Author: Corey Olson, Senior Principal, Systems Engineering, Motivair by Schneider Electric
Water has become one of the newest variables in data center planning—and one of the most regulated.
Across the United States, states are introducing legislation that governs how data centers use, report, and manage water. While early proposals focused on transparency, newer measures regulate cooling methods, requiring hydrologic impact assessments and, in some cases, encouraging or mandating cooling approaches that minimize water withdrawals. For colocation providers, hyperscalers, neocloud players, and the ecosystem of engineers, specifiers, and contractors that support them, cooling architecture has become part of the permitting and compliance strategy.
At the same time, the thermal requirements these facilities must manage continue to rise. NVIDIA’s Vera Rubin platform illustrates the scale of the shift, with rack densities moving up to 227 kW per rack. At those densities, operators need to consider how their cooling architecture affects water consumption and infrastructure integration.
Closed-loop liquid cooling data center strategies can address these challenges together. By circulating coolant through a closed system and removing heat closer to the source, operators can support increasingly dense compute while reducing dependence on consumptive water cooling.
Why water regulation is a data center design issue
For years, discussions around data center water use largely centered on sustainability goals and voluntary efficiency improvements. But a growing number of U.S. states now view water consumption as an infrastructure planning issue, introducing legislation that affects how facilities are designed, approved, and operated. In 2025 and 2026 alone, lawmakers introduced more than 60 bills addressing data center water use.
The individual bills vary but generally fall into three categories with direct implications for cooling system selection.
1. Closed-loop cooling requirements
Some states are moving beyond reporting requirements by regulating cooling technologies themselves. Proposed legislation in South Carolina, for example, would require data centers to achieve zero net water withdrawal and zero net wastewater discharge while prohibiting the use of groundwater or municipal water for cooling. Kansas has proposed restrictions on open-loop cooling systems that expose water vapor to the atmosphere and defines closed-loop cooling data center strategies as sealed systems with no fluid-to-atmosphere contact.
2. Water reporting and disclosure requirements
California, Utah, New Jersey, and others have advanced legislation requiring operators to disclose projected and actual water consumption, water sources, discharge practices, and, in some cases, indirect water use associated with electricity generation. Utah’s enacted legislation also requires advance notification to local water providers before construction and establishes penalties for noncompliance.
3. Hydrologic impact assessments
Minnesota, California, and other jurisdictions have considered or implemented requirements that examine factors such as watershed capacity, drought resilience, water supply, and cumulative regional demand. Policy research has also proposed more detailed watershed accounting based on withdrawal rates, consumptive use, seasonal demand, and drought curtailment triggers.
How closed-loop cooling data center designs can simplify compliance
As more water regulations emerge, cooling architecture becomes a practical way to reduce compliance complexity. Unlike cooling approaches that continually consume water through evaporation, closed-loop liquid cooling recirculates coolant through the system. That distinction becomes increasingly important as states expand requirements around water withdrawals, reporting, and environmental impact.
For operators planning new campuses or expanding existing facilities, closed-loop cooling can help address several regulatory challenges.
First, reducing consumptive water use can make ongoing reporting more manageable. States such as California and Utah are requiring operators to document water sources, withdrawals, treatment, discharge, and other water-use information. Reducing dependence on consumptive cooling can simplify the water profile operators need to monitor and manage over time.
Second, closed-loop architectures are aligned with the direction some state policies are heading. Proposed legislation in South Carolina and Kansas demonstrates that policymakers are beginning to look beyond how much water a facility consumes and toward how its cooling system actually operates.
Finally, closed-loop cooling can support future permitting strategies as hydrologic impact assessments become more common. Systems with low consumptive water requirements may be easier to accommodate within planning frameworks that consider watershed capacity, drought resilience, and competing regional demands.
What this means for existing data center facilities
For data center providers, the transition is complicated because much of today’s capacity was designed before high rack densities or emerging water regulations were anticipated. Those facilities now need to accommodate increasingly demanding workloads that can require changes extending from the chip and rack through coolant distribution and ultimately to facility heat rejection.
