The sluggish growth of data centres is fast falling out of step with the blinding pace of AI adoption.
In the US, nearly half of all data centres slated to open this year are either facing delays or cancellations, per a Slightline Climate report. Meanwhile, in India, while the data centre capacity is set to grow 4–6x, utilisation rates are high,
A major barrier to growth is the traditional ground-up approach to constructing data centres, which typically takes between 18 and 36 months, from planning permission to operational readiness. In technology terms, that is a lifetime. By the time a conventional facility is built, the workloads it was originally designed to serve may evolve beyond recognition.
Now, a growing number of industry leaders are betting on prefabricated and modular data centres: factory-built, pre-integrated infrastructure systems that can be deployed in weeks and scaled to meet the current demand.
From Niche to Mainstream
Modular data centres are growing three times as fast as the overall data centre market, as per research from Omdia. Revenue from facilities using prefabricated components rose from approximately $2.3 billion in 2020 to $3.25 billion in 2023.
Among the biggest areas of growth are high-performance computing and AI, hybrid cloud, intelligent edge, 5G networks, military and defence, and Industry 4.0.
"Modular data centres address three major concerns of data centre deployment: lead time, predictability, and cost. They help ensure that a project will be completed at a predetermined quality and cost," Vlad Galabov, Senior Director of Enterprise Infrastructure at Omdia, wrote in a LinkedIn post.
He noted that modular construction is also sustainable, using less concrete, fewer on-site workers, and measurably lower Scope 3 emissions.
According to Straits Research, the Indian data center market is experiencing intense private investment in modular edge deployments in tier-2 and tier-3 cities to facilitate localised cloud access and AI-based applications.
“Telecom and IT companies are leveraging prefabricated data units to improve rural connectivity and business continuity in the event of power grid outages. New financing models and modular leasing deals are also expanding access to advanced infrastructure among mid-sized and smaller businesses,” the research found.
The AI Pressure Test
The rise of AI workloads has dramatically sharpened the case for modular infrastructure.
A standard data centre rack historically drew 5–10 kW of power. "The bulk of any workload is happily sitting in 10–15 kW racks, and that is either in your own data centre or a rented colocation space or a hyperscaler,” Ian Jagger, who leads services marketing at Hewlett Packard Enterprise, tells AIM.
But generative AI is pushing energy demands towards 150 kilowatts per rack—a jump that many traditional data centres, constrained by legacy cooling and rigid power infrastructure, simply cannot accommodate.
“Because AI is compressing timelines. Infrastructure is evolving faster than traditional energy systems were designed to support. Data centres are moving toward gigawatt-scale campuses, while power projects have historically moved on multi-decade cycles,” says Sridhar Pinnapureddy, Founder and CEO of CtrlS Datacenters.
Prefabricated modular solutions are explicitly engineered to meet these demands.
In July 2024, Vertiv, a digital infrastructure firm, introduced the MegaMod CoolChip, a high-density prefabricated modular data centre solution designed to accelerate the deployment of AI computation, incorporating sophisticated liquid cooling and direct-to-chip cooling for high-power CPUs and GPUs. It claims to reduce deployment time by as much as 50%.
Meanwhile, Schneider Electric, in June 2025, launched a Prefabricated Modular EcoStruxure Pod Data Centre solution built to support high-density racks of up to 1 MW and beyond.
Speed as a Strategic Asset
That agility stems partly from the way modular facilities are constructed. Rather than building sequentially on-site, prefabricated components are assembled in factory conditions, controlled environments that improve quality, reduce weather-related delays, and enable parallel workstreams. Fabrication can proceed simultaneously with site preparation, civil works, and logistics planning, compressing the overall project timeline considerably.
Vertiv, whose prefabricated portfolio spans worldwide deployments, cites time savings of up to 30% compared with conventional builds.
Sai Charan Atelly, Growth and Operations Manager at IT company Zenbyt, explains that factory-based assembly also addresses one of the sector's less-discussed challenges: the global skills shortage.
"By moving construction primarily to a factory environment, you are able to offer consistent work at a consistent location," he articulates, noting that asking skilled engineers to relocate to new sites continuously is neither sustainable nor realistic.
"Modularity helps create cost efficiencies when at the design stage. The lower upfront capital required is matched with further capital investment as power needs escalate. This scalability in capital deployment is completed with lower construction and engineering costs due to the standardisation of designs," he adds.
The "Pay as You Grow" Model
Efficiency is a major selling point of prefabricated data centres. Chilled-water and free-cooling systems are able to achieve a PUE of less than 1.5, setting a new standard for energy efficiency, says Atelly.
The adoption of prefabricated modules can lead to a decrease in construction and operational expenses of upwards of 30% compared to conventional data centre setups, per ENCOR Advisors. Factory assembly allows for accurate cost estimates and reduces costly modifications during deployment.
Vertiv describes this as a "pay as you grow" model—deploying capacity incrementally in line with actual workload growth, rather than committing vast capital expenditure to infrastructure that may sit underutilised for years.
However, some analysts emphasise that modular data centres, for all their advantages, still account for a relatively small share of total data centre revenue.
Saurabh Gayen, Chief Solutions Architect at semiconductor IP company Baya Systems, argues that modular solutions are best understood as complementary to, rather than replacements for, hyperscale infrastructure. "Not every organisation wants to be fully dependent on centralised providers for all of their AI needs. Many are willing to accept worse total-cost ownership if it means owning and controlling their own infrastructure,” he notes.
But in a market where AI adoption trajectories remain genuinely difficult to predict, pre-fabricated data centres offer a stepped approach.
"The future of digital infrastructure is no longer tied to concrete, steel, or multi-year construction cycles,” Jagger concludes.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.