On November 2, Envision Energy's integrated 800-volt high-voltage direct current (HVDC) power supply system successfully achieved end-to-end connectivity at the Envision Ulanqab Xinghe Base. As the world's first integrated 800V HVDC power supply system to enter commercial operation in a large-scale AI data center, the system comprises the Envision EnOS controller, solid-state transformer (SST), direct current battery energy storage system (BESS), and DC/DC converter, pioneering a new power supply and distribution model for AI data centers.
The Envision Ulanqab Xinghe Base commenced production in August this year, with a planned capacity of 3 gigawatts (GW). Once completed, it will be capable of supporting one million GPUs in parallel and a computing power scale of one million petaflops, with a direct green power connection ratio exceeding 80%.
As AI computing power density continues to rise, traditional data center power supply and distribution architectures are facing multiple challenges, including efficiency, space occupation, and dynamic load response. The competitive focus of AI power supply is also shifting from the conversion efficiency of individual equipment to the stability, flexibility, and coordinated control capability of the entire power supply system under real loads.
To address these challenges, Envision has built a novel high-voltage direct current power distribution control and protection system centered on the "EnOS controller + solid-state transformer + DC battery energy storage system + DC/DC converter." Leveraging simulation modeling, stability analysis, and energy control strategy optimization, the EnOS controller coordinates the solid-state transformer and energy storage system to achieve dynamic matching between power supply and computing load, rapidly respond to instantaneous GPU power fluctuations, smooth power surges, and ensure stable operation of the power supply system.
In this commercial application, under real GPU training loads, three core capabilities were validated: first, the solid-state transformer directly converts 10 kV alternating current to 800V direct current and stably adapts to various operating conditions; second, the DC battery energy storage system supports both second-level power smoothing and minute-level backup power supply; third, the solid-state transformer and energy storage achieve millisecond-level coordinated interaction, absorbing in situ the severe power fluctuations caused by GPU training.
In 2026, "computing-power-electricity synergy" was written into the Government Work Report for the first time and incorporated into the outline of the 15th Five-Year Plan. In May, four departments including the National Development and Reform Commission and the National Energy Administration jointly issued the Action Plan for Promoting Two-Way Empowerment of Artificial Intelligence and Energy, explicitly stating "encourage computing facility deployment of grid-forming energy storage" and "improve policies for direct green power connection of computing facilities." According to statistics from the National Energy Administration, the national electricity consumption of computing centers reached 170 billion kilowatt-hours in 2025, and this figure is expected to reach 800 billion kilowatt-hours by 2030.
Huang Zhiyong, Senior Vice President of Envision Energy, stated that the commercial operation of the world's first integrated 800V solid-state transformer HVDC power supply system not only provides an engineering application case of this architecture in a large-scale AI data center, but also provides replicable engineering validation for the large-scale implementation of computing-power-electricity synergy.