Construction Technology

Without more technology adoption, the construction industry’s delivery of data centers could fall short even as global demand for artificial intelligence intensifies, according to a July 2025 white paper.

Unless building methods evolve to match the scale and speed required for these complex jobs, contractors may not be able to keep up with client needs, says the report from Revizto, a Swiss construction software provider.

The construction sector is struggling to keep pace, putting both infrastructure delivery and broader economic ambitions at risk, the paper found.

Globally, capital expenditure on data centers is expected to surpass $1 trillion annually by 2029, according to Redwood City, California-based market research firm Dell’Oro Group. Tech giants such as Microsoft, Google, Amazon and Meta are expected to spend over $300 billion in 2025 alone on data centers and compute power. 

Technology adoption remains a critical barrier to the delivery of these projects, according to Revizto research, which found that many construction teams still rely on static, nonintegrated tools even for complex and fast-moving projects. The firm said 27% of global AEC professionals use email, spreadsheets and PDFs as their primary digital tools. 

However, data center construction demands complex coordination, massive datasets, strict regulatory compliance, high-stakes communication and tight timeframes, with some projects now moving from concept to full design in as little as 10 weeks, according to the white paper.

"The construction industry can't keep pace with demand using static tools like Excel and PDFs,” said Revizto CEO Arman Gukasyan in a news release. “If we're serious about meeting the demands of modern data center builds, we need a fundamental shift in how we deliver the physical infrastructure that supports them."

Beyond technology adoption, U.S. contractors are finding creative ways to meet the demand for data center development. For instance, contractors are consulting with clients on projects earlier than ever in the process to help evaluate site conditions, entitlement risks, energy procurement and utility interconnection timelines.

“Preconstruction services and capabilities have become a true differentiator,” John Arcello, advanced technology core market co-leader at Redwood City, California-based general contractor DPR, previously told Construction Dive. “Early engagement is critical to stay on track.”

Contractors also are turning to prefabrication more aggressively. On remote sites, Arcello said teams are increasingly building components offsite in labor-rich regions and transporting them to projects. That helps ease skilled trade constraints and minimize rework, he said.

Ike Keene, vice president of commercial at Caddell Construction, a Montgomery, Alabama-based general contractor, said that the need for accurate cost forecasting has grown significantly. That’s especially true as President Donald Trump’s tariffs add a fresh layer of volatility to material prices.

“The newly adopted tariffs have become a new variable in cost control,” said Keene previously. “We are adapting day by day for accurate current and future pricing structures to aid our clients in forecasting for the future.”

Continuing on its push to add artificial intelligence throughout its business, London-based builder Balfour Beatty has invested 7.2 million pounds ($9.6 million) into tech giant Microsoft’s AI offering, Microsoft 365 Copilot, according to a July 2025 news release.

Microsoft’s AI-powered assistant embedded within Microsoft 365 applications will operate securely within Balfour Beatty’s compliant and confidential IT environment, which is distinct from publicly available tools, according to the news release.

Alongside the investment, the company plans to develop AI agents to improve its quality, health and safety and assurance processes, with the first trial at an infrastructure project in Scotland, according to the news release. Balfour Beatty announced the AI development push in March 2024, during its full-year earnings call.

Balfour Beatty is running a pilot of the tech at the 185 million pound A9 project in Scotland, a road improvement job where the firm is upgrading a single-lane highway to two lanes, constructing four new bridges and delivering new side roads together with various junction upgrades. 

The AI agent, which focuses on inspection and test plans, identifies common issues that include incorrect or outdated templates reaching technical experts. As a result, the agent enhances the consistency, efficiency and quality of ITPs, according to Balfour Beatty.

By automating the manual review process, which traditionally takes several hours and is repeated thousands of times across the company, the tool accelerates delivery and allows engineers to focus on high-value tasks such as design assurance and technical problem-solving, which ultimately boosts productivity, per the news release.

“This investment isn’t just about embracing technology,” said Jon Ozanne, Balfour Beatty’s chief information officer, in the news release. “It’s about ensuring our business remains at the forefront of competitiveness and cyber security.”

