The data center market globally has some unique characteristics.
It has largely three clients, AWS, Microsoft, and Google, all called as hyperscale clients. The question we get from our investors is if there are only three clients, why don't they buy and build their own data centers as opposed to leasing from third party operators? If you look at Australia, AWS has sixty percent of its capacity leased from third party operators. And if you look at Microsoft, it had hundred percent of its capacity leased from third party operators until last year, and by twenty twenty eight they'll have seventy percent of that capacity leased by third party operators. So the question is, why do third party operators still get majority of data center workloads across the globe and especially in Asia Pacific versus doing the self build themselves?
The first reason is core business focus and expertise.
Data center operators focus exclusively on building data centers and have built this expertise over the last twenty plus years. Hyperscalers on the other hand focus on technology and building data centers is a secondary priority. The second reason is access to land and power. Data center operators have deep expertise and knowledge in identifying land and securing sites.
And due to the scale of their operations, they have strategic relationships with power companies where they can build and accelerate any upgrades required. The third reason is capital and resource allocation. For a data center operator, capital is primarily spent on data center bills. On the other hand, hyperscalers have to balance capital spend across several different priorities including technology, cloud spending, AI, and data centers.
This results in slower decision making and bureaucracy across the hyperscalers.
And the fourth and final reason is strategic relationships. Data center operators have long standing established relationships with construction firms, suppliers, local councils where they're able to accelerate planning, development of pools, but hyperscalers on the other hand have started from scratch. In conclusion, hyperscalers will continue to self build in the long term. However, we see them procuring most of their capacity through data center operators over the next ten to fifteen years.