Owners or managers of industrial, logistics, or office buildings are adopting more efficient and economical cooling technologies, as data center operators are doing, who are switching from air cooling to liquid cooling as rack densities increase, according to a new JLL analysis.

More work, more heat generated by equipment

The increasing workload with artificial intelligence is pushing traditional air cooling systems to their limits. To combat this challenge, both chip manufacturers and data center operators are turning to a more efficient solution: liquid cooling.

Leading the charge is data center operator Equinix, which plans to implement liquid cooling in 100 of its data centers in 45 cities. Similarly, Digital Realty has launched a high-density colocation offering powered by liquid cooling, handling workloads of up to 70 kilowatts (kW) per rack.

Chip giant Nvidia is also on board, designing its next-generation servers specifically for liquid cooling to manage the heat generated by their central processing units (CPUs) and graphics processing units (GPUs).

With rack densities expected to exceed 70kW, the only viable solution for cooling these high-performance servers is currently liquid cooling – primarily direct-to-chip or immersion cooling, according to Andrew Green, Practice Director for Regional Data Centers, Working Dynamics, JLL.

“We’ve reached a point in data center development where rack densities have surpassed what is possible with the physics of air cooling,” says Green. “Liquid cooling is an inevitable development to cope with the increased AI workloads.”

For data center operators under pressure to improve their power usage effectiveness (PUE), the energy efficiency gains achieved through liquid cooling offer a significant advantage. Less energy is required to cool the refrigerant, which translates directly into substantial cost savings.

Embracing liquid cooling

But the benefits of implementing liquid cooling go beyond energy and cost savings.

“A liquid cooling installation eliminates the need for mechanical air cooling equipment, freeing up valuable floor space to be converted into additional data hall capacity,” says Green.

Another advantage, in the case of an immersion cooling bath installation, is the lower floor-to-ceiling height requirement of four meters, similar to a grade A office space.

This major reduction, compared to the standard six meters required for most existing data centers, comes with the trade-off of requiring an increased structural floor load. “If the same amount of computing power is installed in the same area, the weight per square meter increases from 12 to 15 kilopascals (kPa) to at least 20 kPa in a liquid cooling installation,” explains Green.

Despite the clear benefits of liquid cooling, a pivot in cooling technology is a major undertaking for operators, potentially requiring an overhaul of existing infrastructure and design.

One of the key design changes involves plumbing. “Traditionally, we've put a lot of effort into keeping water out of data halls, whether it’s pre-action sprinklers or water detection systems,” says Green.

But in the case of direct-to-chip cooling, for example, small-bore pipes are required to deliver the cooling liquid directly to the chips for heat removal.

Keeping up with change

This reality means that new purpose-built AI data centers are better positioned to more easily adopt liquid cooling. While existing data centers with spare capacity can be partially upgraded, it is unlikely that an existing facility can be upgraded to 100% AI due to infrastructure constraints.

“Any major change in a real-time data center environment is a period of high risk and must be carefully managed,” says Green. “Operators need to consider how, in terms of resilience, those works can lead to service disruptions in other data halls.”

The data center industry generally aims to be cutting edge, not bleeding edge, says Green. “Operators are often slow to adopt changes and usually favor proven solutions.”

But advances in AI and the broader tech industry are forcing data centers to adapt.

“Data center companies are thinking about the best way to implement AI, whether in smaller, older buildings or in the design of their future facilities,” says Green. (Photo: Freepik)