I've been watching TSMC closely lately, and honestly, there's a compelling case for why this semiconductor giant might be one of the better AI plays right now. Hear me out.

So here's the thing about the AI boom — everyone's talking about Nvidia, and sure, they're crushing it. But what people sometimes miss is that someone has to actually manufacture all those chips. That's where TSMC comes in. They're the world's leading foundry, and they're basically the only company that can build high-end chips at the scale and speed the market demands right now.

The numbers tell the story. By Q3, TSMC was sitting on roughly 72% of the global foundry market by revenue. Their closest competitor, Samsung, had just 7%. What's wild is that despite the AI investment cycle flooding the market, TSMC actually increased its share from around 65% mid-2024. That's not luck — it's because chip companies have nowhere else to go. The money's too big, the stakes too high.

Nvidia's been their anchor tenant, and that relationship just got more interesting. Nvidia's next-gen Rubin architecture is coming in 2026, and TSMC is building it on their advanced 3-nanometer process. Meanwhile, Nvidia's sitting on a $500 billion order backlog. That's a massive tailwind for TSMC's foundry business as those orders flow through.

Here's where it gets interesting from a valuation angle. TSMC's trading at roughly 30x 2025 earnings, which might sound pricey. But analysts are modeling nearly 29% annual earnings growth over the next three to five years. Using the PEG ratio — which compares growth to valuation — TSMC's sitting at around 1, which signals the stock is actually attractive. I'm usually comfortable paying up to 2 or 2.5 for quality semiconductor plays, and TSMC is definitely that.

The way I see it, TSMC has a pretty high floor because of how critical they are to the AI infrastructure buildout. Even if growth comes in a bit softer than expected, the long-term returns look solid. The semiconductor industry isn't going anywhere, and TSMC's dominance in the AI chip manufacturing space gives them a structural advantage that's hard to replicate. Worth keeping on your radar if you're thinking about AI exposure beyond just the chip designers.