Facing a 500,000-worker gap, a builder makes the case for modular chip and battery plants

Roughly half a million. That's how many more workers U.S. construction needs in 2026 just to keep up with existing demand, up from 439,000 at the start of 2025. Matt Willson of Holland Construction Services cites the Associated Builders and Contractors figures in an October 15 column for REBusinessOnline. His argument is that the shortfall hits heavy industrial work hardest, and that crews building in the field can't fill it.

In Willson's view, a semiconductor fab or EV battery plant should be seen less as a building than as a precision manufacturing environment housed within one. Its mechanical, electrical, process piping and structural systems are woven together at every level. The plan breaks down when all of those trades have to be on site together and work in the correct sequence.

Those plants are arriving at scale. Willson points to Samsung's $17 billion in semiconductor fabs in Taylor, Texas, TSMC's commitment of more than $65 billion to three fabrication plants in Arizona, and the $5.8 billion EV battery campus Ford and SK On are building in Kentucky. He ties the wave to reshoring, domestic manufacturing investment and landmark federal legislation, and concludes that staffing the work through traditional field sequencing, at the volume this pipeline demands, is increasingly untenable.

His fix is to move whole systems into factories and design them with virtual design and construction (VDC) and building information modeling (BIM). That leaves a harder question for the owner's capital projects director signing off on a new plant: what does the switch actually buy? The independent evidence suggests the answer is time. Cost savings are much less reliable.

Labor gaps that already show up as delays

Willson calls the shortage structural, not cyclical. His figures cover the whole pipeline of workers: who is retiring, who can be hired and what happens to projects when the gap hits the job site.

For a plant owner, the 45 percent figure matters most. Close to half of contractors say they've already had a project run late because they couldn't staff it. The retirement figure points the same direction, since the people leaving are experienced trades workers who can't be replaced quickly.

This hits hardest for owners whose projects need many trades working in the same space at the same time, which describes fabs and battery plants. For them, a contractor's access to labor is worth checking at bid time, alongside price, because Willson's figures suggest it is now one of the more likely reasons a schedule slips.

What moves from the field to the shop

In Willson's description, industrial modular means much more than prefabricated pipe spools or pre-assembled ductwork. Entire mechanical rooms, process skids, electrical distribution assemblies and utility infrastructure modules are engineered as self-contained units. They're built, inspected and tested in a fabrication shop before they ever reach the site.

H+M Industrial EPC, which builds modular process systems, describes the same approach in four phases: front-end planning, where the team defines scope, runs feasibility studies and decides what to modularize and why; detailed engineering; fabrication and testing in the shop; and shipping and installation. For process systems, that means prefabricating entire units complete with piping, instrumentation and control systems. None of it is new, either. H+M traces industrial modular's steady growth back to the 1950s.

The firm also lists disadvantages. Modular design generally requires the finalization of designs early in the process, which reduces the ability to customize solutions in later phases, and design changes can't be made once the modules have been fabricated. H+M adds that while process modules can be highly customized, there are still limits to adapting them to highly specialized needs, a point worth weighing for a fab whose systems Willson describes as intertwined at every level.

The column calls VDC and BIM core components of modular construction projects, and that design limit suggests why. On a field-built job, a late design change usually costs rework on site. On a modular job, the same change may land after the module is already built, which pushes the cost and the decision earlier in the project.

Moving work into a shop relocates part of the labor problem rather than erasing it. A 2025 post from GWGCI claims modular methods cut on-site labor needs by up to 60 percent, yet also notes that traditional construction workers must adapt to the new skill sets off-site manufacturing requires. CII's research, covered below, lists labor at the fabrication area and labor agreements as barriers alongside labor at the construction site. That suggests the bid-time labor check should cover the fabricator's shop as well as the general contractor's field crews.

The nearest commercial parallel may be data centers. Miller Industries, a modular builder, says data center modules are pre-engineered for power, cooling and cabling, and that pre-tested, factory-built environments minimize errors and reduce go-live times. That is the same test-before-it-ships logic Willson applies to electrical distribution assemblies and utility modules in a fab.

