Zankore has secured a senior term loan facility of up to US$3.1 billion to finance artificial-intelligence computing infrastructure in Indonesia and the wider region. Reports by Data Center Dynamics, DealStreetAsia and Reuters describe an initial 100MW deployment using Nvidia systems, with a longer-term ambition to scale the platform to 1GW. The financing moves the project beyond an infrastructure announcement, although it does not by itself prove that capacity is installed, contracted or earning revenue.
The new platform is linked to Indosat Ooredoo Hutchison and Ooredoo Group, with Nokia and Nvidia participating in the infrastructure effort. Citi acted as exclusive debt adviser, while Citi, ING, Natixis, Qatar National Bank and United Overseas Bank were named as senior mandated lead arrangers, underwriters and bookrunners. That lender group makes the deal a material capital-market signal for Southeast Asia’s emerging AI-compute sector.
The distinction between financing and delivery matters. GPU infrastructure requires more than access to processors: sites need dependable power, cooling, network capacity, construction progress and customers prepared to use expensive computing resources. Zankore’s stated scale creates a clear execution sequence to watch, beginning with the initial deployment and extending to the 200MW capacity previously targeted for the first half of 2027.
For Indonesian enterprises and AI developers, local capacity could reduce some of the latency, data-location and procurement friction involved in using overseas compute. It could also broaden the regional supplier base for advanced workloads. Those benefits remain conditional on service availability, pricing, contractual terms and the mix of customers the platform secures.
The facility also shows how capital-intensive the AI infrastructure race has become. A large debt package spreads funding across a banking syndicate, but lenders still depend on credible hardware delivery and future cash flows. The named amount should therefore be read as available financing under the facility, rather than evidence that the entire sum has already been spent or that the full 1GW ambition is funded.
Indonesia already has growing cloud and data-centre activity, but AI factories concentrate power and equipment requirements at another level. The next useful disclosures would include exact site milestones, power arrangements, phased GPU deliveries, committed customer capacity and the proportion of the facility drawn for each stage.
For the regional market, the financing is a stronger signal than another memorandum of understanding. It places banks, a telecom operator and technology suppliers around a defined capital structure. The commercial test is whether that structure converts into available compute that Southeast Asian customers can buy on workable terms.
Financing terms will shape the project as much as the headline amount. A syndicated facility can be drawn in stages and linked to conditions, procurement milestones or other lender protections. Public reporting has not set out the drawdown profile, pricing or security package, so comparisons with the capital cost of a fully operational 1GW platform would be premature.
The first practical checkpoint is therefore the initial 100MW phase. Evidence of energised halls, installed systems and contracted users would show how quickly the capital structure is translating into service. Until those disclosures arrive, the loan is best understood as a major vote of financing confidence with a demanding delivery programme still ahead.
Source note
As reported by multiple credible media reports, Zankore secured a senior term loan facility of up to US$3.1 billion for AI computing infrastructure. SEA Connect treats the amount as the maximum stated facility and separates financing, planned capacity and operational delivery.