Tariffs Split Construction Into Two Pricing Regimes

President Donald Trump’s 50% tariffs on Canadian cement, plywood, and machinery, imposed Saturday under Section 338 of the Smoot-Hawley Tariff Act of 1930, land on a construction industry that was already split between data-center contractors carrying 11.4 months of backlog and everyone else at 7.5 months — one side with pricing power, the other without. This round will sort private real estate capital accordingly.

Bisnow first reported the tariffs, proposed July 20 and joining existing charges of up to 50% on Canadian aluminum and steel from 2025 plus lumber duties raised last year; the administration turned to Section 338 after the Supreme Court in February ruled its preferred tariff method unconstitutional, a reminder that the policy is being assembled with whatever legal tool is at hand.

Construction materials prices were already 7.4% higher than a year ago, per Associated Builders and Contractors data cited by Bisnow, with copper wire and cable up 17.9%, iron and steel up 17.6%, and softwood lumber up 15% — a base of increases the new duties arrive on top of just as developers write pro formas for deliveries 18 to 36 months out.

The cement channel shows how a border tax becomes a national pricing event. Matt Long, a partner at Phoenix-based multifamily developer and contractor Porter Kyle, told apartmentbuildings.com — as Bisnow reported — that cement is bought regionally but repriced nationally, so a tariff on Canadian imports gives domestic producers more pricing power across the country; Canada and Mexico account for 27% of U.S. cement imports and nearly 7% of U.S. cement consumption, according to a 2025 statement from the American Cement Association cited by Bisnow.

Long’s answer to why this round hurts more than the steel duties is the one lenders should underline: “The deeper cost is uncertainty. Production homebuilding, especially build-to-rent, is underwritten 18 to 36 months out. Adding 50% to an input class with 30 days’ notice introduces uncertainty, which is included in every bid as contingency.”

A 50% cost shock delivered with a month’s notice does not simply add a materials line item to a pro forma; it widens the bid-ask between the developer’s required return and the lender’s stabilized-cost estimate, moves the loan-to-cost ratio and the required equity check, and in build-to-rent — a product priced against for-sale housing on a monthly payment basis — leaves less room for the rent growth needed to clear a yield to absorb it.

The backlog spread already shows which side can carry the contingency, and ABC Chief Economist Anirban Basu told Bisnow that the spiraling trade war is “largely negative for contractors on both sides of the border” and will “further expand the cost of delivering construction services, dampening construction starts in the process.”

Backlog is only part of the data-center advantage. As this publication has argued, data center debt is becoming its own asset class; the underwrite leans on power contracts, tenant credit, and escalation clauses, which gives sponsors somewhere to park a cost shock. A build-to-rent pro forma has no equivalent shock absorber, which means the sponsor eats the contingency, the lender reprices the risk, or the deal waits for cheaper inputs.

For private real estate capital, the gap becomes the first real test of whether infrastructure-style data-center underwriting can absorb a hard-cost shock that conventional projects cannot. Contractors with a year of locked-in work and creditworthy tenants likely have more room to pass through costs; developers with a 30-month build-to-rent pro forma can only add contingency and hope. The lenders who price that difference will separate the sponsors who can deliver on time and on budget from those who cannot.

The policy’s speed matters too: the duties were proposed July 20, negotiated for almost a month, and imposed on a Saturday — fast movement even by the standards of a trade war. For an asset class underwritten 18 to 36 months out, the policy’s half-life may be shorter than the construction cycle, which is exactly why uncertainty is priced into every bid. The next ABC materials reading and the next build-to-rent construction loan to close will be more telling than any single tariff rate: the first will show whether the spread between the data-center backlog and everyone else’s widens, and the second will show what that spread costs in contingency.