Reports
IEC Report Calls for Broader Measures of AI Data Center Sustainability

Key PointsPUE reveals overhead efficiency but not absolute demand or the carbon intensity of a…
View more
- PUE reveals overhead efficiency but not absolute demand or the carbon intensity of a facility's power supply.
- The Total Contribution framework spans four dimensions but still lacks a method for making them comparable.
- Demand response flexibility depends on who controls workloads and most operators lack visibility into third-party computing.
The International Electrotechnical Commission (IEC) is proposing a broader way to judge the sustainability of AI data centers. Its Clean Green Data Centres Strategic Position Paper and Leadership Briefing, announced on November 17, introduce a “Total Contribution” framework that considers a facility’s environmental footprint alongside its relationship with the electricity grid, economic productivity and public services.
The central idea is that a data center should be assessed as part of a wider system. Electricity and water consumption remain essential measures, but the IEC wants reporting to capture whether a facility can support grid resilience and deliver useful digital services as well. It has established a Strategic Standards Group on Data Centres to develop consistent metrics, reporting protocols and assurance mechanisms.
This is a strategic roadmap toward further standardization. The announcement does not establish a completed global scoring formula or a binding requirement that every data center adopt zero-water cooling.
Why PUE Cannot Tell the Whole Story
Power Usage Effectiveness, or PUE, measures total facility energy divided by the energy used by IT equipment. A PUE of 1.2 means the facility consumes 20% more energy than its IT equipment alone. Lower overhead can indicate more efficient cooling and power delivery, but the ratio does not reveal the facility’s absolute electricity demand or the carbon intensity of its power supply.
The IEC already publishes ISO/IEC 30134-2:2026, which defines PUE and its measurement, calculation and reporting. The new initiative addresses a broader question: how to put such facility-level indicators into an assessment of environmental and societal outcomes.
A large facility with an excellent PUE can still place substantial demand on a constrained grid. Two facilities with similar ratios can also have very different water requirements or opportunities to reuse heat. Efficiency within a building and sustainability across a region are related, but they measure different things.
The Four Parts of “Total Contribution”
The IEC announcement organizes the proposed framework around four connected dimensions:
- Energy and grid integration: demand response, support for grid resilience, and the integration of on-site or new renewable generation.
- Environmental and resource stewardship: energy consumption, water use, carbon emissions, circularity, advanced cooling and waste heat recovery.
- Economic and digital productivity: the relationship between resource consumption and contributions such as AI-enabled productivity, digital transformation and skilled employment.
- Societal and regional resilience: support for healthcare, communications, government and other critical services, together with regional development.
The difficult next step is making these dimensions comparable. A benefit to a national digital economy does not automatically resolve a local water shortage. A credible assessment needs to show where impacts occur, who receives the benefits and how claims are measured, rather than compressing every trade-off into a favorable headline score.
Grid Flexibility Has Practical Limits
In an earlier IEC contribution to ITU’s AI for Good platform, the organization described the forthcoming paper as a way to identify sustainability and safety gaps in standardization. It emphasized treating data centers as active participants in energy systems and examining cooling, water use, electronic waste and lifecycle emissions.
Demand response can make that relationship more useful: a facility changes electricity demand in coordination with grid conditions. However, the capacity to participate depends on operational constraints and who controls the workloads.
The IEA’s Powering Artificial Intelligence report says active management of computing loads remains rare. Many operators serve third parties and have limited visibility into or control over the work running inside their facilities. The IEA identifies standardized information exchange across computing customers, infrastructure providers and electricity networks as important to unlocking flexibility.
For the IEC initiative, this makes reporting actual capability more meaningful than simply labeling a campus “grid-supporting.” The useful evidence would show what demand can change, for how long and under which service commitments.
Liquid Cooling and Water Consumption Are Different Questions
The IEC’s announcement highlights liquid and immersion cooling, direct-current power distribution and power quality for AI infrastructure as areas for standards development. These technologies address different parts of the design problem: moving heat from dense equipment, reducing conversion losses and maintaining a suitable electrical supply.
Liquid cooling does not by itself establish that a facility consumes no water. The U.S. Department of Energy’s data center water-efficiency guidance illustrates a liquid-cooled system that ultimately transfers heat to a cooling tower. What circulates through a rack and how the facility rejects heat outdoors are separate design choices.
That distinction matters as AI infrastructure expands. A closed equipment loop can coexist with evaporative heat rejection elsewhere in the system. Reporting needs to define its boundaries clearly enough for readers to understand what a water-saving claim covers.
From a Roadmap to Measurable Outcomes
The importance of the IEC’s Clean Green Data Centres initiative lies in the connection it draws between computing infrastructure and the systems around it. Its next phase will determine how the proposed framework becomes consistent reporting and assurance that operators, utilities and policymakers can use.
A broader assessment will be most useful when it preserves clear measures of energy, water and emissions while adding evidence of grid services and public value. The report’s ambition is to make those relationships visible. Turning that ambition into comparable, verifiable outcomes is the work ahead.












