Project Management & Efficiency, Technology & Innovation

AI Is Becoming Construction’s Critical Infrastructure. Your Resilience Plan Should Be, Too.

The firms adopting AI the fastest are also the most exposed to a supply chain risk the industry hasn't faced before — one that looks a lot like lumber in 2021. Here's what the smartest teams are doing about it.

Frank Kalman
April 29, 2026

PDX Reimagined: How ZGF Architects Built a Digital Backbone for Oregon’s Most Ambitious Project

Project Management & Efficiency, Technology & Innovation
Frank Kalman

On the state’s biggest public works project, the hardest part wasn't the engineering but keeping 6,000 sheets — and an entire team — in sync.

When travelers step into Portland International Airport‘s new main terminal, the first thing they see is nine acres of timber soaring overhead — a wood canopy engineered to survive a magnitude 9.0 earthquake, filtering daylight across 72 full-size trees.

The roof was prefabricated in 18 massive sections, each the size of a football field, then rolled across the tarmac and slid into place overnight while ticketing, security and baggage operations kept running below.

Most passengers don’t think about what it took to build it. They just look up.

Behind that canopy sits a different kind of architecture — nearly 6,000 coordinated drawing sheets, thousands of stakeholders, and a documentation effort that became the largest permit set in Oregon history. At $2 billion and 1 million square feet, the Terminal Core Redevelopment was the biggest public works project the state had ever attempted. And it could never, for a single day, shut the airport down.

“Everybody loves Portland International Airport,” said Nat Slayton, principal and senior technical designer at ZGF Architects, the project’s design lead. “It’s a place that belongs to the community. That was the challenge: how do you evolve it while making it something people will love just as much as the original?”

Then COVID hit. And the hardest part of the project got a lot harder.

When the War Room Went Dark

Before 2020, collaboration at ZGF meant proximity. Walls plastered with drawings. Teams shoulder to shoulder, talking through conflicts, marking up together in real time.

“We had entire walls just covered in drawings,” recalled Michael Adams, BIM manager at ZGF. “You’d bring people into the room, talk through a problem and mark it up together.”

COVID eliminated that overnight. The largest design team in Oregon history — engineers, architects, consultants, contractors, Port of Portland stakeholders — was suddenly scattered across home offices. And the project couldn’t pause.

“All of that scale and inertia collided with COVID,” Slayton said. “It was the largest project the state had ever seen — and then COVID hit at the worst possible moment.”

That’s when Bluebeam stopped being a tool and became something closer to infrastructure.

A New Front Door

ZGF moved its entire workflow into Bluebeam Studio Sessions — shared digital environments where dozens of stakeholders could mark up the same drawing set simultaneously, from anywhere. What had required everyone in the same room now happened virtually without slowing the project.

“It quickly turned into my front door,” Adams said.

The team crowdsourced tool sets across disciplines. Color-coded markup standards gave structural engineers in one time zone and architects in another a shared visual language — no confusion about who flagged what or what had been resolved. Sets linked thousands of documents into a single navigable system. Slip Sheeting kept revisions clean. Status tracking made accountability visible to everyone, including owners and contractors.

Review cycles that once took weeks compressed into days. Discrepancies surfaced before they became field problems. Markup histories created a living audit trail that project leads could pull up at any point.

But one of the more unexpected benefits was what it did for the people earliest in their careers.

“You could see how experienced people thought through a problem,” said project architect Christian Schoewe. “That kind of access wouldn’t have been possible in the old room setup.”

In the war room model, junior staff rarely witnessed how senior designers reasoned through complexity. In Studio Sessions, that reasoning was right there in the thread — visible, traceable, instructive. Coordination became mentorship without anyone planning it that way.

Memory, Not Just Efficiency

Years into construction, Schoewe used Bluebeam’s archive to pull a markup that justified a critical roof detail. The digital record was still there. The decision was documented. The team avoided a costly omission.

That moment captures something the speed metrics don’t. Digital delivery isn’t just faster — it’s persistent. When markups, resolutions and revision histories live in one centralized system, institutional knowledge survives personnel changes, project phases and the passage of time.

On a project that stretched across years, across a pandemic, across tens of thousands of daily travelers moving beneath active construction — that kind of continuity wasn’t a nice-to-have. It was operational risk management.

The Part That Stays with You

The PDX terminal opened to the public, and Schoewe walked through the completed ticketing hall and watched passengers look up at the timber canopy for the first time.

“I still get a kick out of seeing people’s reactions,” he said. “You can almost read their lips: How did they do that with all that wood?”

For Slayton, the pride was in who built it. Douglas fir sourced within 300 miles. Timberlab crews assembling the massive roof panels. Local artists filling the concourses with public work. “This was made by the talents and skills of the people they live with in their state,” he said.

For Adams, it came down to something simpler. Every decision — wider security lanes, more daylight, open green space — was measured against one question. “That was the mission,” he said. The passenger.

The lesson extends well beyond Portland. As civic infrastructure grows more ambitious and more constrained by operational realities, the ability to coordinate at scale — without physical proximity, without shutting anything down — becomes the thing that determines whether a project survives its own complexity.

At PDX, that ability didn’t come from a single engineering breakthrough. It came from disciplined information management, built on a digital backbone that held through COVID, construction and everything in between.

Explore the full ZGF Architects case study.

Read case study

Your Takeoff Is Wrong. Here’s Why That Matters More Than You Think.

Project Management & Efficiency, Technology & Innovation
Frank Kalman

Most construction profits don’t die in the field; they’re killed weeks earlier, at a desk, when someone writes down the wrong number.

Most construction mistakes don’t happen in the field. They happen weeks earlier, at a desk, when someone measures 185 cubic yards of concrete and writes it down as fact. Then the crew shows up, the pour comes up short, and suddenly everyone’s scrambling to explain how the numbers were off by 10 yards.

That’s the thing about quantity takeoffs: when they’re right, nobody notices. When they’re wrong, the entire project feels it.

Small Errors Create Outsized Problems

A missed room. An unmeasured run of conduit. A slab thickness that was assumed instead of verified.

Individually, these sound minor. Walk into any project debrief where things went sideways, and you’ll hear the same refrain: “It was just one thing.” But that one thing multiplied across labor, materials and schedule becomes the reason a job that looked profitable on paper ended up underwater.

Because quantities drive almost every other decision. When the takeoff is off, everything downstream compounds:

Pricing lands wrong — either too aggressive to be sustainable or too padded to win.
Labor plans don’t match the real scope, and crews end up standing around waiting for clarity.
Material orders fall short, deliveries get delayed, and the schedule slips while everyone points fingers.
By the time the issue shows up in the field, it’s usually too late to fix it cheaply.

That’s the brutal economics of bad takeoffs: the error is cheap to prevent and expensive to repair.

The Winner’s Curse Starts with Bad Quantities

Underestimating scope is one of the most common ways takeoffs fail, not to mention one of the most dangerous.

When quantities are missed, bids come in low. You win the job. Congratulations. Except you didn’t win because you’re more efficient or better organized. You won because your estimate was incomplete. That’s winning work you can’t afford to build — the winner’s curse in action.

Now you’re locked into a contract where the only way to recover margin is through change orders, value engineering under pressure or eating the cost outright.

Overestimation isn’t harmless, either. Padding quantities to compensate for uncertainty might protect margin, but it makes bids less competitive. On a tight race, that extra 5% contingency buried in inflated scope can be the difference between winning and placing second.

Accurate takeoffs are what allow contractors to bid confidently without hiding behind excessive buffers. You price what’s there, not what might be there if everything goes wrong.

Accuracy Is Also About Trust, Accountability

Project managers rely on estimate quantities to build budgets and schedules. Superintendents use them to plan manpower and logistics. Procurement teams depend on them to stage deliveries and coordinate suppliers.

When those numbers don’t line up with reality, trust erodes quickly. And once trust is gone, every conversation becomes adversarial. The PM questions the estimate. The super questions the buyout. The owner questions the team. Everyone’s defensive because nobody knows which number to believe.

Modern digital workflows make every measurement visible on the drawing and traceable in the data. That transparency isn’t about micromanagement. It’s about making it easier to have productive conversations about scope before the job is awarded, when changes are still cheap.

When someone asks, “Where did this number come from?” you can show them. Not a vague explanation. The actual markup on the actual sheet with the actual measurement tied to it. That’s the kind of accountability that keeps teams aligned.

Accuracy Makes Estimates Easier to Revise

No set of drawings stays static. Addenda happen. Clarifications come in late. Architects change details three days before bid. Scope shifts.

When takeoffs are clean and well-organized, revisions are manageable. You can update affected quantities, isolate the differences, and assess downstream impacts without rebuilding everything from scratch.

But when takeoffs are messy — when assumptions are buried in formulas, or measurements aren’t tied back to drawings, or quantities are scattered across disconnected files — every revision becomes a partial rebuild. You’re not just updating numbers. You’re trying to figure out what the original numbers even meant.

Accuracy at the takeoff stage isn’t just about getting the first number right, but about creating a foundation that can absorb change without falling apart. Because change is guaranteed. The only question is whether your process can handle it.

