OpenAI Signs Multiyear Malaysia Data Centre Deal With Nvidia-Backed Firmus

OpenAI has signed a multiyear agreement with Nvidia-backed infrastructure company Firmus to secure computing capacity from two data centres in Malaysia, the companies confirmed on September 8, 2026, adding Southeast Asia to OpenAI's expanding global infrastructure footprint. The arrangement makes OpenAI a major customer of the Australian-founded data-centre operator as demand for computing resources continues to climb alongside increasingly capable and computationally intensive AI models.

The deal strengthens Malaysia's position within the global AI infrastructure race, a competition that has intensified over the past two years as operators search for markets offering available land, reliable energy supply, fibre connectivity and proximity to fast-growing digital economies. Johor state, in particular, has emerged as one of the region's largest new data-centre hubs, as infrastructure development spills beyond more geographically and energy-constrained markets such as Singapore, which has imposed stricter limits on new data-centre construction in recent years.

For OpenAI, the agreement reflects a deliberate strategy of geographic diversification in its compute sourcing. As the company simultaneously serves consumer products, enterprise API customers, autonomous agent workloads and its own model-training pipelines, relying on a small number of hyperscale facilities or a narrow set of chip suppliers introduces concentration risk that becomes increasingly costly to bear as usage scales. Building a more geographically distributed compute network also brings inference infrastructure physically closer to end users across new regions, which can materially improve response latency for real-time applications.

Firmus, backed by Nvidia, has positioned itself within a growing category of specialised data-centre operators built specifically to serve AI workloads rather than general-purpose cloud computing, a distinction that has become increasingly important as AI training and inference impose different power density, cooling and networking requirements than traditional enterprise data centres. The company's expansion into Malaysia reflects broader investor and operator confidence that Southeast Asia represents one of the most promising remaining regions for large-scale AI infrastructure development, given its combination of available land, growing renewable energy capacity and relatively favourable regulatory environment for data-centre investment.

The agreement is the latest in a series of infrastructure deals OpenAI has struck globally as it seeks to secure the enormous computing capacity required to train and serve its frontier models. Industry analysts note that such long-term capacity commitments have become a standard feature of the AI infrastructure landscape, effectively locking in compute supply years in advance in a market where demand has consistently outpaced the industry's ability to build new capacity.

For Malaysia and the broader Southeast Asian region, the OpenAI-Firmus arrangement adds to a growing list of major international AI infrastructure investments, positioning the region as an increasingly important node in the global AI supply chain — one that extends well beyond the traditional hubs of the United States, China and a handful of established data-centre markets in Europe and East Asia.