TSMC emerges as relative refuge in AI chip selloff

TSMC’s latest quarterly results underscored the point. Revenue rose 51% from a year earlier to almost $47 billion in the third quarter, while net income jumped 69% to nearly $24 billion, a pace of profit growth that outstripped costs and expenses. Yet the stock still fell after the report, showing how quickly investors are discounting the durability of AI infrastructure spending.

The stock’s resilience argument rests on scale and customer stickiness. TSMC controls about 73% of the third-party chip manufacturing market, according to TrendForce, and remains the dominant foundry for advanced chips. Nvidia, Broadcom and other designers still need its manufacturing capacity, regardless of whether the current burst of AI capex cools in coming quarters.

That is why the debate is shifting from whether AI spending slows to which companies can survive a pullback without losing long-term relevance. A temporary bust would hurt demand, but it would not eliminate the need for leading-edge foundry capacity. In that sense, TSMC is closer to the infrastructure layer than to the more cyclically exposed AI software or networking plays.

Valuation also helps frame the case. TSMC trades at about 30 times earnings, a multiple that looks more manageable than many AI-linked stocks given the company’s rapid profit growth. The shares are still up roughly 40% over the past year, but the recent pullback leaves them more attractive to investors looking for exposure without chasing the most expensive names.

Geopolitics remain the main overhang. TSMC has expanded U.S. production, but the company is still tied to Taiwan, and China’s claim over the island continues to weigh on its valuation. The bear case is that a geopolitical shock or a sharper-than-expected AI slowdown could hit both earnings and sentiment at once.

For investors, the narrative is straightforward: the current selloff is pressuring the most crowded AI trades, but TSMC’s earnings power, market position and central role in advanced chipmaking make it one of the few names that can offer both growth and relative safety. The next test will be whether AI capex stabilizes and whether the broader chip sector can recover without another leg lower in rates and risk appetite.