As GPU counts in AI servers soar, rack power consumption has reached record highs, straining data-center energy systems. In response, major European and US power semiconductor makers, including STMicroelectronics (STM) and Onsemi, are rolling out advanced power-management technologies to enhance efficiency, reduce heat, and lower overall system costs.
As Taiwan’s machinery sector enters its 80th year in 2025, the industry is under growing strain from tariffs and currency volatility, forces beyond its control. With regional competition heating up, local manufacturers are reassessing how to stay competitive and sustain growth.
On November 2, local US time, Nvidia made history by sending its powerful H100 GPU into orbit for the first time. Just two days later, Google CEO Sundar Pichai announced that the company’s TPU chips would soon follow, signaling the start of an unprecedented new frontier — the race to build gigawatt-scale AI data centers in space.
No account yet?
Create an Account
