What happens when the same handful of technology giants that dominate equity indexes start dominating the long end of the corporate bond market too? That question has been sitting at the back of my mind for most of this year. The numbers finally forced me to confront it. Artificial intelligence related borrowers now account for nearly half of all year-to-date gross issuance of non-financial US investment-grade debt with maturities beyond fifteen years. The scale is no longer a side story. It is reshaping how credit portfolios behave.
For three years the AI investment boom lived mostly in equity land. Capex announcements moved share prices. Balance sheet strength was treated as a given. Now the financing side has grown large enough that bond investors can no longer treat it as background noise. Hyperscaler dollar-denominated bond issuance alone is expected to approach two hundred fifty billion dollars this year. When you widen the lens to the broader ecosystem of chipmakers, utilities, and data-center developers, estimates stretch between three hundred and five hundred seventy billion. Those are not incremental numbers. They are market-shaping ones.
Why Concentration Risk Suddenly Matters More
I used to view long-dated investment-grade credit as a relatively diversified universe. Different sectors, different balance-sheet stories, different sensitivities to the economic cycle. That assumption is fraying. When almost half of the long end is tied to the same capital-spending wave, correlation across the book rises whether you want it to or not. The traditional duration hedge, longer-dated US Treasuries, has itself been unusually volatile through 2026. So the classic playbook of simply extending maturity or leaning on government bonds feels less reliable than it did two years ago.
Political uncertainty adds another layer. Midterm election outcomes have already sparked debate about potential slowdowns in the AI build-out. Even a temporary pause in spending would hit the same group of issuers that now dominate long-end supply. For portfolio managers sitting with meaningful exposure, that concentration is uncomfortable. Spreads in many corporate sectors already look tight relative to the risks on offer. Financials and energy can provide some relief, yet both currently trade at levels that leave limited room for error.
The practical question becomes straightforward. Where can an investor still find meaningful excess carry without taking on the same AI-driven risks that already sit heavily in the portfolio? Three areas keep surfacing in conversations with multi-asset teams: agency mortgage-backed securities, higher-rated private mortgage bonds, and carefully selected emerging-market debt. None of them is a perfect solution. Each, however, brings a different risk profile that does not move in lockstep with the hyperscaler complex.
Agency Mortgage-Backed Securities as a Quiet Diversifier
Agency MBS has spent years in the shadow of corporate credit. Higher rates and rate volatility punished the sector in 2023. This year the picture looks different. The same instruments have shown more resilience. As of early November, agency MBS offered roughly a thirty-three basis-point pickup over comparable US investment-grade corporate spreads. That is not a massive number, yet it arrives without the direct corporate credit exposure that now feels increasingly concentrated.
The underlying collateral is residential mortgages guaranteed by government-sponsored enterprises. Prepayment behavior still matters, of course. Extension risk is real when rates stay elevated. Yet the credit risk itself sits in a different category from a technology company issuing thirty-year paper to fund data-center construction. That distinction is exactly what many managers are looking for right now.
I have found that high-quality securitized products often get overlooked when equity narratives dominate headlines. Agency MBS and certain non-agency residential MBS can deliver excess carry versus the broader investment-grade market while remaining relatively uncorrelated to the AI capital-expenditure cycle. The correlation is not zero. Nothing is. But it is low enough to matter in a portfolio construction exercise.
In the US, high-quality securitized products such as agency mortgage-backed securities and non-agency residential MBS may provide meaningful uncorrelated exposure to the AI theme while still offering decent excess carry versus the broader investment grade market.
That assessment matches what I have observed in recent allocation shifts. The instruments are liquid enough for most institutional books. The yield advantage, while modest, compounds when you consider the diversification benefit. In a world where corporate credit spreads look tight, any source of carry that does not double down on the same sector story becomes more valuable.
Higher-Rated Private Mortgage Bonds and the Repricing Opportunity
The second area that stands out is the senior, AAA-rated portion of bonds backed by mortgages that fall outside the Qualified Mortgage rules. These non-QM deals have undergone meaningful repricing. Spreads widened enough that the compensation for extension risk and structural complexity looks more reasonable than it did twelve months ago.