How much factory capacity exists

There's real capacity on the supply side. Willson counts 255 modular manufacturers operating in North America, with individual plants able to produce up to 25,000 square feet of assemblies a week, and more than $2.2 billion invested in the sector over the past decade. He cites Astute Analytica's estimate that the global modular construction market was worth about $90 billion in 2024 and will reach $155 billion by 2033, growing 6.2 percent a year. Adoption, he says, is accelerating as owners and contractors recognize the operational advantages.

Market sizing depends a lot on who's counting. The same GWGCI post put the global market at $175 billion by 2025, nearly double Astute's 2024 valuation and above its 2033 forecast. A spread that wide suggests treating any single market number as directional, not as a measure of capacity available to a particular plant.

The manufacturer count also spans very different products. Volumetric Building Companies, for instance, lists product lines in affordable, multifamily, student and refugee housing, healthcare and hospitality. For a fab or battery-plant owner, the relevant pool is the subset of those 255 shops equipped to build process skids and electrical distribution assemblies.

Faster schedules, thin savings

In a 2011 report, the Construction Industry Institute's Research Team 255, building on earlier work by Research Team 232, tested whether shipbuilding's Interim Product Database method could be adapted to construction. In shipbuilding, the method cuts engineering design time by up to 80 percent, mostly by reusing a library of proven, scalable designs instead of starting fresh.

The team's headline finding was about time. RT-255 named schedule reduction the number one benefit of adopting the approach and found that modular execution on the shipbuilding model was driven mostly by owners' need for faster project delivery. Five case studies showed schedule improvements of 15 to 50 percent over original baseline schedules, because the modular approach lets activities run in parallel and lets construction continue out of season. Cost savings were marginal or non-existent.

The mechanism helps explain the split. Paralleling compresses the calendar, since modules come together in a shop while site work proceeds. The fabrication still has to be paid for, and the modules still have to be moved and tied in, which could explain why time saved in the case studies didn't turn into money saved.

Put the vendor marketing next to the research and the gap shows. The schedule claims from Miller Industries and GWGCI sit at the top of CII's measured range. GWGCI's 20 percent cost claim has no counterpart in the case studies. Even H+M, which sells modular process systems, says lower project costs come only "in some cases."

The CII work is now 15 years old, covers only five cases and predates today's modeling tools, so its findings may have shifted since. Still, Willson's own argument rests mainly on labor and schedule, which is the ground the evidence supports best.

CII also catalogued what gets in the way: 22 barriers in six groups.

• Execution (4): integrating reusable modules, customization, scale and repeatability, plant capability
• Design (4): natural hazards, site constraints, codes and standards, inspection and testing
• Transportation (5): travel distance and cost, road restrictions, handling capability, shipping and packaging, transport risk
• Module interfaces (2): interface tolerances, ability to absorb field changes
• Labor (3): site labor, fabrication shop labor, labor agreements
• Long-term (4): market size, initial funding, long-term funding, regulations and codes

Road limits, tie-ins and reusing designs

Several of those barriers come down to physics and logistics. Volumetric Building Companies, a modular manufacturer, says most size constraints come from transportation. H+M flags the cost of moving large modules to remote sites and the work of tying modular systems into existing infrastructure. In other words, the road sets the maximum size of a process skid, and the tie-in to existing systems is where a module that tested fine in the shop has to fit the plant.

Transportation is also the largest group on CII's list, with five of the 22 barriers. H+M says logistics need careful planning to avoid delays or damage, which suggests that route and site access belong in the same front-end planning phase where teams decide what to modularize.

The shipbuilding numbers point to where the bigger savings could come from: reuse. Volumetric Building Companies says manufacturers keep libraries of pre-designed plans and work with clients to decide which elements to standardize and which to customize. Its stated goal is standardization where it adds time and cost benefits in manufacturing, and customization where it adds high value to the design. The shipbuilders got their 80 percent design-time cut from the same habit of reusing proven designs.

That makes multi-plant programs the case to watch. TSMC's three Arizona fabs are exactly the kind of repetition where a reusable module library could pay off. If owners with several sites start carrying standardized module designs from one plant to the next, the cost savings the 2011 case studies couldn't find may finally show up.