与人工智能相关的借贷正在占据美国长期企业债务中越来越大的份额——这促使投资者寻找大型科技股之外的信用分散化工具。根据太平洋投资管理公司（PIMCO）董事总经理兼多资产策略师 Lotfi Karoui 的说法，投资者可以考虑机构抵押贷款支持证券、高评级私人抵押贷款债券和新兴市场债务，作为在日益受人工智能建设影响的企业债务市场中的信用投资组合分散化工具。
Artificial intelligence-related borrowing is becoming a larger share of long-dated U.S. corporate debt — prompting investors to look beyond Big Tech for credit diversification. Investors could consider agency mortgage-backed securities, higher-rated private mortgage bonds and emerging market debt as credit portfolio diversifiers in a corporate debt market increasingly shaped by the AI build-out, according to Lotfi Karoui, managing director and multi-asset strategist at PIMCO.
Karoui 表示，与人工智能相关的债券发行现在对更广泛的债务市场影响沉重。PIMCO 在 11 月 10 日的一份报告中称，此类借款人现在占截至今年的非金融类美国投资级债务总发行量的近一半，这些债务的期限超过 15 年。根据高盛的预测，今年人工智能超大规模企业的美元计价债券发行量预计将达到约 2500 亿美元。
Karoui said AI-related bond issuance now weighs heavily on the broader debt market. Such borrowers now account for almost half of year-to-date gross issuance of non-financial U.S. investment-grade debt with maturities beyond 15 years, PIMCO said in a note Nov. 10. AI hyperscaler debt issuance is projected at around $250 billion in dollar-denominated bonds this year, according to Goldman Sachs forecasts.
今年更广泛的人工智能生态系统债务发行量（包括芯片制造商、公用事业公司和数据中心开发商）的预估范围在 3000 亿至 5700 亿美元之间。人们越来越担心，民主党在美国中期选举中大获全胜可能导致人工智能建设放缓，这可能给寻求分散化的长久期投资者带来麻烦，特别是因为长期美国国债——传统的久期投资组合对冲工具——在 2026 年大半时间里都因波动而动荡不安。
Estimates for the wider AI ecosystem debt issuance this year — which includes chipmakers, utilities and data center developers — range between $300 billion and $570 billion. There are growing concerns about an AI build-out slowdown prompted by the Democrats' sweeping victory in the U.S. midterm elections, which could pose a headache for long-duration investors looking for diversification, particularly as longer-dated U.S. Treasurys — the traditional duration portfolio hedge — have been rattled by volatility for much of 2026.
“过去 3 年中的大部分时间里，人工智能投资热潮一直是一个股票市场的故事，但它正日益成为一个债券市场的故事，”Carmignac 投资委员会成员 Kevin Thozet 表示。“这些数字变得太大，债券投资者无法忽视。”Karoui 表示，在企业信用领域，金融业提供了大量投资级敞口，且对人工智能资本支出周期的直接敏感度低得多，而“旧经济”部门（如能源）在高收益（HY）市场中可以发挥类似作用。“然而，我们认为两者都很贵，”Karoui 说。
"For most of the past 3 years, the AI investment boom has been an equity story, but it is increasingly becoming a bond-market story," said Kevin Thozet, a member of Carmignac's investment committee. "The numbers are becoming too large for bonds investors to ignore." Karoui said that within corporate credit, financials offer a large pool of IG exposure with much less direct sensitivity to the AI capex cycle, while "old economy" sectors such as energy can play a similar role in high-yield (HY) markets. "Both, however, are expensive, in our view," Karoui said.
“它们相对于同期限国债的利差总体看起来偏紧，相对于其风险调整后的回报潜力而言。” “相反，我们看到其他机会，对于对人工智能相关投资主题持谨慎态度的投资者来说，这些机会具有更有利的分散化特征。” 机构抵押贷款支持证券 由房利美和房地美等政府资助企业担保的美国住房贷款支持的机构抵押贷款支持证券，今年对较高利率和利率波动的抵御能力比2023年更强，Karoui表示。
"Their spreads above like-maturity Treasuries generally appear tight relative to their risk-adjusted return potential." "Instead, we see other opportunities with more favorable diversification characteristics for investors wary of AI-related investment themes." Agency mortgage-backed securities Mortgage-backed securities backed by U.S. home loans guaranteed by government-sponsored enterprises like Freddie Mac and Fannie Mae have proved more resilient to higher rates and rate volatility this year compared with 2023, Karoui said.
