Nvidia CEO Jensen Huang Told Investors in Seoul to "Buy at a Discount" During the Recent AI Stock Sell-Off. Here's Whether His Call Has Paid Off.

AI stocks pulled back after valuation worries on May 14, with Nvidia down ~15% and AI hyperscalers Alphabet and Amazon down ~11% each. Nvidia CEO Jensen Huang argued investors should “buy at a discount” in the dip; since June 8 Nvidia is up 5.1% versus the S&P 500’s 4.3% (+0.8% relative). Broader “AI basket” performance since June 8 (equal parts Nvidia, Alphabet, Amazon, Microsoft) is up ~8.5%, suggesting the pullback presented an opportunity for longer-term AI exposure.

Analysis

This is less a “buy AI” signal than a reminder that the trade is now driven by earnings visibility and capex credibility, not narrative alone. The best risk/reward sits with the names that can translate AI spend into nearer-term revenue and margin leverage: NVDA, MSFT, and to a lesser extent AMZN. For the hyperscalers, the market will punish any sign that AI capex is outrunning monetization, so the next quarter’s guide matters more than the current quarter’s beats.

Relative winners also extend down the stack: memory, advanced packaging, networking, and power/cooling suppliers should benefit from the same spend cycle with less valuation risk than the mega-cap platforms. If SKHYV is a Korean memory exposure, it likely offers more convexity than the large-cap software complex because its upside is tied to unit demand and pricing, not just sentiment. GOOG/GOOGL look like the laggard because the AI story is still harder to underwrite in near-term revenue terms, even though the balance sheet can absorb the spend.

The contrarian risk is that the market may be extrapolating another 12-18 months of capex growth without enough evidence. If hyperscaler capex slows, or if enterprise AI usage fails to lift cloud margins, the multiple support for the whole group can compress quickly. Near term, the real catalyst window is the next earnings cycle; structurally, the thesis survives only if AI workloads keep scaling faster than depreciation and competitive pricing pressure.

Key Decisions for Investors

• Long NVDA into the next earnings window, but prefer pullbacks over strength; thesis is a 1-3 month guide-up reacceleration. Falsify if datacenter growth or gross margin commentary slows materially.
• Pair trade: long MSFT / short GOOGL or GOOG for 1-2 quarters. MSFT has clearer AI monetization and operating leverage; GOOG faces more FCF drag from capex with less visible near-term payoff.
• Long AMZN vs QQQ on a 3-6 month horizon as AWS AI demand can offset margin noise if utilization improves. Exit if AWS growth decelerates while capex stays elevated.
• If SKHYV is a liquid memory semi proxy, use it as a higher-beta expression of the AI capex cycle. Best entry is after any sector pullback; thesis breaks if memory pricing rolls over before next procurement cycle.
• Avoid chasing the broad AI basket after a sentiment rebound; the cleaner trade is relative value, not outright beta. If the group rallies on multiple expansion without upward capex revisions, fade the move.