A new centre of gravity is emerging in the global artificial-intelligence race. Ooredoo Group has unveiled Zankore, an ambitious computing platform designed to give companies in Southeast Asia access to the processing power required for building and operating advanced AI systems.
Ooredoo will become Zankore’s founding shareholder and principal investor, taking a 49% stake and committing approximately $800 million toward the platform’s launch and early development.
Indonesia will host the initial deployment. The country offers a rapidly expanding digital economy, a population of more than 280 million people and growing demand for cloud services from banks, telecommunications providers, public institutions and technology companies.
Zankore is being developed alongside Indonesian telecommunications operator Indosat Ooredoo Hutchison, American chipmaker Nvidia and Finnish networking company Nokia. Each partner will contribute a different element of the platform.
Nvidia will provide accelerated-computing technology for operating AI models, while Nokia will bring the networking systems needed to connect large numbers of processors. Indosat will contribute its Indonesian telecommunications infrastructure and knowledge of the local market.
The platform aims to develop as much as one gigawatt of Nvidia DSX AI Factory capacity. Its first major phase is expected to provide approximately 200 megawatts during the first half of 2027.
Zankore will use Nvidia’s GB300 NVL72 computing systems, which are designed to operate demanding generative-AI and high-performance computing workloads. The infrastructure will support services that allow businesses to rent computing capacity instead of constructing their own expensive AI facilities.
The project will also incorporate Nvidia DSX MaxLPS, a system that dynamically manages electricity across groups of processors. According to the announcement, the technology could increase computing output by as much as 40% without requiring an equivalent increase in the available power budget.
This focus on energy efficiency is critical. AI facilities consume enormous amounts of electricity, and access to reliable power is becoming one of the biggest obstacles facing developers around the world. Extracting more work from the same electricity supply could improve operating costs while allowing more customers to use the platform.
Zankore describes itself as a “neocloud,” a newer category of cloud provider built specifically around artificial-intelligence processing. Unlike traditional cloud platforms that support many general computing activities, neocloud services concentrate on specialised processors, high-speed networks and software for AI development.
The platform could also help Indonesian companies keep sensitive information within the country while using advanced computing technology. Local AI infrastructure is particularly important for governments, financial institutions and telecommunications providers concerned about data control, response speed and regulatory compliance.
Ooredoo expects Southeast Asian demand for data-centre capacity to grow approximately three-and-a-half times by 2030, with AI workloads driving much of that expansion. After establishing operations in Indonesia, the company plans to introduce Zankore services in other regional markets.
Important challenges remain. Reaching one gigawatt will require immense investment in electricity generation, cooling, data-centre construction and high-speed connectivity. The project’s environmental impact and ability to secure enough customers will determine whether its proposed scale becomes commercially sustainable.
Nevertheless, the announcement sends a clear message: the global AI infrastructure race is no longer limited to the United States, China and Europe. Southeast Asia is preparing to become both a major consumer and a producer of artificial-intelligence computing power.
