‘Things Are Going to Get Crazier’: AI Debt Deluge Recalls 2007 Risks (Podcast)

Seaport Research Partners warns that “massive borrowing” by AI-focused companies—some reportedly used to fund customers—could inflate a technology-sector bubble. Analyst Jay Goldberg says the debt “is coming on stream,” which may intensify froth in AI-related credit and equity valuations. The note is a risk-off takeaway for investors given potential leverage-driven volatility, though no specific financial figures were cited.

Analysis

The market is still pricing AI capex as an equity story, but once the financing layer becomes part of the demand engine, the trade shifts from “more growth” to “who is taking the credit risk.” That generally helps the nearest counterparties first — lenders, leasing desks, and equipment vendors with pricing power — but it quietly weakens the rest of the stack by turning future demand into today’s receivables and leverage.

The first stress signal should show up in credit, not in headline AI revenue: tech-adjacent high yield, converts, and private-credit marks can reprice within 1-3 months even if equity multiples lag the warning. Over 6-18 months, the second-order risk is that circular financing creates a capex overhang, forcing order pushouts and multiple compression in the more levered AI ecosystem while cash-rich hyperscalers hold up better. Consensus is missing that the bubble risk is less about AI adoption failing and more about capital intensity outrunning cash conversion.

What would falsify the bearish read is evidence that these borrowers can term out debt at stable spreads while utilization and incremental FCF inflect quickly enough to cover financing costs. If that happens, the leverage is just an acceleration mechanism; if it doesn’t, the unwind will start in spreads, then in guidance revisions, then in equity valuation.

Key Decisions for Investors

• Long QQQ / short ARKK for 1-3 months as a quality-vs-speculation pair; if AI financing stress broadens, cash-rich megacaps should outperform levered growth by 5-10% in relative terms.
• Buy 3-6 month puts on SMH or SOXX on any strength; the trade is a hedge against order pushouts and capex repricing, and should be cut if semiconductor bookings re-accelerate for two consecutive prints.
• Short HYG or buy CDX HY protection as a credit-expression hedge; the cleanest catalyst is widening spreads in tech/communications names over the next quarter.
• Set a watchlist on private-credit/BDC exposure to AI infrastructure and vendor-financing structures; no aggressive position until disclosures show concentration, but any reserve build or covenant issue would be an immediate short catalyst.
• If FCD.UN.TO is being used as a leveraged AI/credit wrapper, treat rallies as opportunities to reduce exposure rather than chase beta until the financing channel proves self-funding.