Data centre site selection: criteria for successful development

First published in Environment Analyst’s Corporate guide: Delivering sustainable data centres.

Selecting the right location for data centres is a critical step in ensuring project success. As AI-powered solutions drive unprecedented growth in the sector, careful site selection is more important than ever. Neglecting early-stage assessments can lead to costly delays and wasted resources, making due diligence essential for commercial certainty and long-term efficiency.

The importance of site selection

Global demand for data centre power is surging, with key industry players investing heavily to meet growing needs. From 2023 to 2028, data centre computing power is expected to grow at a compound annual growth rate (CAGR) of 16%. Generative AI workloads are the fastest-growing segment, projected at a 65% CAGR, while traditional enterprise workloads will still account for about 55% of demand by 2028. Hyperscalers are set to generate around 60% of industry growth, increasing their share of global demand from 35% to 45%. The US market will remain dominant, growing at 18% CAGR.

Figure 1: Data centre and related power growth

How to evaluate a site for your data centre deployment

1. Technical considerations

Securing reliable power sources:
Reliable and abundant power supply is the backbone of any data centre. Sites must have access to scalable power grids or on-site generation to support current and future energy needs. Evaluating proximity to power plants and renewable energy sources enhances sustainability and cost-efficiency. Hyperscale cloud data centres typically consume over 100 megawatts (MW), while AI data centres may require upwards of 200 MW. Some clients are seeking up to 600 MW, underscoring the need for robust infrastructure. Innovative solutions include microgrids, battery energy storage systems (BESS), small modular nuclear reactors (SMR), fuel cells, and natural gas plants.

Network connectivity and latency optimisation:
High-speed, reliable broadband is essential for data centre operations. A location’s proximity to major network hubs and fibre routes impacts latency and performance, especially for cloud data centres. For AI data centres, latency is less critical, allowing for more flexibility in site location.

Security measures – physical, network, and data security:
Physical and cyber security are paramount. Sites should offer robust surveillance, access control, and perimeter security, as well as proximity to emergency services. Cybersecurity infrastructure must protect sensitive information.

Proximity to end-users and industry ecosystems:
Strategic site location near urban centres and industry clusters minimises latency and enhances connectivity. Hyperscale and colocation data centres are often built in metropolitan regions but are expanding into secondary and remote areas, benefiting from consistent temperatures, water availability, and renewable energy.

2. Environmental and regulatory considerations

Zoning and permitting:
Zoning laws and permitting processes ensure compatibility with local environments. Developers must navigate regulations on building height, setbacks, air and noise levels, and secure necessary permits early to avoid delays. In Europe, Directive (EU) 2024/2881 focuses on air quality for data centres, influencing site selection and technical design. In the US, new regulations like Virginia’s SB 1046 mandate strict noise abatement and public notification. Cities such as Phoenix have adopted specific ordinances, signalling a trend toward localised regulation.

Proximity to coolant infrastructure and water:
Efficient cooling is vital for data centre performance. Proximity to chilled water plants and sustainable water sources reduces costs and supports long-term operations, especially in regions with water scarcity or strict policies.

Local climate and sustainability:
Cooler climates lower cooling costs, while warmer regions increase energy expenses. Understanding local climate impacts is crucial for operational efficiency and sustainability.

Natural disaster risk:
Sites should be located in areas with minimal risk of earthquakes, floods, storms, and other disasters. Comprehensive risk assessments ensure resilience and uninterrupted operations. Ramboll’s assessments show some regions are unsuitable for hyperscale data centres due to climate risks.

3. Social considerations

Community support:
Community engagement is increasingly pivotal. Data centre projects often face resistance due to environmental concerns, known as “data centre NIMBYism”. In Virginia, $900 million in projects have been blocked, with $45.8 billion delayed. Early stakeholder engagement builds trust and secures approvals.

Stakeholder engagement:
Transparent communication, public consultations, and community benefit programmes foster goodwill. Partnerships with municipalities, especially those transitioning from industrial economies, can facilitate long-term success. Ramboll’s experience shows positive relationships with local authorities can be decisive.