Been looking at the AI infrastructure spending numbers and they're pretty wild. The big hyperscalers - Amazon, Meta, Alphabet, Microsoft - are collectively planning to drop $700 billion on AI data centers this year. That's a jump of over $300 billion from last year alone. The scale is honestly hard to wrap your head around.

Here's what caught my attention though. About 60% of that hyperscaler AI spending is going straight into chips and computing hardware. That's where TSMC comes in. They're manufacturing chips for basically every major AI chip designer - Nvidia, Broadcom, Marvell, Intel, AMD. And they're not just riding the AI wave, they control 72% of the global foundry market. Samsung Foundry is second place with only 7%. It's not even close.

TSMC management said AI accelerators hit high-teens percentage of revenue last year, and customers are already asking for more capacity. They're ramping capex from $40.9 billion to between $52-56 billion this year to handle it. Even more interesting, they're projecting mid-to-high-fifties percent CAGR on AI accelerator revenue through 2029. That's the kind of growth trajectory that doesn't happen often.

But here's the second piece of this. ASML, the Dutch semiconductor equipment maker, is the real hidden beneficiary. They've got basically a monopoly on EUV lithography machines - the equipment that makes advanced chips at 7nm and below. TSMC's 2nm process is in insane demand right now, which means they need more equipment, which means ASML's order book keeps growing.

ASML's net bookings jumped 48% last year to just over 28 billion euros. They're guiding 2026 revenue between 34-39 billion euros, midpoint around 12% growth. But here's the thing - they had almost 39 billion euros in backlog at the end of 2025. With hyperscalers ramping this aggressively, ASML could easily blow past those guidance numbers.

Analysts are calling for 34% earnings growth at TSMC this year. ASML's sitting at a 12-month median price target around $1,675, suggesting maybe 13% upside, but that feels conservative given what's actually happening with hyperscaler capex. The way these two companies are positioned in the AI infrastructure build-out, they're basically printing money right now. If you're looking at where this $700 billion is actually flowing, these are the names to watch.