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

Markets assign a 92.7% probability to a Federal Reserve rate hike at the Sept. 16, 2026 FOMC meeting, the first increase since 2023. Historically, the S&P 500 falls about 2% in the first three months of a new hiking cycle but gains 9% on average over the following 12 months; slow-tightening cycles have delivered a 10.5% one-year gain versus a 3.6% decline under rapid tightening. Persistent inflation tied to the Iran war and tariffs, elevated Shiller CAPE valuations, and rate-sensitive AI infrastructure spending could increase the risk of a sharper equity sell-off if the Fed tightens aggressively.

Analysis

The relevant market variable is not the first hike but the terminal-rate repricing embedded in the front-end curve. A well-telegraphed 25 bp move should be largely discounted; the downside begins if the post-meeting dots, inflation language, or 2-year yield imply a cumulative tightening path materially above current forwards. In that scenario, long-duration equity factors—particularly unprofitable software and AI infrastructure beneficiaries priced on 2028-30 cash flows—should underperform the broad index over the following 1-3 months.

NVDA’s exposure is more nuanced than the headline rate sensitivity suggests. Hyperscaler AI capex is funded largely from operating cash flow, so a single hike is unlikely to alter near-term accelerator orders; the more immediate transmission mechanism is multiple compression and a higher hurdle rate for leveraged AI neoclouds, power developers, and second-tier server vendors. A sustained restrictive cycle would matter over 6-18 months by raising financing costs for incremental data-center capacity, potentially slowing the ecosystem’s demand pull-through before it affects NVDA’s direct revenue.

CME is a cleaner volatility/uncertainty beneficiary than directional financials: rate, Treasury, and equity-index hedging volumes can rise even if cash equities de-rate. GS should benefit initially from FICC activity, but its higher-beta capital-markets and advisory revenue are vulnerable if the policy path induces a growth scare. SCHW faces a different risk: renewed deposit competition and higher cash sorting can delay net-interest-income recovery; its upside requires asset yields to reset faster than deposit costs, not simply higher policy rates.

Consensus appears too anchored to historical average equity returns and insufficiently attentive to valuation concentration. If real yields rise while earnings revisions remain flat, the market can de-rate without a recession, with index-level resilience masking a sharper unwind in expensive mega-cap growth. Conversely, a dovish hike accompanied by stable long-end yields would force a rapid cover of rate-sensitive shorts, making confirmation from the 2-year yield and forward earnings revisions essential.

Key Decisions for Investors

• Use the FOMC as a conditional trade rather than pre-positioning directionally: if the 2-year Treasury yield closes 15-20 bp above its pre-meeting level, initiate a 1-3 month long CME / short SCHW pair. CME captures hedging-volume upside, while SCHW is exposed to deposit-beta and cash-sorting pressure; exit if the 2-year yield retraces below the pre-meeting range.
• Hedge concentrated AI-beta exposure over the next 1-3 months via NVDA put spreads or a long SMH / short IGV-style relative-value position only if the Fed signals multiple additional hikes. The thesis is valuation-duration dispersion, not an immediate NVDA demand collapse; invalidate on unchanged hyperscaler capex guidance and falling real yields.
• Favor GS over SCHW within financials for the first 1-2 quarters after a hawkish repricing, sized modestly. GS has greater trading-volatility sensitivity, whereas SCHW needs observable improvement in deposit flows and net interest margin; reverse the relative trade if credit spreads widen materially, signaling that volatility is becoming risk-off rather than transactional.
• Do not add broad S&P 500 beta solely on an expected initial pullback. Add only after either forward EPS estimates stabilize and the 10-year real yield stops rising, or after a 5-7% index correction; absent those conditions, index downside is dominated by multiple compression rather than an attractive earnings-reset entry.