Data center construction timelines are getting shorter. Cooling technology is getting more complicated.

Caught between a rock and a hard place, artificial intelligence data centers often rely on a range of vendors — including equipment and fluid suppliers and commissioning partners — to help them bring new capacity online quickly and maintain operations with zero unplanned downtime. 

But working with multiple external experts can raise its own challenges, said Jesse Holshouser, corporate account manager for data centers and high-performance compute applications for Ecolab. A data center operator might find that the complexity of coordinating with multiple cooling system vendors can bring operations to a halt due to a simple mistake.

“The interface between all of those pieces can mean equipment is not ready for installation, or the coolant isn't at the site yet, or the systems haven’t been properly cleaned and commissioned, which becomes a schedule issue really quickly,” Holshouser said. “What we're hearing from our customers is that they want simplicity: fewer handoffs, better coordination and clearer accountability across the entire cooling ecosystem.”

With its integrated portfolio of equipment, fluids, analytics and expert services, Ecolab can simplify this process for data center customers and keep projects on schedule, he said. Bisnow spoke with Holshouser to learn how Ecolab and CoolIT Systems, an Ecolab company, help operators accelerate their AI deployments by simplifying liquid cooling management.

Bisnow: How should a data center operator think about outsourcing the maintenance of their cooling systems?

Holshouser: They should look beyond just individual products and ask who's going to help them deploy and operate at scale. To do that, they have to find partners that support the full cooling life cycle.

Obviously, the cooling technology has to actually work, but there are a number of other components that are incredibly important. These include supply chain logistics, quality control, precommissioning and startup services, monitoring and technical expertise to support an operational facility. At scale, even a small logistical or quality issue becomes a huge problem because you're repeating those systems across enormous deployments.

That's where Ecolab's cooling portfolio becomes valuable, because we've taken all those pieces of the puzzle and put them under the Ecolab umbrella instead of focusing on one piece of the system.

Bisnow: You mentioned the supply chain, which is always tricky these days. How does Ecolab avoid logistical hiccups?

Holshouser: We have several supply chain partners for things like fluids, and we also have our own robust internal supply chain network. We have storage strategically placed across the entire United States so that we can simplify logistics to support locations that are seeing a lot of AI deployments. 

Bisnow: How does CoolIT, which Ecolab acquired in July, fit into this integrated approach?

Holshouser: CoolIT manufactures coolant distribution units, the piping and cold plates where the heat transfer actually happens. You can't separate what happens in the coolant from the cooling hardware because it all interacts with one another in one system.

Historically, a customer bought coolant from one supplier, had another supplier for the CDU, another for commissioning and yet another one to help troubleshoot systems. With CoolIT and Ecolab together, we're able to directly connect across each one of those domains, which creates more continuity, streamlined visibility into system health and ultimately a simpler experience for our customers. That's the goal: to make liquid cooling easy to deploy, troubleshoot, scale and operate. 

Bisnow: How else does your company’s vertical integration meet data center customers’ evolving needs?

Holshouser: A year ago, the conversation surrounding data centers was mostly focused on “how do we actually deploy liquid cooling?” Now, the conversation has shifted to “how do we operate reliably at scale?”

There's a tremendous amount of pressure to deploy these systems quickly, but as you move into production, the customers are increasingly focused on uptime, maintainability, reliability and operating cost across a facility’s entire life cycle. 

We've seen real examples of why that's important, especially for equipment. We've seen equipment keep cooling systems offline for weeks or even months at a time, and the customer is just sitting there waiting on repairs or replacement parts. 

Bisnow: That sounds like a very expensive problem for them.

Holshouser: When you're talking about AI infrastructure, that is a huge deal. You're talking about delayed compute capacity, and that represents lost revenue and business opportunity for our customers and for their end users. 

They want reliable and responsive alternatives, and they're focused on service response. Preventive maintenance plays a huge part, as do monitoring and technical expertise. We want to enable customers to focus on their business and not all those little pieces of the cooling loop.

Bisnow: How has AI changed the conversation around cooling?

Holshouser: In traditional glycol applications, the mindset was much closer to “let's fill the system and then set it and forget it.” But AI cooling systems are fundamentally different. They require more interaction with the system, maintenance and component changeouts. There is more opportunity for contamination, dilution or other events that would affect the fluid quality and the health of your system over time. 

Bisnow: What could go wrong in this scenario if cooling is not managed well?

Holshouser: We've seen where racks are supposed to be prefilled with glycol, and they show up to the site with no glycol at all, or at the wrong concentration, or maybe an entirely different fluid altogether. That can be catastrophic, and contamination or dilution issues don't necessarily announce themselves right away.

That's why coolant shouldn’t be considered a commodity purchase. Quality control, logistics and commissioning support, as well as the ongoing service and analytical portion of it, are all important components. That's where Ecolab's fluid management experience really makes a difference. 

Bisnow: Is this what Ecolab means by “cooling-as-a-service?”

Holshouser: The cooling-as-a-service approach allows the customer to focus on the outcome rather than managing each individual piece of a liquid-cooled system. Data center operators just want to deliver reliable compute power. They don't necessarily want to have an internal organization that is well-versed in coolant chemistry, CDU monitoring and analytics. Cooling-as-a-service puts cooling management in the hands of dedicated cooling specialists and gives customers a consistent operating model. 

This article was produced in collaboration between Ecolab and Studio B. Bisnow news staff was not involved in the production of this content.

Studio B is Bisnow’s in-house content and design studio. To learn more about how Studio B can help your team, reach out to studio@bisnow.com.