Introduction: The Shift in Green Data Center Priorities
For more than a decade, the global data center industry has relied on Power Usage Effectiveness (PUE) as the primary benchmark for operational efficiency. However, as AI workloads continue to grow rapidly, the conversation around sustainability is shifting toward Water Usage Effectiveness (WUE) and the future of sustainable cooling infrastructure.
Modern AI systems powered by high-performance GPUs generate significantly more heat than conventional computing environments. As a result, tropical regions such as Indonesia face increasing pressure to improve AI data center cooling efficiency while also reducing environmental impact.
In practice, many hyperscale operators aggressively pursue lower PUE values by relying on evaporative cooling systems such as cooling towers. While this approach reduces electricity usage, it often requires enormous volumes of freshwater to operate efficiently.
As climate risks intensify and urban water stress becomes a growing concern, energy efficiency alone is no longer sufficient to define a truly sustainable green data center.
The industry is now shifting toward a more comprehensive sustainability metric known as Water Usage Effectiveness (WUE), which measures how much water is consumed to support data center operations.
The PUE vs. WUE Dilemma: Why Energy Efficiency Alone Is No Longer Enough
A significant technical trade-off exists between reducing electricity usage and minimizing water consumption. The more aggressively operators push PUE values toward the ideal benchmark of 1.0, the more likely they are to depend on water-based cooling technologies.
Traditional cooling towers work by evaporating water to remove heat from cooling systems. This process can substantially lower electricity consumption because evaporation naturally dissipates heat efficiently. However, it also leads to extremely high levels of water consumption at industrial scale.
Recent discussions surrounding AI infrastructure water consumption have highlighted how next-generation AI data centers may consume massive amounts of freshwater, particularly when relying on open-loop evaporative cooling systems.
This growing concern has elevated the importance of WUE as a critical sustainability metric.
Measured in liters per kilowatt-hour (L/kWh), WUE evaluates how efficiently a data center uses water relative to the energy consumed by its IT equipment. The lower the WUE value, the more sustainable the facility becomes from a water management perspective.
Leading operators are now pursuing near-zero WUE targets to minimize water waste and reduce environmental impact.
Indonesia’s Tropical Climate Creates Unique Cooling Challenges
Operating AI-scale infrastructure in Indonesia presents significantly greater thermal challenges compared to facilities located in colder regions such as Northern Europe or Scandinavia, where operators can leverage free cooling technology using naturally cold outdoor air.
1. Hot and Humid Environmental Conditions
Indonesia’s consistently high humidity levels reduce the efficiency of natural heat dissipation. Because the surrounding air already contains substantial moisture, cooling systems must operate continuously under heavy load conditions to maintain stable server temperatures.
This creates substantial pressure on cooling infrastructure and significantly increases operational demands for hot and humid data center cooling strategies.
2. High-Density AI Computing Workloads
Modern AI servers powered by GPUs generate extremely high rack densities. In many advanced AI deployments, racks now exceed 20 kW and can even surpass 100 kW per rack.
At this scale, conventional air cooling approaches are approaching their physical limitations. Traditional airflow systems are no longer capable of removing heat efficiently from high-performance AI processors.
As a result, the industry is increasingly transitioning toward liquid cooling for AI servers and immersion cooling technology as next-generation thermal management solutions.
Research has shown that liquid cooling technologies can improve thermal stability while delivering significantly higher computational efficiency per watt compared to traditional air-based cooling systems.
Comparison of Data Center Cooling Technologies
Parameter | Conventional Air Cooling | Closed-Loop Cooling | Liquid Immersion Cooling |
Cooling Medium | Cold air from CRAC/CRAH | Water/Glycol in sealed pipes | Dielectric cooling fluid |
Rack Capacity | ~15–20 kW/rack | ~20–40 kW/rack | >40–100 kW/rack |
Water Consumption (WUE) | Low | Near-zero | Zero evaporation |
Energy Efficiency | Higher PUE | Efficient PUE | Extremely low PUE |
AI Suitability | Limited | Moderate-to-high | Excellent for AI workloads |
Modern closed-loop cooling systems are increasingly adopted because they reduce water loss by circulating coolant within sealed systems.
Meanwhile, liquid immersion cooling is considered one of the most promising technologies for high-density AI workloads because liquids transfer heat far more effectively than air.
These systems also utilize specialized dielectric cooling fluids, which allow electronic components to operate safely while fully submerged.
ESG Regulations and Sustainability Pressure
The urgency for adopting WUE standards is not driven solely by technical limitations. Global ESG requirements are also accelerating the transition toward water-efficient infrastructure.
Environmental: Water Footprint Transparency
International sustainability frameworks such as GRI 303 Water and Effluents increasingly require corporations to disclose their water sourcing, water usage patterns, and environmental impacts.
As a result, water footprint reporting is becoming a critical component of sustainability disclosure for digital infrastructure operators.
Social: Community and Resource Conflicts
In many regions, large-scale data center projects are beginning to face public scrutiny over water extraction practices and environmental concerns.
For tropical countries like Indonesia, where urban water resilience is already under pressure, excessive industrial water usage may create long-term social and reputational risks.
Governance: Sustainable Cooling Materials
Modern ESG audits now extend beyond electricity and water metrics into the chemical materials used throughout operational infrastructure.
This has increased interest in bio-based dielectric fluids, which are biodegradable, non-toxic, and environmentally safer compared to traditional petroleum-based cooling liquids.
These fluids also help reduce Global Warming Potential (GWP), making them increasingly attractive for sustainable AI infrastructure deployments.
Conclusion
The AI revolution is pushing the data center industry into a new era. In the past, success was largely defined by achieving low PUE values and maximizing electrical efficiency. Today, sustainability demands a much broader perspective.
For Indonesia, adopting Water Usage Effectiveness (WUE) as a core operational metric is becoming increasingly important due to the country’s tropical climate and rapidly growing digital economy.
Technologies such as closed-loop cooling systems and liquid immersion cooling are emerging as critical foundations for future AI-ready data center infrastructure.
By moving beyond PUE and aggressively optimizing WUE, Indonesia has the opportunity to build digital infrastructure that is not only computationally powerful, but also environmentally and socially sustainable.
