting doesn't happen overnight either. You know, these are sort of decisions that are made over multiple quarters, maybe years. Uh but you know, power
- 5:06
- Unidentified speaker (UNKNOWN): is
- 5:06
- Unidentified speaker (speaker_2): the other constraint we need to keep in mind too.
- 5:08
- Unidentified speaker (speaker_1): Yeah, I think that's right. And the investment in Europe's already been taking place, obviously. It's relatively concentrated. Um, I think about two thirds or more of the um planning in Europe is concentrated around five cities. So there's Frankfurt, London, Amsterdam, um, Paris and Dublin, affectionately known as Flap D.
- 5:29
- Unidentified speaker (speaker_1): And I think both of us know that Europe is uh from an innovation sense, a lot more than just five cities. So there is quite a lot of catch um catch up to do in Europe and and therefore opportunity as well. We know that in Madrid and Milan, um, you know, they're they're they're working on more data center planning. And then some of the Nordic countries are
- 5:49
- Unidentified speaker (speaker_1): energy sufficient, um, you know, to your point about energy or France as well. So, you know,
- 5:55
- Unidentified speaker (speaker_2): Yeah. I think that's gotta be the key differentiator really in terms of the ability to attract that marginal daughter dollar into the build build-out. out. And I would use that same argument for merger markets too.
- 6:05
- Unidentified speaker (speaker_1): Yep.
- 6:05
- Unidentified speaker (speaker_2): They could benefit too, subject to that sort of energy power constraint.
- 6:11
- Unidentified speaker (UNKNOWN): So
- 6:11
- Unidentified speaker (speaker_1): let's move on to the financing aspect of this rather large CapEx uh bill. So we have had high um policy rates and interest rates for about four years Um now, and major central banks are hiking. Um and to add to that, in Europe. yields, particularly in France and Italy, have risen. So, what
- 6:31
- Unidentified speaker (speaker_1): is the impact of these higher funding costs on um the debt
- 6:35
- Unidentified speaker (speaker_1): roller?
- 6:36
- Unidentified speaker (speaker_2): Uh let's unpack that a little bit because yes, you've seen a uh a pretty significant reset higher in global yields, you you know, US and and and the rest of the world. But I would say the drivers of that backup in yields uh were different, actually, across both. sides of the Atlantic. I mean, in the United States, it was mostly a real yield
- 6:55
- Unidentified speaker (speaker_2): story. If you decompose your nominal 10-year yields into real yields, so inflation adjusted and then break-even inflation, which is your compensation for inflation risk, really, the vast majority of the move higher has been driven by real yields. That's really important because that tells you that the the primary driver of that back. backup has been
- 7:15
- Unidentified speaker (speaker_2): actually a reset in growth expectations,
- 7:19
- Unidentified speaker (UNKNOWN): at
- 7:19
- Unidentified speaker (speaker_2): least before October. You look at Europe, if I zoom in on markets that have an active inflation sort of bond market like Euro or Sterling, what you see that the picture has been more mixed. You know, it's been roughly 60-40. 60% is real yields, 40% is a repricing of inflation.
- 7:39
- Unidentified speaker (speaker_2): expectations. That is a somewhat less friendlier backdrop. And so I get asked this question all the time, which is how come risk assets digested up until early October the backup in US yields so well? I think the answer is because the drivers of the backup always
- 7:58
- Unidentified speaker (speaker_2): matter more than the backup itself. And in this case, the drivers were actually. a reset of growth expectations higher. Now, fast forward to today, what happened in October, a major sell-off in tech stocks, and probably a reset lower in those growth expectations. Duration did what it was supposed to do, which is
- 8:18
- Unidentified speaker (speaker_2): rally and give you that protection in a multi-asset portfolio. But here again,
- 8:23
- Unidentified speaker (UNKNOWN): the
- 8:24
- Unidentified speaker (speaker_2): rally value was driven by the real yield component, which again can kind of confirms that you know growth is kind of in the driver's seat a little bit.
- 8:32
- Unidentified speaker (speaker_1): Yeah, I'd agree. So what would you say these higher yields coupled with that um new um you know return of the correlation to equities means for portfolio construction?
