As a designer of micro, prefabricated, and modular data centers – spanning everything from compact Edge sites to high-density HPC and AI environments – Oper8 is well equipped to speak to the future of direct-to-chip cooling as its adoption accelerates across diverse deployment scenarios.

As we in the industry recognize, liquid cooling has rapidly evolved from an experimental concept into a proven, practical response to the challenges operators face today. Andrew Kernebone, technical director for Asia at Oper8, explains:

“It’s the best answer we have right now for those kinds of densities. It’s also a solution that can be deployed on a rack-by-rack basis in certain scenarios. Liquid cooling gives you the ability to address the challenge of having mixed load racks, where you’ve got lower and higher density racks in the same space, to remove heat most effectively at this stage.”

Today’s cooling market

In the data center sector, single-phase liquid cooling remains the most widely deployed method for removing heat from chips. However, two-phase dielectric direct-to-chip solutions – though not yet as widely adopted – are gaining momentum as operators explore alternatives suited to higher densities and specialized environments.

“The scale you’re working with today will probably make a single-phase solution the best fit, whereas if you’re looking at isolated racks or a retrofit, a dielectric solution might be more appropriate. There’s no one answer over the other, and part of the design process – and part of the fun – is figuring out the right solution for each scenario.”

As the industry moves beyond the experimental stage, the past 12 months have marked a clear shift toward recognizing liquid cooling as essential. What was once a largely ‘hydrophobic’ mindset has quickly evolved, demonstrating just how rapidly attitudes and adoption are changing.

Deployment in mixed environments

When deploying a high-density data center with uniform servers, the design process becomes far more consistent. You can apply a more or less single approach across the entire white space – challenging, certainly, but largely uniform. However, organizations that need to run low-density and high-density workloads side by side require a very different approach. This is where expertise in hybrid designs, whether for new facilities or retrofits, comes in handy.

“We’re working with a research organization that has a mixed environment of 25kW and 60kW racks, and they want to scale those 60kW racks to 100kW. That means using different cooling strategies within the same white space,” says Kernebone, before explaining how Oper8 is addressing the challenge.

As Kernebone explains the differing approaches required for hyperscale versus Edge environments, it becomes clear that the nuances in requirements are significant, and that addressing the rising complexity of equipment demands a highly tailored design approach.

Where liquid cooling shines

While air cooling dominated data centers for decades, liquid cooling is increasingly proving itself as a more energy-efficient and space-efficient way to manage today’s demanding IT loads. In hybrid environments, its benefits extend even further – improving the overall efficiency of air-based systems operating alongside it. Kernebone adds:

“For example, in a mixed environment where 80 percent of the load is cooled via liquid, using a fraction of the power that an equivalent air-based system would require, you can design exceptionally low PUE, highly efficient facilities.”

The impact is significant, not only in meeting performance and density requirements, but also in reducing ongoing operational costs and resource consumption.

Watch the full DCD>Talks episode here.