The Uncomfortable Truth

No amount of pricing accuracy can fix bad quantities.

You can have the best cost database in the industry. You can negotiate killer subcontractor rates. You can sharpen your pencil until it’s a needle. None of that matters if the scope you’re pricing isn’t the scope you’re building.

The quantity takeoff is the independent variable. Everything else — pricing, labor planning, procurement, scheduling — depends on it. When it’s wrong, the estimate will be wrong, whether it’s over- or underpriced.

That’s why accuracy matters. Because the margin for error in construction is razor thin. On a good job, net profit might land in the low single digits. There’s no room for compounding errors that start early and ripple through the rest of the project.

So, before you rush to price, before you sharpen that pencil, make sure the quantities are right. Because if they’re not, nothing else you do will matter.

Want to see how modern takeoff workflows hold up when drawings change?

Get the Guide

Why Most Takeoffs Fall Apart When Drawings Change — and How Better Structure Prevents It

Project Management & Efficiency, Technology & Innovation
Frank Kalman

Revisions don't break estimates. Weak takeoff workflows do.

Most quantity takeoffs don’t fail during the first measurement pass. They fail later when the drawings change.

At bid time, everything looks solid. Quantities check out. Pricing feels competitive. The estimate goes out the door. Then an addendum drops. A slab thickens. A wall type shifts. A scope clarification lands late Friday afternoon. Suddenly, what looked airtight starts to leak.

That’s the real stress test of a takeoff — not how fast it was produced, but how well it handles change.

Revisions are unavoidable. Treating them like edge cases is one of the most common — and expensive — mistakes in estimating. The difference between takeoffs that hold up and those that unravel rarely comes down to effort or experience.

It comes down to structure.

A quantity takeoff is the process of measuring and listing material quantities directly from construction drawings — the foundation every estimate is built on. When those drawings change, the takeoff must change with them. If it can’t, the estimate drifts. If it can’t update quickly and accurately, teams end up guessing, and guessing creates risk and lost money. The question isn’t whether drawings will be revised. They always are. The question is whether your workflow was built to absorb it.

Why do drawing changes break so many takeoffs?

Revisions rarely introduce new complexity; they instead expose weaknesses already embedded in how quantities were captured, organized and traced. When takeoffs are built as if drawings are final, even minor changes force disproportionate rework. Failure isn’t about change itself, but about how prepared the workflow is to absorb it.

Addenda don’t create chaos. They reveal it.

Too many takeoffs are built as if the drawings are final, even when everyone knows they aren’t. Quantities get measured fast. Assumptions sneak in early. Data moves downstream before it’s stable. When revisions arrive, teams aren’t adjusting but rebuilding.

That’s when accuracy slips. That’s when confidence erodes. And that’s when estimating turns reactive instead of controlled.

A takeoff that can’t be revised cleanly wasn’t finished. It was fragile.

Why do drawing changes break so many takeoffs in practice?

Most revision failures follow predictable structural patterns: unclear quantity definitions, weak organization and lost traceability between drawings and numbers. These issues stay hidden during initial takeoff but surface immediately when scope shifts. The more assumptions embedded early, the harder it becomes to isolate what truly changed.

Most revision failures follow the same patterns. They stay hidden until the drawings move.

The most common takeoff breakdown triggers include:

Time pressure and last-minute addenda that force rushed, manual updates — where errors sneak in fast and get caught late.
Working from outdated drawing sets when version checks are skipped — the single most avoidable source of rework.
Miscalibrated digital scales — a single wrong calibration can introduce roughly 10% quantity error across an entire sheet.
Decentralized files — takeoffs saved on personal drives with no audit trail mean teams repeat the same mistakes project after project and have no way to prove what changed or when.

Quantities aren’t clearly defined: In many workflows, quantity takeoffs get contaminated early. Waste factors, allowances and procurement logic are baked into measurements before pricing even starts. When a revision hits, it’s no longer clear what was measured and what was assumed.

If slab thickness changes, which number needs to move? The geometric quantity? The waste-adjusted total? The priced value buried three steps downstream? Without clean separation, every revision turns into a guessing game.

Organization comes too late — or not at all: Inconsistent naming, mixed layers and improvised groupings make initial takeoffs harder to review and revisions harder to isolate. Instead of updating a specific system or floor, estimators end up combing through entire datasets trying to figure out what changed.

When structure is missing, small revisions snowball into major rework.

Quantities lose their visual tie to the drawings: Once numbers move into spreadsheets or estimating systems, their connection to the drawing often weakens. Markups stop reflecting current scope. Reviews shift from visual verification to trust-based reconciliation.

At that point, no one is fully confident which quantity is right and proving it burns time teams don’t have.

Data gets copied too early: Manual exports and copy-and-paste workflows introduce version drift almost immediately. When quantities change at the source but not everywhere else, teams spend more time reconciling numbers than evaluating impact.

Revisions should trigger adjustments. Too often, they trigger audits.

What do revision-resilient takeoffs do differently?

Teams that handle revisions well don’t rely on speed or heroics. They design takeoffs to expect change — by keeping quantities clean, visible and layered. Structure limits how far a revision can ripple, turning what could be a rebuild into a controlled update.

Teams that handle revisions without panic don’t have better luck. They have better structure. They design takeoffs for change, not just speed.

They keep the quantity takeoff clean: Revision-resilient workflows treat the takeoff as a stable foundation. Net quantities only. No waste. No pricing logic. No procurement assumptions.

That separation matters. When drawings change, estimators update what the drawings show — nothing more. Downstream logic adjusts without contaminating the base data.

When quantity, material strategy and pricing stay layered, changes stay contained.

They keep quantities visible: Every measurement stays visible on the drawing. Layers are used deliberately to isolate scope by trade, system or phase. Color makes coverage obvious.

Visual verification becomes the fastest revision check. If an area isn’t marked, it likely wasn’t measured — or updated.

This is where digital workflows outperform spreadsheets. Review doesn’t depend on trusting totals. It happens directly on the drawings.

They use overlay and comparison tools to isolate deltas: Overlay tools superimpose a new drawing over the prior version so differences jump out visually — without combing through every sheet manually. Instead of re-measuring the entire plan, estimators can isolate only the areas that changed, which can reduce hours of revision work to minutes.

Tools like Bluebeam include drawing comparison features built specifically for this workflow, letting teams generate side-by-side views of old vs. new sheets and flag only the affected quantities. It’s the difference between a surgical update and a full rebuild.

They organize for change, not just cleanliness: Revision-resilient takeoffs aren’t just tidy but are structured to limit blast radius.

Quantities are grouped so changes affect specific slices of scope, not the entire estimate. A revision to one system doesn’t force a rebuild of everything else.

That upfront discipline can feel slower. It pays off every time drawings shift.

They update quantities at the source: When revisions arrive, disciplined teams update measurements where they live — on the drawing. Downstream systems follow the updated data instead of chasing it across disconnected files.

This “update once, let everything else follow” approach prevents version drift and keeps the takeoff, estimate and budget aligned.

On BIM-enabled projects, that logic goes further: A live BIM link creates a direct connection between the 3D model and the takeoff data, so quantities adjust automatically when the model changes. Drawings and estimates update together rather than requiring manual reconciliation after every design iteration. For teams working on complex or fast-moving projects, BIM integration for takeoffs isn’t a luxury; it’s the only way to keep pace with design changes without burning estimator hours on cleanup.

Why are full takeoff rebuilds a warning sign?

Rebuilding an entire takeoff after an addendum usually signals fragile structure, not unavoidable complexity. When quantities, organization or traceability fail early, revisions feel catastrophic. In resilient workflows, change prompts targeted updates, not a reset.

Rebuilding an entire takeoff after an addendum isn’t normal. It’s a signal.

It usually means quantities weren’t clearly defined, structure was inconsistent or traceability was lost early. Time pressure makes rebuilds feel inevitable, but they’re often symptoms of fragile workflows, not unavoidable complexity.

In resilient workflows, revisions don’t trigger panic. They trigger a process: isolate the change, update the affected quantities, review the impact and move forward.

Adjustments beat rebuilds. Every time.

How do structured takeoffs change the estimator’s role?

When takeoffs are structured for revision, estimators spend less time re-measuring and more time evaluating impact. Judgment replaces firefighting. Experience shows up in understanding scope risk, cost implications and downstream effects — not in scrambling to reconcile numbers.

When takeoffs are structured, revisions shift where estimators spend their time.

Instead of re-measuring everything, estimators focus on validating scope, assessing impact and applying judgment where it matters. Experience shows up not in clicking faster, but in understanding how changes affect cost, schedule and risk.

Modern tools can surface changes quickly. They don’t replace accountability. Estimators still decide what counts, what doesn’t and what needs clarification.

Structure creates room for judgment. Without it, even experienced teams end up firefighting.

What does this mean for teams under constant bid pressure?

Revision-resilient takeoffs change how teams respond under pressure. Faster responses come from clarity, not haste. When quantities are traceable and visible, scope discussions sharpen, pricing adjustments speed up and confidence carries through award and handoff.

