TSMC 3-nanometer revenue tops 5-nanometer

The bull case is straightforward: AI and high-performance computing demand remain strong, TSMC’s capacity additions are still being absorbed, and the company’s technology lead remains intact. Management has expanded 3-nanometer capacity, partly by converting 5-nanometer lines, while the 2-nanometer ramp is running ahead of expectations and remains at least on schedule. That matters because TSMC controls well over 70% of leading-edge chip fabrication, making it the indispensable supplier not only for Nvidia and other AI chip designers but for the broader infrastructure build-out around the data-center cycle.

Investors have also latched onto the technical picture. The stock’s breakout on Friday strengthened a chart setup that points to about 100 dollars of upside from the prior correction range, which would imply a target near 575 dollars. TSMC’s U.S.-listed shares closed at $428.91 on Friday, up from $415.50 on Wednesday, while the Philadelphia semiconductor index rebounded to 519.86 after sliding sharply in late July. TSMC’s relative strength matters because it remains one of the clearest ways to express AI infrastructure demand without taking direct product risk.

Valuation still leaves room for more, at least in the market’s view. TSMC trades at roughly 19 times earnings, below the broader market multiple despite growing faster than the average large-cap company. That discount reflects lingering China-related risk, not a weakening in the core business. For investors, the key question is whether that geopolitical overhang will continue to cap the multiple even as the operational story improves.

The bear case is mainly about expectations and policy risk, not demand. If AI spending cools or the pace of node transitions slows, the premium narrative behind TSMC’s rise could lose momentum. But with Nvidia’s own results still pointing to sustained AI chip demand and TSMC’s advanced-node backlog tightening, the more immediate issue is whether the market has fully priced the combination of better margins, rising capacity utilization and a cleaner product mix.

For now, the message from the stock and the business is the same: TSMC is not just riding the AI cycle, it is setting the pace of it. If 3-nanometer revenue keeps outgrowing 5-nanometer and 2-nanometer stays on track, the path toward 575 looks less like a stretch and more like a function of how long investors are willing to pay for scarce leading-edge capacity.