Lenovo’s Liquid-Cooled AI Servers Push Enterprise Computing Forward

AI and high-performance computing are moving into the same infrastructure conversation, and Lenovo is responding with a broad new portfolio of servers, storage systems, cooling technology, and deployment services. The goal is clear: help organisations turn demanding AI inference, agentic AI, and scientific computing workloads into working systems across on-premises data centres, hybrid environments, and colocation facilities.

At the centre of the announcement are new Lenovo ThinkSystem platforms powered by AMD’s 6th Generation EPYC 9006 SP7 processors. Lenovo is also bringing liquid cooling, flexible server partitioning, high-throughput storage, and infrastructure validation into the same portfolio, giving customers more ways to match computing resources to the work in front of them.

Liquid Cooling Takes Aim at AI’s Power Challenge

The Lenovo ThinkSystem SC755 leads the hardware launch with a dual-socket design and Neptune liquid cooling. The system can remove up to 100% of system heat through water cooling, while Lenovo claims up to 40% higher sustained performance compared with similar air-cooled systems.

That focus addresses one of the biggest infrastructure demands surrounding AI and HPC: improving performance per watt while managing power and cooling constraints. As workloads grow more complex, the ability to control heat becomes part of the performance equation, not a separate data centre concern.

The SC755 also includes Lenovo Dynamic Partitioning, which lets customers convert two dual-socket systems into single-socket configurations and switch them back when workload needs change. That flexibility gives organisations a way to configure server capacity around current demands instead of locking every system into one fixed arrangement.

Lenovo is also positioning the ThinkSystem SR755 for large-scale virtualisation, engineering, financial modelling, electronic design automation, and agentic AI infrastructure. Together, the new platforms target a wide range of workloads, from scientific research and enterprise applications to systems that support AI agents.

Storage and Interconnects Build the Wider Platform

Compute power is only one part of an AI or HPC environment. Lenovo has introduced the ThinkSystem DE6800 and DE6500 storage systems for high-throughput, latency-sensitive block workloads, giving customers new options for moving and managing data alongside the company’s server platforms.

• ThinkSystem DE6800: Designed for applications that require high-throughput storage.
• ThinkSystem DE6500: Focused on performance with simplified deployment and management.
• Cornelis CN5000: Validated as an interconnect option for AI and HPC environments.

The new ThinkSystem platforms use AMD’s 6th Generation EPYC server processors, which AMD says are designed to cover agentic AI, enterprise, cloud computing, HPC, and general-purpose computing through workload-optimised server CPUs powered by the AMD “Zen 6” architecture.

“The new Lenovo ThinkSystem platforms powered by 6th Gen AMD EPYC Server CPUs help customers accelerate demanding workloads, from engineering and scientific research to AI inferencing and agentic AI,” said Dan McNamara, Senior Vice President and General Manager, Compute and Enterprise AI, AMD.

6th Gen AMD EPYC SP7 CPU-based platforms are now shipping from leading server OEMs and ODMs, and the processors are expected to power cloud instances across all major, global cloud service providers. Recent announcements from Microsoft Azure and OCI also showcase leadership agent orchestration, concurrent agent sandboxes, and web and enterprise applications.

Services Extend From Planning to Colocation

Lenovo is pairing the hardware with expanded Hybrid AI Factory services to help organisations deploy and manage AI infrastructure. The services include GPU deployment, configuration, and optimisation capabilities, with multi-node support for Red Hat AI Enterprise and NVIDIA AI Enterprise with NVIDIA Run.

For large-scale AI deployments, Lenovo is expanding its gigafactory services with NVIDIA Infrastructure Services expertise for specialised deployment and cluster validation. The company says the services cover the infrastructure lifecycle, from planning and construction to deployment and optimisation.

A new partnership with Digital Realty adds access to a global network of colocation facilities. Customers can use Digital Realty’s Innovation Lab to validate AI infrastructure, applications, and workload requirements before deployment, creating a step for testing systems before they move into production environments.

Lenovo will provide a single point of contact for colocation requirements, along with infrastructure services and financing options through its TruScale infrastructure-as-a-service offering. That approach connects hardware, services, facilities, and financing in one deployment path.

“AI and high-performance computing are converging around a simple customer need: the ability to turn increasingly complex workloads into real outcomes,” said Scott Patti, Vice President, Infrastructure Solutions Group, Lenovo. He added that organisations need to improve performance per watt, address power and cooling constraints, and reduce the time required to move from infrastructure deployment to operational AI.

That pressure is also visible in Lenovo’s sixth Work Reborn report, The AI-ready workplace. Based on an August 2026 survey of 1,200 IT leaders at organisations with at least 1,000 employees across 12 countries, the report found that 80% of IT decision makers say the digital workplace is critical to successful AI-led transformation, yet only 10% have fundamentally redesigned it for AI.

Infrastructure will remain one part of that transformation, but the direction is already clear: AI systems must deliver measurable results across real workflows. Lenovo’s new servers, storage systems, cooling technology, services, and colocation options aim to move that work from planning into operation as organisations scale AI and HPC across more environments.

Stay connected via Google News

Follow us for the latest AI News