Revision-resilient takeoffs change how teams operate when the pressure is on.

They respond to addenda faster — not because they rush, but because they aren’t untangling their own work. Pricing adjustments are clearer. Scope conversations are sharper. Handoffs to project teams carry fewer question marks.

Confidence improves, too. When quantities are visible, traceable and cleanly separated, teams don’t second-guess themselves after award. They know where the numbers came from and how they changed.

Why should takeoffs be built for change, not ideal drawings?

Drawings will change. Scope will shift. Clarifications will arrive late.

The only real question is whether your takeoff workflow amplifies disruption or absorbs it.

Takeoffs built for speed alone crack under pressure. Takeoffs built with structure, visibility and discipline hold up — and make estimating less reactive, not more.

That isn’t about features but about designing workflows that treat change as expected, not exceptional.

Because in estimating, the work that lasts isn’t the fastest but the work that still makes sense when everything else moves.

Here’s a practical audit to run on your current workflow: Are your quantities cleanly separated from waste factors and pricing logic? Do your layers isolate scope by trade, system or phase? When a revision arrives, can you identify the affected area on the drawing without combing through the whole estimate? Is there a version-controlled document library with a complete revision history, or are takeoffs living on personal drives? If any answer is no, the next addendum will cost more than it should.

Bluebeam is built for exactly this kind of structured, revision-ready workflow — with purpose-built digital takeoff tools, overlay and comparison features, customizable layers and cloud-based collaboration that keeps quantities and drawings in sync.

Bluebeam Takeoff & Revision FAQ

How does Bluebeam help teams manage takeoff revisions?

Bluebeam keeps quantities tied directly to the drawing through visible markups and structured layers, making it easier to isolate changes and update measurements at the source rather than rebuilding downstream data.

Why is visual traceability important during revisions?

Visual traceability allows estimators to verify scope changes directly on the drawing instead of relying on abstract totals. This reduces reconciliation time and increases confidence when quantities shift.

Can Bluebeam separate base quantities from pricing assumptions?

Yes. Bluebeam supports clean quantity takeoffs that remain independent from waste factors, pricing logic or procurement strategy, allowing downstream estimating tools to adjust without corrupting the source data.

How do layers improve revision control in takeoffs?

Layers let teams organize quantities by system, trade, phase or scope segment, limiting how far a revision can ripple and making updates faster and more targeted.

Is Bluebeam suitable for high-volume addenda environments?

Bluebeam is designed for iterative review and revision workflows, helping teams manage frequent drawing updates without losing alignment between quantities, markups and estimates.

Why do small takeoff errors cause major project problems?

Small mistakes in takeoff calculations — misread scales, duplicated items, missed specification changes — compound across project phases. A quantity that’s off by 10% at bid can mean material shortages in the field, on-site adjustments that delay the schedule, and cost overruns that erode the margin a team worked hard to protect. The earlier the error enters the workflow, the further it travels before anyone catches it.

What manual pitfalls most often break takeoff reliability during revisions?

The most consistent offenders are working from outdated plan sets when version checks are skipped, miscalibrating digital scales (one wrong calibration can introduce roughly 10% quantity error across a sheet), and saving takeoffs on personal drives rather than a centralized system. Without shared, versioned storage, there’s no audit trail — which means teams repeat the same estimating mistakes from project to project with no way to learn from them.

How much time and cost do manual revisions typically add?

On mid-sized projects, manual revision workflows can double or triple the time required to update a takeoff compared to structured digital processes. Beyond the direct labor cost, manual updates delay bid responses, increase the risk of pricing errors carrying through to award, and create the kind of version drift that requires reconciliation sessions no one has time for. Construction firms that embrace structured digital workflows — with proper revision controls, centralized documentation and live comparison tools — build in the predictability needed to protect margins and maintain cash-flow clarity, especially as labor shortages and budget pressures intensify across the industry.

Get the full playbook for takeoffs that survive revisions.

Read The Guide

The Physical Economy Is the Bottleneck

Construction & The Economy, Project Management & Efficiency
Frank Kalman

As AI, data centers and advanced manufacturing surge, the real constraint on growth isn’t capital or software, but the skilled labor and physical systems required to build them.

For the past 20 years, we’ve told ourselves a comforting story about how progress works.

Software scales. Capital flows. Innovation compounds. The digital economy, we’re told, floats above the messiness of the physical world — lighter, faster, cleaner. If something matters enough, the logic goes, we’ll fund it, code it and ship it.

That story is starting to fall apart.

Across the United States, billions of dollars are lined up for AI infrastructure, semiconductor plants, grid upgrades and clean energy projects. The money and urgency are real. But the work increasingly isn’t getting done on schedule — or at all. And it’s not because the ideas are flawed or the capital is missing, but because the physical systems required to make those ideas real can’t keep up.

Data centers don’t come online because a model is ready; they come online when someone finishes pulling wire, installing switchgear and energizing the site. Chip fabs don’t run on ambition but on precision installation carried out by people with rare skills and years of experience.

Power grids don’t modernize themselves. They’re rebuilt, mile by mile, by crews aging out faster than they can be replaced.

What’s emerging isn’t a temporary labor shortage or a cyclical slowdown, but a structural bottleneck in the physical economy — the network of skilled trades, construction workflows and coordination systems that quietly underpin everything we call “digital.”

Construction productivity has been flat or declining for decades. Rework, miscommunication and outdated information quietly waste billions of dollars’ worth of skilled labor every year, labor no amount of capital can instantly replace.

The uncomfortable truth is this: the next decade of digital growth won’t be limited by processors, models or funding rounds. It will be paced by the slow, exacting work of building things in the real world, and by how well we support, coordinate and protect the people who do that work.

The signals are already there. Contractors report acute shortages in the trades required to build and power data centers, modernize the grid and bring advanced manufacturing online. Nearly half of the workforce is nearing retirement. Training replacements takes years under the best conditions, and even that pipeline is constrained by instructor shortages and long apprenticeship timelines.

All the while, productivity keeps moving the wrong way. While manufacturing and agriculture increased output per worker, construction has lagged. The systems around them simply haven’t kept pace with project complexity. Rework caused by conflicting drawings, unclear intent and poor coordination consumes skilled hours no hiring surge can quickly replace.

Together, these trends expose the faulty assumption at the heart of the digital economy: that physical execution will always be there when we’re ready for it. Capital can mobilize quickly. Software can iterate overnight.

The physical economy, however, moves at human speed — and right now, it’s being asked to move faster than it’s built to go.

That gap between ambition and execution is where the bottleneck lives.

The assumption everyone is making

For years, the dominant assumption behind economic growth was simple: If demand is real and capital is available, physical capacity will follow. That logic worked when growth was incremental and timelines stretched across decades.

Today’s build cycle is different, however. AI infrastructure, grid expansion and advanced manufacturing compress schedules while increasing complexity.

The digital economy moves at the speed of iteration. The physical economy moves at the speed of people, permits and coordination. Treating those speeds as interchangeable is the mistake shaping the next decade of growth.

The mismatch is already visible. Data centers rise faster than the systems that power them. Buildings go up. Equipment arrives. Then projects stall, waiting on specialized electrical work that can’t be rushed.

Modern AI facilities aren’t server warehouses; they’re dense electrical systems demanding precision installation and careful commissioning. In many regions, the limiting factor isn’t land or capital but the availability of the right crews.

The same pattern plays out in semiconductor manufacturing. Billions have been committed to reshoring chip production, backed by policy incentives and geopolitical urgency.

Yet factories don’t materialize on schedule because funding exists. Semiconductor fabs require installation work performed to exacting tolerances by highly specialized trades. When those teams aren’t available, timelines slip and capital sits idle.

Nowhere is the tension more consequential than the power grid. Every data center, fab and electrified system ultimately depends on transmission infrastructure that has barely expanded in decades. National goals call for rapid increases in grid capacity, but the workforce responsible for building and maintaining it is aging out.

Even fully approved projects hit the same hard limit: without trained line workers and electrical crews, the grid simply can’t grow fast enough.

Not all construction is the same

One reason this bottleneck has been slow to register is that, on the surface, construction looks uneven rather than constrained.

Office projects are slowing. Retail construction is soft. From a distance, it can appear capacity is freeing up — that workers from quieter markets will simply flow to wherever demand is hottest.

That assumption doesn’t hold up.

What’s growing isn’t general construction but mission-critical construction — the high-stakes work required to build data centers, semiconductor facilities and the electrical infrastructure that supports them.

These projects demand a different level of precision and coordination. The skills required to build a speculative office shell aren’t interchangeable with those needed to install high-voltage switchgear, commission backup power systems or work inside cleanrooms.

The result is a misleading picture. Aggregate data suggests slack. On the ground, the trades that matter most to the digital economy are stretched thin. Electricians, pipefitters, instrumentation technicians and line workers are booked months out, even as other segments cool.

It’s a bifurcated market, and digital growth sits squarely on the overheated side.

Why labor isn’t fungible

In theory, labor moves to where demand is strongest. In practice, however, skilled physical work doesn’t behave that way.

