Lenovo puts inferencing at the centre of its AI push
Lenovo sharpens its AI strategy around inferencing, hybrid deployment and AI factory infrastructure with NVIDIA.
Lenovo has expanded its AI push with a broader set of systems aimed at a more immediate enterprise problem, how to run AI inferencing across devices, edge sites, data centres and cloud environments without fragmenting deployment. Announced at NVIDIA GTC, the latest Lenovo Hybrid AI Advantage with NVIDIA package spans workstations, servers, hybrid platforms and large-scale AI cloud infrastructure.
Table Of Content
The company is placing inferencing, rather than model training, at the centre of the pitch. That changes the conversation from raw model capability to operational readiness, response time and infrastructure choices across mixed environments. Lenovo said the expanded portfolio is intended to reduce time-to-first-token and support real-time decision-making in production settings.
Inferencing becomes the centre of the pitch
Inferencing sits at the core of Lenovo’s enterprise AI pitch, shifting attention to execution rather than model development. Its CIO Playbook 2026, conducted with IDC, found that 84% of organisations expect to run AI across on-premises or edge environments alongside the cloud.
That framing gives Lenovo room to argue for a hybrid AI model rather than a cloud-only one. “Together, Lenovo and NVIDIA are uniquely positioned to help organizations operationalize AI—from experimentation to enterprise production to AI cloud gigafactories,” said Yuanqing Yang, Chairman and CEO, Lenovo. “As agentic AI drives exponential growth in inferencing workloads, cost control and performance per token become mission critical. By combining NVIDIA AI Enterprise software with Lenovo’s full-stack hybrid AI platforms and services, we enable customers to scale AI with greater efficiency, lower cost per token, and faster time-to-production.”
A full stack from deskside to data centre
Lenovo’s product set now stretches from mobile and desktop workstations to ThinkSystem and ThinkEdge servers, hybrid AI platforms and AI developer tools. On the device side, the company introduced new NVIDIA RTX Pro Blackwell-based workstation options, alongside ThinkStation configurations aimed at private, on-premises AI development and inferencing.
In the data centre, Lenovo is pairing inferencing-optimised infrastructure with NVIDIA AI Enterprise software and partner platforms from companies including Nutanix, Cloudian and Veeam. The company said its Hybrid AI Advantage with NVIDIA solutions are delivering return on investment in less than six months and up to 8x lower cost per token than comparable cloud IaaS, with the aim of giving enterprises more control over where workloads run and how models are protected.
That emphasis on operational control also carries into industry deployments. Lenovo said the expanded AI Library now includes agentic and physical AI offerings for retail, sports, manufacturing, industrial and mobility use cases, with functions ranging from in-store assistants and broadcast optimisation to inspection, safety and fleet intelligence.
The cloud story now reaches AI factories
Lenovo is also using the announcement to move further into AI cloud infrastructure. As a launch partner for NVIDIA Vera Rubin NVL72, the company is introducing liquid-cooled rack-scale systems for hyperscale and sovereign AI cloud providers, alongside NVIDIA HGX Rubin NVL8 systems.
Here, the argument is focused on economics, deployment speed and lifecycle management for operators building large AI environments. Lenovo said these systems can deliver up to 10x higher throughput and up to 10x lower cost per token than previous generations, while its Hybrid AI Factory Services are meant to help customers move more quickly from infrastructure build-out to commercial use.
“For enterprises, real-time intelligence now depends on infrastructure choices made much earlier in the stack,” said Sumir Bhatia, President, Asia Pacific, Infrastructure Solutions Group, Lenovo. “With NVIDIA, Lenovo is combining its full-stack infrastructure, services, and offerings to give enterprises the agility to scale AI across cloud, data center, and edge.”
