Edged US Builds Waterless, High-Density AI Data Center Campuses at Scale (David Chernicoff)
Data Center Frontier · David Chernicoff
An article profiling Edged US's standardized AI campus model built around closed-loop waterless cooling from sister company ThermalWorks, with design targets for the Aurora and Irving projects.
On this page
Overview
This piece, written by David Chernicoff for Data Center Frontier, is a news report profiling Edged US's standardized campus model for AI data centers. The model is built around closed-loop waterless cooling supplied by sister company ThermalWorks. The figures are design targets as reported by the company and the article, not independently verified operating data.
The portfolio-wide design PUE target is about 1.15, and a water usage effectiveness (WUE) of 0.00 is stated because closed-loop waterless cooling is used. The article also reports standardized Tier III design and integrated monitoring across multi-building campuses. Standardization is presented as an approach that makes the design repeatable and allows comparable targets to be used across campuses.
Project parameters are given for two facilities. The 72 MW facility in Aurora, Illinois uses liquid-to-chip cooling for rack densities above 200 kW. The 24 MW expansion in Irving, Texas has air-cooled densities above 120 kW per rack and liquid-to-chip cooling up to 400 kW. These two examples show that air and liquid cooling can be used together, in different density bands, within the same campus model.
The article also discusses why waterless designs reduce permitting friction: water use is a leading cause of community opposition, especially in the Southwest, so waterless cooling is a permitting advantage. As a limit, this is a company profile; details such as the PUE cost of closed-loop waterless cooling and its climate dependence are not in the record. The figures can be used as benchmark values in AI campus templates.
Key points
- Portfolio-wide design PUE target of about 1.15 with stated WUE of 0.00 using closed-loop waterless cooling.
- Aurora, IL: 72 MW facility with liquid-to-chip cooling for rack densities above 200 kW.
- Irving, TX: 24 MW expansion with air-cooled densities above 120 kW per rack and liquid-to-chip up to 400 kW.
- Standardized Tier III design and integrated monitoring are used across multi-building campuses.
- Water use is a leading cause of community opposition, especially in the Southwest, making waterless cooling a permitting advantage.
- Cooling is based on ThermalWorks' closed-loop waterless solution.
Using it in design
- Provide a waterless closed-loop heat rejection option with WUE of about 0 and its PUE trade-off.
- Flag water-stressed sites where evaporative cooling may face permitting risk.
- Use these values as benchmarks in AI campus templates: PUE 1.15 and racks above 200 kW on direct liquid cooling.
- Report WUE according to ISO/IEC 30134-9.
Monthly newsletter · free
Keep your reading list current
Once a month: new library sources, changes to Turkish regulations and upcoming data centre events.
SoSo CAD
Check cooling capacity alongside your design
Review zone heat loads, available capacity and the capacity balance after the largest unit is lost in SoSo CAD.
Request a demo
Who should read it
Engineers and developers designing AI campuses and choosing sites and cooling in water-constrained regions.
Related standards
- ISO/IEC 30134-9 (WUE)
Publication details
- Authors
- David Chernicoff
- Publisher
- Data Center Frontier
- Year
- 2026
- Document language
- English
- Source type
- Expert / community
- Access
- Free to read
- Licence
- Copyright retained; citation only — the publisher’s terms apply to the original document
Citation
David Chernicoff (2026). Edged US Builds Waterless, High-Density AI Data Center Campuses at Scale (David Chernicoff). Data Center Frontier. https://www.datacenterfrontier.com/design/article/55352525/edged-us-builds-waterless-high-density-ai-data-center-campuses-at-scale
Events on this subject
- SC26 – International Conference for High Performance Computing, Networking, Storage and Analysis
- NVIDIA GTC
- Data Centre World Frankfurt
All content and rights belong to the respective authors and publishers. SoSo CAD provides only the citation, its own summary and a link to the source, and is not responsible for the sources' content.