Aolani Slates Thousands of Advanced Nvidia Chips for New Southeast Asian Data Hubs

A significant expansion of artificial intelligence infrastructure is coming to Southeast Asia, as the Singapore-based cloud firm Aolani recently revealed plans to integrate 22,000 Nvidia Blackwell Ultra graphics processing units into its regional network. Scheduled for deployment in early 2027, these high-performance components will be distributed across two dedicated AI factories located in Malaysia and the Philippines. This initiative is part of a strategic partnership with Nvidia that utilizes a unique model for sharing revenue and supporting credit, ensuring that the rollout of hardware remains synchronized with actual market needs.

Once this project is completed, the total number of GPUs managed by Aolani will exceed 48,000, pushing the company's total processing capacity for AI tasks beyond the 100-megawatt mark. These facilities are designed to assist a wide range of organizations, including established corporations, new startups, and academic institutions, by providing the necessary power for complex AI training and inference operations. While the specific locations within Malaysia and the Philippines have not been named, the move represents a major milestone for the Philippines in particular, which is currently working toward a national goal of establishing 1.5 gigawatts of data center capacity by the year 2033.

This development occurs amidst a massive influx of capital into the Southeast Asian technology sector. Major global players such as Microsoft, Google, and Amazon Web Services have already committed upwards of $50 billion toward cloud and AI-focused infrastructure in the region. Other notable projects in the vicinity include a massive data center planned for Indonesia by Firmus, which could eventually house 170,000 GPUs, and a recently secured multi-billion dollar financing deal by Zankore for Nvidia-based hardware in Indonesia.

Despite the scale of Aolani’s announcement, several specifics regarding the 2027 rollout remain under wraps. The company has not yet detailed how the 22,000 units will be divided between the two nations, nor has it provided information on the total financial investment or when the new capacity will be accessible for commercial purchase. As the regional race for compute power intensifies, these new facilities will join existing Malaysian infrastructure projects, such as those involving SNS Network and Telekom Malaysia, to further solidify the area's position as a growing hub for global AI development.

---

Read original at TechRepublic AI.