
The AI chip arms race gets more expensive
TSMC’s board just greenlit about $31.28 billion in capital spending to expand advanced chip capacity, build new fabs, and install the kinds of factory systems that keep the AI boom fed. In plain English: the company sees enough demand for advanced semiconductors that it’s willing to keep pouring concrete and buying equipment at a pace that would make most CFOs sweat.
The company also approved up to $20 billion in extra investment for its wholly owned Arizona subsidiary, which is another sign the U.S. buildout is still very much on. If you’ve been watching the AI supply chain story, this is the part where the plot gets less about software hype and more about who can actually make the chips fast enough.
Bigger fabs, bigger payouts
TSMC also said it increased its quarterly cash dividend to NT$7 per share from NT$6. That’s the kind of move that tells investors two things at once: management thinks the cash machine is healthy, and it’s happy to share a little more of it.
The company’s first-quarter 2026 results were strong too, with revenue of NT$1.13 trillion, net income of NT$572.48 billion, and diluted EPS of NT$22.08. So yes, the spend is huge — but it’s being funded by a business that’s still printing serious money.
Why investors should care
TSMC is the pick-and-shovel play for AI infrastructure. When it spends more on capacity, that usually means customers like the AI and HPC crowd are still lining up for wafers, not backing away from the table.
Big picture: this is less “TSMC is splurging” and more “TSMC thinks the buffet line is getting longer.”