Workers can have an idea of when they’ll need to retire. But autonomous machinery might push that deadline out, a Granite Construction tech lead says.

Aaron McClellan, construction technology manager for the Watsonville, California-based infrastructure builder, is familiar with the use of the technology. Granite has introduced supervised autonomous equipment, where it has retrofitted existing machines, on the Tucson International Airport in Tucson, Arizona, and on a New Mexico jobsite.

The retrofit kits on those projects came from Bluelight Machines, based in Pittsburgh, which features remote control from any mobile or desktop device and regular software and feature updates delivered over-the-air, per the company’s website.

Here, McClellan talks with Construction Dive about implementing autonomous equipment, the benefits the company has reaped, and tips on how to talk to workers about the new technology.

Editor’s Note: This interview has been edited for brevity and clarity.

CONSTRUCTION DIVE: How did you come across these autonomous kits, and what problems did they solve?

AARON MCCLELLAN: We got introduced to Bluelight Machines through a couple different channels. We ended up going up to Northern California, where we watched a demo, had a conversation with the company there, and it fit well into what we needed to improve operations in the Arizona region.

From there, we started coordinating meetings, had some onsite visits, and then decided to move forward with it. It's been operational for about four or five months.

Dirt rollers for us are a monotonous task that takes a skilled operator away from a more critical path or scheduled activity. For us, the use of a supervised autonomous dirt roller allowed us to use those operators in a more effective manner.

It upskills our labor. It allows them to do more with the equipment that we have, without having to sit inside of a roller that moves up and down a project all day long.

We have two of them right now. We've got one at a project in New Mexico, and we have one at a project in Arizona.

What have you noticed since Granite started using autonomous tech? 

I think the use of supervised autonomous, or remote-controlled equipment allows for safer operation. We would definitely get efficiency out of it. 

There's time savings. If you've got a skilled operator, and they need to be moving material with some kind of rock hauler, they're more critical to the operations to move the material in and around the jobsite than we are flattening out a flat piece of dirt.

It makes a lot more sense to give them the controls. They put the dirt roller in motion, and then they operate the truck. There’s a lot of efficiency we can get out of that.

If we talk more generally, it allows us to scale out our workforce. We know that there's a huge labor shortage. The way that we're looking at it is, if we can scale out to where we've got multiple machines to one operator, then we're allowing our operators to be more effective.

It also takes them out of the seat. In certain cases, we've got very high risk work that, if an employee sits in the operator seat inside the machine and something were to happen, it could be fatal. If we can remove that person from that environment, it allows it to be a safer operation.

How was the learning curve for the technology? 

We expected a little bit of a ramp up period. Like, all right, we'll feel it out for a week and get operators used to it. Maybe another week of getting in and around, but once they got the feel of using the tools that they had, it was a day or two, and then they were off to the races.

I think a lot of it comes down to the partnership that you have with the companies that provide those types of solutions. 

A lot of the tech has moved rapidly over the last couple of years, so it's a lot easier for operators to grasp how to control it. They build in a lot of similar tools that the operators would see in the cab itself. 

A lot of the dashboard that exists in the cabs of the machines is also brought to either the tablets or the seats that they sit in. That way, it replicates what they're used to, and that helps bridge the gap between how fast it allows them to transition to a remote or a supervised role.

How do your workers feel about this technology?

I think there's a lot of work. There's a lot of openness. I think that they see that it's not just coming in to replace anybody.

As long as the messaging is about the safety and the efficiency, and you're extending the tenure of some of these operators that may be seeing retirement on the horizon. It allows them to expand that tenure to maybe even past the dates that they had kind of anticipated for themselves based on physical limitations.

There's always going to be hesitancy and change. We embrace that, and we know that we have to understand where the skepticism comes from and meet them halfway, and understand what that looks like, and how do we get around that, or work through it with you.

What does meeting your workers halfway look like?

A lot of it is understanding what they want. 

