Data Centers Are Keeping US Construction Afloat – But the Rest of the Market Is Struggling
Data center construction by hyperscalers like Amazon and Meta is keeping the US construction industry in positive territory heading into Q1 2026 – but strip that out and the picture gets uncomfortable fast. Material costs are up 45.3% since February 2020, non-residential spending just slipped 0.1%, and residential construction is quietly unraveling. The companies best positioned right now aren’t the broadest. They’re the most specialized.
ABC chief economist Anirban Basu gave a pretty blunt assessment on April 8. Growth is still there. It’s just not evenly distributed – not even close. A narrow band of hyperscaler-funded data center projects is absorbing a lot of the industry’s capacity and headlines, while mainstream contractors face tightening job openings, rising input costs, and a residential market that’s stalling out.
For construction executives and ConTech founders trying to read where the market’s heading, the divergence matters. “Silicon Valley and Wall Street will be fine,” Basu said during the webinar. “Main Street – a little rocky.” That’s not a throwaway line. It’s a useful frame for thinking about where to position a business, a product, or an investment thesis right now.
This piece breaks down what’s actually driving the numbers, what the material price data tells you that the headline figures don’t, and what the data center boom means for firms that aren’t already in that lane.
Hyperscaler data center investment is masking broader softness across US construction. Material costs keep rising, non-residential spending is down year-on-year, and the residential market is locked up. The opportunity is real – but concentrated.
What the Data Center Boom Actually Looks Like on the Ground
Amazon, Meta, and their peers are building at a pace the rest of the market isn’t
When Basu talks about hyperscalers driving construction momentum, he’s describing something structural, not cyclical. Companies like Amazon, Meta, Google, and Microsoft have multi-year capital expenditure commitments running into the hundreds of billions. That spending doesn’t pause when interest rates shift or tariff policy changes overnight. It’s essentially pre-committed demand – and construction firms that have built the capability to execute on those projects are in a genuinely different market right now than firms that haven’t.
The challenge, as we’ve covered in depth before, is that fast-track data center construction demands a very specific operational model. It’s not just about having the workforce or the bonding capacity. Speed requires pre-qualified supply chains, standardized procurement, and contractor relationships built across multiple prior projects. Skanska’s Advanced Technology division is a good example – they compress timelines by starting procurement before design is finalized, and by working with the same specialized subcontractors across different markets. That’s not replicable in 12 months. You either have it or you don’t.
For ConTech founders, this dynamic creates a pointed question about addressable market. If you’re selling to contractors who aren’t in the hyperscaler lane, you’re selling into a much softer environment than the sector-level headlines suggest. And if you are selling to hyperscaler-adjacent GCs and specialty subs, the speed and integration requirements are probably stricter than your current product assumes.
Related reading: How Fast Can You Really Build a Data Center? – BB’s deep dive into the operational model behind fast-track data center construction, featuring Skanska’s Katie Coulson.
The Material Price Problem Isn’t Going Away
Six years of cost inflation have fundamentally reset what “margin” means in US construction
The PPI data from the US Bureau of Labor Statistics is worth sitting with for a moment. Total construction input prices up 45.3% since February 2020. Natural gas up 156.7%. Nonferrous wire and cable up 82.8%. Steel mill products up 73.4%. These aren’t pandemic-era anomalies that have since corrected. They’re the new baseline.
- Natural Gas: +156.7%. The single largest mover, with implications for any energy-intensive construction process.
- Nonferrous Wire and Cable: +82.8%. Critical for data centers and electrical infrastructure – exactly the projects currently driving volume.
- Steel Mill Products: +73.4%. Tariff exposure compounds existing supply pressures on this one.
- Fabricated Structural Metal: +68.5%. A secondary steel effect, hitting commercial and industrial frames hard.
- Iron and Steel: +63.6%. Broad-based, affecting everything from foundations to mechanical systems.
- Concrete Products: +44.1%. Slower than metals, but still dramatically higher than pre-pandemic levels.
