AI and offsite construction could cut costs by 20% and project timelines by 25%

Artificial intelligence could help offsite manufacturers coordinate design, production and installation more effectively, according to a joint white paper from US contractor Suffolk, the MIT Center for Real Estate and MIT Media Lab City Science group.

Construction in the Age of AI identifies offsite manufacturing as one of six areas where AI could improve construction performance, alongside design automation, permitting, scheduling, skilled labour and subcontracting, and procurement.

For prefab and modular businesses, the paper’s central argument is that stronger connections between these activities could make factory production more effective. AI could help identify suitable prefabrication opportunities during design, optimise production schedules and align factory output with site readiness. The authors emphasise that offsite manufacturing itself is not an AI technology. Its inclusion reflects the potential for AI and better data integration to support the coordination required between designers, manufacturers, contractors and trade partners

Repeatable components and early design decisions
The paper identifies early design commitment, repeatable components, fabrication-ready models and integrated logistics planning as foundations for successful offsite delivery.

Standardising components across multiple projects is a particular focus. Rather than designing each assembly from scratch, manufacturers could work from component catalogues that allow variation within established production requirements. The economic benefit would accumulate through repeat orders and higher production volumes.

Volumetric Building Companies founder and CEO Vaughan Buckley describes the commercial calculation underpinning that approach.

“Offsite manufacturing is arbitrage between the efficiency of the labor in the factory that you’re performing and the cost of the labor on the site, minus the transport.”

The paper reports that VBC has achieved schedule savings of up to 47 per cent across different geographies, while stressing that cost outcomes depend on local labour conditions and transport distances. These are company-reported offsite results, rather than savings attributed solely to AI.

It also describes VBC’s transition towards a configurable product platform: maintaining a set module width and structural span while allowing changes to balconies and façades without disrupting factory tooling.

Testing the effect on project feasibility
Suffolk’s modelling estimates that applying the six areas together could reduce total costs by approximately 17–20 per cent and shorten schedules by 22–25 per cent on a sample San Francisco multifamily residential project.

Those figures represent potential outcomes from an integrated delivery approach. They are not measured savings from a completed project using all six capabilities, and the authors caution against adding individual percentage savings together because their effects overlap.

The research draws on academic literature, case studies, interviews, surveys and a roundtable involving more than 50 industry experts. It describes the evidence as directional and calls for further testing across projects, markets and building types.

For Australian manufacturers, the findings offer a useful framework for assessing investment: establish repeatable products, resolve manufacturing requirements during design and connect production with procurement and installation. AI’s contribution will depend on how effectively those underlying processes and project relationships are organised.

Download Construction in the Age of AI HERE (pdf)