TSMC struggles to keep up with AI demand: ‘We can only support so much’

The Verge reports that although Taiwan Semiconductor Manufacturing Co. (TSMC) has been pushing forward factory expansion in the United States, it still faces pressure that makes it hard to fully keep pace with the continuously rising AI demand from US customers. The article cites reports from Reuters and Bloomberg, noting that the world's largest semiconductor manufacturer is currently under intense demand from the customer side, especially as the AI boom drives advanced chips and related capacity to become a key bottleneck in the supply chain. After the shareholders' meeting, TSMC CEO C.C. Wei said customer demand is very high and that the company "can only support so much," a statement that underscores how even the world's most important foundry cannot expand supply capacity quickly without limit. For Taiwanese readers, the significance of this news lies not only in TSMC itself but in how it reflects that the race for AI infrastructure has already extended from models, software, and cloud services down to the more fundamental layer of semiconductor manufacturing capacity. Strong demand from US customers means that investment in AI servers, accelerators, and data centers is continuing; but limited capacity also means that related companies may face challenges in lead times, costs, and supply prioritization. For developers and researchers, this kind of hardware supply pressure may indirectly affect the availability and pricing of cloud GPUs; for investors and entrepreneurs, it shows that the growth of the AI industry still relies heavily on a small number of advanced-process suppliers. The article itself does not provide more complete capacity figures or a specific customer list, so it is unwise to infer which companies are directly affected, but what is certain is that AI demand has left TSMC facing the reality of supply falling short of demand even amid its US expansion.