Lenovo has announced the Lenovo AI Cloud Gigafactory with NVIDIA at Tech World at CES 2026 in Las Vegas. The companies say the program targets AI cloud providers building gigawatt-scale “AI factories,” with an emphasis on reducing time-to-first token (TTFT) and speeding deployment of production-ready AI services for next-generation workloads.
Lenovo says AI cloud providers are under pressure to build data center infrastructure that can support “trillion-parameter agentic AI, physical AI, and high-performance compute (HPC) applications.” The company describes TTFT as a benchmark that reflects how effectively AI factories deliver massive compute, high-performance storage, ultra-low-latency networking, and AI software to meet token demand. Lenovo says the program aims to help providers reach TTFT “in weeks” using ready-to-use components, expert guidance, and “industrialized build processes.”
The announcement centers on an integrated package of infrastructure, manufacturing, and services: Lenovo Neptune liquid-cooled hybrid AI infrastructure, NVIDIA accelerated computing platforms, Lenovo’s global manufacturing, and full lifecycle Lenovo Hybrid AI Factory Services. Lenovo also positions the program as a path to “seamless scaling to millions of GPUs for next-generation workloads,” and as a way for providers to move from concept to build, deployment, and monetization.
On the platform side, Lenovo says the program provides access to NVIDIA Blackwell Ultra architecture, including the NVIDIA GB300 NVL72 system from Lenovo. Lenovo describes GB300 NVL72 as a fully liquid-cooled, rack-scale architecture integrating 72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace CPUs in a single platform. Lenovo also says the program will support the NVIDIA Vera Rubin NVL72 system for AI training and inference, described as unifying 72 Rubin GPUs, 36 Vera CPUs, ConnectX-9 SuperNICs, BlueField-4 DPUs, and Spectrum-X Ethernet in a rack-scale AI supercomputer; networking options listed include NVIDIA Spectrum-6 Ethernet switches and NVIDIA Photonics Ethernet switches.
Lenovo adds that “AI-native platforms” and “repeatable Lenovo AI Library use cases” integrated with NVIDIA AI Enterprise, including the open Nemotron models, are intended to simplify delivery of specialized horizontal and vertical AI workloads at scale. Lenovo also states it powers eight of the world’s top ten public cloud providers and is “the only company offering fully in-house design, manufacturing, integration, and global services for custom AI cloud solutions.”
Source: Lenovo
