IN THE FIELD
Infrastructure
Building at the Speed of AI
The rise of volumetric modular construction & the future of infrastructure
BY MONICA SOTO
The construction industry is bumping up against a wall that traditional delivery methods were never built to handle: the sheer velocity of AI infrastructure demand. As a civil engineer who has spent my career on the procurement and construction management side of complex builds, I’ve watched project timelines that once felt aggressive become almost quaint compared to what owners and developers now ask of general contractors (GCs).
Consider the numbers. Hyperscale data center construction has historically run 24 to 36 months from groundbreaking to ready-for-service, driven by weather delays, permitting, inspections and the sequential nature of on-site trade work. Demand is outpacing the industry’s ability to deliver, forcing a rethink of how nonresidential construction gets built, not just for data centers but for any project where speed, labor availability and site constraints intersect.
Volumetric modular construction has moved from a niche alternative delivery method to a serious business strategy. This involves manufacturing complete, fully integrated building sections off-site and installing them as finished building blocks. For midsized GCs working across commercial, heavy civil, water/sewer and other nonresidential sectors, this shift matters even if you never touch a hyperscale project, because the pressures driving modularization, labor shortages, compressed schedules and rising community scrutiny are showing up on jobsites everywhere.
The Pressure Points Driving Adoption
Three forces are converging to make volumetric modular construction less of an experiment and more of a necessity.
The first is labor. The construction industry needs an estimated 349,000 net new workers in 2026 alone, according to Associated Builders and Contractors, climbing to 456,000 in 2027. Data center builds absorb an outsized share of that strain. The electrical work alone accounts for 45% to 70% of what it costs to build a data center, according to the International Brotherhood of Electrical Workers. When a trade shortage becomes the limiting factor on a schedule, moving that work into a factory, where a smaller crew completes it repeatedly under controlled conditions, stops being optional.
The second pressure is power. Equipment lead times for large power transformers now run two years or more, and medium-voltage switchgear often takes close to a year. When gear delivery dates are fixed, the only variable a contractor can still control is how fast the building comes together. Modular strategies let site work and factory fabrication happen in parallel rather than in sequence, a large part of why highly modularized projects are achieving schedule reductions of 30% to 50% compared to conventional builds, compressing a 24- to 36-month timeline into 16 to 20 months.
The third pressure, which should concern every general contractor regardless of sector, is community pushback. A recent Gallup survey found 71% of Americans oppose new data center construction in their communities, and industry tracking has identified more than 140 activist groups across two dozen states working to block or delay projects. Reducing the duration and disruption of on-site construction, fewer trucks, less noise and a shorter period of active work are becoming a genuine competitive advantage, not just an environmental talking point.
What ‘Volumetric’ Actually Means
It’s worth being precise about terminology, because “modular” gets used loosely and covers approaches with very different capabilities. Volumetric modular construction refers to complete, preintegrated building sections, not shipping containers retrofitted for a purpose, and not flat panels assembled on-site. These are fully 3D structural units with mechanical, electrical and plumbing systems already installed at the factory, often reaching 90% completion before they leave the manufacturing floor.
The engineering discipline behind this is generally called Design for Manufacture and Assembly, or DfMA; it’s a methodology adapted from product manufacturing for construction use. DfMA means a project is architected from the earliest design phase to be built in a factory and transported, rather than value-engineered into a modular format afterward. Projects that start design with DfMA principles see far fewer field changes and requests for information (RFIs) than those retrofitted midstream. Singapore’s Building and Construction Authority has documented this as a core productivity lever, noting that DfMA reduces reliance on manual labor while improving workmanship through the kind of stringent quality control a factory floor allows and a jobsite often cannot.
For contractors evaluating whether a project is a candidate for volumetric delivery, the products involved typically include complete structural steel assemblies, prefabricated electrical and mechanical skids, integrated power and cooling systems, and prefabricated utility modules, all manufactured as turnkey infrastructure blocks ready for rapid installation once the foundation is set.
A New Kind of Construction Ecosystem
Perhaps the most consequential shift for general contractors isn’t the technology itself but the business relationships it requires. Volumetric delivery depends on a genuinely different partnership model between contractors and manufacturers than the traditional subcontractor relationship most GCs manage.
In a conventional build, a general contractor holds the schedule risk and coordinates dozens of trade subcontractors sequentially on-site. In a volumetric model, a significant portion of that risk shifts to a manufacturing partner who owns the factory build, quality control and delivery logistics, all while the GC retains responsibility for site civil work, foundations, utility tie-ins and final integration. That division of scope must be defined with more precision than a traditional trade contract, since both teams are essentially running parallel production lines that must converge on one delivery date.
This is showing up across the industry as deeper, longer-term strategic alliances rather than one-off subcontracts. In South Korea, telecom operator LG Uplus recently formed an alliance with a major engineering and construction firm to speed up modular AI data center construction. With the construction partner providing constructability reviews and process optimization alongside its traditional engineering, procurement and construction capabilities, they are also creating a relationship that looks more like a joint venture than a standard bid-build arrangement. Whereas a construction firm embeds itself into the design process before ground is broken, and it is likely to become more common as owners seek single points of accountability across the factory-to-field handoff.
Resources for Further Reading
Readers interested in going deeper can explore the Modular Building Institute (modular.org), which publishes research and case studies on volumetric construction across sectors, and the DFMA Institute (dfma.com), which offers foundational resources on DfMA. Associated Builders and Contractors (abc.org) publishes ongoing labor market forecasts, and CBRE’s Global and North American Data Center Trends (cbre.com) reports provide detailed data on pressures reshaping this sector. Data Center Knowledge (datacenterknowledge.com) and Data Center Dynamics (datacenterdynamics.com) offer regular coverage of construction methods and labor trends for contractors evaluating this space.
For midsized general contractors, this ecosystem shift creates both opportunity and a learning curve. Firms that build genuine logistics expertise around receiving, staging and setting large, prefabricated modules‚ work that looks more like industrial rigging than traditional trade coordination, find themselves positioned for a growing category of work. Those that treat modular components like oversized standard building materials, without adapting scheduling and risk allocation, tend to run into costly surprises around transportation, crane capacity and site sequencing.
Beyond Data Centers
Hyperscale data centers are the most visible example of this shift because the money and urgency are so extreme. Global spending on AI infrastructure by major technology companies is expected to approach $725 billion in 2026 alone, but the underlying lessons apply broadly across nonresidential construction. Any project facing a skilled labor shortage, a compressed schedule, a remote or challenging site, or heightened community scrutiny stands to benefit from evaluating which components could reasonably move to volumetric delivery.
That doesn’t mean every GC needs to become a modular specialist overnight. It does mean understanding the vocabulary, partnership structures and sequencing implications is necessary in order to have an informed conversation about modular construction. When an owner or developer raises the option, a conversation can lead to the increasing frequency of modular build across commercial, heavy civil and infrastructure sectors alike
Monica Soto, MBA, is a civil engineer and construction executive with more than a decade of experience in construction management, procurement and supply chain strategy across complex infrastructure projects. As director of preconstruction at Mod42, she focuses on collaborative partnerships, design coordination and execution strategies for factory-built data center infrastructure and complex construction programs. Visit mod42.com.
Photo Credit: Jarama-adobestock.com