The electricians needed to energize a data center aren’t interchangeable with crews framing an office building. Semiconductor tool installation can’t be staffed overnight by general labor. These roles require years of training, system-specific experience and precision that only repetition provides.

That rigidity shows up clearly on data center projects. A modern AI facility can sit largely complete — walls up, racks staged, cooling installed — while progress stalls at the electrical layer. High-voltage crews are booked months in advance. Bringing in less specialized labor isn’t an option.

Energizing a data center isn’t about speed; it’s about correctness. One mistake can delay commissioning indefinitely.

So, work waits.

Semiconductor fabs reveal the same dynamic at higher stakes. Installing tools inside a fab requires tradespeople trained for ultra-clean environments and unforgiving tolerances. These aren’t skills borrowed from adjacent projects when timelines tighten. When those teams aren’t available, work simply pauses.

No amount of funding compresses the learning curve.

Why automation hasn’t solved this

It’s tempting to assume automation will absorb the shortage. That logic worked in manufacturing and logistics. Construction — especially mission-critical work — resists it for a reason.

Jobsites are unstructured environments. Conditions change daily. Materials arrive out of sequence. Work happens overhead, underground and inside live systems where errors mean outages or safety risks. Skilled humans adapt. Machines still struggle.

Robotic welding illustrates the gap. In factories, robots thrive. Parts are standardized. Conditions are predictable. On active jobsites, though, that structure disappears. Welds happen in tight chases, overhead, around existing systems. A skilled welder adjusts instinctively. A robot’s advantage collapses.

Automation helps at the margins. Drones speed surveying. Software improves layout and coordination. Robotics reduce physical strain.

But these tools multiply human effort; they don’t replace it. The work that defines mission-critical construction remains stubbornly human.

The hidden drain: wasted labor

If labor can’t be replaced quickly and automation can’t solve the shortage, the most consequential question becomes quieter:

What happens to the labor we already have?

This is where capacity leaks. Rework, miscommunication, outdated information and fragmented workflows quietly consume skilled hours that can’t be recovered. In an environment where experience is scarce, every lost hour matters.

Much of this waste isn’t caused by the work itself, but by the systems around it. Crews aren’t slowed by lack of skill, but conflicting drawings, unclear intent and version confusion. When that happens, progress stalls and work gets redone.

This is where disciplined document control and shared visibility matter. When teams work from a single, current set of drawings — with markups tied directly to scope and intent — fewer hours are lost correcting avoidable mistakes.

Reducing rework doesn’t create new workers, but it effectively gives time back to the ones you already have.

Treating physical labor as strategic infrastructure

Once labor is understood as constrained, the logic changes. Skilled physical work stops looking like a variable cost and starts looking like infrastructure.

The most effective organizations are already adjusting. They invest upstream in training partnerships, rethink sequencing and design workflows that reduce friction on site. They don’t do this because it’s fashionable, but because the economics demand it. When skilled labor is scarce, waste becomes intolerable, and coordination becomes a competitive advantage.

This is where digital tools earn their keep — not by replacing people, but by helping crews spend more time building and less time untangling errors. Clarity, accuracy and shared context become forms of capacity.

The physical premium

As constraints converge, a new reality takes shape. Physical execution — the ability to build, connect and commission systems — is becoming more valuable than the plans that describe them.

This physical premium shows up in subtle ways. Projects delivered on time command outsized value. Existing infrastructure appreciates because replicating it is slower and more expensive. Timelines stretch not because demand is weak, but because execution can’t accelerate without risk.

What makes this moment different is its durability. Demographics are locked in. Training moves at human speed. Automation assists but doesn’t replace.

The pace of the digital economy is increasingly set by the limits of the physical one.

What this means for the next decade

The defining constraint of the next decade won’t be ambition but execution — the physical work required to turn plans into functioning systems. As digital investment accelerates, the gap between what we want to build and what we can build will widen.

Progress won’t stop, but it will become selective. Projects that plan around physical limits — training timelines, coordination complexity and labor scarcity — will move forward. Those that assume the physical economy will bend on demand will struggle.

Over time, value will shift. Skilled labor will be treated less like an expense and more like strategic infrastructure. Reducing waste will matter as much as adding workers. Coordination and clarity will separate projects that deliver from those that stall.

The digital economy will keep pushing forward. Yet its pace will be set by something older, slower and more human: the work of building, connecting and maintaining the systems it depends on.

……

How Bluebeam Fits In: FAQ

How does Bluebeam address labor constraints in mission-critical construction?

Bluebeam helps teams protect scarce skilled labor by reducing rework and coordination friction. When electricians, engineers and specialty trades work from a single, current set of drawings with clear markups, fewer hours are lost to errors, clarification cycles and redo work that no hiring surge can quickly replace.

Why does document clarity matter more when skilled labor is scarce?

As experienced workers become harder to replace, mistakes become more expensive. Bluebeam supports disciplined document control so crews aren’t forced to interpret conflicting drawings or outdated information. Clear intent, shared visibility and version certainty allow skilled workers to spend time executing — not untangling preventable problems.

How does Bluebeam support complex, high-risk projects like data centers and fabs?

Mission-critical projects depend on precision and correctness, not speed alone. Bluebeam enables teams to coordinate electrical, mechanical and systems-intensive scopes in one shared environment, helping ensure installation aligns with design intent before work happens in the field — where errors are slow, costly and risky to fix.

Can digital tools really improve productivity without replacing workers?

Yes — when they focus on coordination rather than automation. Bluebeam doesn’t attempt to replace skilled trades; it helps multiply their effectiveness by reducing rework, shortening clarification cycles and keeping everyone aligned. That recovered time effectively expands capacity without compressing training timelines.

Where does Bluebeam create the most value as projects grow more complex?

Bluebeam delivers the most value at points where complexity and coordination intersect: electrical rooms, commissioning workflows, revisions under schedule pressure and handoffs between design and field teams. These are the moments where clarity preserves momentum and where confusion quietly drains the physical economy.

How does Bluebeam fit into a broader strategy for the next decade of construction?

As physical execution becomes the limiting factor of growth, tools that reduce waste become strategic. Bluebeam fits as coordination infrastructure, helping organizations treat skilled labor as something to protect and optimize, not assume. In an economy paced by human work, clarity becomes a competitive advantage.

Protect your most valuable resource: skilled labor.

Start free trial

Autonomous Equipment Is Here. Construction’s Information Handoffs Aren’t Ready.

Project Management & Efficiency, Technology & Innovation
Steve Smith

AI-ready machines have arrived, but the workflows behind them are still stuck in the trailer.

At CES 2026, construction autonomy stopped being hypothetical.

Equipment manufacturers rolled out machines that don’t just follow commands, but assist operators in real time, flag risks and, in some cases, make decisions on their own.

Caterpillar, for instance, framed its latest AI-enabled equipment as a step toward jobsites where machines don’t just move dirt, but participate in the work.

For an industry that’s spent decades chasing productivity gains that never quite showed up, it was a moment worth paying attention to. Labor is tight. Costs keep climbing. Schedules are under constant strain.

Construction has been ready — borderline desperate — for something to finally bend the curve.

But here’s the part that didn’t make the highlight reels.

The machines are moving faster than the systems that support them.

Autonomous and AI-assisted equipment doesn’t work in a vacuum. It runs on drawings, revisions, approvals, boundaries, utility locations and real-time field conditions. That information doesn’t arrive cleanly packaged. It moves through handoffs — between design and preconstruction, office and field, one trade and the next.

Those handoffs have always been messy. Construction survived by leaning on people to smooth things out. Good operators catch what the plans miss. Superintendents resolve conflicts in real time. Crews adapt when the drawings don’t quite line up with reality.

Autonomy doesn’t have that instinct.

When machines act faster, more precisely and with zero tolerance for ambiguity, the cost of being slightly wrong goes way up. A missed revision or outdated plan doesn’t just slow things down; instead, it sends work in the wrong direction, faster than anyone can react.

CES made autonomy visible. What it also exposed is something the industry doesn’t love talking about: the real bottleneck isn’t the equipment, but the information handoffs holding the jobsite together with duct tape and experience.

Risk Doesn’t Disappear — It Just Moves Earlier.

Construction has always managed risk by keeping it close to the work.

Plans change. Conditions shift. But people in the field act as a constant check on reality. They stop when something feels off. They question dimensions that don’t make sense. They fix problems before they turn into incidents.

Autonomy changes where that judgment lives.

AI-assisted equipment is built to reduce fatigue and inconsistency. That’s the upside. The tradeoff is that many of the informal checkpoints construction relies on disappear. Decisions that used to happen in the cab or on the ground now happen upstream — in models, documents and systems — long before a machine ever starts moving.

Risk doesn’t go away. It moves.

It concentrates in the information itself: whether drawings are accurate, revisions are clear, approvals are real, and field conditions are reflected in time. When those inputs are wrong or outdated, autonomous systems don’t hesitate or “use their best judgment.”

They execute.