- 8:46
- Unidentified speaker (speaker_2): Well, there's a number of implications. The the obvious one is that the ability of fixed income. to give you income, for the lack of a better word, has actually been restored. And so now if you take the dollar market as a benchmark, and you could certainly generalize that throughout the entire G10 universe, but you know, the base yield is around five percent, give or take
- 9:06
- Unidentified speaker (speaker_2): 10 10 basis points. And so the value proposition of high quality fixed income from a risk reward standpoint has improved quite dramatically. I think the other implication of this reset high in yields is that the ability of duration to protect you in a bad state of the world is dramatically
- 9:26
- Unidentified speaker (speaker_2): better because real yields are, you know, north of the two and a half percent today. And October was case in point, in my view. You did have
- 9:35
- Unidentified speaker (UNKNOWN): a
- 9:36
- Unidentified speaker (speaker_2): big sell-off in the equity market, and to your point that negative correlation between bonds and stocks was full and into display. You know, bonds rallied, stocks sold off, and you had that sort of embedded protection in a multi-asset portfolio. And so to me, the lesson from that October sell-off is that
- 9:56
- Unidentified speaker (speaker_2): you have to think more carefully about how you allocate that marginal dollar. You don't have to dislike stocks and in order to own bonds. I think the math looks dramatically different today. You have a starting level of yields that is way more attractive, possibly the most attractive in in 25 years, and that, you know, usually bodes well
- 10:15
- Unidentified speaker (speaker_2): for total returns on a forward basis, but it also gives you that embedded protection in case sort of the cycle ends in an abrupt way.
- 10:23
- Unidentified speaker (speaker_1): Yeah, I I couldn't agree more. Very asymmetric. It at it does at these uh um levels. And what uh you know, this is with regards to portfolio construction. What do the higher yields mean for borrowers?
- 10:35
- Unidentified speaker (speaker_2): Well, someone else is on the other side of that trade and it's the borrower, right? And so I think this is sort of bad news for low quality borrowers across the board. I mean, as you know, we're seeing it in the high-end high-yield market, you know, the triple C part of the market. So that's the low quality uh, you know, part of the high yield market has been under pressure
- 10:54
- Unidentified speaker (speaker_2): pretty much since late April, all the way to today. I think one important driver of that has been the reset higher in the cost of funding. But if you built a capital structure in a world where policy rates were close to zero, and your running assumption was that the base yield would be, um take again the dollar market.
- 11:15
- Unidentified speaker (speaker_2): as an example, between two and a half and three percent, that capital structure does not work in today's world. Right. And so the cost of funding I think plays an important sort of role there. There are other idiosyncratic considerations, but generally speaking, I think you know, look the ability of low quality borrowers to sustain, you know, current leverage
- 11:34
- Unidentified speaker (speaker_2): levels is going to be tested and probably diminished. Um I think there are other rate sensitive parts of the economy that this reset higher rates will be tested. I've had a number of clients asking me what this means for the commercial real estate market, for example. You remember the 2023 episode? Um
- 11:55
- Unidentified speaker (speaker_1): painfully.
- 11:55
- Unidentified speaker (speaker_2): Exactly. So that was a painful reset higher, both on the equity side and the credits. credit side. And both in the US and in Europe, sovereign European credit is another area where you've seen pressure not entirely due to the reset higher in yields, but somewhat due to it.
- 12:16
- Unidentified speaker (speaker_2): You know, when you reset yields higher, you do put pressure on sovereign balance sheets that are sort of suffering from the intersection of weak growth and and and and poor fiscal policy. So, you know, that's that's another area where you could you'll probably see could where the bar for normalization is is is probably quite high.
- 12:35
- Unidentified speaker (speaker_1): Yeah. It I I guess when we say it's not going to become a credit story, we mean it's not systemic, but that's
- 12:43
- Unidentified speaker (speaker_1): name
- 12:43
- Unidentified speaker (speaker_2): correct
- 12:43
- Unidentified speaker (speaker_1): accidents. And we've seen In those already.
- 12:46
- Unidentified speaker (UNKNOWN): So
- 12:46
- Unidentified speaker (speaker_1): it becomes quite important, I think, actually in this in mind.
