Intel Sinks 6% as Profit Taking Hits a Parabolic Run; NVIDIA and AMD Retreat 3%

Intel fell 6% to $99.67 after a 188% year-to-date rally, while NVIDIA and AMD each declined 3% and the SOXX semiconductor ETF dropped 3%, indicating broad profit-taking rather than company-specific deterioration. The selloff coincided with the 10-year Treasury yield reaching 4.84%, its highest level in nearly three years, and crude oil rising above $100 amid Middle East conflict, increasing inflation and valuation-multiple risks for growth stocks. Underlying chip demand remains strong: Intel posted Q2 FY2026 revenue growth of 25.4% to $16.13B, NVIDIA reported 105.85% growth to $96.22B, and AMD's data-center revenue rose 107% year over year.

Analysis

The relevant signal is dispersion, not a single-session decline: INTC’s recent rerating appears more flow-driven than supported by a durable change in earnings quality. A prospective PC pricing action would improve gross margin only if unit elasticity remains benign, while the higher-value question for the next 6-18 months is whether Intel can convert AI/data-center demand into sustained operating leverage without further foundry capital intensity. That leaves INTC more exposed than NVDA to a reversal in retail momentum, short-covering, and expectations embedded in a sharp multiple expansion.

Rising real yields create a near-term de-risking regime for the semiconductor complex because the group’s valuation increasingly depends on cash flows several years out. Within that regime, NVDA is relatively better insulated by backlog visibility and ecosystem lock-in, whereas AMD needs continued accelerator share gains and INTC needs execution credibility; both latter narratives have more estimate-revision risk if enterprise AI deployments elongate. The second-order beneficiary of a broad chip multiple reset could be QCOM and TXN, whose valuation support is more tied to mature-cycle cash generation than AI-capex duration, although neither is a clean substitute for AI beta.

Over the next days to weeks, the key falsifier for a tactical semiconductor short is a meaningful retreat in the 10-year yield combined with another hyperscaler capex affirmation; that would rapidly re-crowd AI leaders. Over 1-3 months, watch whether Intel’s revenue outlook converts into gross-margin and free-cash-flow guidance upgrades rather than top-line growth alone. Consensus is likely too focused on the headline growth rates and too little on the possibility that higher financing costs force customers to optimize AI utilization before committing to the next infrastructure tranche.

Key Decisions for Investors

• Maintain a tactical underweight in INTC versus NVDA for the next 1-3 months: long NVDA / short INTC in beta-neutral dollars. The trade isolates Intel’s greater execution and momentum-unwind exposure while retaining AI infrastructure exposure; exit if INTC delivers a material gross-margin or free-cash-flow guide increase, not merely revenue upside.
• For portfolios needing to reduce sector beta, buy 2-3 month SOXX put spreads rather than shorting NVDA outright. Structure strikes around a further 8-12% ETF decline, funded by selling downside below a 15-18% move; this targets a yield-driven multiple reset while capping loss if AI demand data reaccelerate.
• Do not add directional INTC exposure solely on a return above the psychologically important $100 area. Require confirmation through sustained relative strength versus SOXX and independently verifiable evidence of pricing realization or margin expansion; absent that, rebounds are likely distribution opportunities after a parabolic advance.
• Rotate a portion of AI-semiconductor exposure into QCOM or TXN only if long-end yields remain elevated for several weeks. This is a relative-duration hedge, not an outright bullish call; reverse it if hyperscaler capex guidance pushes NVDA/AMD estimate revisions materially higher or yields decline sharply.