Via alternativecreditinvestor.com
The firm's CIO draws parallels to the late-1990s telecom bust as AI-related debt issuance is projected to hit $570 billion in 2026
Trey Parker, chief investment officer of Sycamore Tree Capital Partners, went on Bloomberg Television to deliver a message that most of Wall Street would rather not hear: the AI infrastructure boom carries the kind of credit risk that has a historical habit of ending badly.
Parker’s core argument is straightforward. Building out AI requires trillions of dollars in capital, and the debt markets absorbing that demand are showing stress fractures that experienced credit investors have seen before. Specifically, he pointed to what he calls “rating-designation risk” in private credit markets, where the sheer appetite from insurance companies for AI data-center debt could force credit rating revisions rather than reflect actual creditworthiness.
The telecom echo #
Parker drew an explicit comparison to the telecom infrastructure boom of the late 1990s. That era saw massive debt-fueled buildouts of fiber optic networks and switching infrastructure, underwritten by rosy demand projections that never materialized at the scale investors expected. Companies like WorldCom and Global Crossing became cautionary tales, their collapses wiping out billions in bondholder value.
The scale of the current buildout makes the comparison worth taking seriously. AI-related debt issuance is projected to reach $570 billion in 2026 alone. That is not a rounding error. It represents one of the largest single-sector debt waves in recent memory, and it’s being absorbed by a market that may not be pricing risk accurately.
The insurance company problem #
One of Parker’s more nuanced points involves the role of insurance capital in this equation. Insurance companies, perpetually hungry for yield and duration-matched assets, have become major buyers of private credit tied to data center construction and operations. Parker warns this creates a feedback loop where competitive pressure can compress spreads and relax underwriting standards. More critically, it can distort credit ratings. If a rating agency knows that downgrading a popular asset class will alienate its largest customers, the incentive to maintain favorable ratings grows. Parker’s concern about “rating-designation risk” is essentially a warning that the grades on these bonds might not reflect reality.
Sycamore Tree Capital Partners, the Dallas-based firm Parker co-founded alongside Mark Okada and Jack Yang in late 2020, manages over $3 billion in assets. The firm launched a credit secondaries platform in April 2026, led by partner Robert O’Connor, specifically targeting distressed and opportunistic investments.
Shorter duration, stronger tenants #
Parker’s prescription for investors who still want exposure to AI data center debt is calibrated rather than apocalyptic. He recommends shorter maturity durations, which reduce the window during which things can go wrong and give investors more frequent opportunities to reassess.
He also advocates prioritizing tenants with strong credit ratings. A facility leased long-term to a hyperscaler like Microsoft or Google carries fundamentally different risk than one dependent on a startup AI company burning through venture capital. Together, these two filters would screen out a significant portion of the riskier issuance flooding the market.
That $570 billion in projected AI debt issuance for 2026 includes everything from investment-grade bonds backing hyperscaler-leased facilities to more speculative private credit deals financing speculative builds in secondary markets.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our