- 12:48
- Unidentified speaker (speaker_2): To be selective. And by the way, dispersion has been the biggest source of of of of alpha generation. And I think that remains the case. I mean, I think the it
- 12:57
- Unidentified speaker (UNKNOWN): this
- 12:57
- Unidentified speaker (speaker_2): is a really important point because when you reset the cost of funding higher, you know, whether it's through the base yield or through the spread component, you essentially reduce incentives to go down in the quality spectrum. All that's at equal. sequel. Again, you know, you can generate a seven. seven and a half percent on a high quality uh uh
- 13:17
- Unidentified speaker (speaker_2): portfolio fixed income. There's no need to take an excessive amount of credit risk. And so this credit jargon, this feels like a decompression environment all at sequel. Um and I think a lot of it is due to that reset high on the cost of uh capital and and the base yield in particular.
- 13:34
- Unidentified speaker (speaker_1): Yeah, there's also well, first of all, the five trillion of Capex over the next five years, in the face of everything you've described, and the lack of potentially attractiveness of the lower part of the capital structure for the wider um investors, then who's gonna buy this five
- 13:54
- Unidentified speaker (speaker_1): trillion?
- 13:54
- Unidentified speaker (UNKNOWN): Well, who's
- 13:55
- Unidentified speaker (speaker_1): gonna sell?
- 13:56
- Unidentified speaker (speaker_2): Yeah, I mean a couple of things I would say. First of all, I'd be a little careful with like long-term projections. of, you know, CapEx. And so I would take things like one year at the time a little bit. I have greater confidence in the 2027 outlook for CapEx as opposed to a multi-year, multi-trillion, you know, type of
- 14:15
- Unidentified speaker (speaker_2): cycle, because that is based on a set of assumptions. And those assumptions may or may not materialize. But the biggest one is that there's a little bit of a path dependency. You know, your ability to sustain those levels of capex.
- 14:27
- Unidentified speaker (speaker_2): depends on your ability to monetize. And so the the the there's a wide range of outcomes. And I think that affects CapEx too. But let's take it at face value, just for the sake of running this exercise. So the question is: who's gonna fund all of this? Um, you know, especially considering the fact that there is a funding gap, that the
- 14:47
- Unidentified speaker (speaker_2): hyperscalers, which are basically the primary source of this capex, have. And so if you have a funding gap, well, you need to rely on external financing. And you do that either by going to the equity market or debt markets. There aren't that many options for you to do that. I really think the better question is at what level, you know, that sort
- 15:07
- Unidentified speaker (speaker_2): of $5 trillion capex, which will translate into a couple of trillion dollars of debt supply, will clear. And I like to look back at the last 12 months as a benchmark. If you look back at the last 12 months, this AI supply shock into debt markets has actually been absorbed at
- 15:27
- Unidentified speaker (speaker_2): its source so far. It's been absorbed via wider spreads for that AI ecosystem itself, as opposed to the broader corporate credit market.
- 15:37
- Unidentified speaker (UNKNOWN): I think
- 15:37
- Unidentified speaker (speaker_2): for the next 12 months. months
- 15:39
- Unidentified speaker (UNKNOWN): or
- 15:40
- Unidentified speaker (speaker_2): 18 months, despite the friction that is introduced in permitting, it's very reasonable to expect the next wave to possibly be
- 15:49
- Unidentified speaker (UNKNOWN): at
- 15:49
- Unidentified speaker (speaker_2): least as large, maybe larger, faster. It'll probably come in more flavors across structures. And so, how would that be cleared? Probably with some additional buildup of verse cream. premium. Now, the difference, I think. between the next wave and the one that we've just experienced is that the next wave will likely become a little bit more
- 16:08
- Unidentified speaker (speaker_2): predictable. And in fact, if you look at CapEx growth, and I want to emphasize on growth, not the absolute number, CapEx growth is already projected to slow down a little bit. So that second derivative in CapEx is going to decline a little bit. And if CapEx growth slows down, your issuance pipeline
- 16:28
- Unidentified speaker (speaker_2): becomes a little bit more predictable. And as an investor, you're sort of able to assess,
- 16:33
- Unidentified speaker (UNKNOWN): to,
- 16:34
- Unidentified speaker (speaker_2): to assign a terminal value as to where this is going eventually. But I do think that part of the reason why the past 12 months
- 16:44
- Unidentified speaker (speaker_2): had
- 16:45
- Unidentified speaker (UNKNOWN): that
- 16:45
- Unidentified speaker (speaker_2): level of reset and spreads is because the pipeline was actually quite quite unpredictable. It was really difficult. I mean, remember, if you look at CapEx consensus expectations for 26, where we are today versus where we started the year, the number more than doubled, basically. And if that number doubled, that
- 17:05
- Unidentified speaker (speaker_2): means that supply expectations probably doubled too. Uh, I think next year it'll feel better, and in 2028, it will likely get even more predictable. So we're not too far from the tip of the mountain, I guess,
- 17:18
- Unidentified speaker (UNKNOWN): if
- 17:18
- Unidentified speaker (speaker_2): you will.
