Google's entry into the AI hardware market with its tensor processing units (TPUs) signals potential growth for South Korean memory manufacturers Samsung and SK Hynix. Industry estimates suggest TPUs can cut costs by up to 80% compared to Nvidia’s H100 chips while maintaining performance. Analyst Chae Min-sook notes that increased TPU adoption will heighten demand for high-bandwidth memory modules, benefiting both firms as AI data center projects expand in the coming years.
Google's entry into the AI hardware market with its tensor processing units (TPUs) signals potential growth for South Korean memory manufacturers Samsung and SK Hynix. Industry estimates suggest TPUs can cut costs by up to 80% compared to Nvidia’s H100 chips while maintaining performance. Analyst Chae Min-sook notes that increased TPU adoption will heighten demand for high-bandwidth memory modules, benefiting both firms as AI data center projects expand in the coming years.