The midterms won’t ease borrowing costs. Here’s what CFOs should do now

EY-Parthenon finance leaders expect sustained higher borrowing costs into 2027, with deficits, capital demand, inflation volatility and policy uncertainty pressuring long-term rates; political gridlock could also tighten credit and raise short-term funding costs. AI and data-center projects face added political scrutiny, with restrictions passed in at least 14 Ohio towns; higher financing costs could increase project expenses and weigh on outstanding AI bond prices as yields rise. CFOs are advised to stress-test investment returns and financing costs, diversify funding, and account for power and infrastructure costs.

Analysis

The investable signal is dispersion, not a blanket “higher rates” call. If fiscal uncertainty lifts term premia or short-term funding spreads, the first-order penalty falls on leveraged, long-duration projects; the second-order effect is tighter underwriting and delayed orders for data-center builders, power-intensive compute, and their contractors. Local opposition to electricity costs makes permitting and power procurement a project-level risk, potentially stranding development spend before it reaches revenue. Conversely, scarce grid capacity can strengthen the position of existing, well-connected power assets—but regulated utilities are not automatic winners if affordability pressure limits cost recovery.

Near term, this is more a catalyst watch than a fresh macro trade: the article supplies no market pricing, refinancing schedule, or project-level economics to establish mispricing. Over 1–3 months, monitor budget/debt-ceiling friction, bank credit standards, short-term funding spreads, and local permitting decisions. Over 6–18 months, higher hurdle rates should favor projects with contracted power, credible utilization, and demonstrable productivity returns; speculative capacity and marginal AI capex face deferrals or repricing. Private-credit lenders may command wider spreads, but higher coupons do not compensate if projects stall or collateral value depends on a single use case.

Contrarian angle: political scrutiny could slow marginal builds while making already-permitted sites and firm power access more valuable. That scarcity benefit is conditional, not yet enough to justify a broad data-center or utility long. The key falsifiers are easing credit spreads and borrowing costs, stable permitting and power costs, or evidence that AI deployment produces measurable cash-flow gains despite financing pressure.

Key Decisions for Investors

• Prefer a relative-quality screen over a broad rate bet: favor firms with low refinancing needs, committed power, and contracted demand over highly leveraged, speculative data-center expansion. Do not size this as a short without checking valuation, debt maturities, and project-level power agreements.
• Treat data-center private-credit exposure as an alert, not a buy-the-yield opportunity. Before adding, verify loan covenants, collateral concentration, borrower liquidity, and whether repayment depends on uncontracted capacity; widen-spread compensation can be overwhelmed by delay or impairment risk.
• Monitor U.S. commercial-paper and bank-lending conditions alongside Treasury term premium. If short-term spreads widen or bank standards tighten, favor issuers with committed multi-bank facilities and longer-dated funding; reverse the defensive tilt if spreads normalize and refinancing access remains broad.
• Watch grid-connected, permitted power assets for relative scarcity value, but avoid assuming utilities can pass through all costs. Reassess on local rate cases, power-price trends, and evidence of cost recovery; weaker affordability outcomes would undermine the beneficiary thesis.