EcoTerra: Novato Yard Base Rock 1/2 price @ $5/ton
View Deal
.png)
Explore
.svg)
Who We Serve
.svg)
Media Hub
.png)
Other Links
.svg)
Most weeks now, someone wants me to explain artificial intelligence to a room. An editor emails about a column. A producer wants twenty minutes on what agents are and whether he ought to be worried about them. I say yes to almost all of it, and lately I've been paying attention to why the asks keep piling up.
I've been around heavy construction long enough to know the software story is more tangled than the version people usually tell. Estimators, planners, and project managers have leaned on serious tools for well over a decade, and the tools themselves held up fine. Plenty of them were still a chore to learn. The price tag shut smaller shops out before they could start, and each program sat in its own bubble, so the data came out fractured and stranded in pockets that never spoke to each other. That mess has dragged on the industry's progress for years.
AI is landing differently. It's the first wave of technology I've watched stir up something like weather in this business. People show up to the same meeting nervous and excited at once, confusion underneath both. A manager will tell me he's sure this changes everything, then admit in the next breath that he has no idea what to do about it.
For decades, this was one of the least digitized industries anywhere. McKinsey has pointed out that construction puts less than 1 percent of revenue into research and development, while the auto and aerospace sectors run between 3.5 and 4.5 percent. Spending on information technology sat under 1 percent of revenue too. Big projects came in around 20 percent behind schedule and as much as 80 percent over budget, and in some markets productivity slipped in the years after the 1990s.
Now look at where the money is heading. One market estimate puts AI in construction at about $ 4.86 billion in 2025, climbing toward $ 35.53 billion by 2034, a compound annual rate near 25 percent, and I believe when the numbers roll in, it will be much higher. Investors feel it too. A 2025 survey of construction tech investors found 56 percent planning to move more money into AI than they had the year before.
What's happening on the ground is even bigger than the spending suggests. A 2025 RICS survey of more than 2,200 professionals found close to half with no AI in their organizations yet, about a third running early pilots, and only a sliver using it across several parts of the business. So the interest has run out ahead of what most companies have in place, which is about where I sit most days, in conversations with people who know they need to move and can't yet see which way is forward.
Why now, and why does it hit this hard? A few things line up at once. The workforce is aging out faster than it's being replaced, so anything that lets a smaller team carry more weight gets attention it wouldn't have drawn fifteen years ago. The tools finally fit the complexity this industry runs on. Our world is plans, specs, change orders, text threads, dusty photos from a phone on a haul road. The older software wanted clean rows and columns we never kept. Agents can read it all and make something of it.
The math pushes too. Margins here run thin enough that a few recovered points of productivity change how the year ends. McKinsey figures digital work can raise productivity 14 to 15 percent and bring costs down 4 to 6 percent. On a heavy civil job, numbers like that decide which bids you can afford to chase. And the cost of trying has dropped through the floor. A superintendent can open a chatbot on his phone at lunch and get a workable answer to a spec question, which was out of reach a couple of years ago. When something useful gets that easy to reach, curiosity turns into habit fast.
This is part of why I've been writing and speaking more. A few pieces of mine have run recently in The Conveyor , CalCIMA's publication, and this month a twelve-part series I've been building with our partners at SEMCO Publishing starts in Rock Products. Rock Products turns 130 this year and just spun up a new network to keep producers current as the ground shifts under all of us. Being part of that feels like a good use of whatever platform I've got.
The field needs voices that have worked in it, people who can stand in front of a room of aggregate producers and say in plain terms what an agent does to their operation, so it lands with someone who's run scales for thirty years. Carrying a new tool the last mile is where I want to spend my time, because that last mile is where a lot of good ideas quietly stall.
None of us gets through this alone, and that's what I keep circling back to. The materials producers, the contractors, the solution partners writing the software, the associations fielding questions from their members, we're all looking at the same shift from inside our own walls, mostly comparing notes with the same people we already know. What's in front of us is bigger than any one company can hold, and it's going to change all of our jobs in ways I can't fully picture. The useful thing we can do is open those walls a little. A contractor who finds something that works can spare a producer three counties over a year of slowly rediscovering it, if word gets around. An honest account of a pilot that went nowhere is worth as much to the next person as any success story. I'd rather work this out in the open, with the people who'll live with the results, while it's still early enough that we get a say in how it goes.
One estimate puts the global AI-in-construction market at about $4.86 billion in 2025, growing to roughly $35.53 billion by 2034, a compound annual growth rate near 25 percent.
A 2025 RICS survey of more than 2,200 professionals found close to half with no AI in place yet, about a third running early pilots, and only a small share using it across multiple parts of the business.
An aging workforce, tools that finally handle messy real-world data like specs and site photos instead of requiring clean spreadsheets, and much cheaper access (a chatbot on a phone) are converging at once.
McKinsey estimates digital work can raise productivity 14 to 15 percent and cut costs 4 to 6 percent, enough to change which bids a contractor can afford to chase on thin-margin heavy civil work.
1. Construction R&D spending under 1% of revenue vs. 3.5–4.5% for auto/aerospace; IT spending under 1%; large projects ~20% behind schedule and up to 80% over budget; productivity decline in some markets since the 1990s — McKinsey & Company — "Imagining construction's digital future" (2016)
2. Digital transformation can raise construction productivity 14–15% and cut costs 4–6% — McKinsey & Company — "Decoding digital transformation in construction"
3. AI-in-construction market valued at ~$4.86B in 2025, projected to reach $35.53B by 2034 at a ~24.8% CAGR — Fortune Business Insights — "AI in Construction Market Size, Share & Industry Report"
4. 2025 survey of construction tech investors: 56% planning to increase AI funding vs. the prior year; 2025 survey of 2,200+ professionals: ~45% with no AI in place, ~34% in early pilots — RICS — "Artificial intelligence in construction" report (2025), citing Zacua Ventures' Contech Investor Survey 2025
Note: the 56% investor-intent figure and the ~2,200-professional adoption survey both come from the same RICS 2025 report — the investor figure originates from Zacua Ventures' Contech Investor Survey 2025, which RICS cites directly.