该报告引用了PIMCO和彭博的数据，显示截至11月5日，机构MBS相对于美国投资级公司债利差有33个基点的额外收益，但没有同样的企业借款人风险敞口。 “在美国，优质证券化产品如机构抵押贷款支持证券和非机构住宅MBS，可能在提供与人工智能主题有意义的非相关风险敞口的同时，仍能提供相对于更广泛投资级市场体面的超额收益，”Karoui说。
The note referenced PIMCO and Bloomberg data highlighting a 33-basis-point pickup in agency MBS over U.S. investment-grade corporate spreads as of Nov. 5, but without the same exposure to corporate borrowers. "In the U.S., 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," Karoui said.
评级较高的私人抵押贷款债券 PIMCO还指出了由不符合联邦偿债能力规则下“合格住房贷款”类别的抵押贷款支持的债券中，高级的、AAA评级部分。 在这里，Karoui表示，在AAA评级的非合格住房贷款分层中，“潜在收益可能超过风险”。 具体而言，风险更多来自偿还时间比预期长，而不是借款人违约的风险。 这种风险推高了回报，而某些较新的交易结构旨在更快偿还高级投资者本金，并提供防范损失的保护。 “我们认为AAA非QM的重新定价创造了一个引人注目的机会，”Karoui说。
Higher-rated private mortgage bonds PIMCO also pointed to senior, AAA-rated portions of bonds backed by mortgages that do not qualify for the Qualified Mortgage category under federal ability-to-repay rules. Here, Karoui said that in AAA-rated non-qualified mortgage tranches, the "potential benefits likely outweigh risks." Specifically, the risk stems more from repayments taking longer than expected, rather than the risk that borrowers will fail to pay. That risk has pushed returns higher, while certain newer deals are structured to repay senior investors more quickly and offer protection against losses. "We think the repricing in AAA non-QM creates an intriguing opportunity," Karoui said.
“更宽的利差在很大程度上补偿了投资者面临的延期风险和结构性风险，这些风险与预期信用损失大幅增加有关，而今天的新发债券在这些风险的缓释措施方面比早期交易更好。”新兴市场债券由新兴市场政府和公司发行，为美国大型科技股提供了另一种分散投资的选择。“尽管新兴市场中有一些板块对人工智能有显著敞口，特别是韩国和台湾，但那主要是一个权益类资产的故事，而不是固定收益类资产的故事，”Karoui说。
"Wider spreads largely compensate investors for extension risk and structural risk tied to materially higher expected credit losses, and today's new issues embed better mitigations against those risks than earlier deals." Emerging market debt Bonds issued by emerging market governments and companies offer another diversification play to U.S. Big Tech. "Although there are pockets of EM with meaningful AI exposure, notably South Korea and Taiwan, much of that is an equity story rather than a fixed income one," said Karoui.
“对于债券投资组合而言，新兴市场债务因此仍是收益率、久期和国家特定风险的一个相对差异化来源。”他聚焦于美元计价的新兴市场投资级“半流动性”债券交易——用于为基础设施投资融资——并指出，与普通新兴市场债券相比，这些交易今年提供了约40个基点的中位数额外收益。他还指出了主要基准之外的机会，包括大宗商品出口国，同时指出在本币债券中，汇率波动可能增加收益，也可能侵蚀收益。
"For bond portfolios, EM debt therefore remains a relatively differentiated source of yield, duration, and country-specific risk." He zeroed in on U.S. dollar-denominated emerging market investment-grade "semi-liquid" bond deals — used to finance infrastructure investment — noting they have offered a median pickup of around 40 basis points this year compared to plain vanilla EM bonds. He also flagged opportunities beyond the main benchmarks, including in commodity-exporting countries, while noting that in local-currency bonds, exchange-rate swings can add to returns or erode them.
“明确地说，并非新兴市场是一个低风险资产类别。相反，新兴市场债务面临的是政治、财政和货币力量的组合，这些力量不同于美国国内市场面临的风险，”Karoui说。
"To be clear, it is not that EM is a lower-risk asset class. Rather, it is that EM debt is exposed to a mix of political, fiscal, and monetary forces that are different than the risks facing the domestic U.S. market," Karoui said.