Overview
Data center cooling has become one of the most critical enablers of AI infrastructure performance.
As rack densities climb and liquid cooling adoption accelerates, thermal management is becoming a critical factor in uptime, operational efficiency, and scalability. While much of the industry’s attention is focused on cooling hardware, CDUs, heat exchangers, pumps, and direct-to-chip technologies, one of the most important drivers of long-term performance is often overlooked: the condition of the heat transfer fluid and how the system is commissioned and maintained over time.
At Apex, we work with operators, EPC firms, and cooling OEMs deploying next-generation liquid cooling environments. Across those projects, one reality consistently emerges:
The most reliable cooling systems aren’t necessarily the most advanced, they’re the systems that are properly commissioned, actively monitored, and managed as critical infrastructure assets.
AI Is Raising the Stakes for Cooling Reliability
Traditional enterprise data centers were often designed around rack densities of 5–15 kW per rack.
Today’s AI environments routinely operate at 50 kW, 100 kW, and even 150+ kW per rack, leaving little margin for cooling inefficiencies.
A partially fouled heat exchanger or trapped air within a cooling loop may only reduce thermal performance by a few percentage points. In an AI environment, however, those losses can translate into reduced cooling headroom, higher operating costs, and constraints on future growth.
Cooling performance is no longer simply an engineering metric.
It’s a business metric tied directly to uptime, utilization, and scalability.
What Leading Operators Are Doing Differently
As AI infrastructure scales, the conversation is shifting from cooling technology selection to cooling system reliability.
The most successful operators recognize that long-term performance isn’t determined solely by CDUs, pumps, or heat exchangers. It’s driven by how effectively the entire cooling ecosystem is commissioned, monitored, and maintained over time.
Across hyperscale, colocation, and enterprise environments, several best practices are emerging as common characteristics of high-performing liquid cooling systems.
Best Practice #1: Treat Glycol as a Critical Asset
One of the most common misconceptions in liquid cooling is that glycol’s primary role is freeze protection. In single phase direct-to-chip glycol fluid cooling loops, glycol is chosen primarily due to it’s higher boiling point and biostatic characteristics.
In reality, modern inhibited glycol formulations provide corrosion protection, scale prevention, materials compatibility, and thermal stability throughout the cooling loop. It is still unknown how long a well-preserved glycol loop fluid will last. Industry experts are saying 6-8 years, but there simply isn’t enough data to tell. Either way, glycol will break down over time so commissioning excellence and on-going monitoring and mitigation is critical to get the most lifespan out of the fluid.
Over time, contamination, oxygen ingress, improper makeup water, and normal operating conditions can degrade these protective properties.
Leading operators increasingly utilize routine fluid analysis to monitor:
- Glycol concentration
- pH and reserve alkalinity
- Conductivity
- Corrosion inhibitor health
- Dissolved metals
- Contamination indicators
Like oil analysis in critical equipment, fluid testing helps identify developing issues before they affect uptime, efficiency, or cooling capacity.
For AI-ready facilities, fluid analysis should be viewed as a reliability strategy, not simply a maintenance activity.
Best Practice #2: Establish a Strong Commissioning Baseline
Many cooling system issues originate long before the first server is powered on.
Construction debris, fabrication oils, welding residue, and improperly prepared piping surfaces can remain within a cooling system if startup procedures are rushed or incomplete.
A best-practice commissioning process should include:
- Pressure testing and leak verification
- System flushing and cleaning
- Glycol charging and verification
- Air elimination
- Baseline sampling and documentation
These activities establish the baseline for long-term reliability, efficiency, and fluid health by ensuring systems begin operation in a clean and chemically stable condition.
Best Practice #3: Don’t Skip Passivation
If there’s one commissioning activity that deserves more attention in the data center industry, it’s passivation.
Proper passivation creates a protective oxide layer on stainless steel surfaces that improves corrosion resistance and helps prevent future fouling and contamination.
While often viewed as a startup task, passivation can have a lasting impact on system cleanliness, fluid health, and equipment longevity.
For operators investing in infrastructure expected to perform for decades, passivation should be considered a standard reliability practice—not an optional step.
Best Practice #4: Eliminate Air and Control Oxygen
Trapped air remains one of the most overlooked causes of cooling system degradation.
Air pockets and oxygen ingress can contribute to:
- Corrosion
- Pump cavitation
- Flow restrictions
- Reduced heat transfer efficiency
Even the best fluid chemistry program can be compromised by excessive oxygen within the system.
Proper venting, air elimination, circulation, and ongoing monitoring should be treated as continuous reliability practices rather than one-time commissioning activities.
Best Practice #5: Adopt a Lifecycle Management Mindset
Leading operators no longer manage cooling fluids reactively. Instead, they apply the same discipline used for other mission-critical assets through structured monitoring and predictive maintenance programs.
