The rise of generative AI has triggered an unprecedented wave of data center construction as companies scramble to secure the computing power needed to train and run large AI models. As hyperscalers race to acquire land, power, and water for massive AI infrastructure projects, CBRE, which is headquartered in Dallas, has quietly become a key player behind the scenes. 

The company says it has secured “dozens” of potential data center sites around the country and is working directly with major tech companies on everything from land entitlements to power and water access. Unlike pure brokerage firms or construction contractors, CBRE increasingly sits across multiple layers of the AI infrastructure buildout. 

The company’s move into critical infrastructure and data center services “is going to be at least as profound as our move into outsourcing in the 1990s and early 2000s, and much faster,” CBRE president and CEO Bob Sulentic said on the company’s April 23 earnings call, in which the company announced the its revenue had jumped 19% year over year. 

The shift is already reshaping CBRE’s business. The company generated more than $3 billion in infrastructure-related revenue in 2025 and nearly $950 million in the first quarter alone. It recently created a dedicated “critical infrastructure services” unit focused on data centers, telecom, and power infrastructure—a business expected to grow more than 60% this year. CBRE also said its data center leasing revenue more than tripled year-over-year in the first quarter.

In an April shareholder letter, the company said critical infrastructure activities—including data center operations and its Pearce Services business—accounted for roughly 14% of core EBITDA in 2025, up from about 3% in 2021. CBRE said it sees “considerable opportunities ahead.”

A company spokesperson said CBRE currently manages about 1,300 data centers globally, serves as project manager for roughly 150 facilities, provides sales, leasing, and financing services for about 250 data centers, and controls more than 30 development sites.

Stephen Sheldon, a financial analyst and partner at William Blair, said CBRE has the most exposure in the data center space and has been more proactive than its competitors, including JLL, Cushman & Wakefield, and Collier. 

“I don’t think it’s slowing down anytime soon,” he said. “I’m sure there’s going to be friction points and bottlenecks over the next five-plus years, but I don’t think anyone would doubt that this is going to be a massive secular investment area [for CBRE]”

CBRE push into data centers came through acquisition

In an interview with Fortune, Sulentic, who has served as CEO since 2012, said CBRE’s push into data centers began several years ago as the company expanded its building management and project management businesses through acquisitions.

“We started to do some work for data center clients,” Sulentic said. “So we started managing data centers, and because of our scale, we started doing a lot of that pretty rapidly.”

That scale—CBRE operates in more than 100 countries and works with 90 of the Fortune 100 companies—positioned it to grow alongside the AI infrastructure boom as hyperscalers rapidly expanded their data center footprints.

“Once data centers appeared on the scene and grew rapidly, we grew rapidly,” Sulentic said.

As the business expanded, CBRE increasingly began viewing data centers and critical infrastructure as strategic growth areas. Sulentic noted that acquisitions have long been central to the company’s growth strategy.

In November 2025, CBRE acquired technical infrastructure maintenance firm Pearce Services for $1.2 billion in cash to expand its digital and critical power infrastructure footprint. The deal followed CBRE’s June 2024 acquisition of Direct Line Global, a provider of technical and installation services for data centers.

Sulentic said the scale of planned AI infrastructure spending over the next five years appears almost unprecedented. Because CBRE works closely with hyperscalers, the company has visibility into their long-term capital expenditure plans and buildout targets—projections he said carry weight because CBRE is directly involved in executing many of the associated projects.

Sheldon pointed out that CBRE is unique in that it is not only providing advice and services around data centers, but also making investments in them through its development arm. 

“They’ve actually procured a good amount of land,” he said. “I think they used some of their insights, knowing what land might be attractive for data center development down the road, looking at infrastructure, power grids, things like that.” Still, he added that Wall Street’s current expectations for CBRE are not heavily dependent on large gains from those land holdings, meaning any significant appreciation would likely be viewed as incremental upside—“icing on the cake” rather than a core assumption built into forecasts.

Challenges include labor expertise and community pushback

The biggest challenges, he said, have evolved. “I would have said a year ago that the hardest part is getting power to sites,” he said. Now, however, finding enough qualified technical labor is “really, really tough.”

To address that shortage, CBRE and Facebook parent Meta recently announced plans to create multiple U.S. training centers expected to teach thousands of technicians and skilled trades workers how to build and maintain the latest data centers. The first two training sites are set to open this summer near airports in Ohio and Indianapolis.

CBRE isn’t the only real estate services provider grappling with the talent crunch tied to data center construction and operations. Rival JLL is also working to recruit a direct pipeline of skilled trade workers, including electricians, HVAC technicians, and plumbers, as demand for AI infrastructure accelerates.

Sulentic also admitted that community opposition to data centers has intensified as the AI infrastructure boom accelerates. As Fortune has reported, communities across states including Louisiana, Arizona, Michigan and Texas have raised concerns over water use, power demand, noise, and land use.

Still, he argued that over time, projects will gravitate toward locations that make the most sense from a power, water, and infrastructure standpoint.

“It’s just like water rising to its own level,” he said. “Over time, data centers will end up in places that work better than other places. We’re going to have to find sources of water and power that work more optimally as the industry evolves.”

Sulentic said public criticism of data centers is sometimes amplified politically and he pushed back on the idea that hyperscalers are indifferent to the communities where they build.

“There’s this sense that corporations are indifferent to human beings,” he said. “They’re not indifferent at all. I’m on the inside with these companies as a provider, and they are working very, very hard to try to solve these problems.”

He added that CBRE expects demand for its data center-related services to continue growing alongside the broader AI infrastructure market.

“The whole business is growing reasonably rapidly,” Sulentic said. “And I think this part of the business will outgrow the rest of the company for some time.”