- 17:19
- Unidentified speaker (UNKNOWN): I
- 17:19
- Unidentified speaker (speaker_1): think that's a fair point about predictability. And also I guess the point we're making is that if a year or two down, which is far away, but if if this is still happening, it's because then the equity market would
- 17:33
- Unidentified speaker (speaker_1): Companies would have shown some monetization.
- 17:35
- Unidentified speaker (speaker_2): And by the way, they did, even though
- 17:38
- Unidentified speaker (speaker_2): third quarter earnings season feels like it was ages ago, but it just ended actually two weeks ago. But what one of the really important takeaways from that earnings season that on the revenue side,
- 17:49
- Unidentified speaker (UNKNOWN): the
- 17:49
- Unidentified speaker (speaker_2): numbers were good. I mean, the issue is that the spending side of that equation continues to move faster, which is, you know, why the market reacted the way it did. But on the revenue side, there's clear evidence of monetization already.
- 18:02
- Unidentified speaker (speaker_1): Yeah. Um, so on the question of you know spreads clearing, they've been relatively contained to your point. And there's been this narrative that investors have been selling government bonds to to make space for this massive um capex spending that's coming from corporates. Do you think that's correct?
- 18:22
- Unidentified speaker (speaker_2): Well, you know, that that is what's usually referred to as sort of a the crowning out theory, which is the this this this idea that this debt funded AI capex cycle is sort of putting pressure on other parts of the fixed income universe. And by the way, you hear it
- 18:37
- Unidentified speaker (UNKNOWN): um
- 18:37
- Unidentified speaker (speaker_2): when it comes to government bonds, treasuries, or or other, you know, government bonds. But you also hear it uh with respect to other corners of credit markets. Um let's start with credit, because that's to me, that's probably the more obvious crowning out channel. For now, there's really little evidence, at least on the
- 18:57
- Unidentified speaker (speaker_2): pricing side, that this AI, you know, debt funded CapEx cycle is actually crowning out other parts of the credit universe. And the best way to look at that, take your investment grade Great index and then divide divided it in two universes, AI debt, non-AI related debt. The non-AI related debt part of the
- 19:17
- Unidentified speaker (speaker_2): market has been remarkably stable. Actually, you know, spreads are still tight. More importantly, they've been really oscillating within a very tight range. And I think, look, government bonds, whether it's treasuries or other bonds, have superior liquidity. You just don't buy them for the income, you buy them for the protection. And I
- 19:37
- Unidentified speaker (speaker_2): struggle to see an AI sort of infrastructure build-out, however big it is, put pressure on
- 19:44
- Unidentified speaker (UNKNOWN): a
- 19:44
- Unidentified speaker (speaker_2): $32 trillion market, which is the Treasury market. Now, the skeptics would say the counterargument to that is that, you know, try to put, you know, supply of long dated AI bonds in the context of supply of long dated government bonds. Yep. Yeah. And the numbers do look big. I don't disagree. And so from a flow standpoint, it does look. you know, a little big. Um,
- 20:04
- Unidentified speaker (speaker_2): but I think it doesn't fit with the definition of a crowding out in my view. Usually the textbook definition of the crowding out runs the other way around. It's usually an oversupply of safe assets that puts pressure on the private sector, not the other way around. We're actually for what it's worth, we put out uh a a piece, I think a
- 20:24
- Unidentified speaker (speaker_2): couple of weeks ago that kind of, you know, runs a little bit of an event study and looks at whether there's any evidence of abnormal returns in the treasury market around days when you had supplies AI deals and we haven't found any. Now, that does not mean that the reset higher in real yields is not related to AI.
- 20:44
- Unidentified speaker (speaker_2): It is related to AI, but I think the channel through which real yields have gone up has nothing to do with post- portfolio rebalancing. It's just, you know, uh an investment equals savings identity. And whether you funded
- 20:58
- Unidentified speaker (UNKNOWN): uh
- 20:58
- Unidentified speaker (speaker_2): the AI CapEx cycle with debt or with return earnings doesn't really change anything.
