Gcore and TengriAI Sign MoU to Explore 100 MW AI Data Center in Kazakhstan

Gcore and TengriAI have signed a Memorandum of Understanding to explore the development of a GPU data center in Kazakhstan with a target capacity of up to 100 MW, to be phased between 2027 and 2029. The agreement, announced jointly from Luxembourg and Astana, aims to establish large-scale AI compute capacity within Central Asia, a region that currently has little domestic AI infrastructure and depends heavily on capacity located in Europe, the Gulf or East Asia — often outside its own regulatory jurisdiction.

The project would be structured so that TengriAI handles land, power, permitting and construction, while Gcore provides the AI cloud stack, operations and customer demand. The platform is the same one Gcore runs across its own GPU clouds as an NVIDIA Preferred Cloud Partner. Under the proposed model, customers could meter, bill and resell capacity, or run isolated workloads that never leave the facility. “Every market we work in is asking the same question: how do we run AI on our own terms?” said Andre Reitenbach, CEO of Gcore. “The project would put large-scale GPU infrastructure in the country, operated from the country, for customers across the region.”

Nurassyl Medeu, General Director of TengriAI, said the goal is to make Kazakhstan a leading AI token hub in Central Asia, adding that Gcore operates GPU clouds at scale and has done so with telecom operators. Gcore positions itself as one of few Europe-headquartered companies that operates its own GPU clouds, owns the global network and secures every deployment. “Software vendors have never operated a platform. Hyperscalers will not hand over the keys. We do both,” Reitenbach said.

Kazakhstan offers competitively priced energy, available land, lower cooling costs and a government that treats AI infrastructure as a national priority. Its location and the Astana International Financial Centre’s common-law framework make it a natural base for neighboring markets. For the region, the phased 100 MW build-out would represent a significant step toward sovereign AI capacity, reducing reliance on foreign compute and giving regional organizations the ability to run AI workloads under local regulatory oversight.