Quick Buzz Feed
The country has the abundant land and energy resources needed to power hyperscale data centers. Can it harness them effectively?
Artificial intelligence (AI) is rapidly becoming as dependent on energy as it is on technology. Training advanced AI models requires vast numbers of specialized processors operating continuously for weeks or months. Even after models have been trained, serving millions of AI queries each day creates a permanent and growing demand for electricity. The newest AI data centers already consume hundreds of megawatts of power, while some planned facilities are expected to consume as much electricity as medium-sized cities. Indonesia possesses several important advantages, including Southeast Asia’s largest economy, a population approaching 300 million people, a rapidly expanding digital economy, growing cloud adoption, and government policies recognizing AI as a national priority. Unlike Singapore, where land and energy are scarce, Indonesia has abundant industrial land, substantial natural gas resources, and one of the world's largest geothermal resource bases. The country currently has approximately 580 megawatts of operational AI data center capacity, with more than 1.3 gigawatts of additional capacity announced or under development. This expansion is driven by demand for AI, cloud computing, and digital services, with investment concentrated around Greater Jakarta, West Java, and Batam. Hyperscale data center developers are now negotiating electricity supply years in advance, seeking dedicated substations, transmission infrastructure, reserved generating capacity, and long-term power purchase agreements. For example, BDx secured commitments for approximately 1.2 gigawatts for future AI campuses in West Java. Energy analysts warn that reserve margins on the Java-Madura-Bali grid could fall below recommended levels by 2027, potentially creating political resistance if data centers are seen as increasing electricity costs or reducing reliability, making alternative power solutions more attractive.
Indonesia holds a strategic advantage with its large natural gas developments and one of the world’s largest undeveloped geothermal resource bases. These resources could form the foundation for a new generation of AI infrastructure, moving beyond their traditional roles as sources of electricity for the national grid or LNG exports. While Indonesia may not manufacture advanced AI processors, it can provide the necessary physical infrastructure and energy. Several major gas developments are underway or expanding, including the Masela LNG project (INPEX), Tangkulo gas development (Mubadala Energy), BP’s Tangguh LNG project expansion, and ENI’s North and South Hub developments in the Kutei Basin, all set to substantially increase Indonesia’s future natural gas production and electricity generation capacity. Traditionally, remote natural gas has been converted to LNG for export or used for fertilizer and petrochemical industries. However, a portion of this gas could be used to generate electricity for hyperscale AI data centers located near LNG facilities, providing reliable, dispatchable power with lower greenhouse gas emissions, suitable for energy-intensive AI infrastructure. This approach could foster a new domestic industry and minimize the need for new infrastructure. Co-locating AI data centers near LNG developments offers advantages like existing industrial land, utilities, and ports, allowing for integrated development of data centers and dedicated independent power producers. This model reduces community disruption, streamlines permitting, and cuts development costs and construction timelines. Furthermore, Indonesia’s vast undeveloped geothermal reserves, often located far from major demand centers, could also power AI data centers built adjacent to geothermal power plants. Geothermal energy offers continuous baseload electricity with high capacity factors, domestic energy security, minimal greenhouse gas emissions, and long-term price stability sought by hyperscalers for low-carbon operations. Unlike natural gas, geothermal has fewer competing commercial uses, making it ideal for large-scale AI infrastructure.
The primary drawback of situating AI infrastructure in remote gas fields or geothermal developments is latency, as many AI applications require near-instantaneous responses and thus need proximity to major population centers. However, AI model training has different operational needs, requiring months of continuous computation across thousands of GPUs and being largely unaffected by modest network delays. Similarly, high-performance computing, scientific simulations, genomic research, and other compute-intensive workloads can operate efficiently from remote locations where electricity is abundant and cost-effective. This suggests a dual-hub strategy: urban data centers could continue to concentrate around Greater Jakarta and other major cities for AI inference and cloud services, while remote AI training facilities could be developed alongside LNG projects and geothermal fields. These remote hubs would support AI training, high-performance computing, scientific research, and other energy-intensive computing tasks, collectively forming the backbone of Indonesia’s future AI infrastructure.
Indonesia faces regional competition for AI investment, with Malaysia positioning itself as a data center hub and Singapore continuing to attract digital infrastructure despite land and energy constraints. Indonesia's competitive edge lies in its combination of abundant industrial land, vast domestic energy resources, a growing digital economy, and the potential to develop dedicated captive power systems at a scale unmatched by many Southeast Asian countries. This presents a strategic opportunity for Indonesia to leverage its natural gas and geothermal resources to power high-value AI infrastructure, while also maintaining support for traditional industries such as LNG, fertilizer, petrochemicals, and domestic electricity generation. Geothermal and natural gas developments could become the foundation for new AI campuses, powered by integrated captive electricity systems developed alongside the energy projects themselves. Realizing this opportunity will necessitate coordinated investment beyond just electricity generation, including high-capacity transmission networks (like high-voltage direct current where economically viable), expanded domestic and international fiber-optic connectivity, reliable water resources (with recycling or desalination as needed), and modern digital infrastructure. Indonesia must also manage competing demands for natural gas, modernize its electricity system, and ensure a stable regulatory environment that fosters long-term private investment. The global race for AI leadership is fundamentally a competition for land, infrastructure, and energy. Countries capable of delivering on these essential factors will attract the next wave of hyperscale AI investment. Indonesia possesses many of the prerequisites, but its success hinges on its ability to translate these advantages into tangible commercial developments.