In a traditional workflow, a bad detail might trigger a pause, call or quick fix. In an AI-driven workflow, that same mistake can propagate instantly. Machines don’t interpret ambiguity. They amplify it.

Autonomy makes construction more precise and far less forgiving. The margin for “close enough” shrinks. The stuff that used to live safely inside a superintendent’s head becomes baked into the system.

The question, then, isn’t whether machines can operate autonomously. They can. The question is whether the information guiding them deserves that level of trust.

The Least Sexy Problem That Matters Most: Handoffs

Construction doesn’t have a data problem. It has a movement problem.

Every project generates a flood of information — drawings, RFIs, submittals, change orders, markups, emails and decisions made under pressure. On paper, it all adds up to a clear picture of what should be built.

In the real world, it’s scattered across tools and formats that don’t talk to each other.

Most of what matters lives in unstructured places: PDFs, inboxes, meeting notes and conversations that never quite make it back into the record. Humans navigate that chaos through experience. Machines can’t.

Information moves through construction by handoff. From design to preconstruction. From office to field. From one trade to the next. Every handoff introduces friction — delays, misreads, missed updates, assumptions that don’t get documented.

For years, the industry absorbed that friction by relying on people. Superintendents knew which plans to trust. Operators knew when something felt wrong. Teams improvised to keep projects moving.

Autonomy removes that safety net.

An AI-assisted machine, however, doesn’t know which drawing is “probably right.” It doesn’t know a late-night call resolved a conflict that never made it into a revision. It only knows what it’s given.

That’s why handoffs become the weak point. A utility update buried in a PDF. A boundary changed in one system but not another. An approval everyone assumes exists, but nobody recorded. All survivable in a human-driven workflow. All dangerous when machines treat them as truth.

From Trusting Operators to Trusting Systems

Construction has always trusted people more than processes.

Projects succeed because experienced professionals know how to work around imperfect information. Judgment isn’t a feature; it’s the foundation.

Autonomy forces that trust to shift.

As machines take on responsibility, confidence moves from individual expertise to the systems feeding them information. The question becomes simple and uncomfortable: can you trust the system enough to let it act?

In human-driven workflows, uncertainty gets resolved socially — a conversation, a walk, a gut check. In AI-driven workflows, uncertainty has to be resolved before work starts.

That’s where pragmatic technology earns its place. Not by replacing people, but by reducing ambiguity — by making it clearer what’s current, what’s approved and what’s changed, and by ensuring that decisions made in one place don’t get lost before they reach another.

This is the layer where construction technology adds value: not at the edge, but in the connective tissue of the jobsite. When information is visible, shared and traceable, both humans and machines make better decisions.

Progress, Without the Confusion

CES 2026 made the technology impossible to ignore. Autonomous and AI-assisted equipment is here.

What’s harder to face is what that technology reveals.

Autonomy doesn’t fail because construction lacks innovation. It stalls when workflows built on informal coordination are asked to support systems that don’t guess.

AI doesn’t forgive. It executes.

The real constraint on autonomy isn’t sensors or horsepower but whether construction can treat information like infrastructure — something solid, trusted and maintained — not paperwork that gets sorted out later.

Autonomy raises the cost of being slightly wrong. Gaps that used to hide inside experience now show up as real risk.

In that sense, autonomy isn’t just a technology shift.

It’s a stress test.

The machines are ready. The opportunity is real.

Still, autonomy will only scale when construction builds systems worthy of the certainty machines bring to the jobsite.

How Bluebeam Fits In

How does Bluebeam fit into AI-driven and autonomous construction workflows?

Bluebeam supports the information layer autonomous systems rely on. It helps keep drawings, revisions and approvals visible, current and traceable, so decisions made upstream remain reliable when work reaches the field or AI-assisted equipment.

Why do information handoffs become a bigger risk as construction becomes more autonomous?

Autonomous equipment executes exactly what it’s given. It doesn’t question unclear plans or resolve uncertainty on the fly. As a result, gaps in revisions, approvals or scope changes shift from minor delays to amplified risk when machines act on incomplete or outdated information.

Why does this matter even if a project isn’t using autonomous equipment yet?

The same information gaps that confuse AI already slow projects, cause rework and hide risk in human-driven workflows. Improving handoffs reduces friction today and prepares teams for a future where systems — not individuals — carry more responsibility for execution.

If machines don’t guess, your documents can’t either.

Try Bluebeam

From Cleanup to Rebuild: Why Wildfire Recovery Needs a Project Mindset

Construction & The Economy, Project Management & Efficiency
Frank Kalman

How fragmented handoffs slow post-fire rebuilding—and what a project mindset reveals about moving recovery forward.

One year after the 2025 wildfires reshaped large swaths of Los Angeles, the physical signs of recovery remain uneven.

In some neighborhoods, rebuilding is well underway. In others, properties have been cleared but still sit idle, or remain caught in layers of review, testing and approval.

The contrast is visible across communities and jurisdictions, and it raises a familiar question for anyone in the architecture, engineering and construction (AEC) industry: Why does recovery slow so dramatically once the immediate emergency ends?

Reporting over the past year points to a range of contributing factors. Coverage from The Wall Street Journal details how insurance challenges, permitting delays and uneven access to capital shape who’s able to rebuild—and when.

The New York Times, meanwhile, has examined how fire behavior, infrastructure failures and post-fire conditions complicate recovery long after flames are extinguished.

Built, in the wake of the fires, explored these issues from the construction side, including the realities of hazardous debris cleanup and the long tail of rebuilding in fire-prone urban areas.

Together, these accounts point to a broader structural issue: Wildfire recovery is often treated as a series of necessary but disconnected actions—cleanup, environmental clearance, permitting, insurance review, rebuilding—rather than as a single, continuous effort.

Without a framework that connects those phases, progress depends less on how much work is being done and more on how effectively one stage hands off to the next.

Why recovery breaks down

Wildfire recovery, as the WSJ and NYT reporting shows, spans multiple, distinct phases, each governed by its own rules, timelines and stakeholders. Hazard mitigation and debris removal give way to environmental testing and clearance, followed by permitting, insurance alignment and reconstruction. Each phase is complex, regulated and essential. Each is also typically managed by different entities using different tools, records and standards.

On their own, these phases often function as intended. Cleanup crews focus on safety and environmental compliance. Regulators verify site conditions before allowing rebuilding to proceed. Insurers require documentation before releasing funds. Contractors wait for approvals before mobilizing.

The breakdown usually doesn’t occur within the work itself, but between phases.

When recovery is managed as a series of discrete tasks rather than as a unified program, handoffs become friction points. Information is recreated instead of transferred. Decisions are revisited because earlier context has been lost. Projects stall not because efforts stopped, but because each transition introduces uncertainty that didn’t need to exist.

For anyone who’s worked on large capital programs, this pattern is familiar. Without shared sequencing, ownership and documentation standards, even well-funded projects struggle to maintain momentum.

Wildfire recovery is no different. The conditions are more volatile and the stakes higher, but the coordination challenge is the same one the industry confronts on complex, multi-stakeholder projects every day.

The issue isn’t a lack of expertise or commitment, but the absence of a program-level approach that treats recovery as a continuous process rather than a collection of isolated actions.

Cleanup is phase one, not a prequel

In urban wildfires, cleanup is often framed as a preliminary step—necessary but separate from the “real” work of rebuilding. In practice, cleanup is the first major construction phase of recovery, and the decisions made during it shape everything that follows.

As Built wrote in March 2025, post-fire cleanup in dense, developed areas involves far more than debris removal. Crews must identify and manage hazardous materials, address contaminated soils and ash, conduct environmental testing, and document site conditions to meet regulatory and insurance requirements.

When those records are incomplete, inconsistent or siloed, the downstream effects are immediate. Environmental clearance slows. Permits stall. Insurance claims linger. In many recovery efforts that struggle to gain traction, cleanup is treated as temporary or transactional—handled quickly, documented loosely and then left behind once debris is cleared.

The result is a reset when rebuilding begins. New teams are forced to re-establish site conditions, reverify earlier work or recreate documentation that no longer exists in a usable form. Time’s lost not because work wasn’t done, but because the continuity of information was broken.

Recovery efforts that move more steadily take a different approach. Cleanup is treated as the first milestone in a longer sequence. Documentation produced during debris removal and environmental testing is designed to carry forward into permitting, insurance review and reconstruction planning. Cleanup outputs become formal inputs to the phases that follow, reducing rework and uncertainty.

For AEC professionals, this dynamic isn’t new. Early site investigations, enabling works and environmental assessments routinely shape scope, schedule and risk on large projects. Wildfire recovery follows the same logic.

When cleanup is treated as phase one of a multi-year effort rather than a standalone task, it becomes a foundation instead of a bottleneck.

What a project mindset looks like in practice

Treating recovery as a project doesn’t require reinventing how construction works. It requires applying principles the industry already relies on—phasing, sequencing, ownership and documentation continuity—to a context where they’re often missing or underdefined.

A project mindset starts with clearly defined phases and intentional handoffs. Each stage of recovery has a purpose, a responsible owner and a set of outputs that enable the next stage to proceed.

