Lenovo expands hybrid AI portfolio to reduce enterprise inference costs by up to 8x through optimized CPU/GPU co-processing.

On June 24, 2026, Lenovo expanded its Hybrid AI Advantage platform with new inferencing and agentic AI capabilities designed to help enterprises deploy AI across devices, edge environments, data centers, and cloud infrastructure. The company said the additions are aimed at improving AI economics, accelerating deployment, and supporting autonomous AI workloads.

According to Ashley Gorakhpurwalla, President of Infrastructure Solutions at Lenovo, 94% of organizations are planning to increase their AI investment over the next year, with enterprises moving beyond AI experimentation and demanding measurable business outcomes. He noted that Lenovo's infrastructure helps businesses deploy AI where it creates the most value, lowering data processing costs, speeding up deployment, and maintaining strict corporate security and governance.

The expanded portfolio is built with technologies from NVIDIA, Intel, Red Hat, and Canonical and forms part of Lenovo's Hybrid AI Factory approach, designed to support AI deployment across hybrid environments. Lenovo introduced two new Hybrid AI Platform configurations. The first is a CPU-only platform powered by Intel Xeon 6 processors and Red Hat AI Enterprise, designed to process approximately 2x more AI requests concurrently and support workloads such as retrieval-augmented generation (RAG), customer service, and human resources applications. The second is the Hybrid AI Platform (221), available with either Canonical Ubuntu and Kubernetes or Red Hat AI Enterprise, deployable in as little as a few weeks and designed to support enterprise AI workloads with governance and data sovereignty requirements.

Industry research shows that 92% of organizations deploying agentic AI report costs exceeding expectations. Lenovo's total cost of ownership analysis found that on-premise deployments can deliver up to 8X lower cost per token than cloud infrastructure-as-a-service environments and up to 18X lower cost per million tokens than model-as-a-service APIs for sustained CPU and GPU workloads.

The company also announced one-click deployment capabilities for AI agents across desktop and data center environments. Lenovo's Knowledge Super Agent use cases, validated through its AI Library, have demonstrated savings of thousands of employee hours per organization. Additionally, Lenovo is co-developing AI agents and skills for NVIDIA NemoClaw with enterprise customers and plans to extend agentic AI capabilities into retail through an AI-powered kiosk designed to assist shoppers in stores.