Trane Technologies has released major enhancements to its industry-first thermal management reference design for gigawatt-scale AI factories and unveiled two Trane Continuum Rubin DSX reference designs.
Engineered specifically to integrate with the NVIDIA Omniverse DSX Blueprint for AI data centres, the new system achieves a nearly 10 per cent improvement in overall thermal management performance compared to the original 1-gigawatt reference design announced in October.
This frees up 22 megawatts of cooling capacity that can be redirected to IT power, helping boost compute output without increasing total energy consumption.
Trane Technologies chief technology & sustainability officer, Mauro Atalla, said since the original launch the team has continued to work hand-in-hand with NVIDIA to push the boundaries of efficiency, scalability and intelligent simulation for gigawatt-scale AI infrastructure.
“These latest advancements represent a major step forward in helping enable the world’s most demanding AI and high-performance computing environments to scale sustainably, reliably and with accelerated digital insight,” he said.
NVIDIA vice president of AI infrastructure, Vladimir Troy, said efficiently scaling gigawatt-scale AI factories requires a fundamental shift in how to approach thermal management and data centre infrastructure simulation.
“Trane Technologies’ integration with the NVIDIA Omniverse DSX blueprint enables the creation of high-fidelity digital twins that help operators optimise cooling performance and maximise energy efficiency for the next generation of AI workloads,” Troy said.
Through continued collaboration with NVIDIA, including as an NVIDIA Partner Network member, Trane Technologies has also expanded its Trane Continuum Rubin DSX reference design portfolio with two additional high-efficiency solutions for large-scale AI deployments:
- 250-Megawatt Duplex Simplified System Design: This new design, available now, supports extended free cooling use and incorporates integrated heat recovery, which helps reduce system complexity and results in a 14% increase in thermal management system efficiency with 10% of the heat rejection load going to heat recovery.
- 1-Gigawatt AI Factory Mag-Bearing Air-Cooled System Architecture: This new design, available soon, features a streamlined air-cooled approach using 3-megawatt units to help reduce equipment count and eliminate the need for integrated waterside economizers. The architecture incorporates Trane’s latest magnetic bearing air-cooled chiller, providing oil-free operation, high efficiency and quieter, more efficient performance through Trane chiller plant controls.
Trane Technologies has also advanced its digital capabilities for adopting the Omniverse DSX Blueprint with more automated, scalable OpenUSD based SimReady assets.
Enhanced with structured metadata, these assets will support upcoming reference design updates, helping improve configurability, accuracy and readiness for high‑scale digital‑twin and AI‑driven simulation workflows.