Fund manager names 3 non-AI stocks to own as Intel, AMD sink amid broader tech rout

Tech stocks are selling off sharply globally, with chipmakers Intel, AMD, Micron and Nvidia under pressure amid a broader risk-off move. The decline is being driven by contagion from a KOSPI crash of more than 10% that triggered a trading halt, renewed fears of US rate hikes, and a valuation reset in the AI trade. The article points to market-wide weakness rather than company-specific fundamentals.

Analysis

This move is less about one bad tape and more about a forced de-grossing event in crowded growth books. When rates reprice higher, the AI complex is hit twice: first through duration compression on far-dated cash flows, and second through an inventory of leveraged and momentum longs that becomes mechanically vulnerable once the benchmark breaks trend support. That makes the initial decline self-reinforcing over the next few sessions, especially if systematic sellers and vol-control funds continue to de-risk.

Relative losers are the highest beta semis with the most consensus ownership and the weakest near-term earnings elasticity. Nvidia should outperform on a fundamental basis versus more cyclical chip names because its demand is still tied to hyperscaler capex rather than consumer PC replacement, but it is also the cleanest proxy for the AI trade and therefore the easiest source of liquidity for managers cutting risk. Intel is less exposed to AI enthusiasm decay and more exposed to macro-duration repricing; AMD sits in the middle but tends to underperform in risk-off tapes because it lacks a strong balance-sheet or monopoly-style narrative to cushion multiple compression.

The second-order winner is not another chipmaker but cash-rich megacap tech and defensives that can absorb a higher discount rate without existential multiple risk. If this is driven by a genuine rate-hike scare rather than a one-day Asia contagion shock, then vendors upstream in the AI supply chain — equipment, memory, and foundry-adjacent names — should see estimate cuts before end-demand names do, because order timing is the first lever CFOs pull. That means the market may be underpricing a second leg lower in the semiconductor equipment group over the next 1-3 months even if the headline AI leaders stabilize first.

The contrarian case is that this is a positioning flush, not an earnings reset. If yields fail to follow through or if macro data soften, the tape can snap back quickly because AI has become the most overcrowded long in growth, and crowded trades often overshoot both ways. The risk to chasing the downside here is that any sign of yield peak, dovish Fed rhetoric, or hyperscaler capex reaffirmation could trigger a violent squeeze in the exact names that are being sold first.

Key Decisions for Investors

• Short NVDA on a 1-3 day horizon only after a failed rebound into intraday strength; use a tight stop above the first post-selloff reclaim of trend support. Risk/reward is attractive because it is the most liquid AI proxy and likely source of further de-grossing if yields keep rising.
• Relative-value pair: long INTC / short AMD for 2-6 weeks. Intel should be less exposed to AI multiple compression, while AMD remains a cleaner high-beta semis expression; this pair isolates valuation risk from broad tape weakness.
• Buy put spreads on SMH or SOXX with 4-8 week tenor rather than outright puts. This captures a potential second leg lower in semis from higher-rate contagion while limiting premium burn if the sell-off becomes a one-day flush.
• If 10Y yields continue higher for another 2-3 sessions, add to short NVDA or short QQQ against long cash/short duration. The trade works best if the market starts pricing a higher terminal rate rather than a transitory risk-off event.
• Avoid initiating fresh longs in AMD/NVDA until yields stabilize for at least 2-3 sessions; the setup favors patience over catching a falling knife because the marginal buyer is likely a systematic flow rather than fundamental dip-buying.