# Indosat Launches Zankore AI Infrastructure Platform in Indonesia > Source: > Published: 2026-08-07 13:00:09+00:00 # Indosat Launches Zankore AI Infrastructure Platform in Indonesia Indosat Ooredoo Hutchison launched Zankore on August 6 with Ooredoo Group, Nokia, and NVIDIA, targeting 1 GW of NVIDIA DSX AI Factory capacity in Indonesia. ANTARA reports that construction is scheduled to start in Batang, Central Java, in the first half of 2027, beginning with roughly 200 MW. The platform is designed to serve Asia-Pacific enterprises, governments, and hyperscalers. Indosat Ooredoo Hutchison launched **Zankore by Indosat** on August 6, 2026, with Ooredoo Group, Nokia, and NVIDIA. The partners are targeting **1 GW of NVIDIA DSX AI Factory capacity** in Indonesia, but that figure is a planned scale rather than installed capacity. Nokia's official announcement says Zankore is expected to deliver about **200 MW in the first half of 2027**, powered by NVIDIA GB300 NVL72 systems. Indonesia's ANTARA news agency reports that construction is scheduled to begin in Batang, Central Java, during the same period, starting at roughly 200 MW before expanding toward 1 GW within three years. Fortune reports that Ooredoo Group is Zankore's lead investor and a 49% shareholder, with an initial commitment of **$800 million over five years**. It says the shareholders' broader commitment is expected to reach multiple billions of dollars and that customers have already been secured for about 200 MW. Those figures describe commitments, forecasts, and contracted demand; they do not establish operational capacity today. ### The proposed infrastructure stack Nokia describes NVIDIA DSX as a reference architecture that coordinates accelerated computing, networking, power, cooling, and operations. Its announcement says the first 200 MW would use NVIDIA GB300 NVL72 systems and incorporate NVIDIA DSX MaxLPS for fleet-level power management. NVIDIA says MaxLPS is designed to enable up to 40% more compute within the same power budget; that is a vendor capability claim, not a measured result from Zankore. The partners assign Nokia responsibility for AI-native networking, NVIDIA for accelerated computing and software, Ooredoo Group for capital and regional scale, and Indosat for local infrastructure and execution. At the announced scale, delivery will depend on power availability, cooling, network fabric, cluster scheduling, and sustained accelerator utilization as much as on the selected GPU systems. ### Local-language models and access ANTARA reports that the project also includes **Sahabat-AI**, an open-source 70-billion-parameter language model supporting Bahasa Indonesia and regional languages including Javanese, Sundanese, Balinese, and Batak. The planned Token-as-a-Service layer is intended to provide AI-compute access to businesses, government agencies, developers, and the public. The announcements do not yet specify customer pricing, developer APIs, service-level terms, detailed data-residency controls, or a completion date for the full 1 GW target. For ML teams, the practical importance of Zankore will therefore depend on how much capacity becomes usable in 2027, which services are exposed, and what operating and governance terms accompany regional access. ## Key Points - 1Indosat, Ooredoo Group, Nokia, and NVIDIA launched Zankore on August 6, 2026, with a target of 1 GW of AI Factory capacity in Indonesia. - 2Nokia and ANTARA say the first roughly 200 MW is planned for the first half of 2027 in Batang, using NVIDIA GB300 NVL72 systems. - 3Fortune reports an initial $800 million Ooredoo commitment, while pricing, service terms, developer APIs, and a completion date for the full 1 GW remain undisclosed. ## Scoring Rationale The announced 1 GW target and first 200 MW deployment are material for Southeast Asian AI-compute availability, with direct relevance to sovereign and enterprise workloads. The score remains moderated because delivery begins in 2027 and the full capacity, service terms, and completion schedule remain targets rather than operating facts. ## Sources Primary source and supporting public references used for this report. Practice interview problems based on real data 1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with. [Try 250 free problems](/problems)