1 min read
Add as a preferred source on Google

NVIDIA CEO Huang Sees Telecom Base Stations, Data Centers Becoming AI Networks

He also estimated that a 1-gigawatt AI data center costs about $50 billion, including land, power, buildings, computing equipment, storage and cooling.

A rooftop base station rises beside a floodlit data center. / TokenPost.ai
A rooftop base station rises beside a floodlit data center. / TokenPost.ai

NVIDIA CEO Jensen Huang forecast Oct. 21 that telecom operators’ base stations and data centers could develop into distributed AI networks, outlining a possible role for telecom companies in providing AI computing capacity.

Speaking at GTC in Berlin, Huang described AI as a five-layer economy: energy, computing, AI factories, models and applications. A presentation slide displayed NVIDIA’s logo in the computing layer. Siemens Energy and Schneider Electric appeared in the energy layer, while Nscale and Firebird were listed among AI factory builders. The examples also included Mistral, Legora and DeepL, which make AI models and tools.

Huang estimated that a 1-gigawatt AI data center costs about $50 billion to build. He said the estimate includes land, power, buildings, computing equipment, storage and cooling.

Simon Yoon

Reporter

Simon Yoon reports on blockchain technology for TokenPost. Send corrections or tips to info@tokenpost.com.

Loading…