The changing composition of the long credit market
For many U.S. corporate pension plans, long credit has become the workhorse of the liability-hedging portfolio. As funded status has improved and glidepaths advanced, plan sponsors steadily shifted from return-seeking assets toward long-duration fixed income. Long credit played a central role in that transition because it helps hedge both interest-rate exposure and the corporate spread component embedded in pension liability discount rates.
Against that backdrop, the recent improvement in the Bloomberg U.S. Long Credit Index’s credit-quality profile appears, at first glance, to be a welcome development. During the first seven months of 2026, the index quietly moved higher in quality: the AA-rated share increased, while the BBB share declined (see Figure 1). For plans still working through the later stages of de-risking, it may be tempting to view this as the market offering a cleaner, higher-quality liability hedge almost for free.
But the story is more nuanced
The long credit market has not broadly migrated higher in quality through a wave of upgrades or generalized balance-sheet improvement. Instead, the improvement has been driven largely by new long-dated issuance from a small number of very large, generally higher-rated technology companies — the hyperscalers. Meta, Amazon, Alphabet, Microsoft, Oracle and SpaceX have become more meaningful contributors to the corporate credit markets as they term out financing needs tied to artificial intelligence (AI), cloud infrastructure and data-center investment.
As shown below, the quality breakdown of the Bloomberg U.S. Long Credit Index has already changed over the past year, with a clear, albeit still small, increase in the AA-rated bucket. While the shift is modest, we expect it to become more visible as issuance continues to grow. JPMorgan Asset Management’s credit research analyst covering the sector estimates $400-450 billion of hyperscaler USD bond issuance (not inclusive of 144a datacenter bonds) for 2026 and 2027 of which approximately ~$250 billion could be eligible for Bloomberg Long Credit Index Inclusion. It is important to keep the magnitude in perspective. While the long credit index has grown materially over the years, reaching ~$2.6 trillion in market value terms as of 7/31/2026, this will still be a material addition to the index, all else equal.
Market-value weight is only one way to measure risk, however. For credit investors, DTS — spread duration times spread — is often more relevant because it captures the amount of spread risk embedded in an exposure. On that basis, the change looks more significant (see Figure 2).
Even after accounting for the fact that only a portion of new issuance is eligible for the Bloomberg U.S. Long Credit Index inclusion, the DTS share of hyperscalers — including Amazon, Alphabet, Microsoft, Oracle, SpaceX and Meta — has increased roughly threefold. It is now larger than the DTS contribution of the Big Six banks: JPMorganChase, Goldman Sachs, Morgan Stanley, Bank of America, Wells Fargo and Citi. This is driven by two factors:
- Hyperscalers have issued heavily at longer maturities with higher spread durations;
- Their spreads have widened recently, partly because of technical supply pressures.
The technical backdrop for hyperscaler and data center credit remains less supportive. Spreads and valuations continue to face pressure from persistent supply concerns and elevated issuance forecasts as the cost-benefit structure of the AI infrastructure build-out comes into focus. We do not, however, view this as an alarming development at this stage. While the volume of issuance creates a technical headwind for hyperscalers—and, more broadly, for the technology sector—it may also open opportunities for active investors to identify relative value. This is particularly true where new-issue concessions or curve dislocations offer attractive compensation, as well as in relative-value trades between hyperscaler and data center bonds in which the same hyperscaler serves as a tenant. Fundamentally, however, the picture is more reassuring. With some exceptions, hyperscalers exhibit favorable credit characteristics: most carry low leverage, generate strong operating cash flows, and hold substantial cash balances that nearly cover their outstanding debt (excluding capitalized leases). This fundamental strength is a key pillar supporting our constructive view on the sector.
Implications for pension liability management
For LDI (Liability Driven Investing) investors, the key point is that a liability-hedging portfolio benchmarked to long credit now carries more exposure to a small group of technology balance sheets than it did a year ago. These companies are all funding the same broad secular theme: AI infrastructure, cloud expansion and data-center buildout. And the AA sleeve of the long credit market increasingly reflects a more concentrated exposure to the ability of the largest platform companies to sustain their credit ratings through a multi-year capital-spending cycle and ultimately reap the benefits and be able to monetize AI technology complex.
We do not view this concentration as outsized or concerning today. It remains much lower than the concentration levels seen in equity indices, for example. Nevertheless, it is a development that pension investors should monitor.
There is also a constructive side to this quality migration. Many corporate pension plans want long-duration spread exposure to help close the duration gap, but they do not want to add significant downgrade or default risk late in the de-risking journey. Historically, increasing long credit exposure often meant accepting more BBB exposure, particularly for plans that needed more spread duration than could be sourced from higher-quality issuers.
