Q3 2026 bank results from JPMorgan Chase, Citigroup, Wells Fargo, and Goldman Sachs (reported October 13) paint a picture of a still-spending U.S. consumer who is drawing down savings, while banks prudently build reserves. This pairs with tight overall credit spreads that are masking stress on the large debt issuers funding the AI infrastructure buildout. Tomorrow’s reports from Bank of America and Morgan Stanley will provide further read-through.[]()
Consumer Spending Holds, but the Foundation Looks Thin
Card and consumer data remain solid. Citigroup reported U.S. card spending up roughly 11% year-over-year (general-purpose cards +13%, with average loans +9% in that portfolio). JPMorgan saw debit and credit card sales volume rise 10% and average credit card loans up 7%. Charge-off rates improved at the major banks (JPM card charge-offs to 2.91% from 3.15%; Citi to ~3.6% from 4.0%; Wells similarly better). Loan growth and spending categories (including travel and gas) continue to show momentum.
The counterweight is household balance sheets. The personal saving rate fell to 4.1% in August 2026 (from a revised 4.6% in July), among the lower readings of recent years as spending outpaced income growth. University of Michigan consumer sentiment has dropped to 46.3 in the preliminary October reading—near historical lows and down from September—with current conditions plunging amid high prices, elevated borrowing costs, and inflation expectations ticking higher (year-ahead ~4.7%). Conference Board measures have also been soft. Higher-income and wealth-effect households are supporting aggregate spending; lower- and middle-income groups look more constrained.
Banks are responding conservatively. Citi and Wells increased total provisions year-over-year; JPM and Citi both recorded net reserve builds (Citi $2.6 billion provision including a $369 million firmwide net reserve build; cards cost of credit elevated). Management commentary (e.g., Jamie Dimon) has flagged risks from higher rates on refinancing and potential spread widening, even as current credit performance remains orderly and NCL guidance is at the low end of prior ranges. This is preparation for possible stress rather than evidence of current deterioration.
Credit Spreads and the AI Debt Wall
Broad investment-grade credit spreads remain relatively contained (in the low-to-mid 80s basis points area in early October data), still tight by historical standards. High-yield and lower-rated spreads have been more mixed, with some widening at the distressed end. The more notable pressure is concentrated in AI-linked issuance. Hyperscalers, chipmakers, data-center players, Oracle, Broadcom, SpaceX and others have issued or lined up hundreds of billions (approaching or exceeding half a trillion in recent tallies, with much more expected) to fund compute and infrastructure. CDS and cash spreads on several of these names have widened relative to the broader IG market; tech credit has underperformed as supply hits investor balance sheets and free-cash-flow/return timelines are scrutinized. Broader IG has largely absorbed the supply so far, but the “parts of the market most in need of future financing” (exactly as you noted) are feeling the strain first.
This creates a classic late-cycle tension: consumer spending and bank earnings still look “America resilient” (echoing Buffett’s long-standing maxim), yet low savings, depressed confidence, rising bank provisions, and a massive, front-loaded AI debt cycle raise the odds of slower growth, higher delinquencies, or credit re-pricing if rates stay elevated, inflation (including energy) persists, or AI returns disappoint.
How to Play It
Positioning should emphasize quality and selectivity rather than a binary bet on the consumer or a full AI unwind:
- **Banks and financials**: Prefer well-capitalized large banks (JPM, BAC, etc.) that are already provisioning and generating NII/fee income. They benefit from higher rates in the near term and have strong capital buffers. Watch tomorrow’s BAC and MS prints for confirmation on NII, provisions, and investment-banking/trading. Quality regional or mid-tier names can be secondary if valuations are attractive and credit quality holds.
- **Consumer exposure**: Favor higher-income-skewed or defensive consumer names and avoid pure leverage-sensitive discretionary. Aggregate spending can stay positive longer than confidence suggests, but the savings drawdown and sentiment readings increase the risk of a sequential slowdown.
- **AI and tech credit/equity**: Equity side—prefer cash-generative enablers (semiconductors with strong balance sheets, select hyperscalers) over pure debt-funded infrastructure pure-plays. Credit side—the relative widening in AI-linked spreads creates selective opportunity for those who believe in multi-year compute demand and can underwrite the specific issuers’ free cash flow and asset flexibility. Broader tech credit remains vulnerable to further supply and any equity-market air pocket. Avoid chasing the tightest overall spreads.
- **Portfolio hedges and structure**: Maintain a core long bias to U.S. equities consistent with “never bet against America,” but raise quality tilt (balance-sheet strength, free cash flow). Consider modest credit hedges (selective CDS protection or underweight high-yield/AI tech credit), some duration if growth risks rise, or equity puts on consumer or rate-sensitive sectors. Monitor unemployment, NCL trends, and AI issuance cadence closely—these are the leading indicators.
The prints so far support another quarter of resilience, but the combination of thin household buffers, bank reserve builds, and concentrated AI financing needs is the setup for greater dispersion and potential volatility ahead. Tomorrow’s results and any further management commentary on loan demand, deposits, and provisions will refine the picture.
Comments