PIMCO's Richard Clarida is putting a $14 trillion frame around the AI investment boom, while the firm's credit team is warning that weaker borrowers are already entering a tougher loss cycle.
The signal coming out of PIMCO this week is worth paying close attention to, because it is two signals at once. Richard Clarida, the firm's Global Economic Advisor and former Federal Reserve Vice Chair, told Bloomberg that artificial intelligence investment has crossed what he called a macroeconomic threshold. The buildout of AI infrastructure, combined with surging defense and energy security spending, could add roughly $14 trillion to global capital expenditure over the next five years. That is not a projection from a startup pitch deck. It is the framework that one of the world's largest fixed-income managers is using as it positions more than $2 trillion in assets.
In the same week, PIMCO's credit team issued a sharper warning. CIO Daniel Ivascyn and colleagues, in analysis published June 10 and reported by Bloomberg, wrote that the default cycle is reasserting itself, with significantly higher losses expected in lower-quality credit, especially leveraged loans and private direct lending. AI spending is widening the gap between companies and capital structures, putting borrowers that cannot service debt taken on during years of cheap money in a more precarious position.
These are not contradictory views. They are the same observation from two different angles.
Clarida's disinflationary argument rests on productivity. If the capital flowing into data centers, power infrastructure, and processing capacity delivers on its promise, efficiency gains could compress costs across the economy. The payoff, he argues, may arrive faster than markets currently expect and prove more powerful as a deflationary force. For the bond market, that calculation matters enormously: a disinflationary AI economy is one where high-quality fixed income has a clearer path to outperform.
Ivascyn's credit warning addresses what happens to the balance sheets absorbing that investment on the way to the productivity payoff. PIMCO's recent analysis noted that capital expenditure tied to AI is surging while free cash flow is moving in the opposite direction. AI-related debt issuance has been running at roughly $100 billion per quarter. At those volumes, the gap between companies that can sustain the load and those that cannot becomes measurable very quickly. The firm also flagged late-cycle stress in direct lending, including rising shadow default rates and increased reliance on payment-in-kind structures, where borrowers pay interest with more debt rather than cash. That shift usually tells lenders that cash generation is no longer keeping up.
Smaller and midsize companies, the primary borrowers in the direct lending market, are doubly exposed. They face rising energy costs and tariff pressures at the same time that AI is disrupting the industries they operate in. PIMCO's analysis is direct on this point: even in a strong economy, AI can pressure old economy companies, and the most leveraged ones face the worst version of that adjustment.
For founders navigating late-stage fundraising and private credit markets right now, this framework maps the terrain clearly. The capital available is not going away, but it is sorting itself. Borrowers with real revenue, defensible margins, and assets a lender can actually value will still have access to credit at workable terms. The cohort that expanded aggressively during the 2021-2023 private credit boom is facing a reckoning that PIMCO is no longer treating as theoretical.
The firm's own positioning reflects that view. PIMCO is rotating toward high-quality investment-grade bonds and pulling back from leveraged and private direct lending exposure. A posture like that from a manager with more than $2 trillion under management does not just describe the market, it can influence it. When a buyer of that size becomes more selective, marginal borrowers feel it first.
The larger question Clarida's framework raises is whether the productivity dividend arrives before the credit strain reaches a breaking point. The late 1990s internet buildout is the obvious historical parallel: disinflationary in theory, brutal for overextended companies in practice, and ultimately resolved through a market washout that erased many leveraged players while the underlying infrastructure kept running. PIMCO appears to be betting on a similar distinction. Own the durable infrastructure and the higher-quality cash flows, but stay cautious on the indebted operators trying to finance their way through the transition. For anyone watching where the smart fixed-income money is positioned right now, that is the clearest forward signal available.
Also read: Adobe's record AI results could not outrun its growing leadership vacuum • MediaTek doubles its AI chip target to $2 billion and enters the data center • CME Group moves gold and oil futures to 24/7 trading as crypto's always-on model wins