For mature plans, the objective is often no longer to maximize excess return from the hedge portfolio. It is often to preserve funded-status gains while reducing unrewarded volatility relative to liabilities. In that framework, the value of long credit is not simply its yield pickup over Treasuries. It is also its ability to align the asset portfolio more closely with the high-quality corporate discount curves used to value pension obligations. If the long end of the market offers more AA and A exposure, sponsors may be able to match liability spread sensitivity with less reliance on lower-quality credit.
That can matter meaningfully for funded-status risk. In risk-off environments, funded status can be pressured from multiple directions at once: credit spreads widen, return-seeking assets may sell off, and weaker credits can face downgrade pressure or fallen-angel risk. A higher-quality long credit allocation does not eliminate spread volatility — AA and A spreads can still widen meaningfully in stress — but it may reduce expected downgrade and default drag in the hedge portfolio. That can help limit forced selling and reduce the likelihood that the LDI book becomes a source of disruption at precisely the wrong time.
There is another important funded-status dimension as well. Because U.S. corporate pension liabilities are typically valued using high-quality corporate bond discount rates, wider AA corporate spreads and higher long-end yields reduce the present value of future benefit obligations. Put differently, when discount rates rise, the liabilities decline mechanically and the funded status improves. However, the higher concentration of a small number of issuers in the AA universe can lead to future increases in liabilities and reductions in funded status if any of these issuers experience rating downgrades that cause them to fall out of the AA universe.
This is one reason the composition of the long credit market matters so much for pension investors. The goal is not simply to own more long corporate credit; it is to own the right long corporate credit in a way that tracks the liability discount curve while minimizing downgrade, default and concentration risks.
Dispersion creates room for active management
This is where active management becomes especially important. Even within ratings buckets that are often treated as high quality, individual bond outcomes can vary meaningfully, particularly at the long end of the curve. Figure 3 shows wide dispersion in second-quarter 2026 excess returns across AA-rated corporate bonds, with outcomes becoming more spread out as spread duration extends beyond roughly 10 years.
For pension plans, that dispersion matters because long credit is not simply a beta allocation. Small differences in spread performance can be amplified by long duration and can have a meaningful effect on funded-status volatility.
Experienced active fixed income managers with deep research teams can potentially add value by separating the winners from the losers within the same ratings cohort. A higher-quality benchmark does not eliminate the need for credit selection. If anything, concentration and dispersion make security selection more important.
Managers can underwrite issuer fundamentals, capital-allocation discipline, new-issue concessions, supply technicals, curve positioning, structural features and relative value across maturities. In a market where long-dated AA bonds can produce very different outcomes over the same period, the ability to own resilient credits, avoid deteriorating stories and manage exposure to crowded issuance can be a meaningful source of alpha.
Active management can also broaden the opportunity set beyond what is reflected in the index. A significant amount of new issuance tied to AI infrastructure and data-center buildout has come through the 144A market, and not all of those bonds are eligible for inclusion in broad public long credit indices. For plans investing only through a benchmark-replicating approach, that means some of the most relevant new long-duration supply may sit outside the index. Active managers with the ability to participate in eligible 144A transactions can evaluate those bonds directly, compare versus public-index alternatives and selectively add exposure where compensation is attractive.
For pension plans, that flexibility can be valuable. The growth of 144A issuance means the long credit opportunity set is broader than the benchmark itself, particularly in sectors connected to hyperscalers, AI infrastructure, power demand and data-center expansion.
Bottom line
The long credit index looks higher quality, and mechanically it is. That can be helpful for pension plans seeking to improve hedge ratios, maintain long-duration spread exposure and reduce reliance on lower-quality credit. But the shape of that quality has changed.
The market has quietly de-risked, but it has also quietly concentrated. For LDI portfolios in the later stages of a de-risking journey, that distinction matters. The benchmark may still be the starting point, but it should not be the full definition of risk. Plans should look through the headline rating improvement, understand the issuer and sector exposures driving it, and ensure that the hedge portfolio still reflects the risks they actually intend to hold.
An active long-duration manager with a proven track record across sectors and deep research capabilities can use today’s elevated supply to source duration and spread carry—without being confined to benchmark names and weights. The opportunity set is broad, with meaningful dispersion across structure, leverage, sponsor quality, tenant profile, and the maturity curve. In this environment, selectivity is essential: returns should be earned through underwriting, not “blind beta,” and exposure should be avoided where asset-level risks fall short of standards.