- 21:03
- Unidentified speaker (UNKNOWN): Um
- 21:03
- Unidentified speaker (speaker_2): so you have to be a little bit more nuanced about it. I think the macro channel is real. Yeah. I mean when you spend more on capex, you mobilize more resources in the economy and real yields go up But even if you did that with retained earnings, you would have seen the same result. And so it has nothing to do with the funding mix.
- 21:24
- Unidentified speaker (speaker_1): Um
- 21:25
- Unidentified speaker (speaker_1): so we mentioned a few things that um could be um, you know, Europe could be hamstrung by. One is obviously regulation and the, you know, the varieties of regulation we have across geographies, the space limitations.
- 21:39
- Unidentified speaker (speaker_1): um geographically, and then being not self-sufficient in energy, and then also finally having um, I guess, fiscally strapped balance sheets for the government.
- 21:50
- Unidentified speaker (UNKNOWN): Is
- 21:50
- Unidentified speaker (speaker_1): there low-hanging fruit? What what would you be optimistic about in Europe?
- 21:55
- Unidentified speaker (speaker_2): Well, I would say we do have fiscal capacity in Germany. Uh I mean, by and large, you're absolutely right, but if you kind of disaggregate things a little bit. there is, you know, decent fiscal space in in Germany, which is still the largest economy in Europe. So that is good news. So you have the ability to stimulate if you want to. I would say, look, the silver lining is
- 22:15
- Unidentified speaker (speaker_2): that Europe uh has not been at the epicenter of this AI infrastructure build-out. And so that's a bad thing and a good thing, but it could be a good thing in a state of the world where, you know, maybe this first. friction in permitting ends
- 22:28
- Unidentified speaker (UNKNOWN): up
- 22:29
- Unidentified speaker (speaker_2): being something bigger than just a temporary slowdown in the cycle. You know, maybe you know it it could cause a change in the
- 22:37
- Unidentified speaker (speaker_2): ultimate size a little bit. And so Europe in some ways does provide diversification uh, you know, for investors that are looking for ways to sort of hedge uh, you know, the AI infrastructure build-up. build up. That's a good thing. Now, of course, with we've seen some issuance in the Euro market from the hyperscalers, uh, but it's not nearly close to the levels that you've seen in
- 22:57
- Unidentified speaker (speaker_2): in in the dollar market. So that's one I think an an an interesting point in in in in my view. Yeah, I think you know, deregulation and fragmentations, those are sort of structural themes that will take time to unfold. But generally as a whole, you know, the European economy uh
- 23:15
- Unidentified speaker (speaker_2): you know, could have bigger capital markets, could have deeper capital markets. So there's plenty of room, I think, to grow those. Um, you know, the laundry list is, however, quite quite quite expansive and there's quite a lot that needs to be done there. But I think if you bring it closer to home, to me, the pitch, so to speak, is, you know, diversification against the
- 23:35
- Unidentified speaker (speaker_2): AI theme, uh if if that's a concern.
- 23:39
- Unidentified speaker (speaker_1): Yeah. I mean on the planning side as well, if there's been there's been some uh acceleration. I know in Spain they relaxed some of the requirements around using um renewables, so they've allowed non-renewables out to 2029. So that's been very supportive for the data center build out over there. And then you know, as we discussed earlier, there are
- 23:58
- Unidentified speaker (speaker_1): regions, France and some of the Nordics where there they are energy self-sufficient. There's there's bright spots. Um
- 24:06
- Unidentified speaker (speaker_1): I guess if compute does become scarce and Europe takes longer, and then um I guess the US is in a slower path, is that overall a positive for credit investors?
- 24:21
- Unidentified speaker (speaker_2): It depends what part of it AI uh financing chain or supply chain for that matter you look at. Um
- 24:28
- Unidentified speaker (UNKNOWN): I
- 24:28
- Unidentified speaker (speaker_2): think there's no question that if you slow down the infrastructure build out, you do create scarcity, basically. And the question becomes where is the premium for that scarcity value? Uh in my view, I think it's a more positive story for hyperscaler credits relative to hyperscaler equities, because in credit, there's a technical sort of
- 24:47
- Unidentified speaker (speaker_2): tailwind that doesn't exist on the equity side, or not at least not. to the same degree, and that's supply. And so if you slow down capex, you slow down supply, et cetera. I think it's probably a negative for picks and shovels, um, you know, semis, memories, et cetera, because that is demand destruction that will have to
- 25:07
- Unidentified speaker (speaker_2): happen now. If you slow down the capex, that means that you're buying less of whatever goes into the shell, which is the data center. And I think for you know, further down the chain, neo clouds in particular, it's almost on a case-by-case basis. You could argue that scarcity means that strong demand is going to linger for a
- 25:27
- Unidentified speaker (speaker_2): little longer.
