Picture this: 1908. Over 240 car companies are fighting for scraps, dealers can’t move inventory, and newspapers are calling automobiles a “rich man’s fad.” Sound familiar? Because we’re watching the exact same movie play out right now with AI infrastructure stocks—and most investors are about to make the same mistake those 1908 skeptics did.
Here’s the thing: AI chip stocks just got hammered. We’re talking 20-30% drops in a matter of weeks. And everyone’s convinced it’s because demand is about to crater. But here’s what actually happened: Samsung just reported operating income up 19 times year-over-year, described memory demand as “white-hot,” and the market… sold the stock anyway.
Yeah. You read that right.
The problem is that investors are watching the wrong scoreboard. They’re freaking out about supply hitting the market while ignoring the fact that the five biggest tech companies—Amazon, Meta, Google, Oracle, and Microsoft—are about to spend roughly $800 billion on AI infrastructure this year. That number could hit $1 trillion annually by 2027. Amazon just raised another $25 billion in bonds specifically for AI. These aren’t companies pulling back. They’re emptying their coffers and hitting the debt markets because they’ve already spent everything they had.
Meanwhile, OpenAI raised $122 billion, Anthropic grabbed $60 billion, and SpaceX pulled in $85 billion while becoming a serious compute player. None of that screams “demand is collapsing.”
The Setup
Here’s where it gets interesting: the semiconductor ETF (SMH) is sitting about 14% below its highs. Historically, that’s exactly where AI semiconductor stocks bottom out. Since late 2022, every routine pullback in this group has bottomed in the 10-15% range (except for two freak-out moments). We’ve had thirteen double-digit corrections since the AI boom started, and these stocks are still up over 565%.
Zoom out to the 1990s tech boom, and you see the same pattern. Semiconductor stocks got crushed multiple times—including two near-40% bear markets—between 1995 and 2000. They still climbed over 1,100% by March 2000. Sharp pullbacks aren’t a sign the party’s over. They’re the price of admission for triple-digit rallies.
Where to Look
If you’re thinking about buying this dip, look for stocks that have crushed it over the past 6-12 months, dropped more than 10% from their highs, but are still trading above their 200-day moving averages. That means the long-term uptrend is intact even during the panic.
Names like Aehr Test Systems (AEHR), Ultra Clean Holdings (UCT), Axcelis Technologies (ACLS), and ACM Research (ACMR) fit that profile. These are the unglamorous picks—the equipment and materials companies that make chip manufacturing possible. They’re not sexy, but they’re essential.
The Bottom Line
Volatility is baked into every hypergrowth bull market. The 1908 car panic felt like proof the industry was doomed. It was actually proof the industry was working exactly as it should: in violent, uneven bursts that reward patience and punish panic.
Late July, when the hyperscalers report earnings and confirm their capex plans, is when this trade wakes back up. Until then, the gift is sitting right in front of you.
