Chris Evans keeps a whiteboard in Timberlab's Portland fabrication plant with two columns: steel projects that stayed steel and steel projects that converted to mass timber after the tariff math changed. Six months ago the columns were roughly equal. Now the conversion column is three times longer, and Evans, who runs the Swinerton-owned mass timber fabricator, told Construction Dive in May 2026 that steel tariffs are doing what carbon policy never could. "When steel prices go up and tariffs caused that throughout the whole system, then what that will really do is help make mass timber cost neutral or better in more markets across the U.S."
He is right about the cost. He may also be accidentally right about the carbon.
Nobody in the tariff conversation is talking about embodied carbon. Trade representatives are not thinking about lifecycle emissions when they set duty rates. Builders converting from steel to cross-laminated timber are chasing cheaper materials, not lower global warming potential. But the conversion produces a carbon dividend whether anyone intended it or not, and the scale of that dividend, if the substitution pattern holds across enough projects, is large enough to matter at the national level.
The problem is that another tariff is simultaneously working against it.
Two Tariffs, Opposite Carbon Vectors
Steel and aluminum face 50% tariffs implemented June 2025. Canadian lumber imports carry a combined 35.2% duty that could rise to 45%. Construction input prices climbed 7.1% annually as of January 2026, according to Associated Builders and Contractors. Forty-three percent of general contractors in an AGC-NCCER survey reported projects canceled, postponed, or scaled back because of material costs.
Steel tariffs push builders toward timber. Lumber tariffs push them away from it. Cement and concrete tariffs, 25% on Canadian and Mexican imports, discourage the third major structural option. The tariff structure is internally contradictory on carbon, and nobody writing these policies appears to have noticed, or cared, that the net effect depends on which tariff bites harder in a given project's cost model.
For mid-rise multifamily and mixed-use buildings where steel framing is common, the steel tariff currently bites harder. Timberlab's conversion whiteboard confirms that. Mass timber is gaining share specifically in project types where steel was the default structural system, because CLT and glulam now price at or below tariff-inflated steel for buildings up to about 18 stories, a ceiling that keeps rising as code bodies update allowances under the 2021 International Building Code's Type IV-A, IV-B, and IV-C classifications.
For single-family residential, the picture is muddier. Steel framing accounts for less than 5% of US single-family construction. Most houses are already wood-framed. The steel tariff's residential impact runs through components: connectors, joist hangers, fasteners, reinforcing bar in concrete foundations. Those add $2,000 to $5,000 per home, according to NAHB estimates, but the embodied carbon of connectors and fasteners is a rounding error against the concrete foundation and insulation envelope. In single-family, this is mostly a cost story. In multifamily, it is both.
What MIT, Yale, and the Field Data Actually Show
MIT's Climate Portal estimates that cross-laminated timber reduces building emissions by approximately 40% compared to steel and concrete construction. A Yale study published in Nature modeled CLT adoption at 30-60% of new urban buildings through 2100 and projected reductions of 25.6 to 39 gigatons of global CO₂ emissions, alongside expanded forestland of 36.5 million hectares from managed harvesting cycles. Under Armour's headquarters in Baltimore, built with mass timber, documented a 69% reduction in embodied carbon compared to a conventional steel-and-concrete equivalent, per Trellis.
Those numbers are real. They are also best-case.
A Frontiers in Built Environment study found that transport distance can increase CLT's embodied energy by up to 24%, which erodes the carbon advantage for projects far from domestic fabrication plants. Timberlab operates two facilities, with a sawmill coming in 2027. The domestic supply chain is thin. The World Resources Institute has argued, in a position that the mass timber industry would rather you not read, that "mass timber construction likely increases emissions for decades" if forest harvesting exceeds sustainable yields. Biogenic carbon accounting, the practice of crediting timber buildings for the CO₂ stored in their wood, is contested methodology. Trees store carbon as long as the building stands. If the building is demolished at 60 years and the wood goes to a landfill, the sequestration calculus changes, and not in timber's favor.
AI Tools That Could Navigate This, and Why Builders Ignore Them
At least six computational tools exist that can quantify embodied carbon during design, before procurement decisions lock in material choices and the carbon trajectory of a building crystallizes into physical reality.
EC3 is the industry standard: a free, open-access calculator built on digitized Environmental Product Declarations, launched by Building Transparency at Greenbuild 2019. It lets you compare the carbon intensity of specific steel products against specific CLT panels from specific manufacturers, accounting for supply chain geography rather than relying on generic category averages that obscure the variation within material classes.
University of Bath's AI tool, published in 2025, does something genuinely novel. It predicts embodied carbon from conversational text descriptions of buildings using machine learning and natural language processing, meaning an architect can type a paragraph describing a proposed structure and receive a carbon estimate without producing a detailed BIM model first, collapsing what was previously a weeks-long analysis into minutes.
