Envision Commissions Initial Galaxy AI Data Center
Envision commissioned the initial phase of its Galaxy Campus AI data center in Ulanqab, Inner Mongolia, on August 6. Bloomberg reports the 120-megawatt phase has been allocated to two unnamed local technology companies, while Envision said the renewable-powered campus is designed to scale beyond 2 gigawatts.
Envision has commissioned the initial phase of its Galaxy Campus AI data center project in Ulanqab, Inner Mongolia. Bloomberg reports that the operating phase provides 120 megawatts of capacity and has been allocated to two unnamed local technology companies.
According to Envision's August 6 announcement, the campus is designed to scale beyond 2 gigawatts and is directly supplied by renewable energy. The company described the site as a next-generation AI infrastructure campus combining renewable generation, transmission infrastructure, energy storage, and its proprietary AI Power System.
Scale and power design
Envision said the campus's central AI supercomputing building covers 120,000 square meters. At full build-out, the company said the facility is designed for one million PFLOPS of AI compute and up to one million AI accelerators. Those are design targets rather than reported current operating capacity.
Ricky Zheng, general manager of Envision's AIDC business, said supporting one million accelerators requires an AI power system integrating renewable generation, dedicated transmission, and large-scale storage. He said this configuration can provide stable, low-cost green power and up to 10 times more compute output per square meter than conventional data centers.
Envision also called the building the world's largest single AI data center building. That comparative characterization has not been independently substantiated in the supplied reporting.
Ulanqab's infrastructure role
Bloomberg described the project as an unusual expansion by a renewable-energy company into AI infrastructure. The outlet noted that Ulanqab is roughly 350 kilometers northwest of Beijing and has substantial wind-power resources.
For AI infrastructure practitioners, the project illustrates the coupling of accelerator deployment with power-system engineering. At gigawatt scale, cluster capacity depends not only on GPU availability and networking, but also on generation interconnection, transmission, storage, and controls that can manage variable renewable supply. Companies pursuing comparable deployments often face the practical challenge of matching high-density compute utilization with power availability and network performance.
Key Points
- 1Envision has activated a 120-megawatt initial phase, while the Galaxy Campus is designed to exceed 2 gigawatts of capacity.
- 2The proposed full build-out targets one million accelerators and one million PFLOPS, figures Envision presents as facility design goals.
- 3Comparable gigawatt-scale AI projects can combine compute deployment with generation, transmission, storage, and power-management infrastructure.
Scoring Rationale
A 120-megawatt operating AI data center phase and a build-out designed to exceed 2 gigawatts are notable infrastructure developments for AI workloads. The story is especially relevant to practitioners tracking the convergence of accelerator clusters, grid capacity, and renewable-energy integration, although the largest capacity figures remain design targets.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
