Project Background
The energy optimization and intelligent control project jointly developed by DeepCtrl and Infinaxis for an AI data center in Malaysia recently passed acceptance testing against international Tier III (T3) standards. A significant milestone in Southeast Asia’s green AI computing infrastructure, the center deeply integrates DeepCtrl’s proprietary PhyAI technology and DeepSYS platform to create a next-generation AI-powered energy management and control system, advancing its power usage effectiveness (PUE) to a new level of industry performance.
A Green Foundation for Southeast Asia’s Computing Hub
AI data centers are a new form of shared computing infrastructure, providing the computing power, data services, and algorithm services required by AI applications. They play a crucial role in accelerating the intelligent transformation of industries and strengthening regional competitiveness. Located at a key computing hub in Malaysia, the newly commissioned Infinaxis AI data center spans approximately 17,000 square meters. Built to Tier III standards with an IT capacity of 15 MW, it is one of Southeast Asia’s leading examples of energy-efficient AI data centers.
From the outset, the project adopted “green, efficient, innovative, and recyclable” as its guiding design principles, embedding ESG considerations throughout its development. With computing density and energy consumption expected to rise after operations began, conventional control technologies could not meet Infinaxis’s demanding energy efficiency requirements. An AI-powered energy management and control platform with predictive capabilities and system-wide optimization therefore became essential to delivering green computing.
Bringing Expertise from China’s Leading Supercomputing Facilities to International AI Data Centers
Drawing on extensive experience from projects with leading clients—including the National Supercomputing Center in Wuxi, Tencent, and Chindata—and continuous advances in PhyAI (Physical AI) technology, DeepCtrl integrated computing and power resources to enable coordinated, system-wide energy management and dynamic optimization at the Infinaxis AI data center.
At the hardware level, DeepCtrl deployed intelligent devices that combine computing and control capabilities to collect data across the center’s cooling infrastructure, including cooling-source equipment, precision air conditioning, and fresh-air ventilation systems. Powered by PhyAI, the DeepSYS platform builds high-fidelity digital twin models that simulate the center’s electricity supply and demand and heat dissipation in real time.
Acting as an “AI commander,” the system proactively optimizes operations around the clock, dynamically coordinating critical equipment such as chilled-water plants, cooling towers, and precision air conditioning units. This enables intelligent decision-making and coordinated operation across both the central cooling plant and precision air conditioning units, while effectively smoothing power fluctuations caused by AI computing workloads to maintain power quality and system stability.
AI-Driven Unattended Operation and Intelligent Control Across Operating Scenarios
During testing by NV5 Mission Critical, an internationally recognized commissioning specialist, the system demonstrated the following:
Environmental parameters remained within service-level agreement (SLA) limits across simulated failure scenarios involving fan walls, chilled-water systems, cooling towers, and pumps.
PUE remained stable throughout testing, with precise temperature and humidity control at 25°C ± 2°C and 30–70% relative humidity.
Overall system operation met design standards and operating requirements, demonstrating outstanding resilience and intelligent control capabilities.
These results demonstrate that the Infinaxis AI data center has achieved fully automated, AI-powered energy management and control, enabling unattended operation, intelligent operations and maintenance, and dynamic optimization driven by AI predictions. While ensuring reliable computing services, the project represents a generational advance in intelligent control for AI data centers.
A representative of the Infinaxis AI data center stated that DeepCtrl’s AI-powered energy efficiency system directly enables intelligent, coordinated control across multiple types of equipment, providing a robust foundation for energy-efficient, high-density computing.
As demand for AI data center capacity expands beyond Singapore, Malaysia is rapidly emerging as a major computing hub in Southeast Asia. The successful acceptance of this project marks a significant acceleration in DeepCtrl’s global commercialization of solutions for AI data centers. Building on its pioneering PhyAI technology and extensive experience in the sector, the company is expanding across key regions, including Southeast Asia, the Middle East, and North America. Through the continued deployment of green AI data centers worldwide, DeepCtrl aims to drive energy efficiency innovation with AI and lead the transformation of energy performance across the global AI data center industry.
