Book Demo
Book Demo

At its Frankfurt data center site in Germany, Telehouse Deutschland GmbH implemented AI-based cooling optimization using the etaONE® platform from etalytics. The project focused on optimizing the existing cooling system through a digital twin and data-driven setpoint optimization, without additional hardware and while maintaining full operational reliability.
Telehouse Germany and etalytics have been working together on a pilot project to optimize cooling operations at the Frankfurt data center site. The project leverages the AI-based energy management platform etaONE®, supported by a digital twin of the cooling system. The objective is to achieve measurable energy savings without additional hardware and full operational reliability.
The overarching goal is to enable data-driven operational optimization, reducing the energy required for cooling while maintaining stable conditions for the data center. At the start of the test phase, autonomous control was intentionally deactivated to validate all system functions under controlled conditions. etaONE® is already generating optimized setpoints in live operation. These are currently implemented manually by the operations team during structured test runs. Various operating modes were analyzed, including chiller mixed operation, free cooling, and part-load conditions. External influences such as temperature profiles and time-dependent load fluctuations were also factored into the analysis.
All scenarios showed clear improvements in efficiency. Measured energy savings ranged from 13 to 30 percent in cooling generation, depending on the operating conditions. The greatest reductions were observed during free cooling operations.
These results were achieved under real-world conditions but covered only a limited period. Based on the available data, both project partners expect double-digit percentage savings in cooling operations over a full operational year. System stability was maintained throughout all tests.
We have an increasing number of AI customers, and as a logical consequence we now use AI to reduce our energy consumption and improve substantially our energy efficiency.
