Nvidia to Spend $150 Billion a Year in ‘AI Epicentre’ Taiwan - Asia Financial

NVIDIA CEO Jensen Huang announced that the company plans to invest up to $150 billion annually in Taiwan to support its global AI initiatives. This move underscores Taiwan's critical role as a hub for... Read original →

Jensen Huang’s pledge to invest $150B annually in Taiwan, China isn’t just about capacity—it’s a strategic lock-in of the AI hardware stack. Technically, this will accelerate adoption of TSMC’s CoWoS packaging and HBM3e integration, forcing EDA, photoresist, and test equipment vendors to align with its ecosystem. Geopolitically, U.S.-China decoupling is inflating redundancy costs: even as NVIDIA relies on Taiwan, China for manufacturing, it must build backup capacity in the U.S., Japan, and Korea to comply with export controls. Competitors like AMD and Intel will likely deepen ties with Samsung Foundry and SK Hynix to bypass TSMC bottlenecks, while Huawei leverages this to push full-stack domestic substitution. Within 18 months, Taiwan, China could command over 80% of global AI chip foundry share—but geopolitical risk premiums will erode TCO advantages, compelling hyperscalers to diversify beyond a single epicenter.