Southeast Asia’s AI compute race is set to add 22,000 more Blackwell Ultra GPUs.
Singapore-founded AI cloud provider Aolani said Sept. 21 that it plans to deploy 22,000 Nvidia Blackwell Ultra GPUs across two AI factories in Malaysia and the Philippines in early 2027. The expansion would take Aolani’s total GPU deployment above 48,000 and its AI factory capacity past 100 MW.
For enterprises evaluating regional AI capacity, the plan would add a sizable pool of accelerated compute. In its strategic compute collaboration with Nvidia, Aolani did not disclose the exact sites, country-by-country GPU allocation, project cost, or customer availability timeline.
Aolani sets early 2027 Blackwell rollout
Aolani plans to use the systems for cloud services supporting AI training, inference, and other compute-intensive workloads for enterprises, startups, developers, and institutions across Southeast Asia.
Nvidia’s Blackwell Ultra platform is built for large-scale AI reasoning and inference workloads. Aolani’s arrangement with Nvidia also includes a revenue-sharing and credit-support model intended to align infrastructure deployment with customer demand. Financial terms were not disclosed.
Malaysia already hosts Nvidia-backed AI infrastructure. SNS Network has launched an AI Factory at Telekom Malaysia’s Klang Valley Data Center in Cyberjaya, while OpenAI has agreed to anchor two planned Firmus AI Factory sites in Malaysia. The projects add to the country’s growing AI-compute footprint.
The Philippines is still building out large-scale accelerated-compute infrastructure. Aolani says its planned facility would be among the country’s first large-scale AI infrastructure platforms. The government also released a $34.4 billion AI infrastructure plan in September targeting 1.5 GW of data center capacity by 2033.
Malaysia and the Philippines join a wider GPU buildout
Aolani says the deployment would bring its total GPU count above 48,000 and AI factory capacity above 100 MW. Those are projected totals tied to the expansion, not capacity already operating in Malaysia and the Philippines.
The project joins a wider wave of regional investment. More than $50 billion has been invested by AWS, Google, and Microsoft in AI-ready data center and cloud infrastructure across Southeast Asia, according to research from Singapore’s Economic Development Board and McKinsey.
Indonesia is drawing similar capital. Zankore secured up to $3.1 billion in financing in September for Nvidia-powered AI infrastructure, another sign of accelerated-computing investment spreading across Southeast Asia.
Key details of Aolani’s project remain undisclosed, including the GPU split between the two countries, exact facility locations, pricing, project cost, and customer commitments. Deployment is expected in early 2027, but Aolani has not said when customers will be able to buy the additional capacity.
For APAC IT leaders, the next milestones are the site locations, commercial availability, and how much of the 22,000-GPU deployment lands in each market.
Also read: The regional GPU buildout extends beyond these two markets, with Firmus planning a 360 MW Nvidia AI data center in Indonesia that could use as many as 170,000 GPUs.