Fed Chair Kevin Warsh Raises Interest Rates for the First Time in Over 3 Years. These AI Stocks Will Benefit

The Federal Reserve raised the federal funds rate by 25bps on Sept. 16 and signaled another hike before year-end, pushing 30-year mortgage rates close to 7%. Morgan Stanley argues that higher rates will favor profitable AI hyperscalers with dependable cash flow while pressuring unprofitable, capital-dependent companies as the industry targets an estimated $7 trillion in data-center infrastructure spending by 2030. Alphabet illustrates the advantage, with roughly $73 billion of annual free cash flow and $132 billion of net profit supporting about $200 billion in planned AI-focused capex, versus C3.ai's $288 million net loss and negative free cash flow.

Analysis

The investable effect is less "AI up versus down" than a widening cost-of-capital dispersion. GOOG, MSFT, AMZN and META can internally fund infrastructure while preserving strategic optionality; smaller model/application vendors must either slow product investment or issue equity at lower valuations. That can shift enterprise AI workloads toward integrated platforms, improving hyperscaler cloud attach rates and customer lock-in over the next 6-18 months.

The near-term risk for hyperscalers is not financing availability but capital discipline. A higher discount rate raises the hurdle rate on AI infrastructure, so investors will increasingly punish capex that fails to translate into cloud growth, ad/productivity monetization, or accelerating free-cash-flow conversion. The next 1-3 months of earnings guidance matter more than the policy move itself: an upward capex revision without a corresponding revenue or margin bridge would compress multiples even for cash-rich incumbents.

For AI, the key falsification variable is liquidity runway rather than reported profitability alone. If cash balances and operating-cash-burn trends indicate no external capital need for at least 18-24 months, a blanket short thesis is vulnerable to contract wins or a lower-rate reversal. Conversely, a widening gap between bookings rhetoric and cash collections would expose the equity-financing risk quickly, particularly if the stock rerates before fundamental improvement.

Consensus may be too sanguine on NVDA: its direct balance-sheet sensitivity is low, but its demand is exposed to customers' return thresholds. If hyperscalers rationalize deployment schedules rather than cancel projects, the impact would initially appear as slower accelerator lead-time/order visibility and a mix shift toward lower-cost inference hardware, not an abrupt revenue collapse.

Key Decisions for Investors

• Initiate a 3-6 month beta-neutral pair: long GOOG versus short AI, sized by 60-day beta rather than equal dollars. The thesis is relative funding capacity and distribution advantage; cover the short if AI demonstrates two consecutive quarters of improving free-cash-flow burn and material backlog-to-revenue conversion.
• Maintain selective hyperscaler exposure through GOOG and MSFT, but add only ahead of earnings if management provides an explicit revenue/operating-income bridge for incremental AI capex. Reduce if capex guidance rises while cloud growth or consolidated free-cash-flow guidance weakens; this is a multiple-risk trade, not a rates-duration hedge.
• Avoid adding directional NVDA exposure solely on the view that cash-rich customers are insulated from higher rates. Set an alert for hyperscaler capex guidance and accelerator lead-time commentary over the next two reporting cycles; evidence of deferred build schedules would favor trimming NVDA or hedging with 3-6 month put spreads.
• Monitor a basket of cash-burning AI software names (AI, PATH, SOUN) against IG credit spreads and equity issuance activity. A sustained widening in spreads or a discounted secondary offering is the actionable confirmation for a broader short basket; absent those signals, treat the funding-stress narrative as a watch item rather than a standalone trade.