The primary risk here is not that borrowers stop paying. It is that repayments take longer than models assume. That extension risk has been the main driver of higher expected returns. Newer transactions also incorporate structural features designed to repay senior investors more quickly and to offer stronger protection against losses. Those improvements matter. They reduce the chance that earlier structural weaknesses reappear.
In my view the potential benefits in AAA non-QM currently outweigh the residual risks for investors who understand the collateral. Wider spreads largely compensate for extension risk and for the higher expected credit losses that sit further down the capital structure. Today’s new issues embed better mitigations than many deals from earlier vintages. That combination of yield and improved structure is hard to ignore when traditional corporate credit feels expensive.
These bonds are not as liquid as agency MBS. Position sizing has to reflect that reality. Yet for the portion of a portfolio that can tolerate modest liquidity risk in exchange for diversification and carry, the sector offers a genuine alternative to simply adding more technology or utility exposure at tight spreads.
Emerging Market Debt Beyond the Usual Suspects
Emerging-market debt brings an entirely different set of drivers. Political, fiscal, and monetary forces in many of these countries operate independently of the US technology capital cycle. That independence is the point. Pockets of EM do have meaningful AI exposure, particularly in certain Asian economies, but much of that exposure lives in equity markets rather than fixed income. For bond portfolios the asset class still functions as a differentiated source of yield, duration, and country-specific risk.
US dollar-denominated investment-grade deals used to finance infrastructure have been especially interesting. This year they have offered a median pickup of around forty basis points relative to more conventional EM bonds. That premium is not enormous, yet it arrives with a different risk signature. Commodity-exporting countries also present selective opportunities for managers willing to look beyond the main benchmarks.
Local-currency bonds add another dimension. Exchange-rate movements can enhance or erode returns, sometimes dramatically. The volatility is real. Still, for portfolios that already carry substantial US duration and credit risk, a measured allocation to local-currency EM can reduce overall dependence on the same set of domestic factors.
It is worth stating clearly: emerging-market debt is not a lower-risk asset class. It simply exposes investors to a different mix of political, fiscal, and monetary forces than those facing the domestic US market. In a year when AI-related issuance has crowded the long end of the US corporate curve, that difference itself becomes an attractive feature.
Putting the Pieces Together in Practice
None of these three diversifiers should be treated as a complete replacement for corporate credit. They work best as complementary exposures. A practical approach might start by trimming some of the most AI-sensitive long-dated corporate holdings and reallocating a portion into agency MBS for liquidity and modest excess carry. A smaller sleeve of AAA non-QM can then add incremental yield and structural diversification. Finally, a measured EM allocation, focused on dollar-denominated investment-grade infrastructure paper and selective local-currency positions, introduces genuine geographic and policy diversity.
Risk management remains essential. Extension risk in mortgages, liquidity constraints in private deals, and currency volatility in EM all require active monitoring. Position sizes should reflect those characteristics. Yet the alternative, simply accepting higher concentration in a handful of AI-related issuers because the rest of the market looks expensive, carries its own dangers.
I have watched portfolios that leaned too heavily on a single theme before. The correction, when it arrives, tends to be sharper than models predict. Spreading exposure across instruments whose primary risks sit outside the AI capital cycle is one of the cleaner ways to reduce that vulnerability without sacrificing too much income.
How Spreads and Carry Compare Across Alternatives
Numbers help ground the discussion. Agency MBS currently offers a meaningful pickup over investment-grade corporates without the same issuer concentration. AAA non-QM spreads have widened enough to compensate for extension and structural risks that newer deals have partially mitigated. Dollar EM investment-grade infrastructure paper has delivered a consistent forty-basis-point median advantage over vanilla EM bonds this year. Each of those figures is modest in isolation. Together they form a set of options that look more attractive than adding another layer of technology or utility exposure at historically tight levels.
| Asset Type | Primary Risk Driver | Relative to AI Cycle | Liquidity Profile |
| Agency MBS | Prepayment and extension | Low correlation | High |
| AAA Non-QM | Extension and structure | Very low correlation | Moderate |
| EM IG Infrastructure | Country and policy | Low to moderate | Moderate |
| Long IG Corporates | Issuer and sector | High concentration | High |
The table is deliberately simplified. Real portfolios require more granular analysis. Still, the directional message is clear. Moving even a portion of exposure into the first three rows changes the overall risk composition of a credit book in a useful way.
