Back

China Builds 800MW AI Data Center Campus in Inner Mongolia
SiTech AI Team3 min read

China Builds 800MW AI Data Center Campus in Inner Mongolia

Envision has completed the first 120MW of the Galaxy AI data center campus in under a year, at a cost researchers say is one-fifth of comparable US projects, while planning up to 800MW by 2027.

China is building an 800-megawatt AI data center campus in Inner Mongolia that could challenge the economics of America's Stargate project, according to a SemiAnalysis report. Envision has built the first 120 MW of the Galaxy AI data center campus in less than a year, at a cost the researchers put at one-fifth of comparable US projects after visiting the site in September. The completed campus is planned to provide up to 800MW of IT capacity by the end of 2027, close to the roughly 820MW of critical IT capacity at the Stargate campus in Abilene, Texas.

Power and construction

Galaxy is located in Ulanqab, Inner Mongolia, and is being developed across four phases for two unnamed internet companies. Envision is building its own power supply rather than relying only on the local grid, installing wind farms, battery storage, and dedicated power infrastructure alongside the computing facility. The company has installed more than 100GW of wind turbines and wants its wind farms to provide over 60% of the electricity needed by the campus's first two phases. Those farms, located about 80 kilometers apart, have a combined capacity of 420MW, above the 240MW load expected from the first two phases, although wind turbines rarely run at full output. Envision is adding four-hour battery storage to keep electricity flowing when the wind dies down.

Power hardware and cooling

In one 5MW hall, solid-state transformers take 10kV AC electricity and convert it directly into 800V DC for server racks. The design removes several intermediate components found in conventional data center power systems, including low voltage transformers, uninterruptible power supplies, and rectifiers. Fewer conversion stages can reduce equipment requirements and energy losses, although SemiAnalysis did not provide a measured efficiency improvement for Galaxy's pilot. The 5MW installation is only a small part of the campus, but it puts the technology into operation at a live facility, and Envision plans a larger deployment in the next construction phases. That move aligns with Nvidia's push for 800V DC distribution in future high-density AI systems. Galaxy also uses a cooling arrangement in which liquid-cooled and air-cooled servers share a single water loop, eliminating the need for conventional central chiller plants. Removing central chillers can simplify infrastructure, although the design depends on local climate conditions and the facility's ability to reject heat through its water loop.

Wider buildout and constraints

Galaxy is the first major project in the developer's wider Inner Mongolia data center plans, which SemiAnalysis says add up to 3GW, with the rest of the campus still a busy construction site. The surrounding Chahar High-tech Zone is also expanding rapidly, with nearly 30,000 construction workers spread across 22 data center projects, according to a local report cited by the researchers. Alibaba, Huawei, and ByteDance are among the firms expanding infrastructure in the region, which benefits from relatively cheap electricity, available land, and abundant wind resources. However, access to electricity and data center buildings does not automatically translate to available AI computing power. China's access to advanced accelerators remains a separate constraint, particularly as US export restrictions limit its ability to acquire Nvidia's most powerful chips.

Sources: Tom's Hardware

SiTech

SiTech — AI-powered web development

We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.