InfrastructureNews US - America's infrastructure news for industry decision-makers
United States
Schneider Electric backs liquid cooling software for AI

Schneider Electric backs liquid cooling software for AI

Wed, 18th Nov 2026 (Yesterday)
Amir Ibraheem
AMIR IBRAHEEM Cooling Solution Architect Schneider Electric

For AI factories, liquid cooling has become a thermodynamic necessity rather than a design choice. As rack power densities continue to rise, the challenge is no longer whether to deploy liquid cooling, but how to orchestrate cooling infrastructure at scale to optimize cost, efficiency, resilience, and GPU utilization.

Liquid cooling orchestration offers the ability to manage cooling as a system rather than as a collection of individual assets, which helps ensure that cooling capacity efficiently matches AI workloads at any given time, while minimizing waste. In doing so, it can deliver cost savings by improving power usage effectiveness (PUE) and water usage effectiveness (WUE). What's more, orchestration software can help optimize data center design and construction costs by giving owners more accurate data on how much power and water they'll need for cooling in the facility in any specific location.

Liquid cooling orchestration transforms cooling from a collection of interdependent systems managed through separate local controls into a coordinated thermal management platform for the AI factory. By maintaining the required heat rejection and thermal headroom as workload conditions change, operators can maximize GPU performance, improve cooling-capacity utilization, and reduce unnecessary energy or water use.

Orchestration enables operational intelligence, which improves power usage effectiveness (PUE) and water usage effectiveness (WUE) and drives real-time analytics and diagnostics across cooling, power, and workload conditions. It also enables autonomous optimization of thermal performance, energy efficiency, water consumption, and capacity utilization.

The result is a more resilient, sustainable, and economically optimized AI factory capable of supporting today's workloads as well as next-generation GPU deployments at scale.

What is liquid cooling orchestration and what are its benefits?

AI factories typically use a hybrid thermal architecture that combines liquid-to-chip cooling infrastructure for high-density AI workloads with air-cooling systems for building and other IT equipment. The cooling ecosystem may include coolant distribution units (CDUs), pumping systems, liquid distribution networks, CRAHs (computer room air handlers), fan walls, chillers, dry coolers, cooling towers, and adiabatic heat rejection technologies.

AI workloads can fluctuate rapidly, but the thermal, hydraulic, and mechanical systems that support liquid cooling operate on much slower timescales. As a result, liquid cooling orchestration aims not to react to individual GPU load changes, but to continuously optimize the end-to-end cooling ecosystem so sufficient cooling capacity is available when needed while minimizing energy, water, and infrastructure overhead.

The idea is to maintain thermal headroom and optimize supervisory setpoints over seconds-to-minutes and minutes-to-hours, depending on the equipment involved. Fluid volume, metal mass, and allowable temperature rise provide a temporary thermal buffer while slower equipment responds.

Orchestration software also continuously learns from historical and real-time data, then coordinates chillers, heat rejection equipment, CDUs, pumps, and airflow systems to achieve system-wide outcomes. Predictive analytics can help anticipate sustained thermal demand, identify abnormal conditions, and recommend supervisory adjustments to maximize overall facility efficiency.  

How do I manage mixed air and liquid environments?

Cooling orchestration software ingests real-time telemetry and operational data from both liquid and air-cooling infrastructure. It provides a unified operational view of the thermal infrastructure, enabling coordinated control and optimization across both air and liquid cooling domains. This is crucial because AI environments often consist of multiple server platforms with different thermal profiles and cooling architectures.

How do I optimize cooling across multiple sites?

As AI factories expand across regions and geographies, cooling optimization increasingly becomes a fleet management consideration. Cooling orchestration software provides a centralized view of thermal performance, infrastructure utilization, energy efficiency, and water consumption across sites, enabling operators and stakeholders to benchmark these AI data centers, identify underperforming assets, and standardize best practices. By applying lessons learned from high performing data centers across the broader fleet, operators can improve operational consistency, maximize cooling infrastructure utilization, and reduce the total cost of operating large scale AI deployments.

How do I coordinate thermal and compute workloads?

Cooling orchestration software uses AI to add intelligence around cooling capacity decisions. For NVIDIA GPUs, for example, the software considers MAX-LPS, MAX-P, and MAX-Q Profile configurations to meet the desired performance requirements with "just right" cooling capacity.

Intelligent software continuously monitors real-time GPU and rack power telemetry as a leading indicator of heat load, correlated with manifold return temperature and flow. Using predictive AI models, it anticipates workload driven thermal demand and proactively recommends or automatically adjusts CDU and chilled water plant setpoints, so cooling capacity is in place before heat reaches the coolant loop.

How can liquid cooling orchestration improve data center efficiency?

As liquid cooling becomes more prevalent in AI data centers, hyperscalers, colocation providers, neoclouds, and other data center operators have an opportunity to improve efficiency by managing cooling as a coordinated system rather than as individual components.

Local CDU control loops hold supply temperature and differential pressure within expected operational ranges, while orchestration operates at the supervisory layer. Fully automated, or operator approved, orchestration can continuously reset setpoints across the heat rejection plant, chillers, CDUs, and secondary fluid network to keep adequate thermal headroom at the cold plate with the least energy and water use. Managing the full heat transfer chain as one system, from the cold plate to heat rejection, lets operators run warmer supply temperatures, gain more economizer hours, cut pumping energy, and recover stranded cooling capacity without compromising thermal resilience or GPU performance.

Liquid cooling orchestration can also extend beyond day-to-day operations. By providing greater visibility into actual cooling requirements and performance across sites, cooling orchestration software can help operators optimize capacity, identify best practices, and make more informed infrastructure investments. This can help avoid unnecessary capital expenditures, reduce operating costs, and support more efficient, scalable AI infrastructure.

Learn more about how Schneider Electric is helping data center operators address evolving AI cooling requirements on our liquid cooling resource page or contact us to discuss your cooling strategy.