Luckily, liquid cooling does not necessarily require a complete facility rebuild. A phased or hybrid strategy can introduce liquid cooling where density demands it while maintaining existing air-cooled infrastructure elsewhere. Schneider Electric supports this approach through a comprehensive, end-to-end liquid cooling portfolio that spans heat capture through heat rejection.
At the IT and rack level, direct-to-chip accessories such as cold plates, manifolds, hoses, and related components transfer heat from high-power CPUs and GPUs into the technology cooling system. Coolant Distribution Units (CDUs) then manage coolant flow between the IT loop and facility infrastructure.
Where bringing facility water to the IT space isn’t practical, Heat Dissipation Units (HDUs) provide another transition path by rejecting heat from liquid-cooled computing systems into the surrounding air. For other retrofit scenarios, a Rear Door Heat Exchanger (RDHx) can provide rack-level cooling for high-density equipment without requiring access to facility water and necessitating a complete rebuild.
At the facility level, chillers and heat rejection complete the thermal path by moving heat out of the data center. Taken together, these technologies illustrate why liquid cooling should be considered as a system rather than as an individual rack-level component.
This flexibility matters particularly for colocation operators. Different customers may arrive with different server platforms, rack densities, cooling requirements, and deployment schedules. Rather than choosing between an entirely air-cooled or liquid-cooled facility, operators can build an architecture that supports a mix of technologies and evolves with tenant requirements.
Building a compliance-ready liquid cooling strategy
Whether developing a new campus or modernizing an existing facility, operators should consider several things early in the design process:
- Understand the regulatory landscape. Evaluate both current requirements and proposed legislation in jurisdictions where capacity is planned.
- Consider future reporting requirements. Determine how the cooling architecture affects the water sources, withdrawals, consumption, and discharge information the facility may need to track.
- Design for flexibility. Consider how CDUs, HDUs, rear-door heat exchangers (RDHx), direct-to-chip technologies, and hybrid architectures could support phased adoption as tenant and workload requirements change.
- Plan the entire thermal path. Cooling does not stop at the chip or rack. Consider coolant distribution, facility piping, chillers and heat rejection, controls, monitoring, and maintenance as part of one engineered system.
- Coordinate power and cooling. Ultra-high-density racks require electrical and thermal infrastructure to scale together. Designing them independently can introduce integration constraints later.
- Evaluate deployment models. For new capacity, options such as a Prefabricated Modular IT Pod can provide an alternative to conventional construction while supporting hybrid cooling architectures.
For data center providers, that end-to-end view can help shift the conversation from Which cooling product do we need? to a more important question: What architecture will allow this facility to support higher densities while remaining adaptable to changing operational and regulatory requirements?
Closed-loop cooling for data centers is an infrastructure strategy
As states introduce new reporting requirements, permitting criteria, and cooling regulations, water use is an important consideration in how data centers are planned, built, operated, and expanded.
At the same time, NVIDIA Vera Rubin and other next-generation accelerated computing platforms are pushing rack densities higher than ever. These two trends are converging; data center providers need to remove substantially more heat while also becoming more deliberate about how their facilities use water.
Closed-loop liquid cooling provides a practical response to both pressures. It can support high-density computing while reducing dependence on consumptive water cooling and helping operators prepare for increasingly complex reporting and permitting requirements.
The strongest strategy goes beyond any single cooling component. From direct-to-chip accessories and CDUs to HDUs, rear-door heat exchangers (RDHx), chillers and heat rejection, and prefabricated infrastructure, the thermal system needs to be designed as an integrated whole.
Schneider Electric brings those elements together through a comprehensive, end-to-end liquid cooling portfolio designed to support the transition from today’s data center environments to the higher-density infrastructure now arriving.
Explore Schneider Electric’s Liquid Cooling solutions and guide to Liquid Cooling Architectures to learn how to plan a cooling architecture that can support higher-density workloads, existing-facility constraints, and evolving water requirements.
Add a comment