Authored by Simon Harris, Director of Critical Infrastructure at BCS Consultancy
One of the questions we’re increasingly asked by operators, investors and asset owners is whether an existing data centre is still fit for purpose.
The question is usually prompted by a specific challenge. Sometimes it’s driven by AI deployments and higher rack densities. Sometimes it’s part of a due diligence exercise during an acquisition. Sometimes it’s because cooling systems are beginning to struggle during periods of sustained high temperatures. Occasionally it’s simply because the facility is ten or fifteen years old and someone has started referring to it as a ‘legacy’ site.
What’s interesting is that the age of the building is rarely the factor that determines the answer. Many of today’s existing data centre facilities have already adapted to several generations of technology. They’ve accommodated changing server platforms, increasing power densities, evolving cooling strategies and different customer requirements throughout their operational lives. In many cases, they’re well positioned to do exactly the same again. Our work increasingly centres on helping operators upgrade, expand and modernise existing facilities rather than replacing them entirely.
Which begs the question: are we as an industry becoming too quick to label facilities as legacy? The term itself has gradually become shorthand for something that’s outdated or approaching the end of its useful life. I don’t think that’s a particularly helpful definition anymore.
Technology now moves at a very different pace to buildings. A data centre may take four or five years to progress from concept through planning, design and construction before it becomes operational. During that same period, processor architectures change, rack densities increase, cooling technologies develop and customer requirements continue to evolve. If we define a legacy facility simply as one that wasn’t designed around today’s technology, then almost every operational data centre is, to some extent, a legacy asset.
That isn’t a criticism of the building. It’s simply recognising that digital infrastructure has always been designed to evolve. What has changed over the past two years is the pace of that evolution.
Much of the industry’s attention has understandably focused on the construction of new AI-ready facilities. However, a significant proportion of the work happening across the market today isn’t taking place on greenfield sites. It’s happening inside live operational facilities.
Across the industry, operators are assessing how existing electrical infrastructure can support greater loads, whether cooling systems can be upgraded for higher-density deployments, where operational efficiencies can release additional capacity and how established assets can continue supporting changing customer requirements. Those conversations reflect a growing recognition that existing estates often represent some of an operator’s most valuable assets. That value extends well beyond the building itself. Increasingly, the conversation is moving beyond whether an existing facility can be upgraded, to whether doing so represents the stronger commercial proposition.
Many existing facilities occupy locations that would be extremely difficult to develop today. They already have planning consent, established fibre connectivity, experienced operational teams and valuable power allocations. In some cases, they also contain electrical capacity that has never been fully utilised. Given the increasing difficulty of securing new power, identifying that untapped capacity can significantly increase the value of an existing facility. Unlocking it through targeted engineering improvements can create expansion opportunities that offer a significant speed to market advantage over new build developments, allowing operators to achieve faster revenue commencement while others are still progressing through planning, utility connections and construction.
Looking at existing assets differently also changes the sustainability conversation. Much of the discussion around sustainable data centres quite rightly focuses on renewable energy, operational efficiency and lower carbon operations. Those issues remain critically important. However, there is another aspect that deserves greater consideration: making better use of the infrastructure that already exists.
Every operational facility represents a significant investment in materials, engineering and construction. Extending the life of those assets through targeted upgrades and refurbishment can avoid unnecessary demolition, reduce embodied carbon and improve operational performance at the same time. Modernisation should not be viewed as a compromise to new development. In many cases, it represents the most efficient use of both capital and resources.
Perhaps this is where the industry’s thinking needs to evolve. Data centres no longer need to be viewed as either new or legacy, as though those are fixed categories. In reality, every operational facility sits somewhere on a continuous cycle of improvement. Cooling systems are upgraded. Electrical infrastructure evolves. Mechanical plant is replaced. Operational strategies change. None of those activities suggest a building has reached the end of its useful life. They simply reflect the pace at which the industry develops.
The question therefore isn’t whether a facility is legacy. The more useful question is whether it still has the ability to evolve. The experience of operators across the market suggests that many existing assets have considerably more potential than their age alone would imply.
As demand continues to grow, the industry will undoubtedly need to build substantial new capacity. That isn’t in question. What deserves more attention is the role existing infrastructure will play in meeting that demand.
Some of the most commercially attractive opportunities over the next decade will not necessarily come from new sites. They will come from existing assets that can be expanded more quickly, brought to market sooner and begin generating revenue significantly earlier than equivalent greenfield developments. Operators who understand the potential within their existing estates will be better positioned to respond to demand while others are still progressing through planning and construction. The question is no longer whether existing assets have a role to play in the industry’s future, but how much value remains to be unlocked within them.