The Role of Financials and Old-Economy Sectors
Some managers still prefer to stay inside corporate credit and simply rotate toward financials or energy. Both sectors offer large pools of investment-grade paper with less direct sensitivity to the AI capital cycle. Energy, in particular, can play a similar diversifying role inside high-yield allocations. The difficulty is valuation. Spreads in both areas generally appear tight relative to their risk-adjusted return potential. Paying up for diversification inside the same broad asset class often leaves less room for error than stepping into securitized or emerging-market alternatives that start with wider cushions.
That does not mean financials or energy should be avoided entirely. They remain useful tools. It does mean that treating them as the primary solution to AI concentration risk may prove incomplete. Combining a core holding in those sectors with the three diversifiers discussed earlier produces a more balanced outcome in most scenarios I have modeled.
Looking Ahead: Volatility and Opportunity
The AI build-out is unlikely to reverse overnight. Capex plans already announced will keep issuance elevated for some time. At the same time, political and regulatory developments can alter the pace. Longer-dated Treasuries remain volatile. Corporate spreads in many sectors leave limited margin of safety. Against that backdrop, the search for genuine diversifiers is not optional. It is necessary portfolio hygiene.
Agency MBS, higher-rated private mortgage bonds, and selected emerging-market debt will not solve every problem. They do, however, offer practical ways to reduce dependence on the same group of issuers that now dominate long-end supply. The carry they provide is not spectacular. The correlation benefit is the more important feature.
I keep returning to a simple observation. For most of the past three years the AI story belonged to equity investors. It is now firmly a bond-market story as well. The numbers have grown too large to ignore. Investors who adjust their credit diversification approach accordingly will likely sleep better when the next wave of volatility arrives. Those who treat the concentration as temporary may find the adjustment more painful later.
The good news is that alternatives exist. They require more work than simply buying the next large technology or utility issue. They also reward the effort with a risk profile that does not move in perfect step with the AI capital cycle. In a market increasingly shaped by that cycle, that independence is worth paying attention to.
Practical Steps for Portfolio Adjustments
Start by measuring the true AI-related exposure already sitting in the long-dated corporate book. Many portfolios underestimate the cumulative weight once hyperscalers, related utilities, and supporting infrastructure issuers are aggregated. Once that number is clear, set a target range that feels comfortable given the rest of the risk budget. The gap between current and target exposure becomes the amount available for redistribution.
Agency MBS can absorb a meaningful portion of that redistribution because of its liquidity. AAA non-QM works better as a smaller, higher-yielding sleeve. Emerging-market allocations should be sized with currency and country limits firmly in mind. Rebalance periodically. The relative attractiveness of these sectors will shift as spreads and issuance patterns evolve.
None of this requires abandoning corporate credit. It simply means refusing to let one powerful theme dominate the risk profile of the entire long-duration book. That discipline has served investors well in previous cycles. There is little reason to believe the current cycle will prove different.
The AI debt boom has already changed the shape of the investment-grade market. Credit diversification strategies that ignore that change risk becoming outdated faster than many expect. The three alternatives outlined here will not eliminate risk. They can, however, change its character in ways that make the overall portfolio more resilient. For investors who still care about sleeping at night, that distinction matters.
Markets rarely reward complacency for long. When nearly half of long-dated non-financial investment-grade supply is tied to a single investment theme, complacency becomes especially expensive. The tools to address that concentration already exist. Using them is simply a matter of recognizing that the bond market story has changed and adjusting accordingly.