The goal is to move from reactive maintenance to predictive reliability.
A fluid lifecycle management strategy typically includes:
- Routine laboratory testing
- Performance trending
- Predictive maintenance
- Corrective treatment mitigation
The objective is to maximize cooling capacity, extend equipment life, reduce operational risk, and maintain confidence in mission-critical infrastructure.
As liquid cooling adoption accelerates, fluid lifecycle management is quickly becoming an industry best practice.
Cooling Is No Longer a Utility Function
As AI workloads continue to increase cooling demands, the distinction between mechanical systems, fluid chemistry, and operational reliability is rapidly disappearing.
Organizations that invest in proper commissioning, fluid quality management, routine testing, and lifecycle optimization will be better positioned to maximize uptime, improve efficiency, reduce operational risk, and support the growth demands of AI infrastructure.
At Apex, we believe the future of cooling reliability lies at the intersection of engineering expertise, fluid science, and operational discipline.
As an authorized DOWFROSTâ„¢ service provider and long-time distributor, Apex supports data centers throughout the cooling system lifecycle: from commissioning and startup to fluid management, performance optimization, and long-term reliability programs.
The organizations that treat cooling fluids as strategic assets, not consumables, will be best positioned to maximize uptime, improve efficiency, and support the next generation of AI infrastructure.
Because in an AI-driven world, thermal performance isn’t just an engineering metric. It’s a business outcome.
Frequently Asked Questions
What is glycol used for in AI data center cooling?
Glycol-based heat transfer fluids circulate through closed-loop liquid cooling systems to remove heat from high-density AI servers. Modern inhibited glycol formulations provide freeze protection, corrosion inhibition, thermal stability, and long-term system protection while supporting efficient heat transfer in direct-to-chip cooling applications.
What is the difference between propylene glycol and ethylene glycol in data center cooling?
Both propylene glycol (PG) and ethylene glycol (EG) are effective heat transfer fluids used in liquid cooling systems. Ethylene glycol offers slightly better heat transfer performance and lower viscosity, while propylene glycol provides significantly lower toxicity and is generally preferred in occupied facilities where personnel safety and environmental considerations are important.
Many modern AI data centers utilize propylene glycol-based fluids, such as DOWFROSTâ„¢ LC, because they combine excellent thermal performance with long-term corrosion protection and a lower toxicity profile. Regardless of the fluid selected, following OEM recommendations and implementing a routine fluid monitoring program are essential to maximizing cooling performance and system reliability.
How long does glycol coolant last in a liquid cooling system?
There is no fixed lifespan for glycol coolant. A well-maintained inhibited glycol solution may perform effectively for 6 to 8 years or longer, but its actual service life depends on system commissioning, operating conditions, contamination, oxygen ingress, and ongoing maintenance. Rather than replacing coolant on a fixed schedule, leading data center operators rely on routine laboratory testing to monitor glycol concentration, pH, inhibitor health, conductivity, dissolved metals, and other indicators. This condition-based approach helps maximize coolant life while protecting cooling performance and system reliability.
What causes glycol coolant to degrade?
Over time, oxygen ingress, contamination, improper makeup water, thermal stress, and normal operating conditions can reduce inhibitor effectiveness and degrade coolant performance. Without monitoring and corrective treatment, degradation can increase corrosion, fouling, and maintenance costs.
How often should glycol coolant be tested in a data center?
Leading operators typically perform routine laboratory analysis throughout the life of the cooling system. Testing commonly includes glycol concentration, pH, reserve alkalinity, inhibitor health, conductivity, dissolved metals, and contamination indicators to identify problems before they impact system performance.
Why is commissioning important for glycol cooling systems?
Proper commissioning establishes the foundation for long-term reliability. Flushing, cleaning, glycol charging, leak testing, air removal, passivation, and baseline fluid testing help ensure the cooling system begins operation in a clean, chemically stable condition and reduce future operational risk.
What are the benefits of glycol lifecycle management?
A proactive glycol management program helps:
- Maximize cooling efficiency
- Reduce corrosion and fouling
- Extend coolant life
- Lower maintenance costs
- Improve uptime and operational reliability
- Protect mission-critical cooling infrastructure
Rather than reacting to failures, lifecycle management enables predictive maintenance based on fluid condition and system performance.
Brett Robison
Director of Corporate Accounts
Brett leads business development initiatives focused on data centers, liquid cooling infrastructure, and mission-critical thermal management solutions. He works closely with data center operators, EPC firms, cooling OEMs, hyperscalers, and technology partners to develop scalable cooling strategies that support AI-driven growth, operational resilience, and long-term infrastructure performance.