Fixed Income Conversation Corner with Lotfi Karoui (PIMCO) and Leslie Falconio (UBS CIO)

Lotfi Karoui · PIMCO

AI-linked investment-grade debt has widened steadily while the rest of IG stays resilient, with no crowding out yet.

Splitting investment grade into AI-related and non-AI debt shows AI-related debt has moved steadily wider over the past year, while non-AI IG has grinded tighter. Karoui sees little pricing evidence of crowding out so far. He expects crowding out inside the AI financing chain instead. The roughly $1.2-1.3 trillion of 2027 capex will require more supply from more issuers, which should push investors up in quality over the next couple of quarters. That will crowd out financing for the most leveraged capital structures. Dispersion is also appearing across the layers of the chain, from senior hyperscaler debt to neoclouds, and he thinks it is here to stay.

“AI sort of related debt has been moving steadily wider over the last year or so. The rest of the market has been a lot better behaved.”at 7:48

Lotfi Karoui · PIMCO

Higher borrowing costs won't meaningfully slow hyperscaler AI capex; AI issuers absorb it via higher risk premiums.

Higher cost of capital pushes the market to scrutinize financing models and move up in quality, including in the AI financing chain. Karoui doubts it will slow capex. Capex is highly concentrated in a handful of firms that can shape the whole business cycle, and free cash flow is now negative for some. A strategic race keeps spending going even when economics are less compelling. Since hyperscalers began relying on debt about 18 months ago, the AI supply shock has been absorbed at its source. The market charges AI issuers more, but not the rest of the market, so the adjustment comes as extra risk premium in the AI financing chain, with possible limited spillover to the index.

“I think the answer is no. There's also a strategic element to this race that may sort of force you to accept economic outcomes that are slightly less compelling... So I'm skeptical that this will have a meaningful impact on sort of the forward trajectory of CapEx”at 22:25

Lotfi Karoui · PIMCO

AI data-center monetization outcomes are too uncertain; credit investors should favor downside-protected structures.

AI compute looks sold out for two to three years, so spot data-center economics look great today. The credit question is how those economics will look in three to five years, and estimates of monetization vary very widely. A credit investor underwriting the leverage needed for the build-out captures no upside in a good state of the world. Karoui says PIMCO therefore looks for structures that pay back in a bad state of the world. He argues the AI debt-funded capex cycle is not one big trade and must be underwritten deal by deal.

“The problem I think from a credit perspective is how will those economics look like three, four, five years from now? And if you look at most estimates of monetizations, range of outcomes is incredibly wide.”at 24:40

Lotfi Karoui · PIMCO

A Fed with a stricter inflation focus is paradoxically good for duration by reducing de-anchoring risk.

The rise in yields mostly reflects a repricing of the Fed's forward path because the economy is healthy, not a risk-off rate sell-off. Karoui argues the Fed's changed reaction function, with a stricter interpretation of the inflation side of the mandate, is good for duration. A new leadership committed to inflation removes the risk that inflation expectations de-anchor and pushes intermediate-to-long yields sharply higher. He also explains that yields fell after a hot CPI print because a hike path eliminates that left-tail inflation scenario, which supports duration.

“Paradoxically, you could argue that that's actually a good thing. The fact that you have a new leadership that is committed to the inflation side of the mandate sort of eliminates the risk that inflation expectations get de anchored”at 3:41

Leslie Falconio · UBS Chief Investment Office

UBS RCAO expects Fed hikes in December 2026 and March 2027, with cuts starting late 2027.

Falconio relays the UBS RCAO view on the Fed path. Hikes are expected in December 2026 and March 2027, followed by cuts starting near the end of 2027. She describes the market argument that pricing reflects the stronger inflation print, and that the market may be underestimating the amount of cuts needed later rather than the number of hikes.

“[From RCAO,] we do believe the Fed hikes in December. We believe the Fed hikes in March 2027. And then by the near the end of twenty seven, it starts to cut.”at 27:27

Lotfi Karoui · PIMCO

Private credit's fundamental issues are unlikely to become a systemic shock, though a recession would expose weakness.

