Advanced cooling technologies enable data centers to reduce energy and water consumption while supporting higher compute densities associated with AI and cloud workloads. Instead of depending primarily on conventional air conditioning systems, approaches such as free-air cooling, liquid cooling and immersion cooling transfer heat directly from chips or racks, improving thermal efficiency: 

• Free air cooling uses filtered outdoor air in suitable climates to reduce mechanical cooling loads.  
• Liquid cooling circulates coolant directly to servers or chips, improving heat transfer and lowering energy use.  
• Immersion cooling submerges IT hardware in thermally stable dielectric fluids to remove heat with very high efficiency and minimal water use. 

The International Energy Agency (IEA) reports that cooling accounts for about 7% of electricity use in very efficient hyperscale sites, compared with over 30% in conventional data centres.[1] The International Renewable Energy Agency (IRENA) similarly highlights opportunities to reduce environmental impact by adopting high-efficiency cooling and capturing waste heat for productive use.[2] Together, these technologies support the expansion of digital infrastructure in a more resource-efficient way. 

Main outcomes of the solution  

• Sustainability: Reduces energy and water consumption associated with thermal management and enables productive reuse of data-center waste heat where appropriate infrastructure exists.  
• Affordability: Lowers cooling overheads and reduces the operational costs of high-density compute environments, helping delay or avoid expensive expansions of conventional cooling systems.  
• Business Value: Supports higher compute densities and more thermally stable operation of AI and cloud workloads, improving uptime and enabling more efficient use of data-center space. 

Key implementation and scaling factors 

1. Pilot deployment in selected data-center areas to validate thermal performance, safety and integration with existing infrastructure. 
2. Use modular cooling units and standardized interfaces to support replication across facilities with varying rack densities and computing environments. 
3. Assess power, water, airflow and climate conditions—including humidity thresholds—to ensure technology suitability and avoid retrofit challenges. 
4. Align cooling systems with heat-recovery infrastructure to maximize the value of captured heat for potential district-heating or industrial applications. 
5. Collaborate closely with equipment vendors and operators to maintain compatibility with evolving server architectures, high-density AI workloads and sustainability targets. 
6. Evaluate material compatibility, maintenance needs and lifecycle performance for liquid or immersion-based systems. 
7. Integrate monitoring and data-analytics tools to track thermal performance and support predictive maintenance. 

Case Study

Polar Data Centres and Vertiv (2024), Norway

Polar Data Centers partnered with Vertiv to design and deliver DRA01, a high-density data center in Norway optimized for AI and high-performance computing workloads. The project’s design phase was completed in Q4 2024, with modular infrastructure shipped and installed in Q1 2025 and commissioning planned for completion later in 2025. The facility is designed in line with Tier III–type principles, including N+1 redundancy and concurrent maintainability. 

Powered by local hydropower, the site integrates liquid cooling and free-air cooling approaches to support high-density compute while reducing energy and water use. The cooling architecture includes glycol-free free-cooling chillers and a chilled-water network with thermal storage, alongside direct liquid cooling designed to support advanced GPU-based systems. The modular setup is designed to support rack densities of up to 110 kW, enabling a total initial IT capacity of approximately 12 MW, with provision for future expansion. 

Based on design targets and commissioning plans, the facility is expected to achieve high energy efficiency, with a projected power usage effectiveness (PUE) as low as 1.15, benefiting from Norway’s cold climate and advanced cooling design. The project illustrates how advanced, scalable cooling technologies can enable energy-efficient, high-density data-center development in regions with favorable climatic conditions and low-carbon electricity.[3] 

References: 

1. Energy demand from AI I IEA (April 2025)
2. Waste heat recovery from data centres I IRENA (June 2023)
3. Polar and Vertiv drive innovation with modular AI solution for their DRA01 data center in Norway I Vertiv (2025)