Deep Dive
The Nvidia paradox: buying the dip makes no sense
Nvidia has cratered nearly $1 trillion from its May 14 peak, and Wall Street's response is predictable — 95% of analysts still rate it a buy or strong buy. Kim breaks the consensus: he's not betting against Nvidia, but he's also not treating the dip as a screaming opportunity. The core issue is valuation. At 22x forward earnings, Nvidia remains expensive despite being at its cheapest level since 2019. Kim sets a hard floor: if it drops below 20x, he'll add. But until then, he's looking elsewhere in the AI ecosystem. Nvidia is the hub, he says, but the real growth is in the spokes.
Memory is the secret chokepoint nobody's pricing in
Here's the insight Kim leads with: the AI revolution stops without memory and power. Micron is the play here, trading at just 6x forward earnings while publicly admitting it can't keep up with demand. Every time Micron has to turn away a customer, SK Hynix benefits. The timing is perfect. SK Hynix launches as an ADR on Nasdaq this Friday, ending the friction of buying on Korean exchanges. Kim will be buying on day one, but with a caveat he borrowed from the SpaceX and Facebook IPO playbooks: expect initial euphoria to fade. Companies trading at wild valuations on launch tend to cool within weeks or a month. SK Hynix had a $4 billion operating loss last year yet traded at rich valuations at IPO, signaling irrational demand that usually deflates.
The AI 7 outpaced Mag Seven by 75% in the first half
Kim's core thesis hinges on a simple rule: instead of asking who's spending on AI, ask who's receiving the spending. Mag Seven stocks fell 1.7% on an equal-weighted basis in H1, except Alphabet, which was up double digits. His AI 7 basket — anchored by Nvidia and Alphabet but filled with TSMC (the world's largest dedicated chip foundry, now reshoring production to Phoenix), Micron (onshoring to Syracuse), Vertiv Holdings (cooling infrastructure), and American Electric Power (power and a 2.5% dividend yield, up 20% year to date) — returned 75% over the same period. The philosophy is capital efficiency: why own seven momentum stocks when you can own the entire ecosystem getting paid by their spending spree?