I have seen this movie too many times. Wall Street decides there is a macro theme—this time, it is AI infrastructure and economic resurgence—and suddenly every company that can spell "data" or "materials" gets lumped into the same absurd bucket. This is stupid and here's why. If you are buying everything from semiconductor designers and concrete mixers to toy makers and gas shippers in one go, you are not diversifying; you are just being lazy. It is time to look at who is actually doing the work, and who is just asleep at the wheel.

Let's start with the companies doing the literal heavy lifting. Vulcan Materials (VMC.US) and Martin Marietta Materials (MLM.US). You think you just plug an Nvidia GPU into the sky? You need physical data centers, and that requires concrete and aggregates. Vulcan just posted Q1 2026 EPS of USD 1.35, beating expectations thanks to aggressive data center build-outs. Their stock has comfortably outperformed the broader market this year. Martin Marietta isn't playing around either, posting a massive 17% surge in Q1 revenue. This is what real infrastructure looks like. Without them, your AI dreams are just hallucinations.

Then you have Fastenal (FAST.US). It is not a sexy business, selling nuts, bolts, and fasteners. But with gross margins historically holding steady around 45%, they are quietly raking in the cash and maintaining a solid stock performance. They are executing flawlessly while flashy tech startups burn capital. Good for them.

On the tech side, Rambus (RMBS.US) is surfing the AI memory wave, pulling in USD 180.2 million in Q1 2026 product revenue. Shares have soared this year, but seriously, why aren't you moving faster? The window for DDR5 upgrades is closing, and they need to capitalize on this bottleneck immediately before someone else eats their lunch.

Kingsoft Cloud (KC.US) finally showed some teeth, with Q1 2026 revenue up 37% to RMB 2.7 billion and AI billings surging, giving its stock a much-needed bounce. But they are still bleeding money, with net losses widening to RMB 343.7 million. In a winner-takes-all cloud market, time is running out to prove you belong at the adult table. Meanwhile, Japanese legacy giant Fujitsu (FJTSY.US) is still trying to pitch its "digital transformation." With annual revenues hovering around JPY 3.7 trillion, the stock has traded sideways. Turning this bureaucratic ship around is going to take a miracle. Good luck with that.

Then we have the geopolitical casualties and the oddballs. WuXi Biologics (WXXWY.US) has been battered by U.S. policy risks, sending its shares into a tailspin. When over 40% of your historical revenue comes from North America, you are uniquely vulnerable to Washington's moods. Conversely, PayPay (PAYP.US) is dominating the Japanese mobile payment space with over 60 million registered users, its stock quietly grinding higher. Dorian LPG (LPG.US), operating a fleet of over 20 VLGCs, continues to print cash and trade near historical highs. Boring, but incredibly effective.

Finally, what on earth is Mattel (MAT.US) doing in this cycle? Pushing USD 5 billion in annual sales by selling plastic toys in an era dominated by TikTok and generative AI. The stock has been sluggish, and honestly, they need a better script than hoping for another movie miracle.

Bottom line: If you are placing bets in 2026, buy the companies laying the actual groundwork. Stop funding the illusion.

This article does not constitute investment advice.