Stock futures are little changed as traders eye rising oil prices and rates: Live updates

U.S. equity futures were nearly flat ahead of Wednesday's Fed decision after the S&P 500 fell 0.5%, the Nasdaq declined 0.6%, and the Dow lost 0.3% amid an AI-stock selloff and rising yields. The 10-year Treasury yield briefly exceeded 5% for the first time since October 2023, while fed-funds futures imply a 92% probability of a 25bp hike, taking the policy-rate upper bound to 4.0%. Nvidia fell 3%, Corning dropped 13%, and the AI-focused BAI ETF lost nearly 4%; meanwhile, Saudi pipeline disruptions pushed Brent above $105 per barrel and WTI above $101.

Analysis

The AI drawdown is more consequential for second-derivative infrastructure than for platform incumbents if customers merely defer, rather than cancel, model-training spend. GLW has materially higher operating leverage to optical-connectivity volume and lower valuation support than NVDA; any hyperscaler capex pause can drive a disproportionate estimate reset in its Optical Communications segment. Conversely, NVDA's near-term revenue is protected by backlog and supply allocation, but its multiple remains vulnerable if the market shifts from a compute-scarcity narrative to a slower utilization/ROI narrative; that is a 6-18 month valuation risk rather than an immediate revenue cliff.

The policy decision is largely priced at the headline level, making the statement, projections and press-conference reaction the tradable event. Elevated energy costs raise the odds that policymakers retain a restrictive bias even after a move characterized as discretionary; higher real yields would pressure long-duration AI equities and levered communications suppliers simultaneously. In the next 1-3 months, the key falsifier for the AI-caution thesis is hyperscaler capex guidance holding or increasing alongside sustained NVDA data-center order visibility; absent that, consensus 2026 infrastructure assumptions likely still have downside.

The contrarian case is that safety-focused rhetoric slows frontier-model releases but increases spending on inference efficiency, data governance and enterprise deployment. That would hurt speculative AI-beta baskets more than established semiconductor leaders, and could make a broad short in AI ETFs an inferior expression to a selective GLW underweight. Avoid extrapolating a one-day de-risking move until options-implied volatility and positioning show whether the selling is systematic or fundamentally sponsored.

Key Decisions for Investors

• Initiate a 1-3 month pair: short GLW / long NVDA in dollar-neutral size. The thesis is divergent estimate risk from deferred networking build-outs versus NVDA backlog resilience; stop if GLW Optical Communications guidance is maintained or raised, or if the pair underperforms by 8%.
• Reduce unhedged long-duration AI exposure into the policy event; use a tactical long XLE / short QQQ hedge for 2-6 weeks rather than adding outright index shorts. Persistent higher energy and real yields favor the relative trade; exit if Brent retraces below $95 or the 10-year yield falls below 4.6%.
• Do not chase a broad BAI short after the initial move. Reassess after the policy conference and the next hyperscaler capex datapoints; initiate only if AI ETF volatility normalizes while NVDA and GLW estimate revisions turn negative, creating a cleaner 3-month downside setup.
• For existing NVDA longs, retain core exposure but buy 2-3 month downside puts or trim trading exposure ahead of guidance-sensitive earnings. A sustained break in data-center revenue guidance or evidence of customer order deferrals, not safety commentary alone, would invalidate the near-term earnings resilience thesis.