Cleanup establishes verified site conditions. Environmental clearance confirms readiness to rebuild. Permitting and insurance alignment provide scope and funding certainty. Reconstruction advances with fewer unknowns because earlier decisions were made deliberately rather than reactively.

Across recovery efforts examined by government auditors and infrastructure agencies worldwide, coordination often matters more than raw funding in determining how quickly this sequence moves.

Programs with significant financial resources still stall when approvals, standards and documentation are fragmented across agencies and timelines. Others progress more smoothly by aligning expectations and sequencing early, even under tight constraints.

Documentation is the connective tissue that makes that alignment possible. In long-duration recovery efforts, records aren’t administrative byproducts. They’re the infrastructure that allows work to continue as teams, contractors and public officials change over time.

When documentation persists across phases—tied to the site rather than to a single stakeholder—projects spend less time revisiting past decisions and more time moving forward.

None of this is foreign to the AEC industry. Large capital programs, campus expansions, transportation corridors and utility upgrades rely on the same fundamentals. They succeed because early phases are designed to support later ones, and because information’s structured to survive complexity.

Wildfire recovery becomes more predictable when it’s managed with the same discipline.

What AECO teams already know and can apply

For AEC professionals, the mechanics of recovery-as-a-project aren’t new. The industry routinely manages multi-year efforts that involve layered approvals, regulatory oversight and changing teams.

Wildfire recovery introduces additional pressures, but the underlying coordination challenge remains the same. When cleanup aligns with downstream needs, when documentation is designed to persist and when stakeholders work from a shared sequence, recovery efforts move with greater predictability.

Built’s coverage in February 2025 on rebuilding in Los Angeles underscores that technical capability isn’t the limiting factor.

The opportunity lies in applying existing project discipline more deliberately, and earlier, in the recovery process.

Looking ahead

As wildfires grow larger and recovery efforts stretch over longer periods, the line between disaster response and capital construction continues to blur. Recovery increasingly resembles a multi-year construction program, whether it’s managed that way or not.

The lesson from Los Angeles isn’t that recovery is uniquely difficult—but that recovery works best when it’s treated as a continuous effort, guided by the same discipline that governs complex projects across the built environment.

For the AEC industry, that perspective offers a practical path forward: By applying familiar project principles to an unfamiliar context, recovery can move with greater clarity, fewer resets and a stronger foundation for rebuilding what comes next.

Bring project clarity to complex recovery efforts.

Start trial

The Efficiency Mandate: Why 2026 Is the Year Construction Runs Out of People

Construction & The Economy, Project Management & Efficiency
Frank Kalman

As megaprojects surge and the workforce thins, builders will have to create capacity through efficiency, not headcount.

Capital isn’t the problem. Projects aren’t the problem.

The problem is bodies.

Over the next decade, the U.S. will need roughly 650,000-725,000 construction and extraction workers every year just to fill open roles and replace people retiring or leaving the industry.

That’s not to grow capacity. That’s just to keep the lights on.

At the same time, demand is tilting toward the most labor-hungry, skill-intensive projects the industry has ever seen:

AI-driven data centers.
Grid and transmission buildouts.
Clean-energy and storage projects.
Semiconductor fabs and advanced manufacturing.
Plus, the unfinished business of housing and traditional infrastructure.

In 2026, those curves intersect: an aging workforce, a smaller pipeline of young workers and a wall of megaprojects all competing for the same electricians, linemen, pipefitters and supers.

There’s no plausible hiring plan that closes that gap.

That’s why 2026 isn’t just going to be “another busy year.” It’s the start of what you could call the efficiency mandate: If each worker isn’t effectively doing the work of 1.2-1.5 traditional workers — without burning out — projects will slip, get de-scoped or never break ground.

This is what that means in practice.

Why is the labor problem structural, not just “a hot cycle”?

This isn’t just another tight market that will ease after a rate cycle. Structural forces — demographics, replacement needs, immigration dependence and a thin pipeline of young workers — mean the industry is running out of experienced people faster than it can bring new ones in. That imbalance defines the next decade.

Is this different from every other “skilled labor shortage” headline you’ve seen for 30 years?

Yes. For a few reasons.

Replacement demand dwarfs new job growth

U.S. construction employment today sits around 8.3 million workers, including roughly 3.4 million in residential. The raw growth story doesn’t look explosive; the Bureau of Labor Statistics (BLS) projects only single-digit percentage job growth over the next decade.

But that’s not the real issue.

The real issue is replacement demand:

The BLS expects about 650,000 openings per year in construction and extraction roles through the mid-2030s, mostly to replace people retiring or leaving the occupation.
NAHB/HBI’s labor market analysis pegs it even higher: around 723,000 construction occupational openings per year right now, implying more than 2.1 million hires needed just in 2024-26.
ABC’s modeling says the industry needed about 500,000 additional workers in 2024, and a similar order of magnitude in 2025-26, on top of those replacement needs.

The math is simple and ugly: Replacing today’s workforce is a much bigger job than adding new positions.

“Openings are down” is not the good news it sounds like

If you look at job openings data, you’ll see a story that, at first glance, looks like relief. Open construction job postings have fallen from roughly 375,000 in mid-2024 to about 245,000 in mid-2025. That’s a big drop. It’s also misleading.

At the same time:

Overall construction employment remains near record highs.
The unemployment rate in construction is hovering near historic lows.
National contractor surveys still show 70-80% of firms struggling to fill hourly craft roles, especially in mechanical, electrical and civil trades.

In other words, we’re close to full employment for skilled craft labor. Openings are dropping not because there’s suddenly plenty of talent, but because many contractors are posting fewer jobs they know they can’t fill and stretching the people they have.

Demographics are destiny

The age profile is even more telling:

More than 1 in 5 construction workers are 55 or older, a share that has nearly doubled since the early 2000s.
Various workforce studies estimate 30-40% of today’s construction workforce will retire or age out over the next decade.
Only a small slice of workers — about 1 in 10 — are under 25, and the pipeline of new entrants is nowhere near large enough to backfill what’s leaving.

These aren’t interchangeable heads, either. The workers retiring are often your most experienced supers, foremen and specialist trades. When they walk off the job for the last time, you don’t just lose a pair of hands; you lose institutional memory and productivity that took decades to build.

Immigration is the quiet keystone

On top of that, construction is highly dependent on immigrant labor:

Immigrants make up roughly 25-30% of construction workers nationally.
In key trades — roofers, drywallers, laborers, carpenters — immigrants account for a third to more than half of the workforce in many markets.
In states like California and Texas and fast-growing metros, those shares are even higher.

Any tightening or uncertainty in immigration policy isn’t an abstract political debate for this industry but directly caps the maximum achievable headcount, especially in the trades that already feel tightest.

Put all that together and you get a simple conclusion: This isn’t just a hot cycle where “we’ll hire once rates fall.” The constraint is structural and baked into demographics and policy for the next decade.

How are four megacycles colliding over one shared talent pool?

Over the next several years, multiple policy- and technology-driven buildouts hit at once: data centers, grid upgrades, clean energy and advanced manufacturing. Each needs overlapping trades in overlapping regions. Instead of balanced cycles, contractors face stacked megacycles that all pull from the same shallow talent pool at the same time.

If the labor side of the equation weren’t bad enough, look at what’s arriving on the demand side.

1. Data centers and AI’s power appetite

You don’t need to be in the tech world to feel the ripple effects of AI. Data centers already used about 176 TWh of electricity in 2023, roughly 4.4% of total U.S. power demand. Updated federal and independent studies now project that number could reach 325-580 TWh by 2028, or 6.7-12% of total U.S. demand.

Private-sector forecasts like Goldman Sachs are even more aggressive, projecting data centers could hit about 8% of U.S. power demand by 2030 and require tens of gigawatts of new generation capacity.

All of that must be designed, permitted and built:

Hyperscale and colocation campuses
Substations, high-voltage lines and interconnections
Cooling infrastructure and high-density MEP systems
Supporting roads, water and utilities

These are complex, coordination-heavy projects with intense demands on mechanical, electrical and civil trades.

2. Grid modernization and transmission

At the same time, the grid those data centers rely on is being rebuilt in real time. The U.S. Department of Energy’s transmission needs analysis concludes that to meet reliability and clean-energy goals, the country must effectively double regional transmission capacity and increase interregional transfer capacity fivefold by 2035.

That translates into:

Tens of thousands of new line miles over the next decade
Hundreds of billions of dollars in capital expenditures
Thousands of substations, towers, foundations and associated civil work

Federal programs are already moving money: multibillion-dollar grid resilience grants, transmission facilitation loans and direct federal support for marquee lines. Those aren’t hypothetical white papers; they’re construction pipelines.

3. Clean energy and storage

Then layer in the clean energy buildout: utility-scale solar, onshore and offshore wind, storage, hydrogen hubs and more.

Analysts tracking the Inflation Reduction Act estimate:

Hundreds of new clean energy projects announced in its first couple of years.
Hundreds of thousands of construction job years generated during buildout alone.

