South Korea focuses on refrigerant recycling to cut greenhouse gas emissions
July 28, 2026
Artificial intelligence is set to play a growing role in data center cooling management, with the aim of improving energy efficiency and reducing power consumption. This is the goal of a project launched by Daikin Industries and NTT DATA Japan, which will begin a Proof of Concept (PoC) in July 2026 to develop an intelligent platform capable of optimizing the operation of HVAC systems in data centers.
The initiative comes in response to the rapid expansion of generative AI, which is driving a sharp increase in servers dedicated to AI workloads. These systems consume significantly more power than conventional servers, generate greater amounts of heat, and experience highly variable workloads, making efficient cooling management increasingly critical.
The solution under development will use artificial intelligence algorithms to predict temperature changes inside servers and automatically adjust cooling equipment accordingly. One of its most innovative features is the ability to estimate server thermal conditions without accessing internal server data. Instead, the platform will rely on indirect indicators such as power consumption and outdoor temperature. This approach helps overcome potential cybersecurity and operational constraints that often limit access to server-level information in data centers.
The AI-generated insights will be used to coordinate air conditioning units, chillers, and liquid cooling systems as a single integrated platform, maintaining optimal operating conditions while minimizing energy waste and preventing overcooling.
Combining Expertise for Greater Efficiency
The collaboration brings together the complementary expertise of the two companies. NTT DATA will contribute its experience in data center operations, server performance data, and its Green DC Energy Management System (GDCEMS), while also providing the infrastructure for the pilot project. Daikin, meanwhile, will supply its HVAC, chiller, and liquid cooling control technologies, together with its AI-based automation solutions.
The demonstration project will take place at an NTT DATA data center in Japan between July 2026 and March 2027. During this period, the companies will assess the platform’s ability to improve energy efficiency, reduce electricity costs, and increase automation in cooling system management. If the results meet expectations, commercial deployment is planned for fiscal year 2027.
The project is built on three key pillars: an AI model capable of predicting server thermal behavior using available operational data, coordinated control of the entire cooling infrastructure to maximize overall system performance, and the integration of NTT DATA’s operational expertise with Daikin’s advanced cooling technologies.
Following the testing phase, the two companies plan to further refine the platform before initially introducing it to the Japanese data center market and later expanding internationally. The long-term objective is to help data center operators reduce energy consumption, lower operating costs, and cut greenhouse gas emissions.
Industrial refrigeration, commercial refrigeration, cold chain logistics, freezing technologies, and cryogenics will be among the key topics at REFRIGERA, the international trade exhibition for the refrigeration industry, which will take place at Bologna Exhibition Center from 10 to 12 November 2027.
