September 10, 2025 | BOMA International, Ella Krygiel
By 2030, the electricity consumption of data centers is projected to more than double, driven by the need to support sophisticated AI applications, Scientific American reports. This is happening because AI data centers are fundamentally different from traditional ones—they’re built to meet the high-performance computing and power demands that AI requires. These facilities have some key features that set them apart, according to the Kaizen Institute:
- High-performance infrastructure
- Advanced cooling systems
- Automated management
- Low latency and high connectivity
- Optimized software
AI data centers were developed largely in response to the AI platforms that are now part of our lives. Chat GPT, for example, was only created in November 2022, which led to an urgent sprint for developers to create new data centers models that could handle these demands.
“An AI data center is a completely new model,” Sean Farney, Vice President, Data Center Strategy, JLL explains. “They use GPU power servers, which were previously used in video games like Pac-Man.” GPUs, or graphics processing units, are electronic circuits designed to speed computer graphics and image processing on various device (IBM). Essentially, they help computers run multiple programs faster, which is exactly what consumers want today.
Farney explains that because of AI’s rapid development, these specialized data centers were needed almost overnight due to high demand. One of the biggest differences from traditional data centers is that that they are five to ten times more power dense, which means they produce a lot more heat.
This is where liquid cooling comes in. Unlike air, liquids can carry away more heat more efficiently due to their higher heat capacity, according to Flexential. Traditional data centers used air cooling, but as server densities increased, liquid cooling became the primary method to handle the “greater thermal loads of modern data centers.”
There are three different types of liquid cooling worth knowing about, according to Inova:
- Rear Door Heat Exchanger: This system draws ambient air into the cabinet, removes heat using a heat exchanger door attached to the back of the rack, and reintroduces the cooled air in the room.
- Direct-to-Chip Cooling: Direct-to-Chip cooling involves circulating liquid coolant directly over the heat-generating components of individual chips, ensuring precise and targeted cooling.
- Immersive Cooling: Immersive cooling submerges entire servers or hardware components in a dielectric fluid, creating a direct and uniform cooling effect.
Farney mentioned that direct-to chip cooling is likely the favored method for technicians, as it presents an efficient way to cool the equipment. As Inova describes, Direct-to-Chip cooling technologies can remove 70-75% of the heat generated by the equipment in the rack, leaving 25-30% that must be removed by air-cooling systems.
These data centers aren’t just revolutionizing technology—they’re creating new opportunities in the job market too. According to a recent Pittsburgh Public Source article, the rise in AI data centers will impact communities, especially those that are “resource hungry” like Western Pennsylvania. A large-scale data center in the area would bring many temporary construction jobs, and 80 to 100 long-term jobs, which would help close the workforce gap. They also discuss how community colleges could offer programs to students interested in joining the industry.
This is hugely beneficial, especially since Farney acknowledged there’s a tremendous shortage in AI data center technicians. “We can’t find enough talent,” he says. “We need a lot more and so does the entire industry.” However, companies like JLL are partnering with junior colleges and technical colleges by providing data center certification programs, so they can hire folks after they complete the training. “I would implore anyone, particularly the younger folks who haven’t entered the workforce yet to get really involved in this industry because the growth is endless and it’s a really awesome career choice,” Farney says.
As for the future, Farney has faith that AI will not take over the world. “I think the future is very bright. AI has the ability to be a force multiplier and take care of some administrative and operational overhead for ourselves, personally and for companies,” he emphasizes. “The next ten years will be mind-blowing for where data centers will be, especially as we consider the different types of technology advancing like quantum computing. We’re already seeing quantum computing applications applied to solving cancer, for example,” he points out. “It’s a great time for anyone to be in the digital infrastructure industry.”
The facilities that seemed impossible just a few years ago are now becoming the backbone of our AI-driven future, and the opportunities, both technological and professional, are just getting started.
Interested in more content like this? Read our recent articles, Cold Storage’s Industrial Real Estate Impact or Industrial Real Estate Goes Electric. You can view all this content and more when you click here to sign up for our Industrial newsletter!