Particle.news
Download on the App Store

At CES 2026, Lenovo and Nvidia Launch Gigawatt-Scale AI Gigafactory

The effort pairs hyperscale build-outs with supply-chain hedges, coinciding with CES demos that blur lines between concepts and production.

Overview

  • Lenovo and Nvidia introduced the Lenovo AI Cloud Gigafactory to speed deployment of gigawatt-scale infrastructure for AI service providers.
  • The plan integrates Lenovo Neptune liquid-cooled systems with Nvidia platforms, highlighting GB300 NVL72 with 72 Blackwell Ultra GPUs and the new Vera Rubin NVL72 for training and inference.
  • Lenovo CEO Yang Yuanqing set a target to quadruple the companies’ collaboration within three to four years to compress time to first token and scale to 100,000 GPUs.
  • Nvidia executives said large prepayments, direct DRAM procurement and long-term supplier agreements have largely insulated its AI hardware supply from ongoing memory shortages.
  • CES showings underscored commercialization and caution: Lenovo displayed a 45-gram, 8-hour AI glasses prototype and a personal AI hub, ByteDance’s Doubao denied imminent AI glasses shipments, Bosch pledged over €2.5 billion for AI through 2027, and Chinese firms RoboSense and Black Sesame demonstrated autonomous robotics and automotive AI platforms.