Fed’s Musalem says more rate hikes likely needed to curb inflation

St. Louis Fed President Alberto Musalem said further rate increases will likely be needed, arguing the current 3.75%-4.00% policy range remains accommodative despite the Fed's 25bp hike last week. July PCE inflation was 3.7% year over year, up from 2.3% in April 2025, as tariffs, the US-Israeli war with Iran, higher fuel costs and AI-driven commodity demand added price pressure. Markets price in three additional 25bp hikes through April, while Fed officials' median projection indicates one more increase this year.

Analysis

The market is pricing an unstable combination: multiple additional hikes and continued AI-led multiple expansion. That is most problematic for long-duration, capex-intensive AI beneficiaries, where valuation depends on discount rates falling while incremental infrastructure spend is rising. META is relatively insulated versus smaller AI names because its cash generation can fund data-center investment internally, but a higher terminal-rate regime still raises the hurdle rate for buybacks and makes advertising multiple expansion harder to sustain.

The less-obvious transmission is from AI infrastructure into physical-input inflation: power, copper, grid equipment and cooling constraints can extend the capex cycle while simultaneously tightening the policy backdrop that de-rates it. This favors asset owners and bottleneck suppliers over application-layer AI equities over the next 6-18 months. Energy and materials exposure also becomes a hedge against a scenario in which nominal growth stays firm but inflation fails to normalize, while consumer discretionary and low-margin industrials face a lagged margin squeeze from freight, fuel and insurance costs.

Near term, a strong risk-on tape can overwhelm rate sensitivity for days or weeks, particularly if AI earnings revisions remain positive. Over 1-3 months, the relevant catalysts are inflation prints, wage-insensitive services inflation, and any upward revision to the expected policy-rate path; a sustained rise in real yields would likely force dispersion within mega-cap tech rather than a uniform selloff. The contrarian view is that the policy concern may be underpriced in broad indices, but not necessarily in META specifically: its operating leverage and cash balance make it a better relative long than unprofitable AI infrastructure peers if rates remain elevated.

Key Decisions for Investors

• Initiate a 3-6 month pair: long META / short ARKK, sized beta-neutral. META offers self-funded AI capex and a resilient ad engine, while ARKK has materially greater long-duration valuation exposure; target 10-15% relative return, with thesis invalidated if real yields decline materially and unprofitable-growth earnings revisions turn positive.
• Add a 6-12 month inflation-barbell hedge: long XLE or XOP against an underweight in XLY. Persistent input-cost inflation supports upstream cash flow while pressuring discretionary purchasing power and retailer margins; reassess if core inflation momentum rolls over for two consecutive monthly releases and crude/gasoline retreat materially.
• Prefer AI bottleneck exposure over broad AI software beta: accumulate GRID and selectively ETN on pullbacks rather than chase high-multiple application names. Grid investment is less dependent on near-term model monetization, though entry should be reduced if copper prices reverse sharply or utility capex plans are deferred.
• Do not add outright META index-like exposure after a sharp momentum move; use any 5-8% pullback or post-earnings volatility to enter. Watch for upward revisions to 2027 capex without corresponding advertising revenue acceleration—this would weaken the cash-return thesis and warrant reducing the long leg.