Too much haste. Not enough speed: why data centre delivery needs a strategic reset through modularisation

Data centre delivery is under real strain.

Growth is accelerating, driven by AI and rising digital demand. Power demands are increasing, pushing more capacity into the same footprint, and adding heat and complexity to delivery. Teams are then revisiting designs mid-project to keep pace. Access to power is becoming a major constraint, shaping what can be delivered and when. Despite this, assets are still expected to come online quickly and with a high degree of delivery certainty.

The industry has responded by trying to move faster within the same delivery model. Projects start quickly, teams work at pace and decisions are made continuously. Yet delivery timelines have not improved.

Most data centres still take between 20 and 36 months to build. Throughout the build, costs continue to rise, skilled labour is limited in many markets, and lead times remain unpredictable. These pressures are not easing, and they are beginning to expose the limits of current delivery approaches.

The issue sits earlier in the process, in how projects are set up and how decisions are made.

Busy is not the same as fast

Most projects feel fast, at least on the surface. But, much of this activity is driven by decisions that were not made early enough. In a bid to get moving as quickly as possible, key issues get pushed downstream, design continues to evolve during construction, and problems are solved on site rather than prevented at the start.

Changes that would have been straightforward early on become more complex once delivery is underway, affecting multiple trades, interfaces and workstreams at once. The programme absorbs the impact, often without fully recovering.

Projects can move constantly and still fall behind. Over time, this pattern becomes embedded, with project teams adapting to change rather than removing it.

The industry is protecting certainty in the wrong place

The sector is focused on certainty, and with good reason. Data centres are high-value assets, and delivery failure carries significant consequences. Established delivery models offer reassurance, and therefore remain in place, but they are not designed for the speed, complexity and changing demands of today’s technology projects.

That instinct for familiarity and drive for certainty is understandable, but I would argue it is being applied in the wrong place.

The delivery environment has changed. Requirements are changing faster, labour is harder to secure and supply chains are less predictable as geopolitical forces continue to shape the operating environment. In this context, holding on to familiar delivery models can introduce risk rather than reduce it.

Instead of rethinking the model, teams force flexibility elsewhere. They adjust designs, allow scope to evolve and delay critical decisions. From a career spent working on complex projects, I know that by the time those issues are resolved, they carry cost and delay that is difficult to recover.

The industry is holding onto familiar delivery models and absorbing the consequences, often without realising the cumulative impact. Certainty comes from early, well-informed decisions and a clear delivery strategy, not from preserving the status quo.

Modularisation is already in use, but not early enough

Modularisation is already part of the data centre sector. Prefabricated elements such as power skids, cooling units and plant rooms are common on many large projects and bring clear efficiency gains. Increasingly, systems are assembled and tested off site before delivery, reducing commissioning risk and simplifying installation on site.

Other sectors have taken this further. Semiconductor facilities rely on prefabricated systems to deliver complex environments with fewer specialists on site, while healthcare and life sciences projects use modular approaches to meet tight timelines in demanding conditions. In oil and gas, large-scale modules are routinely assembled off site and shipped or floated into position, reducing on-site complexity in some of the world’s most challenging environments.

However, in many data centre projects, modularisation is limited to individual components rather than shaping the delivery strategy as a whole. By the time it is considered, decisions on design, procurement and sequencing have already been made.

Projects that introduce modularisation late tend to see only marginal gains. By that stage, the delivery model has already been set. Greater control comes when modularisation is considered early and used to define the delivery approach. To do this, its important to understand the implications early. With the right data and analysis, teams can assess how modularisation will affect cost, programme, risk and flexibility before delivery begins, making it easier to commit to a clear approach.

The case for change is clear

The pressures on delivery are building across the system.

Currie & Brown’s Construction Certainty Index shows that the most significant risks to project delivery are spread across multiple fronts. Material cost inflation (73%), supply chain disruption (70%) and energy price volatility (68%) are all cited by technology sector decision makers as having a high impact on their ability to meet project goals.

Labour remains a critical constraint. Across the sector, 66% of respondents identify labour and skills shortages as having a high impact on their ability to achieve delivery goals. Within data centre projects, that pressure is more acute. 74% of data centre leaders report a high impact from labour shortages on their ability to deliver. A third of projects have already been delayed as a result, and more than half of industry leaders expect the situation to worsen over the next two years.

Taken together, these pressures are reshaping how projects must be delivered.

They influence decisions day-to-day, increasing reliance on reactive adjustments and making it harder to maintain control as delivery progresses. Changing the design during construction does not resolve these issues. In fact, it makes it harder to keep the project on track.

Modularisation changes the point of control

Modularisation is often described in terms of speed or efficiency, but its real value lies in control over delivery.

