Hot or not? The economy’s fate rests on Kevin Warsh’s answer to one key question

Markets have priced a 25bp Federal Reserve rate hike with near certainty as the 10-year Treasury yield exceeds 5%, oil hovers near $110 per barrel, inflation remains above 3%, and nominal GDP growth accelerated above 6.5% in Q2. The policy outlook hinges on whether inflation reflects overheating demand—supported by a roughly 6% of GDP federal deficit, AI investment and prior 175bp of rate cuts—or temporary tariff and supply shocks that tighter policy cannot address. AI and data-center financing has reached $308B in U.S.-dollar debt issuance through July, while U.S. national debt has crossed $40T, increasing pressure on long-end yields and complicating Fed-Treasury policy coordination.

Analysis

The investable issue is not a 25bp move; it is whether policy coordination suppresses term premium long enough to preserve risk-asset multiples while fiscal financing and private AI capex compete for the same marginal dollar. A hike paired with continued long-end Treasury support would mechanically favor curve flattening, but it would also make the market increasingly vulnerable to a disorderly bear-steepener if investors lose confidence in the durability of that support. That regime is negative for long-duration equities and leveraged real-estate/utilities, even if headline equity indices initially absorb the decision.

AI infrastructure is the clearest second-order pressure point. Hyperscalers can fund incremental capex despite a modestly higher policy rate, but their suppliers, data-center developers, independent power producers and smaller infrastructure borrowers face materially higher refinancing costs. Credit issuance tied to this buildout may therefore remain robust while spreads stay deceptively contained; the eventual stress is more likely to appear in lower-quality borrowers and project-finance structures over 6-18 months than in mega-cap tech earnings over the next quarter.

GS is directionally supported by elevated issuance, trading activity and rate volatility, but its upside is conditional on orderly markets rather than a rates shock that impairs underwriting and risk appetite. The contrarian read is that a one-off hike could initially be bullish for long duration if it restores inflation-fighting credibility and reduces the probability of a larger tightening cycle; the more damaging outcome is a dovish hike or no credible fiscal/term-premium response, which would validate inflation-risk pricing rather than contain it. Falsify the bearish-duration thesis if the 10-year yield retreats and holds below 4.75% while inflation expectations and high-yield spreads both decline.

Key Decisions for Investors

• Use the policy decision as a curve-volatility entry point: initiate a modest 2s10s flattener only if the Fed hikes and communicates a data-dependent, non-committal path. Target a 15-25bp flattening over 1-3 months; exit if 10-year yields close above 5.25%, signaling fiscal/term-premium stress is overwhelming long-end support.
• Maintain a 3-6 month long XLE / short XLRE pair. Higher energy prices and persistent nominal growth support energy cash flows, while commercial real estate remains the most direct listed-equity casualty of elevated long-end yields. Reassess if oil falls below $90 or the 10-year yield sustains below 4.75%.
• Avoid adding broad duration-sensitive utility and data-center infrastructure exposure through XLU or leveraged REITs until refinancing schedules and contracted power-price pass-through are clear. The AI-demand narrative does not eliminate the valuation and funding-cost risk; treat any sharp post-Fed rally as an opportunity to reduce exposure.
• Keep GS as a tactical, not structural, long: buy only after evidence that investment-grade and leveraged-finance calendars remain open following the meeting. A 1-3 month trade works if volatility lifts trading revenue without a meaningful widening in credit spreads; abandon if high-yield spreads widen by more than 75bp or underwriting activity slows.