- 25:28
- Unidentified speaker (UNKNOWN): Uh,
- 25:29
- Unidentified speaker (speaker_2): but that doesn't solve any of the funding issues and that doesn't solve any of the question marks that have a secular flavor to them. So, i.e., what happens four or five. years from now. But it certainly buys you time. But to me, the the the the key message here is that yes, you're absolutely right. There's scarcity value that gets
- 25:49
- Unidentified speaker (speaker_2): unlocked immediately because you gotta make that imbalance between too much demand for compute and not enough supply persist a little longer. Not everyone will benefit from it. To me, the clear winner or the clear beneficiary, at least in relative terms, is probably high quality balance sheets in general because they can afford sort of waiting.
- 26:10
- Unidentified speaker (speaker_1): That was um a lot of really interesting um insights, um, Lotfi. So it sounds like high funding costs um and regulations will impact companies differently over the next few years, and corporates will have different fates. and therefore s corporate selection or active
- 26:30
- Unidentified speaker (speaker_1): selection is going to be important. And I guess secondly, Europe is lagging, but um there are bright spots and therefore fertile ground for potential opportunities in the near future.
- 26:44
- Unidentified speaker (speaker_2): Yep.
- 26:45
- Unidentified speaker (UNKNOWN): Well
- 26:45
- Unidentified speaker (speaker_1): thank you very much um Lotfi.
- 26:48
- Unidentified speaker (speaker_2): Thank you for having me that's
- 26:49
- Unidentified speaker (UNKNOWN): great. It's great seeing you
- 26:49
- Unidentified speaker (speaker_1): in London. It's a pleasure as always.
- 26:51
- Unidentified speaker (speaker_2): Thank
- 26:52
- Unidentified speaker (UNKNOWN): you.
- 26:55
- Unidentified speaker (speaker_0): Thanks for joining us on Fixing Your Interest as we explored what the AI investment boom could mean for global credit markets. Stay with us for more perspectives on the forces shaping markets and investment opportunities. For more insights and analysis, please visit PimC.com.
- 27:14
- Unidentified speaker (speaker_0): PIMCO.com.
- 27:17
- Unidentified speaker (UNKNOWN): The
- 27:17
- Unidentified speaker (speaker_0): discussion and content provided within this podcast is intended for informational purposes only and may not be appropriate for all investors. Reliance upon information provided in a podcast is at the sole responsibility of the listener. The information included herein is not based on any particularised financial situation or need and is not intended to be and
- 27:37
- Unidentified speaker (speaker_0): should not be construed as a forecast, research, investment advice, or a recommendation for any specific PIMCO or other security, strategy, product or service. Past performance is not a guarantee of future results. All investments contain risk and may lose value. Investors should speak to their financial advisors regarding the investment
- 27:57
- Unidentified speaker (speaker_0): mix that may be right for them based on their financial situation and investment objective. Podcasts may involve discussions with non-PIMCO personnel, and such content contain the current opinions of the speaker, but not necessarily those of PIMCO. Other podcasts may consist of audio recording of an existing PIMCO article and such material contains the current opinions
- 28:17
- Unidentified speaker (speaker_0): of the manager.
- 28:18
- Unidentified speaker (UNKNOWN): The
- 28:18
- Unidentified speaker (speaker_0): opinions expressed in all podcasts are subject to change without notice. Information contained herein has been obtained from sources believed to be reliable. but not guaranteed. PIMCO as a general matter provides services to qualified institutions, financial intermediaries, and institutional investors. This is not an offer to any person in any jurisdiction where
- 28:38
- Unidentified speaker (speaker_0): unlawful or unauthorized. For additional important information, go to www.pimco.com dot com forward slash GBL forward slash EN forward slash general forward slash legal hyphen pages forward slash podcast hyphen disclosures