China Builds Massive 800-Megawatt AI Data Center Campus to Rival US Stargate Project

>_ mC0de INSIGHTS / mC0de TechINSIGHTS ONLINE
CLOUD INFRASTRUCTURE

China Builds Massive 800-Megawatt AI Data Center Campus to Rival US Stargate Project

China is constructing a massive 800-megawatt AI data center campus in Inner Mongolia to challenge the economics of American facilities. The Galaxy project integrates wind farms, battery storage, and an innovative 800-volt DC power distribution system.

November 9, 2026  ·  mC0de Tech

A massive 800-megawatt artificial intelligence data center campus is taking shape in Inner Mongolia, designed to challenge the scale and economics of prominent American infrastructure efforts like the Stargate project. Developed by Envision in Ulanqab, the Galaxy AI campus has delivered its initial 120 megawatts of capacity in less than a year. Researchers note the construction cost sits at a fraction of comparable US projects. However, the facility's most notable innovation lies in how it secures and distributes massive amounts of electrical power, pairing dedicated renewable generation with novel solid-state transformation architecture.

800 MWPlanned IT Capacity
120 MWFirst Phase Delivered
5 GWMission Gobi Target
01 / THE BIG PICTURE

The Galaxy Campus Infrastructure

The Galaxy AI data center campus is located in Ulanqab, Inner Mongolia, and is being developed across four distinct phases for two unnamed internet companies. When fully completed by the end of 2027, the campus is planned to provide up to 800 megawatts of critical IT capacity. This scale puts the development close to the approximately 820 megawatts of critical capacity found at the Stargate campus in Abilene, Texas.

Envision has constructed the first 120 megawatts of the facility in less than a year at a cost estimated by researchers to be one-fifth of comparable American projects. The wider Chahar High-tech Zone is expanding rapidly alongside this development, featuring thousands of construction workers spread across multiple data center projects as major Chinese technology firms scale up their regional footprints.

Ulanqab Location

Situated in Inner Mongolia, taking advantage of abundant land, local climate conditions, and high wind resources.

Rapid Deployment

Envision built the initial 120 megawatts of the sprawling campus in less than a single year.

02 / INFRASTRUCTURE

Renewable Power Integration

AI data centers require enormous amounts of continuous electricity, prompting Envision to take an unusual route by building dedicated wind farms and battery storage alongside the computing facility. While US developers often turn to on-site gas turbines to work around grid constraints, Envision is utilizing nearby wind generation. The company wants its wind farms to provide over sixty percent of the electricity needed for the campus's first two phases.

The wind farms are located about 80 kilometers apart and possess a combined capacity of 420 megawatts, comfortably exceeding the expected 240-megawatt load of the initial phases. Because wind turbines experience fluctuating capacity factors between 35 and 40 percent, Envision is incorporating four-hour battery storage systems to maintain continuous electrical supply when wind generation drops.

03 / THE TECHNOLOGY

Direct Current Power Distribution

Inside a 5-megawatt operational hall at the facility, solid-state transformers take 10-kilovolt alternating current from the grid and convert it directly into 800-volt direct current for server racks. This architecture departs significantly from conventional data center power distribution, which typically requires multiple intermediate conversion stages involving low-voltage transformers and uninterruptible power supplies.

By removing several intermediate components, the design reduces equipment requirements and potential energy losses, while also allowing power equipment to occupy less physical space. The system includes battery storage and a specialized controller designed to respond within milliseconds to sudden swings in electrical demand caused by dense artificial intelligence workloads.

Solid-State Transformers

Convert 10kV grid electricity straight into an 800V DC supply, cutting out intermediate conversion steps.

Rapid Response

Advanced controllers manage transformers and batteries within milliseconds to handle volatile AI workload demands.

04 / BUSINESS IMPACT

Cooling and Broader Ambitions

The Galaxy campus utilizes a cooling arrangement where liquid-cooled and air-cooled servers share a single water loop, eliminating the need for conventional central chiller plants. This approach simplifies infrastructure as high-density servers generate increasing amounts of heat, though effectiveness relies on local climate conditions and heat rejection capabilities.

Envision's ambitions extend beyond this initial site, with plans for the Galaxy facility to eventually grow to three gigawatts to serve up to one million AI accelerators. The company also intends to introduce similar renewable-powered computing concepts through Mission Gobi, an initiative aiming for five gigawatts of desert-based capacity by 2030.

THE mC0de TAKEAWAY

Why It Matters

The development of the Galaxy campus highlights how infrastructure bottlenecks are driving distinct regional engineering solutions in the artificial intelligence sector. While American builders wrestle with gas turbine shortages and conventional grid limitations, the Inner Mongolia project demonstrates an aggressive integration of dedicated wind generation and direct-current distribution. However, significant questions remain regarding independent verification of efficiency gains and performance under sustained full loads. Furthermore, while physical power and building capacity are expanding rapidly, China's broader artificial intelligence ambitions continue to face separate constraints related to export restrictions on advanced semiconductor accelerators.

← Explore More Insights