# China Deploys First Commercial 800V DC Data Center Power System Envision's new architecture aims to boost AI facility efficiency using solid-state transformers. - **Source:** [Latest from TechRadar](https://www.techradar.com/feeds.xml) - **Published:** 2026-11-09 - **Topic:** [technology](https://brifing.si/c/technology) - **Original article:** - **On Brifing:** https://brifing.si/s/719f32f4-7302-4a08-97f8-95bea347804b ## Commercial Deployment Milestone Envision has launched the first commercial 800-volt direct-current power system at the Galaxy AI campus in Ulanqab, Inner Mongolia. Trials for the proprietary Dubhe controller and its integrated energy model officially commenced in August 2026. This installation marks a significant shift in how large-scale AI facilities manage their incoming electrical supply. ## Direct Power Conversion Mechanism The architecture utilizes solid-state transformers to convert 10-kilovolt medium-voltage alternating current directly into an 800-volt direct-current line. By bypassing conventional multi-stage conversion processes, the design minimizes energy losses and reduces the physical footprint of power equipment. This streamlined approach allows for higher-density server racks within the same floor area. ## Managing AI Workload Volatility AI workloads often cause rapid, substantial swings in electricity consumption, placing immense pressure on traditional infrastructure. Envision’s controller coordinates solid-state transformers and battery storage within milliseconds to stabilize supply. This rapid response capability ensures that power delivery remains consistent even when GPU demand fluctuates unexpectedly during intensive computing tasks. ## Material and Efficiency Gains Higher-voltage distribution networks significantly reduce the volume of copper required for cabling across large data centers. As computing installations expand globally, lowering material requirements becomes a critical factor for sustainable growth. While Envision reports improved efficiency, the company has not yet released independent performance figures to verify these specific gains. ## Integrated Battery Storage Role The system incorporates battery storage designed to provide immediate support during peak demand and temporary backup during grid interruptions. These batteries can maintain operations for several minutes, bridging the gap during power instability. This integration is essential for maintaining the uptime required by modern, high-density AI computing hardware environments. ## Ambitious Scaling Plans The Galaxy AI facility is projected to grow to 3 gigawatts, potentially supporting one million AI accelerators. Furthermore, Envision’s Mission Gobi initiative aims to establish 5 gigawatts of renewable-powered computing capacity by 2030. These desert-based projects rely on local renewable generation and energy storage to power future large-scale computing developments. ## Unverified Operational Metrics Despite the commercial launch, Envision has not provided independently verified measurements regarding installation costs or actual energy savings. The announcement lacks data on how the architecture performs under sustained full-load conditions. Consequently, the real-world impact of these claimed operational advantages remains subject to future validation by third-party industry experts. ## Future Industry Implications The deployment serves as a commercial milestone, yet the broader scalability of this 800V DC configuration awaits further evidence. Future performance across diverse hardware setups will determine if this model becomes a standard for AI data centers. Observers should monitor upcoming data from the Ulanqab site to assess long-term reliability. ## Sources covering this story - https://www.techradar.com/pro/has-china-cracked-the-ai-data-center-power-problem-new-800-volt-dc-power-system-is-a-world-first-offering-greater-efficiency-and-conversion-than-ever — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._