Coordinate or Stall: Trane Technologies and Eaton Make the Case for Integrated Design Amid Data Center Opposition

In the face of increasing state regulations and mounting power constraints, Trane Technologies and Eaton have announced a reference design collaboration that addresses the growing power and thermal demands on high-density Artificial Intelligence (AI) facilities through a unified infrastructure. Trane’s Continuum Rubin DSX provides the thermal-management architecture for high-density AI workloads, incorporating advanced cooling, free cooling, heat recovery, and air-cooled configurations to improve thermal efficiency and allow for more power to be directed toward compute. Eaton’s Beam Rubin DSX provides the grid-to-chip infrastructure layer, combining medium and low-voltage power distribution, modularity, advanced cooling, and digital capabilities in a scalable architecture. The collaborative design is included in both platforms; together, the offerings create a coordinated architecture aligned with the NVIDIA Omniverse DSX Blueprint.

The collaboration comes as policymakers and regulators increasingly focus on the grid and consumer impact of data center expansion. States are responding with a mix of large load tariffs, interconnection requirements, ratepayer protections, resource disclosure rules, and temporary development moratoriums. In the face of these regulatory uncertainties, operators need to demonstrate that new capacity can be added without shifting disproportionate infrastructure costs onto utilities and surrounding communities. New York is enforcing a 1-year pause on facilities consuming 50 Megawatts (MW) or more while developing standards around ratepayer and environmental impacts. ComEd, the major utility serving Chicago and northern Illinois, is tightening grid access by requiring new data centers to post a financial commitment tied to their requested power to filter out developments and protect consumers from the cost burden. Subsequently, technologies and designs that improve data center efficiency are becoming important not only for complete infrastructure performance, but also for enabling AI infrastructure to develop more efficiently in regions where restrictions on utility capacity and customer bills are expanding.

At its core, Eaton and Trane’s reference design collaboration integrates electrical and thermal systems into a pre-coordinated infrastructure architecture that can be repeatedly deployed across AI data center projects. In moving toward medium-voltage architectures, this collaborative design can support the higher power densities required for AI capacity and reduce some of the complexity and material requirements created by traditional low-voltage designs. By optimizing the design for efficiency and material use, the reference architecture can also lower copper and installation costs while enabling capacity to come online more quickly.

NVIDIA has extended this model across its DSX ecosystem. Schneider Electric is contributing a validated Vera Rubin power-and-cooling reference design plus ETAP-based power distribution simulation, Vertiv is providing simulation-ready power and cooling assets for pre-build validation, and Siemens is integrating electrical power, controls, and automation into a Vera Rubin DSX reference architecture. These collaborations bring NVIDIA’s compute requirements together with the specialized facility expertise needed to turn specifications into repeatable and operationally validated infrastructure.

However, the potential value of reference designs extends beyond system coordination into site strategy and stakeholder trust. A validated architecture can give utilities, regulators, developers, and local communities a common framework for understanding expected power demand, cooling requirements, efficiency, and infrastructure impact before a facility is built. This can improve credibility concerning site-selection and permitting discussions, reduce uncertainty around gird and resource requirements, and help operators communicate more transparently with communities that are apprehensive about energy use and local utility capacity. This approach would enable reference designs to become not only deployment blueprints, but also validation tools that help secure stakeholder confidence and reduce friction around new capacity.

To leverage reference designs as a means to build trust, validate infrastructure plans, and improve engagement with utilities, regulators, policymakers, suppliers, local communities, vendors, and data center operators must convey both the technical and broader operational benefits. Reference designs can translate technical coordination into two immediate operational advantages for data center operators:

• Clear expectations for energy demand, cooling performance, deployment timelines, equipment requirement, and expansion potential before construction begins.
• Reduced procurement complexity, limited redesign work, and condensed timeline for securing a site and bringing compute capacity online.

The immediate benefit is greater capacity, certainty, and time to value. Enhanced coordination can support high sustainable rack densities, reduce wasted energy and thermal risk, and improve the utilization of compute infrastructure. However, vendors should recognize that technical efficiency alone will not remove the regulatory and community constraints surrounding data center expansion:

• Reference designs should increasingly provide measurable evidence around power consumption, infrastructure efficiency, resilience, and resource utilization that operators can communicate to utilities, regulators, and communities.
• Water is incorporated into reference architectures at the cooling system level but should extend beyond to quantify water performance, including expected consumption and Water Usage Effectiveness (WUE) as water availability and reporting are critical for site selection and permitting.
• To reduce uncertainty during site evaluation and interconnection planning, reference designs can give utilities a more consistent view of a facility’s load profile, flexibility, and infrastructure requirements so projects can demonstrate predictable demand and operational responsiveness.
• This can strengthen project trust and developers can demonstrate that new capacity is being added responsibly rather than simply increasing demand on already-constrained grids.
• Credibility comes from independently measured results under real operating conditions. Third-party validation can turn projected performance into a documented record that strengthens confidence in future deployments.

The vendors that broaden the role of reference designs beyond the technical coordination of power, cooling, compute, networking, controls, and facility infrastructure into validation, stakeholder engagement, and site strategy will be better positioned to overcome the infrastructure and community constraints influencing the next phase of data center growth.