Basu was direct about the tariff connection: “I don’t think it’s a coincidence that many of the categories that have experienced the largest price increases are subject to tariffs.” That’s a significant point for anyone trying to separate structural inflation from policy-driven inflation in their cost models. The two are now deeply entangled, and there’s no obvious near-term path to unwinding either.
For contractors working on fixed-price or GMP contracts, this environment has been brutal. The industry’s average operating margin already sits somewhere around 2-3%. A 45% increase in input costs over six years, landing on top of margins that thin, doesn’t leave much room. The firms that have survived reasonably intact are generally those that renegotiated contract structures, built stronger escalation clauses, or moved toward cost-plus arrangements where possible.
Non-Residential Spending Is Softer Than the Headlines Suggest
Manufacturing leads the decline, and the residential market is structurally locked
Total non-residential construction spending fell 0.1% between January 2025 and January 2026. That sounds modest. But manufacturing took the largest hit, down 15% over the same period – a sharp reversal from the reshoring-driven surge that characterized 2023 and early 2024. Some of that pullback was inevitable as the initial wave of IRA and CHIPS Act-funded projects moved through the pipeline. Some of it reflects genuine uncertainty about trade policy and long-term demand.
Residential construction is a more complicated story. Basu’s framing was direct: “People talk about the housing affordability crisis – we’re not actually solving the problem.” Millennials are still largely priced out of the single-family market, which would normally be the driver of new construction activity. The underlying demand is there. The purchasing power, at current mortgage rates and current land and materials costs, largely isn’t. That gap has been a persistent drag on residential starts, and there’s no clear policy signal that it’s about to change.
“I think there’s enough to keep the economy afloat. But there’s still real risk of recession in this current volatile market.”Anirban Basu, Chief Economist, Associated Builders and Contractors – April 8, 2026 webinar
The labor picture adds another layer. Basu had expected job openings in construction to tick up as immigration policy changes pushed employers toward domestic or documented workers. That uptick hasn’t materialized. With roughly 25% of the US construction workforce foreign-born, any sustained reduction in that labor pool – without a corresponding training and recruitment pipeline for domestic workers – puts capacity pressure on an industry that’s already stretched.
What This Means If You’re a Contractor, a Founder, or an Investor
Three different positions, three different reads on the same data
If you’re running a contracting business, the data center opportunity is real – but access isn’t free. Breaking into that market requires capital investment in specialized workforce, supply chain pre-qualification, and in many cases, a track record that takes years to build. That doesn’t mean it’s closed. It means the entry cost is high and the timeline is longer than a typical business development push. The more immediate lever for most contractors is cost management: tighter escalation clauses, more aggressive value engineering during pre-construction, and a harder look at which project types still generate acceptable returns at current input prices.
If you’re building a ConTech product, the bifurcated market matters for your go-to-market. Selling speed and cost control into the hyperscaler lane requires deep integration with procurement workflows and the ability to demonstrate ROI in compressed timelines. Selling into the broader market – where spending is soft and margins are thin – requires a very clear and fast payback story. Vague productivity claims don’t land right now. Specific, defensible numbers do. For a fuller view of where construction tech investment is heading this year, our 2026 trends piece is worth a read: Our Take on 2026 Construction Trends.
If you’re an investor or allocator, the question is whether data center construction exposure is already priced into the GCs and specialty subs you’re looking at – and whether those businesses have the operational depth to sustain that exposure as hyperscaler capex cycles inevitably shift. The companies most worth watching are the ones with repeatable, relationship-driven execution models, not just headline revenue tied to a single client vertical.
Get the construction briefing that doesn’t sugarcoat the numbers
Every week, 2,500+ construction execs, ConTech founders, and AEC investors read Bricks & Bytes for the takes they’re not getting anywhere else.
Join 2500+ ReadersThe Recession Question
Basu isn’t calling a crash – but he’s not dismissing the risk either
Basu’s position is measured: enough momentum to keep the economy moving, but real and present downside risk. That’s not a contradiction. It’s an honest read on an environment where a narrow set of sectors – data centers, AI infrastructure, energy – are carrying a lot of weight. If hyperscaler capex pauses, or if tariff escalation tips key material categories into another leg up, the buffer disappears quickly.