If we move to an autonomous machine, for example, there are certain aspects that they like to feel right. If we can have relationships with the vendors that are providing these retrofit kits that do these awesome things through the use of technology, we also have to consider, what is the operator so used to that it's going to be hard for them to transition without it?

One of the things that we had conversations with a provider about a while back was the natural feel of the machine and how the controls work. That is something that is inherently developed over years and years of operating that machine.

Taking that feedback, and then providing that to those solution providers, helps bridge the gap on the field versus the tech being able to be used.

Nearly two-thirds of survey respondents described their construction companies’ artificial intelligence adoption maturity level as “average,” while 22% referred to it as “above average,” according to BuiltWorlds’ 2025 Annual AI Benchmarking Report.

When asked if AI was a core strategic focus companywide, 35% of respondents reported that they were exploring its potential but had yet to fully embrace it, according to a news release highlighting the report. Of the respondents, 4% said, “No, AI is currently not a core strategic focus for us.”

BuiltWorlds, a research firm headquartered in Chicago, based its data on its 2025 AI Benchmarking Survey, which featured responses from 53 individuals representing 30 organizations, BuiltWorlds told Construction Dive via email. The majority of participating organizations — 82% — were general contracting firms, and 40% of respondents generated more than $5 billion in revenue.

When asked why they were adopting AI, respondents were nearly unanimous — almost 100% of respondents said that “improving operational efficiency and productivity” was a primary driver of their work, according to the news release.

With that in mind, the report flags that companies are using AI to strengthen core business functions rather than pursue cutting-edge innovation, per the release.

However, some of these respondents — which also featured subcontractors, specialty firms and engineers — had reservations about the tech. For example, 62% cited “concerns about data privacy and security,” 58% pegged a “lack of internal expertise or skilled personnel” and a further 56% called out “limited availability or quality of data.”

Despite those hurdles, builders’ enthusiasm for AI is high, said Audrey Lynch, BuiltWorlds’ director of research and the report’s author, in the release. Meanwhile, 2% of respondents said they were “skeptical” of the tech.

“It’s important to note that these challenges likely stem from the technology’s relative immaturity and the industry’s historical resistance to widespread innovation,” Lynch said.

In addition, companies with revenue over $5 billion are 45% more likely to describe their AI maturity level as “above average” compared those with an annual revenue under $5 billion, Lynch told Construction Dive via email.

The report fits into a growing body of studies and surveys that arrive at similar conclusions — while most builders are optimistic about AI and its potential benefits, adoption remains low. 

Larger contractors, however, are engaged in an AI arms race to gain an edge over their competitors. Swedish-builder Skanska and Watsonville, California-based Granite Construction, have touted their own AI offerings. Skanska’s Safety Sidekick combined its EHS Manual, OSHA construction standards and supplemental safety documentation into one platform, and Granite’s Asphalt Guru helps the contractor train its asphalt paving team with a collection of informational and safety documentation, such as DOT standards and industry articles.

Humanoid robots might not be a fixture on the jobsites of today, but the time is now for builders to start gameplanning for their future usage, according to an October 2025 report from consulting firm McKinsey and Co.

The report paints a scenario that contractors across the country are all too familiar with: Fewer young people entering the field, skilled workers retiring and a physically demanding environment have created a skilled labor crisis in the industry. The consulting firm also pointed to dismal gains in productivity over the last two decades, far below manufacturing and the overall economy.

In response, McKinsey claims, humanoid robots might be the answer. With the advent of embodied artificial intelligence, better physical component technology and increasingly sophisticated AI models, these robots represent what McKinsey calls a “potentially transformative solution” for the industry’s productivity crisis.

To start, contractors should begin to figure out where humanoid robots could fill in the gaps and aid their businesses the most, either by closing productivity gaps, reducing hazardous risks and supporting large-scale infrastructure construction, per the report. 

Current deployments focus on repetitive, moderately complex tasks in low-variability environments such as mapped grocery aisles or staged interiors with identical layouts, according to McKinsey. But future applications could include pipe installation in tight spaces, sensor installation or earthwork.

