As Asia Pacific races to build 100MW+ campuses at unprecedented speed, bespoke engineering is no longer fit for purpose. Pramod Deshmukh, Country Head for Hyperscale Data Center Design & Implementation at Ramboll, makes the case for a fundamentally different way of building at scale.
The narrative in recent years that micro and Edge data centres would dominate was largely shaped by early assumptions around latency-sensitive applications and distributed computing. Yet in Asia Pacific, the market is moving decisively in the opposite direction. Across key markets such as Malaysia, Australia and India, the industry is scaling up at pace, with 100MW campuses increasingly becoming the norm rather than the exception.
India, where I am based, is a great illustration of why mega-scale is becoming essential. With over a billion connected users, rapid 5G expansion and AI now accessible in multiple local languages, digital demand is accelerating across the population. Coupled with data localisation requirements, this demand must be met within the country, driving a surge in hyperscale development. The result is a structural shift towards large, high-density campuses, with industry projections pointing to hundreds of hyperscale facilities required by the early 2030s.
At this scale, mega-scale is not a choice, it is a necessity. It is fundamental to delivering the performance, efficiency and economics that next-generation digital infrastructure demands. However, this raises a more pressing question: how can these increasingly ambitious, high-density developments be delivered in practice? As project sizes grow and timelines compress, traditional approaches to data centre design and construction are being pushed to their limits.
The changing approach
The challenge is not only to build bigger, but to build faster, more reliably and with greater adaptability. In this context, the industry is beginning to shift from bespoke, project-by-project engineering towards more repeatable and scalable models. Modular design is emerging as a critical enabler, offering a pathway to translate mega-scale ambition into deliverable reality.
Traditionally, data centres were engineered as individual projects. Every substation, generator yard, cooling plant and distribution system was designed specifically for a single facility. While this approach delivered reliable infrastructure, it often resulted in longer engineering schedules, complex procurement activities and extensive construction efforts on site.
As demand continues to grow, this approach is becoming increasingly difficult to sustain. The industry needs a different mindset – one that focuses on repeatability, scalability and adaptability. At Ramboll, we describe this philosophy as ‘Design Once, Deploy Many’.
This does not mean creating identical facilities. Rather, it means developing proven engineering solutions that can be replicated, expanded and adapted across multiple projects while maintaining the flexibility required by different customers, locations and technologies.
Power at scale
The journey towards modularisation begins with power infrastructure.
Electrical substations have traditionally been treated as unique engineering assets. Every transformer arrangement, protection philosophy, cable routing strategy and equipment interface was developed independently. A modular approach allows these systems to be standardised into repeatable building blocks. Transformer bays, GIS systems, protection panels and cable interfaces can be designed once and deployed many times. This significantly reduces engineering effort while improving procurement efficiency and construction certainty.
The same principle applies to generator yards and fuel systems. Generator foundations, fuel storage systems, exhaust arrangements and control interfaces can be developed as modular packages with standardised connection points. Instead of assembling hundreds of individual components on site, project teams can install integrated systems that have already been manufactured, assembled and tested in factory-controlled environments.
Power distribution presents perhaps the greatest opportunity for modularisation. High-voltage to medium-voltage distribution, medium-voltage to low-voltage distribution and low-voltage distribution within the data hall can all be developed using repeatable architectures. Standardised switchgear lineups, busway systems, power skids and rack power distribution modules allow facilities to scale rapidly while reducing installation complexity.
What we are increasingly seeing across projects is that some of the largest schedule reductions occur when power infrastructure is viewed as a scalable platform rather than a collection of project-specific solutions. Standardisation creates predictability and predictability creates speed.
Modular cooling solutions
The same transformation is taking place within cooling infrastructure.
The rapid growth of AI workloads is driving unprecedented increases in rack densities, placing greater demands on cooling systems than ever before. Traditional cooling plants often relied on extensive site
installation of chillers, pumps, valves, heat exchangers and piping systems. Today, modular cooling solutions are enabling a much more efficient approach.
Chillers can be deployed as repeatable capacity blocks. Pump skids can arrive fully assembled and tested. Ring main distribution systems can be designed with standardised connection points that simplify both installation and future expansion. Entire cooling modules can be manufactured, tested and commissioned before arriving on site.
