Here’s the thing about market panics: they’re usually just a gap between what insiders know and what regular people feel. Last week, pop star Lorde literally told a stadium full of fans to reject Meta’s AI glasses while celebrities were getting paid to promote them. The crowd saw surveillance. The billionaires saw the future. Same tech, completely different vibes.
That disconnect is exactly what’s happening with AI infrastructure stocks right now, and it’s creating a genuinely interesting opportunity.
Samsung just reported operating profits of nearly $60 billion—up 19 times year-over-year—almost entirely from AI memory demand. Meanwhile, IBM pre-announced a revenue miss and tanked 20% when CEO Arvind Krishna revealed that enterprise clients were panic-buying servers and memory before expected price hikes. The selloff spread like wildfire, dragging down software and consulting stocks in sympathy. Samsung’s collecting the ransom. IBM just showed Wall Street who’s paying it.
Here’s the kicker: Wall Street’s conviction on AI infrastructure hasn’t budged. The crowd just hasn’t caught up yet.
The Five Plays:
SpaceX (SPCX) is the most argued-about stock in the market, but ignore the Elon drama. What matters is that SpaceX is the only vertically integrated company combining rocket launches, frontier AI through xAI, and a live data feed from X. Wall Street’s penciled in $330 billion in 2030 revenue. At a 10x multiple (reasonable for this growth rate), you’re looking at a $2-3 trillion company. At $150 a share, the math works.
TeraWulf (WULF) pivoted from Bitcoin mining to leasing power. They just signed a 20-year, $19 billion deal with Anthropic for an AI campus in Kentucky. Revenue’s expected to grow 89% this year, then 210% next year, with gross margins expanding from 50% to 70%. The stock’s sitting right at its historical support level—classic entry point.
Amazon (AMZN) just borrowed $25 billion for AI infrastructure and launched 29 more satellites, bringing their total to 396. They’re moving from concept to commercial deployment on satellite broadband. Trading at 22.6x forward earnings and 11x forward EBITDA—basically five-year lows—while margins expand. This is a company that’s cheap, dominant, and growing profits faster than sales.
Palantir (PLTR) got caught in the software selloff, down 26-27% from highs. The chart looks rough, but the fundamentals are wild: 73% revenue growth expected this year, then 46-52% after that, with 86-87% gross margins. Once it reclaims its 200-day moving average, this could be a 70-80% compounder.
Micron (MU) is down 22% from highs, and bears say memory’s peaked. Samsung’s blowout quarter says otherwise. The real question gets answered in three weeks when hyperscalers report earnings and either confirm or hike 2026 capex plans. Micron’s past pullbacks have bottomed at 20-30% drawdowns. We’re at 22.6% right now.
The pattern repeats every cycle: scary headlines, broken charts, then earnings roll in and fear becomes the entry point. This time might be different, but the math says it probably isn’t.