It brings key decisions forward. Teams are required to define what must be fixed, where design efficiencies can be achieved, what can be standardised and where flexibility should sit, before delivery begins. That changes how projects unfold.

More work can move off site into controlled environments, reducing congestion and reliance on scarce labour. At the same time, some modular packages can be designed and procured earlier, allowing manufacturing and site activities to progress in parallel. This supports faster delivery without increasing complexity.

Working in controlled manufacturing environments can also improve consistency, reduce material waste and limit some of the safety risks associated with highly congested construction sites.

Fewer unknowns reach the construction phase, and teams spend less time reacting to issues that could have been resolved earlier. That reduces the need for adjustments during delivery. Digital tools such as BIM can further support modular delivery by helping teams align designs earlier and resolve issues before components are manufactured and delivered to site.

Supply chains can also be structured more deliberately, with production located in logistical hubs where skills and capacity exist. This reduces pressure on local markets and improves resilience across the project.

Modular approaches also support phased delivery. Capacity can be brought online in stages, allowing investment to be deployed with greater precision and reducing the risk of overbuild. This allows infrastructure to scale progressively as demand grows.

What we increasingly find is that modular approaches can reduce delivery timelines by 10–30%. But achieving those gains depends on making earlier decisions and structuring the supply chain effectively from the beginning.

Certainty and speed can work together

There is a perception that modularisation limits flexibility, particularly where projects are complex or requirements are still evolving.

In practice, stronger outcomes come from deciding early what needs to remain flexible and what can be fixed. When those boundaries are clear, delivery becomes more predictable and easier to manage.

Continuous adjustment during construction tends to have the opposite effect. It introduces uncertainty, disrupts sequencing and makes progress harder to maintain.

Modularisation supports a more disciplined approach. It encourages earlier planning, clearer scope and stronger alignment across teams. It also brings suppliers into the process sooner, when their input can shape the delivery strategy rather than react to it.

This creates a more stable path through delivery, with fewer surprises and greater confidence in outcomes.

Modularising for power certainty

Small modular reactors are one example of how this thinking is beginning to extend beyond the building itself and into wider infrastructure strategy.

Power availability is becoming a major challenge for data centre delivery. Securing enough capacity, quickly enough, is shaping decisions on location, scale and programme, while also increasing pressure on local infrastructure and communities. Modular approaches may offer a more flexible and scalable way to respond.

Some data centre developers are already exploring modular captive power solutions, including small modular reactors, as part of longer-term capacity planning. These approaches may support phased growth, improve resilience and provide greater flexibility in locations where power infrastructure is constrained or still developing.

Evidence from other sectors

I’ve seen this approach deliver strong results in projects where speed, quality and delivery certainty are critical.

One example is InnoCell in Hong Kong, the city’s first multi-storey building delivered using Modular Integrated Construction (MiC). By moving major construction activities off site, the project reduced exposure to labour shortages, weather delays and on-site inefficiencies.

The building was completed in 13 months, around five months faster than traditional construction methods would typically allow. Projects like this show how earlier planning and more off-site delivery can create a more controlled path through construction.

Where modularisation is most effective

The value of modularisation depends largely on when it is considered.

It works best where there is repeatability, where scope can be defined early and where speed, labour availability or location are key constraints. It is particularly relevant in markets where access to skilled labour is limited or where delivery needs to be scaled quickly.

It is less suited to highly bespoke developments or projects where late design change is unavoidable.

The key is to assess it early, when it can still influence the delivery strategy. Teams can then decide whether it supports the specific needs of the project.

Where to start

Better outcomes start with the right questions, asked early enough to shape the project approach. Answering these questions with robust data and analysis gives teams the confidence to make early decisions and move forward with certainty.

• What matters most: speed, cost certainty, flexibility or scale?
• What needs to be flexible, and what can be standardised?
• Are design, procurement and construction teams working to the same delivery strategy?
• Where are the biggest risks around labour, power and supply chain capacity?
• Can delivery be phased and scaled to support investment and future growth?
• How early can suppliers be engaged to test logistics, manufacturing capacity and programme assumptions?
• What data and analysis are needed to reduce change during delivery?
• What decisions can be made early to support both speed and delivery certainty?

Projects that answer these questions early are better placed to maintain control as delivery pressure increases.

A more deliberate approach to delivery

Delivery pressure is increasing. Labour shortages, power availability and supply chain constraints are already shaping what can be built, where and at what pace.

Pushing harder within the same delivery model will not resolve those pressures. Better outcomes come from making earlier decisions, testing them with real data and maintaining a clear strategy from planning through to construction.

Modularisation should be considered more often, and much earlier. Projects that do this gain control sooner and avoid carrying unnecessary complexity through the build.

Real speed comes when modularisation shapes the delivery strategy from the outset, not just individual parts of the build.