3 Resilient Artificial Intelligence (AI) Stocks That Did Well the Last Time the Fed Raised Interest Rates

The article assesses Nvidia, Palantir and Broadcom as AI beneficiaries that previously outperformed after the Fed's July 2023 rate hike, but warns that prospective further hikes could restrain AI spending and equity-market gains. Nvidia generated $127B of trailing-12-month free cash flow and more than doubled most recent-quarter sales, while Palantir posted 93% revenue growth but trades near an 80x forward P/E. Broadcom grew revenue 48%, trades at an 18x forward P/E and is down more than 30% from its $495 52-week high, making it the author's relatively more attractively valued choice despite rate-risk concerns.

Analysis

This is not a standalone catalyst; the relevant mechanism is whether higher real yields force hyperscalers to ration AI capex. NVDA remains most exposed to a discrete reduction in accelerator orders, but its cash-rich customer base makes a one-quarter digestion cycle more likely than an immediate demand collapse. The more important 1-3 month indicator is not Fed rhetoric but capex guidance from MSFT, AMZN, GOOGL and META; aggregate AI infrastructure commitments below consensus would compress semiconductor multiples before reported revenue weakens.

AVGO offers the cleaner relative setup because custom silicon adoption is a cost-optimization response to AI spending, not merely an incremental spending category. If customers shift part of training workloads from merchant GPUs to ASICs, AVGO can retain AI-content growth even under a fixed hyperscaler capex envelope; its software cash flows also reduce duration sensitivity versus pure semis. The risk is customer concentration and program timing: any delay in a major ASIC ramp can create a sharp estimate reset despite the lower headline valuation.

PLTR is the vulnerable leg if yields rise because its valuation requires sustained growth and expanding enterprise adoption simultaneously. A modest deceleration in commercial deal conversion, rather than an outright AI-demand decline, could drive disproportionate multiple compression over 6-18 months. Contrarianly, a broad AI selloff may be underdone for lower-quality infrastructure suppliers but overdone for NVDA/AVGO if hyperscalers explicitly reaffirm capex; the market has repeatedly confused spend timing with the end of the compute buildout.

Key Decisions for Investors

• Initiate a 3-6 month pair: long AVGO / short PLTR, sized beta-neutral. Target relative upside from ASIC share gains and PLTR duration-multiple compression; exit if AVGO discloses a material customer ramp delay or PLTR sustains commercial growth above consensus while raising full-year margin guidance.
• Maintain NVDA as a core long only with a defined hedge into hyperscaler earnings: buy 1-2 month QQQ puts or reduce gross exposure rather than shorting NVDA outright. The key falsifier is aggregate 2027 capex guidance from MSFT, AMZN, GOOGL and META falling meaningfully below sell-side expectations.
• Add AVGO on evidence of design-win conversion or post-earnings confirmation of AI semiconductor backlog, not on rate headlines. Risk/reward improves if management demonstrates that custom-chip revenue offsets any networking or broadband softness; avoid adding if forward AI revenue commentary depends on a single customer program.
• Set an alert for a sustained rise in 10-year real yields alongside downward revisions to hyperscaler capex estimates. That combination warrants reducing NVDA and AVGO exposure; yields alone are insufficient evidence of an AI-demand impairment.