Envision has commissioned a GW-scale AI data center campus in China.

Envision has commissioned its Galaxy Campus, a gigawatt-scale AI data center in Ulanqab, China. The project marks the first flagship milestone of Mission Gobi.

The campus includes a 120,000-square-metre AI supercomputing facility designed to deliver one million PFLOPS of computing capacity. At full build-out, it will support up to one million AI accelerators, and the campus could scale beyond 2 gigawatts (GW).

Mission Gobi aims to develop 5 GW of green AI data center capacity in desert and arid regions worldwide by 2030. Envision Founder and CEO Lei Zhang first presented the initiative at VivaTech in Paris in June 2026.

**Energy-first design**

The campus power system integrates renewable generation, dedicated transmission infrastructure, large-scale energy storage, and intelligent controls. The approach is intended to maximise output and deliver stable power to AI clusters. Envision treats energy as part of the computing architecture rather than as a utility connection.

Ulanqab, in Inner Mongolia, offers abundant renewable resources and an annual average temperature of about 4.3°C. This natural "free cooling" reduces energy use and operating costs. The site is approximately 300 kilometres west of Beijing.

Envision is entering the AI infrastructure sector as a renewable-energy systems company, which allows it to design generation, grid, storage, and compute as one integrated system. Ricky Zheng, Head of Envision's AIDC business, said that power availability, network performance, and energy efficiency are the limiting factors for AI buildout.

Other international players are also prioritising energy-first solutions. US-based Crusoe AI Data Centers and Portugal's Start Campus, which is engineered for a 1.2-GW renewable-powered AI campus, are among them. If replicated, Mission Gobi would export a systems approach: building compute where renewable energy is plentiful and orchestrating delivery with innovative power systems.