The next step is then determining, as a builder, how fast to move toward adoption. 

Doing so could help position the company as:

• A first mover that pilots projects in collaboration with original equipment manufacturers.
• An early adopter that scales proven models rapidly.
• A selective deployer that concentrates on implementing use cases with high return on investment.

Each type has its own benefits — a first mover could be seen as a trailblazer and attract talent, but could incur risks on the ROI front. By contrast, selective deployers face the lowest risk but gamble potentially falling behind peers.

“With the industry’s long-standing labor and productivity challenges likely to intensify, construction leaders would do well to begin considering potential uses for humanoids now,” the report reads. “If humanoids do become a cost-effective solution, companies will want to move as fast as possible. Those who prepare now will be best positioned to seize the opportunities ahead.”

Robots in the construction industry are not a totally new phenomenon. For example, Watsonville, California-based Granite Construction uses supervised autonomous machinery on its jobsites to upskill its heavy machinery operators. Thus far, contractors have used Boston Dynamics’ Spot the Dog to take photos and track jobsite progress, Dusty Robotics’ layout machine to perform layout work six times faster than without the tech and Canvas’ drywall finishing robot.

Contractors are also more receptive to their own companies’ forays into robotics. Positive evaluations of the tech rose from 74% in 2024 to over 95% in 2025, according to the Equipment & Robotics Benchmarking report from Chicago-based BuiltWorlds. However, the number of firms reporting active robotics use fell from 65% last year to 46% now.

Turner Construction, the largest contractor in the country by revenue, has embarked on a deal with OpenAI to give every Turner employee access to ChatGPT Enterprise, according to a November 2025 announcement.

The partnership, which will last for two years, came together as the company made the decision over the summer to go “wall to wall, floor to ceiling” with ChatGPT, OpenAI’s generative artificial intelligence product, according to Jim Barrett, the head of global innovation for the New York City-based contractor.

Turner already uses the tech in its work, from AI-powered assistants and photo analysis tools to digital tracking systems and autonomous drones, per the release. Barrett told Construction Dive that AI reaches “literally every function in the company.”

The announcement came amid Turner’s Innovation Summit in Nashville, Tennessee, which ran from Oct. 21-23, where attendees participated in AI learning sessions led by experts from OpenAI, according to the release. 

Attendees then put those lessons to the test at Turner’s largest-ever AI agent build session, a collaborative hackathon, where teams created more than 100 custom AI agents, according to the news release. Participants designed these tools to address current, on-the-ground project challenges from streamlining contract reviews to enhancing safety protocols.

The work builds on a vision the company shared going into 2025. The New York City-based contractor has at least tripled or quadrupled its investment in AI over the past couple years, Barrett told Construction Dive in January.

Barrett emphasized that there will be experts embedded throughout the company that he called “AI champions” to help employees get up to speed on the tech. Barrett also mentioned the importance of guardrails for emerging uses and making sure what’s generated by the tools is accurate and hallucination-free.

“We'll probably have 20 to 30 people that will be dedicated to helping coach our people. We want to make sure that we're getting everybody up on that learning journey, and that they're really leveraging the capabilities of ChatGPT and other tools,” Barrett said.

Because the deal is for two years, it allows Turner to be flexible down the road, Barrett said. While the company views ChatGPT as the strongest enterprise solution for the entity at the moment, it’s still early in the AI arms race.

“We're always exploring the different solutions that are out there,” Barrett said. “It’s a constantly changing landscape, and we want to be at the forefront of that.”

With that in mind, Turner is being selective when it comes to its investment. The advent of AI at the company, and the ability of its employees to build its own agents, changes the buy versus build calculus, Barrett said.

Procurement agents, Barrett said, are building tools to review and level bids, while estimators are building agents to read through building specs and cross reference them to changes in drawings.

Turner is also training its employees to build their own agents, he added. 

As AI providers knock on Turner’s door, with, “startups in particular, the question now becomes, if we can build something ourselves over a weekend,” Barrett said, “then why would we necessarily pay for someone to build it for us?”