GMI Cloud, an AI cloud provider, has partnered with AI transformation firm Magna AI to launch a sovereign AI factory initiative. Under the agreement announced on June 11, the two companies will collaborate on designing and deploying a global network of sovereign AI factories for governments, enterprises, and national AI programs.
Separately, Singtel and WEKA are collaborating to build sovereign AI infrastructure across ASEAN and Singapore. – GMI Cloud. The initial deployments are scheduled for Malaysia, Belgium, and Romania, after which the companies intend to expand throughout the Middle East and Africa.
The “AI factories” will be constructed based on Nvidia’s Vera Rubin NVL72 architecture. “Governments are increasingly viewing AI infrastructure as a critical national asset,” said Alex Yeh, CEO of GMI Cloud. Our mission is to support countries in designing, constructing, and running their own sovereign AI platforms to drive innovation, ensure security, and strengthen long-term competitiveness.
We believe that sovereign AI will emerge as a core foundation for future economic growth. Sovereign AI is emerging as a cornerstone of future economic and technological competitiveness,” said Dr. Moataz BiAali, CEO of Magna AI.
“In partnership with GMI Cloud, we are helping nations accelerate AI adoption by providing secure, resilient, and scalable sovereign AI infrastructure.” The companies have started collaborative planning efforts focused on site assessment, infrastructure design, regulatory compliance, and rollout phases, with construction activities slated to commence later this year. Earlier this year, GMI Cloud unveiled a $12 billion sovereign AI infrastructure project in Japan, under which it will establish an AI Factory in Kagoshima for large-scale physical AI applications in collaboration with Wistron.
Separately, in Singapore, Singtel Digital InfraCo’s AI cloud unit RE:AI has partnered with WEKA to develop and provide sovereign AI infrastructure solutions across Singapore and the wider ASEAN region. The collaboration will integrate RE:AI’s AI cloud infrastructure and network orchestration capabilities with WEKA’s NeuralMesh, a high-performance software-defined storage platform.
The two aim to eliminate storage bottlenecks and minimize idle GPU time. Singtel will also incorporate NeuralMesh into its GPUaaS platform, which caters to governments and enterprises. Bill Chang, CEO of Singtel Digital InfraCo, said, “As sovereign AI adoption accelerates across ASEAN, governments and enterprises need infrastructure that is secure, compliant, and engineered for high-performance at scale.” Today’s AI systems have moved well beyond being constrained solely by computational resources.
