Executive Summary
- Written by Aaron Oddy, Sales Manager, Centiel, this article explores the notion that successful capacity planning and growth depend on meeting the current demand.
- Global AI computing capacity is doubling every seven months, recent studies suggest, so guaranteeing sufficient power is becoming the biggest hurdle operators must face.
- The role of UPS is continually evolving and they could potentially act as energy hubs, storing power and peaking shaving when energy costs are highest.
AI is accelerating data centre growth at an unprecedented rate. Goldman Sachs’ research forecasts that global power demand from data centres will increase by 50% by 2027, and by as much as 165% by the end of the decade. For data centre operators, capacity planning has become more important than ever. Yet ultimately, successful capacity planning still comes down to one simple thing: meeting demand.
The Demand for Power
With current estimates suggesting global AI computing capacity is doubling every seven months, ensuring sufficient power from the source is fast becoming the biggest challenge.
Uninterruptible power supplies (UPS) need to play their part. UPS systems have traditionally been used to protect critical loads from outages while delivering clean, continuous power. However, they are power hungry, therefore it is essential to maximise a UPS’ energy efficiency to help balance increasing power demand.
Right-sizing for current demand is the starting point; oversized UPS systems can be inefficient. At the same time, AI workloads are driving increases in load, meaning operators must also plan for future expansion. This is where true modular UPS systems come into their own, offering the flexibility to scale while removing single points of failure and maintaining high availability. Systems with higher power density also take up less space, helping operators make the most of existing – and often expensive – building infrastructure as they grow.
True modular UPS systems have a further benefit: they can also help facilities manage energy usage more dynamically. By adding or removing modules, or by using intelligent energy management modes, operators can optimise UPS efficiency as demand changes without impacting their availability.
Demand for Rapid Deployment
Speed of deployment is another key consideration. Data centres are under pressure to grow quickly, placing increasing strain on their supply chains. Manufacturers that can offer short lead times and support efficient builds are better positioned to help operators keep pace with demand.
To support this, UPS systems and data centres can be provided as a containerised solution, assembled off-site and delivered as modular, bolt-on designs. In this way, both the modular UPS and the wider data centre infrastructure can be designed with flexibility and rapid deployment in mind.
Tailor Systems to Meet Demand
Despite these advancements, power protection is never a one-size-fits-all solution. Every data centre has unique requirements based on its location, load profile, and growth plans. Effective capacity planning, therefore, also relies on designing solutions tailored to each site, rather than adopting a standardised approach. Although availability must always remain a high priority, depending on the tier classification for the site.
Looking ahead, the role of UPS will continue to evolve. For example, there is growing potential for UPS systems to act as energy hubs, storing power and peaking shaving when energy costs are at their highest.
Ultimately, as data centres race to support AI and other emerging technologies, capacity planning success depends on aligning infrastructure with demand, both today and over the longer term. Those who can strike that balance will be best placed to succeed in an increasingly power-hungry world.



