Two asset managers are positioning their portfolios to withstand a potential credit shock related to the wave of corporate spending on artificial intelligence, according to reporting that cited comments from their portfolio managers.
At the Bloomberg Global Credit Forum, Robert Cohen, a portfolio manager at DoubleLine Capital LP, warned that markets could become overheated in the months or years ahead as technology companies commit large sums to AI. Cohen said investors should favor credits that can endure stress either through conservative deal structures or because the borrowers maintain robust balance sheets.
Cohen highlighted particular challenges tied to AI-related lending. Many securities being put up for sale now will not mature for decades, which raises the prospect that the technologies underpinning those assets could be obsolete long before maturity. He also pointed to an elevated risk of overbuilding in data-center markets, noting that those facilities can take a long time to construct and that many projects are currently moving forward simultaneously.
The note cited a Barclays report dated May 21 that found large U.S. technology companies - often called hyperscalers - have sold more than $155 billion of unsecured bonds globally, an increase of more than 45% compared with their total issuance the prior year.
Christina Lee, co-portfolio manager in private credit at Oaktree Capital Management, said the market for data-center financing is still in its early stages. Oaktree is being selective, she said, because it is not yet clear which projects or operators will prevail.
Bloomberg Intelligence has estimated that companies will invest roughly $5 trillion in capital expenditures for AI over the next five years, and that much of that investment will be financed with debt.
Implications and market focus
- Credit managers are prioritizing structure and balance-sheet strength when buying AI-linked debt.
- Data-center financing and unsecured bond issuance by hyperscalers are central to the changing credit picture.
- Large-scale AI capital expenditure plans are expected to be largely debt-funded.
Outlook
Portfolio managers cited in the reporting are preparing for a market environment that could grow frothy as AI investment accelerates. Their approach reflects caution around long-duration securities and concentrated build cycles in the data-center sector.