Then there is BEAM from the Rocky Mountain Institute for whole-building lifecycle assessment, tallyCAT from RMI as a free LCA tool, RESNET 1550 as a new standard simplifying embodied carbon measurement specifically for residential new homes, and the ASPEC project in the UK, an £800,000 initiative led by Winvic Construction to predict embodied carbon directly from BIM data using AI.
None of these tools have meaningful penetration among residential builders. EC3 has traction in commercial construction, especially on projects pursuing LEED v5 or BREEAM Version 7 certification, both of which now require whole-life carbon assessments. But the general contractor building 200-unit townhome communities, exactly the project type where tariff-driven material substitution is actually happening right now, has almost certainly never opened EC3, has never heard of the Bath AI tool, and is making material substitution decisions based entirely on the latest quote from their steel supplier versus their timber supplier, with carbon as a factor in precisely zero of those conversations.
Running the National Numbers
Approximately 1.5 million residential building permits are issued annually in the United States. If tariff-driven material substitution shifts even 5% of projects from steel framing to mass timber, that represents 75,000 buildings. At a conservative estimate of 20 metric tons of CO₂e avoided per building, accounting for both reduced steel emissions and biogenic carbon storage in the timber, the aggregate reduction reaches 1.5 million metric tons annually.
For context, that is roughly equivalent to taking 325,000 gasoline cars off the road for a year, using the EPA's standard conversion of 4.6 metric tons per vehicle. It is also approximately 0.03% of US annual greenhouse gas emissions, which means it is meaningful in the context of building sector decarbonization but irrelevant as a national climate strategy.
The calculation has significant assumptions baked in. It uses published EPD data for steel framing at 15-20 metric tons CO₂e per typical 2,000-square-foot structure and CLT framing at 5-8 metric tons, plus 5-10 metric tons of biogenic carbon storage. The 5% substitution rate is a projection based on Timberlab's reported conversion patterns; no national dataset tracks tariff-driven material substitution in real time. And the numbers apply primarily to multifamily and mixed-use projects where steel was the baseline, not single-family homes where wood was already the dominant material.
Buy Clean Acts and the Policy Gap
California, Colorado, New York, and Washington have enacted Buy Clean Acts that require lifecycle carbon assessments for materials used in publicly funded buildings. These laws create industry momentum by forcing manufacturers to produce EPDs and contractors to consider embodied carbon in procurement, but they apply to public buildings and have no jurisdiction over the private residential construction market where tariff-driven substitution is occurring.
LEED v5 and BREEAM Version 7 certification now require whole-life carbon assessments. But fewer than 3% of new residential units in the United States pursue LEED certification. The policy instruments that could connect tariff-driven material substitution to intentional carbon tracking do not reach the projects where the substitution is actually happening.
This is the gap that AI tools could fill, not as regulatory instruments but as decision-support layers that make embodied carbon visible at the moment builders are already reconsidering their material choices. A GC who would never voluntarily run a lifecycle assessment might run EC3 if the output told them that the CLT option they are already considering for cost reasons also generates a carbon credit or satisfies an increasingly common green building requirement from their institutional client. The AI tool from Bath could make that analysis trivially easy by accepting a natural-language project description rather than requiring structured BIM data that small builders do not produce.
What You Should Do
If you are a builder evaluating steel-to-timber conversion on a current project: Run the numbers in EC3 before committing. It takes less than an hour for a first pass on a mid-rise project, the tool is free, and the output gives you a quantified carbon reduction that institutional clients, municipal incentive programs, and future certification requirements will increasingly ask for. The tariff is already pushing you toward timber on cost. Knowing the carbon number lets you capture additional value from that decision without changing the decision itself.
If you are a homeowner or buyer: Ask your builder what structural system they are using and why. If they switched from steel to timber because of tariff economics, ask them to quantify the embodied carbon difference. They probably cannot. That gap between the material decision and the carbon awareness is where the industry needs to move.
If you are an architect or designer: The Bath AI tool and EC3 together cover the full design lifecycle, from early conceptual descriptions through detailed material specification. Use them. The carbon data will be required by code within five years in most major jurisdictions if the trajectory of Buy Clean legislation continues, and building the analysis workflow now is cheaper than retrofitting it later under deadline pressure.
Limitations
Embodied carbon data varies widely by source, region, and manufacturing process. Published EPDs from different manufacturers for nominally equivalent products can differ by 30% or more, which means that any aggregate calculation is a range estimate, not a precise measurement. The 20-25 metric ton per-home figure used above applies to projects where steel framing was the baseline structural system. For single-family construction where wood framing was already standard, the tariff-driven carbon impact runs through component substitution at much smaller magnitudes. Mass timber supply chain capacity in the United States is immature: Timberlab's two fabrication plants and forthcoming sawmill represent a significant fraction of domestic production, and scaling to meet the demand implied by widespread steel-to-timber conversion would require supply chain buildout that takes years. Biogenic carbon storage in timber is credited in this analysis using the convention that the building stands for its design life, but end-of-life emissions from demolition and disposal are not modeled, and the WRI's critique of net carbon claims for harvested wood products remains unresolved in the scientific literature.