Asia’s Data Center Market Shifts to Disaggregated Ownership Model
Rising AI demands and massive capital requirements are pushing the region away from vertical integration toward a specialized ecosystem of land owners and infrastructure funds.
Asia’s data center landscape is undergoing a fundamental structural shift as the industry moves from vertical integration toward a disaggregated ownership model. This transition allows companies to decouple the ownership of real estate and power infrastructure from the operation of the facilities and the consumption of capacity.
Driven by the urgent need for high-density power and cooling to support AI and cloud computing, the market is increasingly relying on colocation, leasing, and joint ventures. This shift is evidenced by massive capital movements and strategic partnerships. For instance, Blackstone and CPP Investments acquired AirTrunk for AUD 24 billion, approximately USD 16.1 billion. Additionally, the regional data center pipeline reached 26.5 GW in the first half of 2026, according to Cushman & Wakefield.
The Rise of Specialized Partnerships
The move toward disaggregation is most visible in how land and infrastructure are being secured. Rather than building from the ground up, hyperscalers are partnering with local developers to accelerate deployment. In Malaysia, EcoWorld Malaysia sold land in Johor to Microsoft in deals totaling RM 402.3 million and RM 693.96 million. Similarly, YTL Power and JLand Group announced a partnership to develop a campus in Johor with a planned capacity between 1.3 GW and 1.5 GW, pairing infrastructure capital with land ownership.
International operators are also using joint ventures to navigate complex local markets. In 2025, Digital Realty formed a 50:50 joint venture with Bersama Digital Infrastructure Asia (BDIA) to develop data centers in Indonesia, leveraging local expertise to establish a footprint in the region.
Why the Model is Shifting
This evolution transforms data centers into a distinct, investable infrastructure asset class, making them attractive to pension funds and sovereign wealth funds. For the companies consuming the capacity, the primary benefit is financial flexibility. By shifting from a model of owning the entire stack to leasing or using colocation, enterprises can convert massive capital expenditures (CapEx) into flexible operating expenses (OpEx).
This financial pivot allows hyperscalers to scale their AI capacity rapidly. Because greenfield construction can take years, leasing existing or planned capacity removes the time-to-market barrier, enabling firms to deploy hardware as soon as the power and cooling are available.
Future Outlook
As the region continues to expand, the interdependence between land owners, infrastructure funds, and operators will likely deepen. The industry is now watching how these disaggregated partnerships handle the increasing power demands of next-generation AI chips, which require more specialized cooling than traditional cloud workloads. While the trend toward joint ventures is clear, the speed at which these partnerships can deliver the massive gigawatt-scale capacity required for the AI era remains the critical variable for the region's digital growth.