ConstructConnect’s summer forecast lands on a number that reframes the year: stripped of data centers, nonresidential building spending is projected to decline 3.8% in 2026. Data centers alone are what keep the sector from contracting. The “Offices” category — which now folds data centers into the same bucket — is forecast to rise 63% to more than $26 billion in starts, while the segment’s year-to-date spending has reached $58.1 billion, roughly four times the record pace set a year earlier.

The contrast with traditional product is stark. Conventional office starts totaled just $9.1 billion in 2025 — the lowest since at least 2020 and down 36% from the prior year. Total nonresidential put-in-place is still projected near $1.3 trillion (+5.4%) for 2026, but that headline masks a sector leaning hard on one demand source.

There is a counterweight forming on the other side of the ledger. Gallup finds 71% of Americans oppose AI data centers in their local area. Opposition groups have doubled to 833 across 49 states, and grassroots resistance blocked 75 projects worth roughly $130 billion in Q1 2026 — matching the full-year 2025 total in three months.

➤ Implications / Our Read

The healthy way to read a +5.4% construction year is broad recovery. That reading is wrong. This is not a rising tide; it is one tower of demand standing in a draining pool. When a single, highly cyclical end-use is the difference between growth and a 3.8% contraction, the industry’s apparent strength is really a concentration bet.

That matters for anyone pricing construction risk. Contractors, materials suppliers, lenders, and trades that have repositioned around data-center work are now correlated to one capital cycle — hyperscaler AI capex — and to the willingness of utilities to deliver power and communities to permit the sites. Diversified backlog has quietly become a competitive advantage again, because the backlog that looks busiest is the least diversified.

The entitlement data is the part the spend figures hide. A project can be fully financed and still never break ground if the jurisdiction says no — and increasingly it does. $130 billion blocked in a single quarter is not noise; it is the supply side of construction’s one growth engine meeting organized local resistance. Speed-to-power was last cycle’s constraint. Speed-to-approval is becoming this one’s.

Developers (non-data-center): expect competition for trades and long-lead equipment from hyperscale projects; lock pricing and capacity early, and watch for relief if AI starts slow.

Contractors & trades: concentration is margin today and fragility tomorrow. A diversified backlog is worth a lower bid.

Lenders: construction-loan exposure is more correlated to one capex cycle than portfolio labels suggest. Stress the data-center scenario explicitly.

Communities & site selectors: entitlement, not financing, is now the binding constraint on the sector’s growth engine. Local approval risk deserves first-page underwriting.

Whether AI infrastructure demand holds long enough to carry the sector through 2027 — and whether organized local opposition slows starts before any demand softening does. Both risks point the same direction: each would expose how little is holding up nonresidential construction underneath the data-center line.

Nonresidential construction looks like a +5.4% growth year. Take out one building type and it’s a 3.8% decline — the whole industry is riding a single, contested engine.

Twelve outside-the-deal reads shaping the capital backdrop this week:

1 — The Big Shift in CRE Capital Markets — banks and private credit are now partnering, not competing, across the capital stack.

2 — CMBS Stress Set to Intensify in 2026 — RXR targets delinquent loans as Chicago leads CMBS distress at 22.7%.

3 — Deloitte’s 2026 CRE Outlook — recovery with a reckoning; delinquencies stay above decade averages.

4 — CBRE U.S. Real Estate Outlook 2026 — investment volume seen rising 16% to $562B, near the pre-pandemic norm.

5 — JPMorgan’s 2026 CRE Trends — multifamily leads, industrial and retail hold, office still lags.

6 — Cushman & Wakefield U.S. Outlook 2026 — the prime-versus-secondary divergence widens across every asset class.

7 — The Fed’s June Banking-Resilience Report — regulators flag emerging private-credit and real-estate risks.

8 — CBS News: The 2026 Economy in Five Questions — affordability, rates, and jobs frame the year ahead.

9 — FOMC June 17 Projections — the Fed’s own dot plot on where rates and growth land.

10 — Capital Economics U.S. Housing Chart Pack — listings up, prices down a seventh straight month.

11 — PwC 2026 Real Estate Deals Outlook — where M&A and platform consolidation head next.

12 — Stanford SIEPR: The U.S. Economy in 2026 — the macro signals worth watching into the second half.