As liquid cooling becomes increasingly common, modularisation becomes even more valuable. Coolant Distribution Units, liquid cooling manifolds, fan wall systems and rack-level cooling interfaces can all be delivered as integrated solutions. These systems reduce installation risk, simplify commissioning and support the rapid deployment of high-density AI environments.
At the same time, modular cooling contributes to another critical objective: sustainability.
Built-in sustainability
Modern data centres must be designed not only for speed but also for long-term environmental performance. At Ramboll, sustainability is integrated into every stage of the design process. By shifting significant portions of construction into factory environments, modular systems reduce material waste, improve manufacturing efficiency, minimise rework and reduce the environmental impact associated with site construction activities.
Cooling infrastructure also plays a major role in determining a facility’s operational efficiency. Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE) remain critical measures of data centre performance. Modular cooling systems create opportunities to standardise highly optimised designs that have already demonstrated strong energy and water efficiency. By repeatedly deploying proven cooling solutions, operators can reduce engineering variability while improving both PUE and WUE performance across multiple facilities.
Inside the data hall, modular thinking continues to create value.
Hot aisle containment systems, cold aisle containment systems, rack layouts, power distribution systems and cooling interfaces can all be standardised to support rapid deployment. Instead of treating each data hall as a unique environment, operators can establish repeatable deployment models that balance speed, efficiency and flexibility.
This flexibility is becoming increasingly important as technology evolves.
Designing for the future
One of the greatest advantages of modular design is future adaptability. Rack densities continue to increase. Cooling technologies continue to evolve. Power requirements continue to change. Facilities designed around modular principles can respond to these changes much more effectively than traditionally designed facilities.
Rather than requiring extensive redesign, operators can introduce new modules, expand existing systems, or replace specific infrastructure components while maintaining ongoing operations. This flexibility helps safeguard long-term investments while supporting continuous technological evolution.
The rapid adoption of AI provides a clear example of why adaptability matters. Many facilities originally designed for rack densities of 10 to 20 kW are now being asked to support workloads exceeding 100 kW per rack. At Ramboll, we increasingly design infrastructure with future retrofits and technology transitions in mind. Modular architectures provide clear upgrade pathways, allowing power and cooling capacity to evolve alongside customer requirements without requiring wholesale replacement of existing infrastructure.
Standardising critical systems
The same philosophy extends to fire protection and digital infrastructure systems.
Fire detection, suppression systems, control panels, Building Management Systems, access control systems, video surveillance platforms, environmental monitoring systems and Very Early Smoke Detection Apparatus systems can all be developed using repeatable architectures. Standardisation simplifies engineering, reduces commissioning effort, improves operational consistency and enables future expansion with minimal disruption.
At Ramboll, we believe that true modularisation extends beyond individual equipment packages. It requires a coordinated design philosophy across every engineering discipline. Power systems, cooling systems, fire protection systems, digital infrastructure, architecture and structural engineering must all work together as part of a unified strategy.
As the industry moves towards larger AI campuses and increasingly compressed delivery schedules, this integrated approach will become even more important. Future developments may be measured in hundreds of megawatts and eventually gigawatts. Such facilities cannot be delivered efficiently through repeated custom engineering. They require repeatable building blocks, scalable infrastructure and engineering teams that understand how to transform engineering complexity into scalable simplicity.
Delivering scale successfully
Ultimately, modularisation is not simply an engineering strategy. It is a business strategy that enables organisations to respond faster to market demand, accelerate AI adoption and future-proof critical digital infrastructure investments.
As Asia Pacific continues to lead the global expansion of digital infrastructure, the shift towards mega-scale, high-density campuses is no longer a question of if, but how fast. Delivering 100MW+ campuses at speed, with the resilience and flexibility demanded by AI workloads, requires a fundamental shift in how data centres are conceived and built. Modular design provides the bridge between vision and delivery, enabling operators to translate scale into reality without compromising on time, quality or performance. In this new era of digital infrastructure, success will belong to those who cannot only build big, but build smart, repeatable and ready for what comes next.
