AI infrastructure platform Zankore has signed a senior term loan facility of up to $3.1 billion to finance the acquisition and deployment of Nvidia-powered GPU and cloud-computing infrastructure, as it expands AI capacity in Indonesia and across Southeast Asia.
The financing will support Zankore’s rollout of advanced Nvidia GPU infrastructure and comes shortly after the platform was launched in partnership with Indosat Ooredoo Hutchison, Nokia and Nvidia.
Qatari telecom group Ooredoo is Zankore’s founding shareholder and lead investor, holding a 49% stake in the venture.
Financing targets large-scale AI compute buildout
Zankore said proceeds from the debt facility will be used to acquire and deploy advanced Nvidia GPU infrastructure.
The platform is initially developing 100MW of Nvidia AI infrastructure and is positioning itself as a regional provider of cloud and AI compute services for AI-native companies in Indonesia and Southeast Asia.
The company has set significantly larger ambitions beyond the initial deployment.
At its launch on August 6, Zankore said it was targeting 1GW of Nvidia DSX AI Factory capacity and approximately 200MW of initial capacity in the first half of 2027.
The newly announced financing provides substantial capital backing for that expansion, although Zankore did not disclose how much of the facility has already been drawn, the maturity, pricing or deployment schedule tied to individual tranches.
Ooredoo moves deeper into Southeast Asian AI infrastructure
Ooredoo’s 49% stake gives the Qatar-based telecom group a major position in Zankore and represents a broader push into AI infrastructure beyond its traditional connectivity markets.
The investment gives Ooredoo exposure to one of Southeast Asia’s fastest-growing digital markets while extending its infrastructure strategy into GPU-based AI computing.
For Zankore, the partnership combines telecom infrastructure, cloud and networking expertise with Nvidia-based accelerated computing.
Indosat Ooredoo Hutchison provides a strong local market position in Indonesia, while Nokia contributes network and infrastructure expertise.
Indonesia positioned as initial expansion market
Indonesia is central to Zankore’s initial strategy.
The country’s large digital economy, growing cloud demand and expanding AI ecosystem are creating increased requirements for local computing capacity.
Zankore is seeking to serve companies that require access to high-performance GPU infrastructure without building and operating their own AI data centres.
The neocloud model typically focuses on providing accelerated computing capacity and specialised infrastructure for AI workloads, particularly training and inference applications that require large clusters of GPUs.
Zankore is positioning that model around Southeast Asian demand rather than relying exclusively on global hyperscale cloud platforms.
Initial 100MW could scale toward 1GW
The difference between Zankore’s initial 100MW buildout and its stated 1GW target illustrates the scale of its longer-term ambitions.
At launch, the company also outlined plans for around 200MW of initial capacity during the first half of 2027.
Reaching 1GW would place Zankore among significantly larger AI infrastructure developments, but that figure remains a target rather than deployed capacity.
The pace of expansion will depend on several factors, including financing drawdowns, data centre construction, power availability, GPU supply and customer demand.
No specific sites, customer commitments or commissioning dates beyond the previously stated capacity targets were disclosed in the latest announcement.
Citi advises on debt and equity structure
Citi acted as exclusive debt adviser on the senior term loan facility.
The bank had also served as exclusive financial adviser to Indosat on the creation of Zankore and its equity fundraising earlier in August.
Its involvement across both the debt and equity components indicates that the platform has been structured around a significant capital programme from the outset.
AI infrastructure projects typically require large upfront investment because of the combined cost of GPU hardware, power systems, cooling, networking and data centre capacity.
Source: MEA Tech Watch Press Reporter
