Nvidia CEO warns about China's big AI advantages: construction and energy (www.techradar.com)

Nvidia CEO Jensen Huang has raised alarms about China's rapid advancements in AI, particularly highlighting its superior infrastructure for building data centers and managing energy demands. At a recent event with the Center for Strategic and International Studies, Huang noted that while establishing an AI supercomputer in the U.S. takes approximately three years, China can construct equivalent facilities in a matter of days. He pointed out that China's energy capacity is double that of the U.S., further fueling its AI ambitions. Despite this, Huang asserted that Nvidia maintains a technological edge in AI chip development. In addition to Huang's remarks, a new study revealed that Chinese open-source language models (LLMs) have dramatically increased their share of global AI usage, climbing from just over 1% to 30% of token processing in a year. While proprietary models like ChatGPT still dominate the market, Chinese models such as DeepSeek, Qwen, and Kimi K2 are emerging as significant players, illustrating a swift shift in the competitive landscape. This growth trajectory in the AI sector underscores the urgency for the U.S. to enhance its investment in AI infrastructure to remain competitive on a global scale.