OpenAI closed a multi-year agreement this Tuesday (8) with Australian Firmus to contract dedicated computing capacity in two new artificial intelligence data centers in Malaysia, expanding the presence of advanced AI infrastructure in Southeast Asia. The partnership was announced by Firmus itself and turns the ChatGPT developer into an anchor client for the projects.

The amount of capacity reserved specifically by OpenAI and the financial value of the contract were not disclosed. With the new agreement, however, Firmus says it has surpassed 900 megawatts of contracted capacity among all its clients.

The centers in Malaysia are part of a network of seven “AI factories” planned by the company in four Asia-Pacific countries. Two facilities are already operational, in Australia and Singapore, while another five are under development and are expected to enter service over the next 24 months.

The infrastructure tied to the expansion will use systems based on the next generation of accelerators, Nvidia Vera Rubin, which Firmus plans to deploy at scale in the region. Nvidia is also an investor in the company, alongside names such as Coatue, Jane Street and Blackstone funds.

Malaysia gains clout in global AI infrastructure

The choice of Malaysia comes amid an acceleration in data center investments in the country. Official data from the Malaysian Investment Development Authority (MIDA) show that data center and cloud computing projects totaled RM 95.8 billion in approved investments in the first half of 2026, nearly 44% of all investment approved in the period.

The agency says the country is already among the ten leading global destinations for data center projects, according to UNCTAD data, and has been prioritizing facilities geared to AI workloads. The advance also led the government to increase control over new projects, requiring proven availability of energy and water and environmental criteria before approval.

For OpenAI, the contracted capacity in Malaysia adds to the race for infrastructure needed to train and operate ever-larger models. Sachin Katti, the company’s vice president of compute strategy, said the new centers should help meet growing demand for OpenAI’s products in the region and in other markets.

The move also deepens Southeast Asia’s transformation into a region that not only consumes AI services but is beginning to host a larger share of the physical infrastructure that supports these systems. Firmus is already developing a facility of 360 MW in Batam, Indonesia, planned in partnership with Nvidia, in addition to its existing operation in Singapore.

More from Radar