BEIJING, July 2, 2026 - Chindata today announced the commencement of commercial operation of one of the industry's first Solid-State Transformer (SST)-based intelligent DC power systems at Chindata's Sangyuan Cloud Computing Campus near Beijing.

The project marks an important milestone in the evolution of power distribution for hyperscale AI infrastructure, advancing data center electrical architecture from conventional power-frequency systems to high-frequency power conversion, and from AC distribution to intelligent DC power delivery. The project further strengthens Chindata's position in AI infrastructure and provides a practical foundation for the next generation of AI data centers, enabling higher efficiency, greater power density, and improved system reliability.

As AI workloads continue to increase rack power density, conventional power distribution architectures are becoming increasingly constrained by their large physical footprint, complex electrical topology, and limited ability to respond to rapidly fluctuating GPU workloads. Purpose-built for AI computing, the SST integrates voltage transformation, electrical isolation, and power quality management into a single platform, delivering a simplified power architecture with higher power density, improved space utilization, and intelligent power management.

The project was jointly developed by Chindata, Meituan, HEC Group, and Delta Electronics. It introduces an innovative 10kV triangular SST architecture that enables direct one-step conversion from 10kV utility AC power to 800V DC, eliminating multiple intermediate conversion stages found in conventional power systems.

The system delivers up to 98.5% power conversion efficiency. Compared with conventional architectures, it reduces the electrical infrastructure footprint by more than 50%, effectively doubling space utilization. Its integrated energy storage design supports both uninterrupted backup power and energy optimization through peak-valley electricity management. Combined with an intelligent energy management system, the platform dynamically balances fluctuating AI workloads, further enhancing system stability and operational resilience.

Throughout the project, Chindata played a leading role in advancing the application of SST technology for hyperscale AI infrastructure. Drawing on real-world data center deployment requirements, the company established a comprehensive technical specification framework for SST-based power systems and led the development of an end-to-end integrated system architecture. Following 34 rigorous performance and safety validation tests, the project successfully entered commercial operation, providing an important reference for broader SST adoption across the data center industry.

Looking ahead, Chindata will continue expanding the deployment of SST-based intelligent DC power systems across its AI infrastructure portfolio. The company will also work closely with equipment manufacturers to accelerate the development and localization of key SST components, while sharing project experience and technical best practices with industry partners to help advance the next generation of AI data center power infrastructure.