Effective site selection begins with understanding and defining the operational requirements of a data centre. This involves quantifying the infrastructure needs, such as network density and cabinet counts. This is necessary to design a facility that can operate and scale without future disruption.

Working with an experienced firm that can help judge the feasibility and stability of different locations can be critical. This process involves considering land-use planning, zoning constraints, seismic risks and access to energy and water infrastructure. The JLL Global Data Center Outlook notes that "speed to power" has become the primary site selection criterion, followed by “community support, latency and proximity to customers.”

It’s important to consider all these factors and more when scouting locations for data centres. Aligning the master plan with an organisation’s broader business goals offers higher chances of success.

It’s often necessary to navigate complex local legislation to identify and obtain various project approvals. For example, this might involve securing high-capacity electrical connections through licensed electrical workers and managing the long lead times associated with power utility negotiations.

These lead times can be a major source of frustration for project managers and leadership. As an example, the estimated lead times for a new 66kV connection to an existing substation in Singapore are 26 months, and 28 months if a new substation is required. However, a whole-of-project, strategic approach to the approvals process can help manage risks and prevent them from hindering the project timeline.

Likewise, a constructive approach to stakeholder engagement can be hugely beneficial. For example, working with a team that has experience in consulting communities and addressing environmental and social impacts can help in the approvals process and avoid a potentially negative public relations scenario.

Data centre operations aren’t stagnant, and neither is the demand for data storage, cloud computing and AI applications. When designing data centres, it’s advantageous to anticipate the rapid growth of high-density AI workloads. By designing with modularity in mind, facilities can incrementally add compute and cooling capacity, avoiding wasted capital expenditure. 

By embracing a smart-building philosophy, companies can benefit from integrating adaptable modular systems, enabling a responsive facility that can react to real-time changes in energy demand and occupancy.

A cloud provider’s services may be ethereal and intangible, but while the cloud is virtual, the facilities that support it are very real. Data centres aren’t islands operating in their own worlds. They require access to, and protection from, a laundry list of physical structure and environmental elements.

That’s not to say, though, that self-sufficient systems are impossible or not beneficial. For example, the Airways Wellington Control Tower integrates independent data centre infrastructure with a self-sufficient water supply and standby power. Beca supports Smart Buildings that link previously isolated systems, such as HVAC, lighting and power, into a cohesive, data-driven platform that facilitates the more efficient usage of both energy and water.