Construction has a $40 trillion productivity problem. Over the past two decades, global constructionproductivity grew just 10% in total; one-fifth the rate of the broader economy, one-third that of manufacturing.Between 2020 and 2022, US and European construction productivity did not stagnate. It declined by 8%. Everyyear this gap persists, the world pays in unbuilt housing, deferred infrastructure, and compoundinginefficiency.
The industry has known this for a generation. The question is why the response has been so slow and why that isabout to change.
Three Forces That Have Run Out of Road
The case for modularization and prefabrication is not new. What is new is the simultaneous maturity of threestructural forces that have removed the industry’s ability to defer action.
The labor market will not recover. The US construction industry needs 439,000 to 723,000 additional workersannually. 41% percent of the skilled workforce will retire by 2031. Ninety-two percent of contractors cannotfind qualified workers, despite average hourly earnings reaching $38.76 in early 2025. This is not a cyclical gap.It is a demographic cliff with no site-based solution. Factory construction requires 40–50% fewer on-site workers for equivalent output, not as a future efficiency, but as a current operational reality.
The housing deficit has become a political emergency. Goldman Sachs estimates the US is short 3–4 millionunits. The UK delivered 208,600 homes in 2024–25 against a stated need of 300,000. Germany faces aprojected shortfall of up to one million apartments. India, at 36% urbanization, has decades of housing volume ahead of it. Site-based construction, which isweather-dependent, labor-intensive, project-fragmented, cannot close these gaps. The mathematics do notwork, regardless of political will.
AI infrastructure has created a demand category that has already chosen prefabrication. This is themarket signal the industry is underreading. Traditional data center construction takes 24–36 months. AI infrastructure requires deployment in 8–12. A prefabricated 2MWdata center costs approximately $8 million versus $14 million built conventionally. Schneider Electric reports67% of new edge deployments already use modular designs. Microsoft, Google, and Amazon have all expanded modular programs. The most technicallydemanding, capital-intensive, time-sensitive infrastructure category in the modern economy has concluded thatprefabrication is the default; not a pilot, not a procurement preference. That logic applies to hospitals,schools, and housing with equal force. The only difference is that those markets carry more regulatory friction,not less commercial rationale.
What Operational Commitment Actually Looks Like
The shift from industry narrative to operational reality is visible in how leading firms are restructuring, notexperimenting.
DPR Construction, a $7 billion contractor, has embedded prefabrication as a core delivery pillar alongsidevirtual design and construction, not as a project-level option. Its new Silicon Valley headquarters contains adedicated Prefabrication Assembly Facility. On its Meta data center programs, DPR deploys MOFEcomponents (factory-built MEP modules that Meta contracts directly) reducing field labor, compressingschedule-critical paths, and buffering supply chain risk simultaneously. The Boldt Company has developedSTAAT MOD: a repeatable volumetric product line for hospital ICU and observation bed configurations,designed for predictable delivery across a recurring client base. These are not innovation showcases. They arepermanent operating models from firms with deep balance sheets and long client relationships.
The technology frontier makes the ambition vivid. In January 2024, China’s BROAD Sustainable Buildingassembled a 26-storey residential tower in five days using its 16th-generation Holon modular system (90%factory-prefabricated), generating 1% construction waste versus the 15,000 truckloads of concrete aconventional build would have required. In Germany, Gropyus, founded by the co-founder of Delivery Hero andbacked by over €400 million in equity, runs an 86%-automated factory in partnership with KUKA robotics that produces wall elements in 17 minutes. In the US, Sekisui House (Japan’s largest modular homebuilder),ranked 5th nationally in US homes sold in 2024, is targeting 20,000 annual US deliveries by 2031, explicitlyexporting six decades of Japanese factory-construction discipline into the world’s most supply-constrained housing market.
The geographic spread of this commitment is telling. These are not regional experiments. They are convergingbets on the same structural thesis across three continents.
The Failure Archive
No honest treatment of this sector avoids Katerra. It raised over $2 billion, employed 8,500 people, and filed forbankruptcy in June 2021. Nexii followed in 2024 with over $109 million in debt. Ilke Homes collapsed in the UKshortly after.
The instinct is to read these as validation of skepticism toward modular construction. That reading is wrongand dangerously convenient. These were not failures of the industrialization thesis. They were failures of the same execution error, repeated: factory capacity built aheadof confirmed demand; vertical integration across too many value chain nodes before any single one wasproven; a Silicon Valley scaling model applied to an industry that rewards operational depth, not platformambition.
McKinsey’s August 2025 analysis of 700+ modular companies across 50 countries makes the success patternexplicit. The highest-performing companies share three characteristics: a well-defined, deeply optimized building system; coordinated (not necessarily owned) valuechain control; and geographic or asset-type focus held until unit economics are proven. Profitability is not afunction of whether a company produces 2D panels or 3D volumetric modules. It is a function of simplicity, standardization, and the discipline to build demand before building capacity. The failures happened because founders built factories and then tried to fill them. The successes happened in the opposite order.
The Barriers Are Real
Intellectual honesty requires stating what the commercial case does not resolve on its own.
Regulatory fragmentation is the largest US systemic barrier. A module manufactured in Ohio requiresseparate inspections and approvals in New York, Pennsylvania, and California. There is no federal modularconstruction standard. President Trump’s Executive Order 14394 (March 2026), directing re-examination ofrestrictions that discriminate against modular housing by method rather than performance, is the first meaningful federal signal in a generation but policy signal and permitting reality are separated by years of implementation.
Financing misalignment persists. Traditional construction lending advances funds against a titled,permitted property. Modular construction creates value in a factory – off-title, off-permit, in a form mostlenders cannot underwrite. Canada’s CMHC Rapid Housing Initiative (i.e., C$3.2 billion committed, 10,000modular units delivered in 24 months) and the UK’s BOPAS mortgage assurance scheme demonstrate that thefinancing gap is solvable. Neither has been replicated at meaningful scale in the US.
The 2025 tariff shock compounded these pressures. US steel and aluminum tariffs reached 50% underSection 232, driving steel mill product costs up 20.7% and aluminum mill shapes up 33% year-on-year byJanuary 2026. Steel-framed volumetric operators absorbed input cost spikes that eroded the cost certainty advantage they had spent years building. Cross-laminated timber and mass timber hybrid systems carry structurally lower tariff exposure, which partly explainsthe accelerating capital flow into timber modular approaches from investors and developers alike.
The Only Question That Remains
The global modular and prefabricated construction market stands at approximately $160 billion and isprojected to reach $200–$230 billion by 2030. The modular data center segment alone is growing at a near-18% CAGR. Global ConTech investment reached $6.57 billion in 2025. Ninetypercent of investors plan to increase or maintain deployment in 2026.
But market size is not the point. Structural irreversibility is. An industry facing simultaneous labor scarcity, agenerational housing deficit, binding carbon mandates, and AI-driven infrastructure demand that operates ontwelve-month timelines cannot continue building the way it has built for a century.
Modularization reduces on-site labor by 40–50%, compresses delivery timelines by 30–50%, cutsconstruction waste by half, and with material discipline, reduces embodied carbon by up to 54%. The companies capturing value from this transition are not predicting a market shift. Theyare building the factories, DfMA workflows, digital-to-physical interfaces, and supply chain integrations thatmake it real.
The question is no longer whether construction industrializes. It is who leads it and who spends the next decadeexplaining why they waited.