For construction businesses, the practical implication is that scenario planning matters more right now than it has in a while. Not catastrophizing, but building genuine operational flexibility. Which cost lines can move quickly if project volume drops? Which client relationships are durable across a softening cycle? Which technology investments reduce fixed cost exposure rather than just adding capability? These are the questions that separate businesses that navigate a volatile year cleanly from those that don’t.
| Segment | Current Spending Trend | Key Driver | Labor Exposure | Material Risk | Outlook |
|---|---|---|---|---|---|
| Data Center / AI Infrastructure | Strong growth | Hyperscaler capex commitments | High (specialized workforce) | High (wire, cable, steel) | Positive near-term; cyclical risk if capex pauses |
| Manufacturing | Down 15% YoY | Post-IRA/CHIPS project pipeline normalizing | Moderate | Moderate to high | Uncertain; trade policy sensitivity |
| Residential (Single-Family) | Declining | Affordability crisis, mortgage rates | High (foreign-born workforce exposure) | High (lumber, concrete) | Weak; no near-term structural fix visible |
| Broader Non-Residential | Flat to slightly down | Mixed; office weak, some infrastructure stable | Moderate | Moderate | Cautious; margin compression ongoing |
| Energy Infrastructure | Growing | Grid expansion, data center power demand | Moderate | High (natural gas inputs) | Positive; tied to AI infrastructure cycle |
Hyperscaler companies – Amazon, Meta, Google, Microsoft, and others – have made multi-year, multi-billion-dollar commitments to AI and cloud infrastructure buildout. That spending is largely pre-committed and continues regardless of short-term economic volatility. It’s creating sustained demand for specialized construction capacity at a time when much of the broader non-residential market is flat or declining. According to ABC’s Q1 2026 forecast webinar, this hyperscaler investment is the primary driver keeping total construction activity in positive territory. (Source: ENR, April 2026)
According to the US Bureau of Labor Statistics, total inputs to construction industries rose 45.3% between February 2020 and February 2026. Individual categories moved far more dramatically – natural gas up 156.7%, nonferrous wire and cable up 82.8%, and steel mill products up 73.4%. ABC’s chief economist Anirban Basu directly linked many of the sharper increases to tariff exposure, noting the overlap between high-inflation categories and tariffed goods. These aren’t figures that are expected to reverse quickly. (Source: ENR / BLS, April 2026)
Total construction job openings are down from recent highs. Basu noted that he had expected a rise in openings as immigration enforcement pushed employers to recruit more domestic or documented workers – given that about 25% of US construction workers are foreign-born – but that surge hasn’t materialized. The labor market picture is therefore one of softening demand meeting constrained supply, which creates ongoing capacity risk, particularly in specialized trades. (Source: ENR, April 2026)
Basu stopped short of calling a recession but was clear that the risk is real. His read is that there’s enough activity in high-growth sectors – data centers, energy, AI infrastructure – to keep the industry moving, but that the concentration of that growth in a narrow set of segments creates fragility. If hyperscaler capex slows, or if material cost increases accelerate further, the buffer disappears. He described the current environment as “volatile,” with Main Street facing a harder path than Silicon Valley or Wall Street. (Source: ENR, April 2026)
Residential construction continues to decline. The core issue, as Basu framed it, is that the housing affordability crisis isn’t being solved – millennials remain largely priced out of the single-family market due to high mortgage rates, high land costs, and materials inflation. Demand exists in aggregate but lacks purchasing power at current price points. There’s no near-term policy intervention that appears likely to materially change the trajectory. (Source: ENR, April 2026)
The bifurcated market creates two distinct selling environments. In the hyperscaler-adjacent lane, buyers are active but have high expectations around speed, integration depth, and demonstrated ROI in compressed timelines. In the broader market, spending is soft and decision cycles are longer, which means vague productivity claims don’t cut through. The clearest path for most founders is a sharper ICP – understanding precisely which segment of the market you’re selling into and tailoring your value proposition accordingly rather than trying to serve both simultaneously.
Related Articles
How Fast Can You Really Build a Data Center?
Our Take on 2026 Construction Trends