Again, these need line workers, civil crews, steelworkers, electricians and commissioning specialists — the same people AI data centers and the grid are trying to hire.

4. Semiconductors and advanced manufacturing

Finally, there’s the semiconductor wave. CHIPS-backed fabs in Arizona, New York, Texas and Ohio are already confronting labor shortages severe enough to delay timelines. We’ve seen:

High-profile fabs pushing production dates out by several years.
Public commentary from project sponsors citing a lack of skilled construction workers, especially for high-purity process piping, power distribution and controls.

Fab projects, like data centers, demand the best of the best: highly experienced mechanical, electrical and process trades, plus tight QA/QC and commissioning.

Now put all four together: data centers, grid, clean energy, fabs — plus ongoing housing and infrastructure backlogs. They all want the same people, in the same timeframe, often in the same regions.

That’s the 2026-30 collision.

Why doesn’t “just pay more” solve the labor crunch?

Raising wages helps but can’t overcome time, geography and policy. Apprenticeships still take years, workers can’t instantly relocate to every hot market and immigration rules sit outside contractors’ control. Compensation becomes table stakes, not a silver bullet, in a market where the total pool of skilled labor is capped.

In a textbook market, high demand and short supply should mean one thing: Pay more. Problem solved. Reality isn’t that simple.

Yes, wages have moved:

Construction wages have grown faster than the overall economy since the pandemic.
Some trades have seen double-digit percentage increases in just a few years.
Contractors are offering hiring bonuses, retention perks and more training.

And yet the shortages persist, for reasons that aren’t fixable with a line item in a budget:

Training takes time: You don’t turn a new hire into a journeyman electrician in 18 months, no matter what you pay.
Work is geographically sticky: Projects don’t neatly line up where the workers are. Convincing specialized trades to move across the country at scale is slow and expensive.
Immigration policy is out of contractors’ control: The industry can’t unilaterally expand the pool of eligible workers.

There are also early signs of cooling in a few regions — more applicants here, fewer job openings there — but that’s cyclical noise on top of a structural trend. If your plan is simply “we’ll pay up when things get tight,” you’re already behind.

How are rework and bad data draining hidden capacity?

Even before the crunch peaks, many projects effectively operate with smaller crews than they think. Time lost to rework, poor information flow and mismatched documents quietly burns a double-digit share of available hours. In a world where new people are scarce, recovering that wasted capacity becomes existential.

Even with today’s workforce, the industry is leaving a massive amount of capacity on the table.

Productivity has flatlined

Global construction productivity has grown at about 1% per year over the past two decades — roughly one-third the rate of manufacturing and well below the broader economy. In many advanced economies, including the U.S., construction labor productivity has stagnated or declined since 2000.

That would be annoying in a balanced market. In a market with structural labor tightness, it’s lethal.

Rework is a phantom workforce

Look at rework and bad data:

Rework alone is estimated to account for around 5% of U.S. construction spending, or tens of billions of dollars a year.
Studies tying cost overruns to data problems estimate that bad data and communication failures drive well into the triple-digit billions in avoidable cost annually, globally.
Field and office staff regularly report losing a full workday every week to low-value tasks like searching for information, reconciling versions and clarifying miscommunications.

Translate that into people: If an average project team is losing 10-20% of its time to rework, hunting for documents or fixing coordination errors, that’s the equivalent of phantom crews you’re paying for but not actually getting. In a world where you can’t conjure up an extra 10% headcount, the only rational move is to stop wasting the 10% you already have.

What does the efficiency mandate look like in practice?

The efficiency mandate is less about heroic overtime and more about redesigning how work flows. Firms that standardize, digitize and industrialize — through BIM, coordination, prefab and lean planning — unlock more value from every hour on site. Those choices determine who can still deliver complex work when the talent pool tightens.

“Be more efficient” is meaningless. The question is: How? The data and the leading case studies point to a clear answer: standardized, digital, industrialized workflows that unlock more output per worker without asking people to simply sprint harder.

BIM and model-based coordination

When BIM is used consistently — not as a one-off experiment — contractors report:

Dramatic reductions in clashes and RFIs
Fewer constructability problems in the field
Lower defect rates at handover
More predictable schedules

That is pure capacity. Less time fixing what shouldn’t have been built in the first place means more time building what matters.

Prefabrication and modular

Industrialized construction isn’t theoretical anymore. On the right types of projects, the numbers are well established:

20-50% faster delivery for suitable projects.
Up to 20% cost reductions in some modular case studies.
Hospital projects that moved more than 150,000 work hours off site, cut more than two months from the schedule and still reduced overall cost once you count rework and safety benefits.
Data center and health care jobs where 70% of complex piping or MEP assemblies were prefabricated, shrinking onsite headcount and congestion.

Again: That’s what making each worker “count for more” looks like in the real world.

Lean/IPD and digital planning

Lean construction and integrated project delivery aren’t just management buzzwords. In projects where they’re taken seriously, documented results include:

Schedules 30% faster than traditional delivery
Double-digit reductions in total labor hours
Lower peak onsite crew counts
Higher safety performance

When pull planning and Last Planner systems move from sticky notes on a trailer wall to digital environments tied to actual model and schedule data, those gains become repeatable instead of a one-off success story.

Put it all together and you get the heart of the efficiency mandate: Firms that combine BIM, prefab, lean/IPD and structured data can realistically get 1.2-1.5 times the effective output per worker on complex projects. In a structurally tight labor market, that isn’t a nice differentiator. It’s survival.

How should construction really think about automation and AI?

Robotics and AI are best understood as amplifiers sitting on top of strong digital foundations, not magical replacements for crews. Where data is clean and scopes are repetitive, they can meaningfully shift labor curves. Where workflows are messy, they mostly expose underlying problems instead of solving them.

Then there’s the current obsession: robotics and AI. They matter. But not in the way the marketing suggests.

Where robotics is paying off

Real projects — not glossy concept videos — show robotics moving the needle in specific scopes:

Rebar-tying robots cutting the schedule for a bridge deck scope roughly in half.
Layout robots that not only increase layout speed but drive better upstream coordination, leading to measurable profit improvements.
Autonomous or semi-autonomous scanning robots walking jobsites at night, feeding reality capture data into BIM and progress-tracking tools.

The pattern: Robots do well on repetitive, physically demanding tasks where there’s a strong digital model and clear tolerances.

Where the hype runs into the wall

You don’t hear as much about the pilots that stall out. But they’re common:

Robots that outperform humans on a per-day drilling metric but require so much BIM prep and coordination that net productivity is barely better — or worse.
Firms excited by robotics on paper, but unable to justify scaling beyond pilots because ROI is fuzzy and site teams weren’t brought in early.
Vendors that disappear or fail to integrate with existing systems, leaving contractors burned.

Survey data is telling: Optimism about construction robotics is high, but actual adoption has dipped in some studies, as contractors pull back to a smaller number of well-chosen use cases instead of chasing every new demo.

AI as a force multiplier for knowledge work

AI is already proving its worth in less glamorous but more fundamental ways:

Progress tracking: comparing 3D scans to BIM to automatically flag deviations, delays and billing issues — something that would otherwise soak up scarce VDC staff.
Predictive scheduling: using historical performance, weather and resource data to surface likely schedule risks weeks before a human would see them.
Estimating and document search: reducing the time preconstruction and field teams spend digging through drawings, RFIs and emails to figure out what’s current and what’s not.
Safety and quality monitoring: computer vision systems that spot PPE noncompliance or installation defects at scale.

The common denominator is obvious: None of this works without clean, standardized, current project data. AI doesn’t rescue bad workflows; it amplifies whatever you feed it.

How are leading builders already closing the efficiency gap?

Large builders are already operating on a blunt assumption: they can’t simply hire their way through the next decade.

Instead, they’re quietly redesigning how work gets delivered. That means shifting hours offsite, tightening coordination through BIM, standardizing data environments and focusing automation on a small number of high leverage use cases that move schedules and margins. Their project results offer a preview of what’s becoming the new baseline.

If all this still sounds theoretical, look at what’s happening on the industry’s most complex work:

On large, multi-building data center campuses and similarly fast-moving programs, leading builders are increasingly leaning on scan-versus-BIM comparison and AI-assisted deviation detection to maintain quality and schedule when internal VDC capacity can’t keep pace with field progress.
Automated reality capture handles monotonous documentation, allowing superintendents and project engineers to focus on coordination and problem-solving instead of clerical work. In preconstruction, AI-assisted estimating and standardized data environments are reducing friction and compressing timelines before crews ever mobilize.

The motivation isn’t trend-chasing but structural. These firms can’t simply triple their VDC staff or double their superintendent bench.

The same logic shows up in how industrialized construction is being applied across data centers, health care and hospitality.

Multi-trade prefabrication is shaving weeks off schedules. Hundreds of thousands of labor hours are being shifted offsite, reducing peak headcount, congestion and safety exposure. Volumetric modular systems are delivering finished components faster and with far less onsite disruption.

Again, the through-line is clear: when you can’t find more labor, you change where and how the work happens.

