World’s Largest Single-Body AI Computing Supercomplex Launches at Envision Ulanqab Star River Base
On August 6, Envision Group announced the official commissioning of its Ulanqab Star River Base, marking the successful launch of the world’s largest single-body AI computing supercomplex. Covering 120,000 square meters, equivalent to 20 standard football fields, the mega facility adopts an ultra-high proportion of direct green power connection. With million-level parallel computing card capacity and millions of PetaFLOPS computing power, it has become the world’s top single-building AI data center with outstanding Token output capability, setting a new global density record for AI infrastructure. Empowered by professional AI data center liquid cooling technologies, the supercomplex achieves stable and efficient heat dissipation under ultra-high computing density.
As a landmark supercomputing hub built by Envision in the Gobi region, the Ulanqab Star River Base boasts a planned total capacity of over 2GW, aiming to grow into China’s largest AI computing park and Token production base. As the first flagship project of Envision’s “Gobi Mission”, the base precisely meets the domestic demand for million-scale GPU interconnection. Equipped with reliable data center liquid cooling CDU systems, it delivers a replicable and mature Chinese solution for large-scale domestic AI computing cluster construction.
China’s AI computing industry is accelerating toward hundred-thousand and million-level GPU interconnection to compete with global leading computing enterprises. The Envision Ulanqab Star River Base accommodates massive million-P computing resources in a single supercomplex, greatly enhancing the global competitiveness of domestic AI chips. Supporting such ultra-dense computing loads relies on high-performance high density server liquid CDU solutions, which resolve core heat dissipation challenges for fully deployed million-card clusters.
A world-class AI supercomplex relies not only on oversized construction scale, but also on a robust dedicated AI power system for stable power transmission and energy storage. Envision’s exclusive AI power system builds a reliable “energy expressway” and “power granary” for the super base, enabling 10 times higher computing output than traditional data centers of the same area. Cooperated with precision AI GPU liquid cooling CDU equipment, the entire facility maintains continuous, high-efficiency full-load operation and continuously refreshes global AI infrastructure density records.
Large-scale million-level computing clusters are extremely sensitive to network latency, where tiny delays will be exponentially amplified. As one of the eight core nodes of China’s “East Data, West Computing” strategy, Ulanqab enjoys innate low-latency network advantages to support large-scale real-time computing services. Benefiting from local wind and solar resources with a 67% regional green power ratio, the base realizes direct green power supply via self-built wind farms and dedicated power lines. Supported by energy-saving coolant liquid CDU for data center systems, it builds an ultra-solid green and low-carbon foundation for the world’s largest green AI supercomplex.
Zheng Zihao, General Manager of Envision AIDC, stated that domestic AI chips are achieving industrial breakthroughs through large-scale clustered development. Envision’s supercomplex provides ultra-large deployment space, sufficient green power supply and complete energy guarantee for domestic computing clusters. Adopting adaptive AI server air liquid cooling CDU tailored for AI business scenarios, the base strongly empowers the large-scale expansion of China’s independent computing industry.
Targeting the multiple demands of domestic AI chips for expanded deployment space, higher interconnection efficiency and stable green power supply, Envision Star River Base adopts the innovative “supercomplex + AI power system” integrated architecture. Professional layout of AI data center liquid cooling pipe networks ensures balanced coolant circulation and consistent thermal control across the entire mega computing cluster. This unique Chinese development model effectively breaks domestic computing bottlenecks and pioneers a new path for global green AI data center construction.
Officially launched at France’s VivaTech Summit in June 2026, Envision’s “Gobi Mission” targets the construction of 5GW global green AI computing centers in Gobi and desert regions by 2030. As the first practice of this global strategy, the Ulanqab Star River Base serves as a model benchmark for Chinese AI infrastructure going global. Matched with high-precision AI data center liquid cooling manifold components, the base realizes rack-level precise cooling and full-link thermal stability, fully adapting to the iterative upgrade of ultra-large-scale intelligent computing scenarios.
Focusing on the booming demand for large-scale domestic computing deployment among leading tech and AI enterprises, Envision is committed to creating exclusive “new energy + AI computing infrastructure” Chinese solutions. In the future, the company will promote replicable and mature Chinese AI data center construction and liquid cooling solutions to global regions with urgent computing demands and insufficient energy infrastructure, contributing to the high-quality development of global green intelligent computing.
As a professional global liquid cooling solution provider, EXTRCOOL offers full-set thermal management products including dedicated AI cooling CDUs, precision manifolds, liquid cooling pipelines and supporting systems. Our scenario-based solutions perfectly match ultra-high-density AI computing, hyperscale data center and HPC cooling demands, delivering efficient, energy-saving and stable full-link thermal protection for next-generation green AI infrastructure worldwide.
For detailed product specifications and customized liquid cooling solution consultations, please visit our official website: https://www.extrcool.com
