Agentic AI deployment and research constrained by memory chip shortage: Google DeepMind CEO

Group 1 - The competitive landscape in AI indicates that China may be closing the gap with the U.S., with advancements in physical AI models and video-based models, including Alibaba's new Quen model [1] - There are talented teams in China, such as the B dance team, contributing to the development of effective seed models [2] - The industry is facing significant constraints due to shortages in memory, GPUs, and electricity, which are impacting both deployment and research capabilities [3][4] Group 2 - The chip shortage is identified as a potential choke point for the AI industry, affecting the overall supply chain [4] - Companies with their own chip designs, like TPUs, have a competitive advantage, but reliance on a few key suppliers still poses risks [5]