Private credit and direct lending have real fundamental issues, but the bar is very high for those issues to become a systemic shock like the run-up to the global financial crisis. Karoui says the systemic debate is largely settled. Headline fears have passed their peak because the market understands how private credit works better than in late 2025. The caveat is that the view depends on the cycle. A recession would expose the weakness again, and older vintages could still disappoint if the cycle keeps going.

“Private credit and direct lending in particular has fundamental issues for sure. But I think the bar is very, very high for those fundamental issues to morph into something that looks like a systemic shock”at 13:34

Lotfi Karoui · PIMCO

Direct lending will bifurcate: older 2021-23 vintages likely to disappoint while newer vintages look better.

The 2021-2023 vintages were underwritten during record fundraising, when underwriting standards loosened and capital outran deployable assets. Their sector mix leans heavily on software, which is vulnerable to AI disruption. Newer vintages should have better underwriting and possibly better valuations. Early evidence is tentative: spreads on newly originated loans widened even as the broader market tightened, leverage multiples on deals have declined, and lender-friendly terms are returning. Karoui says this shift will take several quarters to affect the whole stock of debt. He compares the situation to 2023 commercial real estate, where regional bank retrenchment created survivorship bias in later vintages.

“I think the theme in direct lending is going to be bifurcation. Newer vintages are gonna look better, both from a fundamental standpoint and a performance standpoint. I think some of the older vintages will probably disappoint from a performance standpoint.”at 15:49

Lotfi Karoui · PIMCO

The US cycle is unlikely to turn over the next twelve months; the economy remains healthy in aggregate.

The economy is uneven, with a K-shaped pattern across many dimensions, but in aggregate it is still pretty healthy. Karoui therefore sees a fairly low likelihood of a cycle inflection over the next twelve months. Credit is an anti-recession asset class that does poorly in bad states of the world. What would change his view is evidence of labor-market deterioration. That would spill over from CCC into broader high yield, and then widen investment grade.

“But in aggregate, this is an economy that's still pretty healthy in my view. And so the likelihood that the cycle inflects over the next twelve months is still fairly low.”at 9:27

Lotfi Karoui · PIMCO

US inflation should trend lower beyond the next couple of months as one-off price shocks fade.

Beyond the next couple of months, Karoui expects the direction of travel for inflation to be lower. Part of this is mechanical, because one-off factors that pushed inflation higher in 2025 and 2026 will naturally fade. Some decline should also come from productivity gains and benefits from broader AI adoption. He agrees with the view that the Fed's hikes are being priced as a price-stability response.

“I think we broadly agree with your assessment, which is if you look beyond sort of the next couple of months, think the direction of travel for inflation is probably lower.”at 28:21

Lotfi Karoui · PIMCO

Bonds are regaining their diversifier role; starting yields offer the strongest income and equity protection in 25 years.

Last week, the Nasdaq fell about 2.5% while the 10-year Treasury rallied roughly 25 to 30 basis points, showing the negative stock-bond correlation that duration is supposed to provide. Karoui calls this the first real protection from bonds in a multi-asset portfolio in a long time, though he says one data point does not make a trend. He argues the higher starting yield level changes the math for multi-asset portfolios. Bonds should provide income and protection during the 5-10% equity drawdowns he expects, and this is why he sees the recent bond bid as a structural shift.

“everywhere you look, you're being paid probably the strongest level of yield support in now twenty five years”at 29:57

Lotfi Karoui · PIMCO

High yield bonds beat loans on index quality, but alpha lies in bottom-up single-name credit selection.

Higher yields have reset the forward outlook for total returns, so investors do not need to move far down in quality to earn them. At the index level, high yield bonds have better quality than the leveraged loan market and an advantage on carry versus duration. Loans have sharp sector-composition differences. Karoui argues that the index comparison understates the opportunity set, because most alpha comes from bottom-up credit work. The period from the global financial crisis to COVID offered little dispersion, but that world has ended, so most opportunities are at the single-name level.

“at the index level, I think there's better quality and high yield relative to the loan market, but I think that understates, I think the opportunity set, because again, most of the Alpha generated capabilities are probably coming from bottom up work”at 31:53