Data centers could account for 44% of U.S. electricity load growth by 2028 and consume up to 9% of the country's electricity supply by 2030, raising concerns about U.S. electricity availability and costs. According to the National Renewable Energy Laboratory, up to 40% of a data center's electricity consumption is used for cooling, making improved cooling efficiency one of the key strategies for reducing energy consumption. Cooling systems are also a core part of data center design, directly affecting how these facilities are developed, built, and retrofitted.

In the second half of 2024, several significant announcements related to data center cooling systems were released. These systems protect high-performance processors and servers that support the advanced computing required for artificial intelligence. In December, Microsoft and Schneider Electric each unveiled efficient liquid cooling system designs to support increasingly powerful AI chips. Microsoft's water-based design operates in a closed loop, eliminating waste from evaporation, while Schneider Electric's data center reference design uses a non-water refrigerant. Earlier in 2024, Vertiv and Compass Datacenters showcased their "first-of-its-kind" liquid-air hybrid system, with deployment expected in early 2025.

Here is what data center cooling experts see for trends and developments in 2025 and beyond.

Two-phase liquid cooling technology will move into mainstream implementation

According to AFCOM's 2024 State of the Data Center Industry Report, most data center professionals are dissatisfied with current cooling solutions. 35% of respondents said they regularly make adjustments due to insufficient cooling capacity, and 20% are actively seeking new scalable systems.

Many data center cooling experts predict that developers and operators will increasingly turn to two-phase direct-to-chip cooling technology to improve cooling performance. These systems typically use non-water refrigerants as the working fluid, switching between liquid and vapor states, a process that "plays a critical role in heat removal," according to Accelsius, a designer of direct-to-chip liquid cooling systems.

Josh Claman, CEO of Accelsius, said in an interview that 2025 will be the "year of implementation" for two-phase systems, as data center professionals gradually become familiar with the technology. Claman noted that data centers with more complex and advanced computing demands are more likely to adopt two-phase cooling.

Traditional air cooling hits its physical limits when server rack density reaches about 70 kilowatts, a density that is the benchmark for today's most advanced AI training facilities, said Sarah Renaud, vice president of advisory services at ENCOR Advisors, a commercial real estate firm serving data center clients.

With higher rack densities ahead, "two-phase is the way forward," Renaud said. "It can handle higher power densities and heat fluxes, meaning it's better suited for AI workloads."

Hybrid cooling solutions will expand, but supply chain risks loom

According to research released in March 2024 by Chemours, Syska Hennessy Group, and cooling system designer LiquidStack, two-phase immersion cooling offers data center operators a lower 10-year total cost of ownership than direct-to-chip cooling or single-phase immersion cooling. But experts say its high upfront costs, the long lifespan of traditional cooling systems, and the varying cooling needs within a single data center mean two-phase technology will coexist with other technologies for some time.

"Almost no new data center will be entirely air-cooled or entirely liquid-cooled, because not all applications require high-intensity liquid cooling—think of rarely accessed archived data versus generative AI," Renaud said. "Those lower-demand racks can be air-cooled, which is more cost-effective."

Microsoft's closed-loop water-based cooling system "seems to align with a gradual strategy," supporting its near-term needs while "allowing its infrastructure to easily pivot to advanced cooling technologies, such as direct-to-chip two-phase cooling, when the time is right," said Nick Schweissguth, director of product and commercial enablement at LiquidStack.

But data center operators' hybrid cooling plans could be complicated by supply chain issues, which anticipated Trump administration tariffs could worsen, Schweissguth said. He specifically pointed to direct-to-chip coolant distribution units (CDUs)—used to keep processors submerged in fluid—as being particularly at risk.

As CDU demand surges in 2025, "companies vying for the direct-to-chip cooling market will ultimately depend on their ability to scale production and build strong relationships with suppliers," Schweissguth said.

Building and system design will evolve to support round-the-clock operation

Steven Carlini, vice president of innovation and data centers at Schneider Electric, said operators' expectations for the most advanced AI data centers far exceed those of previous generations of facilities. Early facilities' workloads might fluctuate between 30% or 40% of total processing capacity, while AI facilities typically run at 100% capacity for weeks or months when training models, requiring more robust and redundant designs, Carlini said.

"This eliminates variability, but you have to ensure the cooling system design can support that," he said.

Carlini described a near future where higher rack power densities require heavier cooling infrastructure, adding extra physical demands on data center design. For example, a recent design his team worked on involved "enormous" pipes with "large steel cages covering superclusters," or a two-story floor layout with the first floor flush with the concrete slab to bear the extra weight.

"All that water has to go somewhere," he said.

A technician inspects data center servers in an immersion cooling tank.
Experts predict that developers and operators will increasingly adopt two-phase direct-to-chip cooling technology (pictured)—which switches the working fluid between liquid and vapor states—to improve cooling performance, a process that "plays a critical role in heat removal," according to cooling system designer Accelsius.
Image credit: Halbergman via Getty Images

'Slow but steady' retrofit activity is coming

Accelsius's Claman said retrofitting operating data centers to accommodate more powerful processors is a huge technical and logistical challenge, leading some to believe new construction is easier. But he noted that new buildings consume significantly more resources, complicating corporate sustainability goals. Existing data centers often have more robust power supplies. "That's why they're located where they are; they don't move easily," he said.

According to JLL's 2025 Global Data Center Outlook, most of the asset value of operating data centers lies in their power supply and infrastructure, such as electrical, plumbing, and other technical systems. Given the challenges of securing power for new developments, these assets are especially valuable. Therefore, retrofits such as transitioning existing data centers to liquid cooling will "become a viable solution and an opportunity to increase asset value," JLL's outlook said.

Renaud noted that Meta is transitioning its existing data centers to liquid cooling, "because they say 'we have to do it,'" and colocation giant Equinix said in December 2023 it would extend liquid cooling to its 100 data center facilities. Claman predicted retrofits will proceed "slowly but steadily," with a "more balanced discussion" around their benefits. Schneider Electric is also betting on this trend, recently partnering with Nvidia to release three retrofit reference designs to help data center operators boost performance without redesigning facilities from scratch.

Carlini said the rapid growth of computing power means that data centers at the cutting edge today can quickly fall behind, further compounding the daunting challenge of designing facilities with both air and liquid cooling infrastructure.

"Ten years ago, you would try to design a data center with excess capacity and scale up gradually, but now you don't know what power density to build for," he said.

Cold northern climates may gain an advantage

Carlini said that even in newer data centers, air still handles 20% to 30% of the cooling load. This is prompting efficiency-focused developers to locate more facilities in the "attic"—the industry's informal term for colder northern regions, Renaud and Claman said.

"The market often talks about 'free cooling zones,'" Claman said, referring to the northern United States, Northern Europe, and Canada.

In colder weather, energy consumption for air cooling systems can drop by up to 95%, Renaud said. "We're seeing a trend toward hybrid colocation strategies, where data that doesn't need frequent access can be stored in more remote and colder locations," while high-access facilities continue to operate in warmer, more established data center hubs like Northern Virginia, she said.

Claman said cold-climate sites are also less likely to need water-intensive evaporative cooling systems, which are common in warm, dry climates and have raised concerns about data centers' environmental impact. He predicted a shift toward closed-loop cooling systems that can leverage seasonal free cooling.

"There's a lot of scrutiny around draining aquifers to cool data centers," he said.