The initiative is bifurcated into two distinct projects: one in the state of Sarawak and another in Peninsular Malaysia. The structure of the investment reflects the specific strengths and roles of the three principal partners: SoftBank Group Corp, Grab Holdings Ltd, and PETROS. In Sarawak, the project is led by SoftBank, which is contributing the majority of the funds. The proposed initial investment for this region is US$1 billion, composed of US$800 million from SoftBank, US$100 million from Grab, and US$100 million from PETROS. SoftBank remains the majority shareholder in this entity, though the agreements include an option for Grab to increase its stake in the future, which would correspondingly reduce SoftBank’s holding.
Conversely, the project in Peninsular Malaysia is led by Grab as the founding shareholder. Grab is committing US$750 million of its own seed capital to this endeavor, with SoftBank contributing an additional US$200 million, bringing the total initial investment for the peninsula to US$850 million. This division of labor underscores a strategic alignment: PETROS brings local energy and infrastructure development experience, essential for powering data centers; SoftBank provides global AI and technology capabilities; and Grab contributes digital and AI expertise from its operations in the region.
The technological significance of this investment lies in its scale and the “five gigawatts” target. In the context of AI infrastructure, gigawatts measure the electrical power capacity available to sustain compute operations. As AI models grow in complexity and size, the demand for electricity and cooling becomes the primary bottleneck. By targeting multi-gigawatt capacity, the partners are aiming to create a hyperscale environment capable of supporting not just training, but massive inference workloads. The partnership intends to serve a broad range of clients, including governments, multinational companies, and small businesses, suggesting a public or semi-public utility model for these compute resources rather than a private, siloed system.
However, the deployment of this US$1.85 billion is not immediate or unconditional. The agreements specify that the funding will be deployed in stages, subject to the projects achieving key development and customer contract milestones, as well as necessary closing conditions. This staged approach mitigates financial risk for the investors, ensuring that capital is released only as the infrastructure physically materializes and commercial demand is verified. Furthermore, the current seed capital is intended to kickstart the process; further development is expected to be financed through third-party equity and non-recourse project financing at the project level. The partners have explicitly stated they will seek additional investors and “hyperscaler” customers to support the long-term expansion of the facilities.
The commercial and regulatory context of this project is further complicated by the need for international collaboration. The partners plan to work with global hyperscalers, indicating that the infrastructure will likely integrate with or serve the existing cloud networks of major global tech providers. This raises questions about data sovereignty and access, particularly as the infrastructure is located within Malaysia. For local businesses, the promise is access to high-performance AI compute that was previously only available through expensive global cloud services. For the developers and engineers in the region, this could catalyze a local AI ecosystem, reducing dependency on external infrastructure for training and deploying models.
Leadership of the project is centralized under Anthony Tan, Grab’s co-founder and chief executive, who is expected to oversee both projects. While Tan will continue to lead the ride-hailing and delivery group, he has indicated that dedicated management teams will be appointed for the day-to-day operations of the AI infrastructure. Masayoshi Son, chairman and CEO of SoftBank Group, framed the investment as a response to the “next wave of growth,” stating that Malaysia possesses the talent and ambition to lead in this sector. “Together with Grab, we will build the capacity needed to support Malaysia’s digital ambitions and accelerate AI innovation across Southeast Asia,” Son said.
Despite the substantial commitment, uncertainties remain. The projects are still subject to further development and financing, and the realization of the five-gigawatt target depends on sustained commercial demand and successful execution of the staged funding. For shareholders and investors, this represents a large commitment of capital and future spending. The success of this initiative will not be measured solely by the signing of the agreements, but by the subsequent securing of hyperscaler contracts and the physical completion of the data center facilities. As the projects move from agreement to construction, the next confirmed development will be the announcement of specific customer contracts and the commencement of physical infrastructure work in both Sarawak and the Peninsular region.
2 Comments