The Fed Could Be About to Raise Rates for the First Time Since 2023. Here's What History Says Happens to Stocks Next.

CME FedWatch priced a 92.7% probability that the Fed will raise rates at its September 16, 2026 meeting, ending a three-year period without hikes. Historically, the S&P 500 declines about 2% in the first three months of a new hiking cycle but delivers an average 9% return over the following 12 months; slow-tightening cycles have produced average 12-month gains of 10.5%, versus a 3.6% decline during rapid tightening. Persistent inflation, Iran-war-related pressures, tariffs, elevated S&P 500 valuations, and rate-sensitive AI infrastructure spending could force a more aggressive cycle and increase downside risk.

Analysis

The market is unlikely to trade the first hike itself; it will trade the implied terminal rate and the dispersion between policy-sensitive equities and nominal-growth beneficiaries. With broad-index valuation support thin, a 25 bp move accompanied by a higher-for-longer dots path could produce a disproportionate de-rating in long-duration software, unprofitable AI infrastructure suppliers, and highly levered small caps. The key 1-3 month transmission channel is real yields: a sustained rise in the 10-year Treasury real yield above its pre-meeting range would matter more for NVDA’s multiple than the policy rate headline.

AI capex is the non-obvious vulnerability. Hyperscaler spending plans are funded largely from strong operating cash flow, so an initial hike should not immediately curtail orders; however, higher discount rates raise the hurdle rate for marginal data-center projects and can shift spend from accelerator purchases toward utilization optimization, power efficiency, and networking. That creates a 6-18 month risk of order-duration compression across the AI supply chain even if headline capex remains elevated.

Financials are not a clean directional long. GS benefits from volatility, financing activity and potentially stronger trading revenues in the days around policy repricing, but a faster tightening path can choke off issuance and M&A over subsequent quarters. SCHW’s upside depends on deposit beta and asset-gathering stability; a renewed money-market fund migration would offset any benefit from higher client cash yields, making weekly deposit data and net interest revenue guidance decisive.

Consensus appears overly anchored to historical average equity returns after first hikes. Those episodes are not directly comparable if inflation persistence prevents the Fed from validating an easing path, while elevated starting multiples leave less room for earnings to absorb discount-rate pressure. A benign outcome requires inflation expectations to remain anchored and the forward curve to price a shallow cycle within the next two meetings.

Key Decisions for Investors

• Maintain a tactical 1-3 month long XLF / short IWM pair rather than a broad market short: regional/small-cap balance-sheet and refinancing sensitivity should widen if real yields rise. Exit if the 2-year Treasury yield falls below its pre-FOMC level for two consecutive sessions or if the Fed explicitly signals a one-and-done hike.
• Buy 2-3 month QQQ put spreads, financed selectively with XLF call spreads, only if post-meeting 10-year real yields break higher versus the prior month’s high. Target 2:1 minimum payoff; this expresses multiple compression without requiring a recession. Do not initiate if yields fall despite a hike.
• Use NVDA as an AI-duration hedge: reduce incremental exposure into the meeting and consider a 3-month NVDA/SMH downside hedge if hyperscaler guidance begins to emphasize return-on-investment or power constraints. Falsifier: unchanged or raised aggregate capex commitments from MSFT, AMZN, GOOGL and META alongside stable real yields.
• Tactically overweight CME over SCHW for the immediate policy-volatility window: derivatives volumes and collateral activity are more direct beneficiaries of rate uncertainty, with less deposit-flight risk. Reassess after the next CPI release; a downside inflation surprise would rapidly compress the volatility premium.
• Treat GS as an event-driven watch, not a rate-cycle long: initiate only if credit spreads remain contained and equity issuance backlog improves. A widening in high-yield spreads or weaker investment-banking backlog commentary would negate the 6-12 month operating-leverage thesis.