This article first appeared on GuruFocus.
Taiwan Semiconductor Manufacturing (NYSE:TSM) is striking a more cautious tone on artificial intelligence even as the chipmaker remains one of the biggest financial beneficiaries of the AI boom. A senior executive warned that the technology is powerful but still immature, underscoring a growing tension for investors: TSMC is supplying the chips driving AI adoption while remaining reluctant to deploy the technology deeply inside its own most sensitive operations.
Is TSM fairly valued? Test your thesis with our free DCF calculator.
AI is like a three-year old superman. It can be quite powerful, but it doesn't know what kind of damage it can inflict, TSMC co-chief operating officer Y.J. Mii told students at National Taiwan University. And it doesn't know right from wrong, he added.
Mii said TSMC remains wary of using AI with confidential research and development data because improper deployment could lead to information leaks or data loss.
His comments come as the broader technology industry debates whether frontier AI development should slow while safety controls catch up. Recent incidents involving rogue AI agents accessing external systems have intensified those concerns.
For TSMC, however, the issue is not only safety.
Mii also said current AI systems provide limited assistance in developing cutting-edge chipmaking processes, where progress depends heavily on physics, manufacturing constraints and access to increasingly advanced equipment.
That distinction matters because TSMC sits at the center of the AI supply chain. Nvidia (NASDAQ:NVDA), Apple (NASDAQ:AAPL) and other major customers depend on its leading-edge manufacturing capacity, making protection of proprietary process technology critical to maintaining the company's competitive moat.
Mii's comments do not weaken the near-term AI demand story for TSMC, but they highlight how differently AI may affect chipmakers internally versus externally.
Investors should watch whether Nvidia and other customers continue accelerating advanced-node demand, as well as TSMC's capacity expansion, pricing and margins.
The bigger risk would be a slowdown in AI infrastructure spending, not TSMC's cautious internal adoption. Continued strong demand for advanced chips would keep the investment case intact, while weaker hyperscaler spending or slower model development could eventually pressure utilization and growth expectations.