On major infrastructure and complex building projects, builders are also combining lean delivery models, BIM and digital twins to tighten feedback loops between design and construction. By continuously comparing as-built conditions to design intent using drones, sensors and model-based workflows, teams are reducing rework, improving material efficiency and compressing project durations without adding headcount.

Why isn’t this pure doom — and what’s still different this time?

Short-term signals can be confusing — local slowdowns, softer openings data, mixed technology results — but they sit on top of deeper trends that don’t reverse quickly. Leaders must read both layers at once: acknowledge regional cooling where it exists without mistaking it for a return to the old, labor-abundant normal.

To be fair, there are countersignals:

Some regions report that hiring is getting slightly easier.
Job openings are off their peaks.
A few robotics pilots have underwhelmed or been shelved.
Forecasts for AI and data center power demand vary widely, with some scenarios showing slower growth if efficiency improves.

All true.

But those nuances don’t change the underlying structural picture:

Demographics point one way.
Policy-driven capital expenditures (CHIPS, IRA, infrastructure) are already committed.
Data center and grid projects in the 2026-30 window are locked into the pipeline.

You might get temporary pockets of relief. You won’t get a return to the world where you could always solve problems by “adding a few more workers.”

What hard choices does 2026 force construction leaders to make?

As projects and people diverge, 2026 becomes a forcing function. Owners, general contractors and trades all must decide whether they will privilege partners and practices that create capacity — through digital coordination, prefab and smarter planning — or hope the market loosens. Those choices shape who can even bid certain work.

In 2026, the stories you tell yourself about staffing will collide with reality. Practically, that means a few hard choices.

If you’re an owner or developer

You can’t just pick the lowest bidder and assume they’ll “figure it out.” You need to ask:

How standardized and digital are their workflows?
How do they handle coordination, rework and data?
Can they realistically staff this project in this market, or are they gambling?

Soft factors like BIM maturity and prefab capability are now directly tied to your schedule and risk profile.

If you’re a GC or EPC

You must decide whether you’re going to be a capacity creator or a capacity victim. That means:

Treating BIM, structured data and digital collaboration as core operations, not side projects.
Identifying where prefab and modular can be standard practice, not an exception.
Choosing a small number of automation and AI use cases tied to real bottlenecks — progress tracking, scheduling, layout, documentation — and doing the change management to scale them.
Investing in training so your people can operate confidently in this environment.

The firms that do this will bid — and deliver — projects their competitors literally can’t staff.

If you’re a trade contractor

Your choice is stark:

Become the partner who can integrate with model-based workflows, prefab assemblies and digital QA/QC, or
Become the shop that only makes sense on smaller, less time-sensitive work.

There’s a lot of business in both lanes. But you can’t pretend they’re the same.

Where does Bluebeam fit in the efficiency mandate?

Bluebeam doesn’t manufacture robots or design fabs; it quietly shapes how information moves. When drawings, markups and reviews live in a single, structured environment, teams waste less time chasing clarity and fixing preventable errors. That document layer is often the fastest, least disruptive way to unlock real capacity.

None of this is about a single tool solving a structural problem. The firms winning the efficiency game are doing it with systems: people, process, data and technology working together.

But if you strip away the buzzwords, a few foundational needs show up repeatedly:

Teams need clean, current documents everyone trusts.
They need standardized markups, layer conventions and workflows so data can be reused — not recreated — across scopes and phases.
They need fast, transparent review cycles that don’t leave junior staff guessing which version is “real.”
They need digital guardrails that help a less experienced engineer, coordinator or foreman perform closer to how a veteran would.

That’s where a platform like Bluebeam sits: not as the robot or the AI “brain,” but as the collaboration and data-quality layer that makes those bigger moves possible.

If rework and bad data are burning the equivalent of whole crews off your projects, then tightening up how drawings are shared, reviewed, marked up and standardized is one of the fastest ways to create capacity without hiring a single extra person.

What’s the bottom line for construction in 2026 and beyond?

The industry isn’t running out of projects or capital; it’s running out of time and people. Firms that treat efficiency as a strategic mandate — re-engineering how they coordinate, document and deliver work — will still have room to grow. Everyone else will find that the real constraint is no longer negotiable.

In 2026, the industry’s binding constraint isn’t going to be money. It isn’t going to be projects.

It’s going to be people.

You won’t hire your way through a decade where:

A third or more of your workforce retires.
Immigration inflows are uncertain.
Data centers, the grid, clean energy and fabs are all demanding the same scarce trades you need.

The only lever left with enough throw is efficiency — real, structural efficiency, not just working longer hours. The companies that treat 2026-30 as an efficiency mandate — and industrialize how they plan, coordinate and build — will get to say yes to the best projects and deliver them.

Everyone else will be stuck bidding work they can’t reliably staff.

Create capacity without adding headcount.

Start free trial

Why Construction’s Supply Chains Still Run Like It’s 1999 — and How AI Might Finally Fix It

Project Management & Efficiency, Technology & Innovation
Jessica Brown

Manual processes are still draining time and money from projects, and AI may finally give teams the edge they need.

Across construction, one complaint echoes from project to project: the workload is climbing while the workforce is shrinking.

The labor shortage already stretches teams thin — and supply chain chaos piles on more pressure. A May 2025 industry poll found that 71% of respondents cited material availability and supply chain issues as the leading cause of construction project delays. No wonder owners and project managers scramble daily to keep things moving.

Something has to give. And for some, that means turning to agentic AI — not to replace people, but to relieve pressure on human teams and squeeze more value out of the resources they have.

That’s where Ojonimi Bako and Nick Selz come in.

From Walmart and Google to Construction

Bako, a mechanical engineer, spent years refining Walmart’s e-commerce strategy and operations before starting his own construction business. That’s when he ran headfirst into the industry’s messy supply chain reality.

His idea: merge his expertise in retail logistics with Selz’s background in systems design at Google. Together, they built Kaya AI, a platform aimed at fixing construction’s most painful bottleneck.

“Between our tech and construction backgrounds, we saw a massive problem in the construction supply chain space,” Selz said. “So many processes are manual, time-consuming and prone to human error. Meaningful insights that could have a measurable impact on projects often go unnoticed.”

AI That Thinks Like a Project Team Member

Kaya AI is designed to facilitate better collaboration and communication between stakeholders — general contractors, project managers and executives alike.

“The thing I love and find so interesting about the supply chain is it’s an incredibly collaborative workstream,” Selz said. “The different stakeholders on projects are actually on the same team.”

The stakes are real: if a generator lands on site four weeks early, nobody benefits. “Better collaboration and coordination are in everyone’s best interest.”

Here’s how it works:

Kaya AI digests construction data: drawings, specs and equipment lists.
It cross-checks for missing items and connects equipment lists to scheduling and submittals.
The result: a holistic view of what needs to be onsite, when and with which approvals.

And instead of asking crews to learn yet another system, Kaya uses autonomous AI agents that communicate by text, phone or email. To suppliers and contractors, it looks like the usual lead-time confirmation requests, but behind the scenes, AI is handling the heavy lifting.

Meet Jarvis, the AI Assistant

One example is Jarvis, Kaya AI’s project management agent.

“Jarvis helps customers identify schedule risk sooner,” Selz said. Project managers often miss the dependencies between fabrication, shipping and the submittal approval process. Jarvis surfaces those risks in real time.

“For example, when the lead time changes, Jarvis gathers that data and alerts you via text with a new submittal approval date.”

While the platform includes a web-based app and dashboards, Selz says most stakeholders still interact through everyday channels.

“It works with the communication channels they’re already using, meaning they don’t have to learn a new system or download another app.”

Kaya also integrates directly with scheduling and submittal software, cutting down on re-entry and manual work. Users can even generate calls, emails and texts to release project data or validate lead times. “That is saving folks a tremendous amount of manual work.”

From Pilot to Billions in Active Projects

Founded in 2023, Kaya AI was accepted into the Suffolk BOOST Accelerator and quickly found traction.

“We’re now the most quickly adopted software in Suffolk’s portfolio,” Selz said. Client projects span everything from single-family homes to data centers. “Everyone has issues with the supply chain, and we’re grateful we’re able to help.”

Following its official 2024 launch, Kaya now manages supply chain coordination across billions of dollars in active construction projects.

Selz sees it as more than a business opportunity. “Ultimately, I think integrating tools like AI can enable teams to do more with the same number of workers. That’s going to be imperative to the survival of the industry.”

The Human Factor

Still, Selz is quick to note: AI won’t replace people in construction.

“There’s too much complexity and risk in construction to turn any project over to AI. This is about how to capitalize on the strengths of AI, such as its ability to analyze data, recognize patterns and expand your team’s capabilities. That gives humans time to focus on the higher-order strategic work and relationships that this industry is built on.”

The Hard Truth

Supply chain headaches are crushing projects. AI alone won’t solve them. But platforms like Kaya AI point to a smarter path forward — one where machines crunch the numbers and humans focus on building.

Because if construction keeps running supply chains like it’s 1999, the industry’s survival is what’s really at risk.

See how Bluebeam can streamline your projects.

Start free trial