WHILE OTHER FACTORS SUCH AS DIGITISATION OF RURAL AREAS AND CONTINUING 5G ROLLOUTS ARE SPURRING SOME DEMAND , THE GROWING USE OF CLOUD REMAINS THE TOP FACTOR FOR GROWTH IN THE SECTOR .
DATA CENTRE PREDICTIONS
WHILE OTHER FACTORS SUCH AS DIGITISATION OF RURAL AREAS AND CONTINUING 5G ROLLOUTS ARE SPURRING SOME DEMAND , THE GROWING USE OF CLOUD REMAINS THE TOP FACTOR FOR GROWTH IN THE SECTOR .
are spurring some demand , the growing use of cloud remains the top factor for growth in the sector . By 2027 , almost 50 % of revenue in the space will be captured by hyperscalers – cloud providers such as Amazon AWS , Google Cloud , Microsoft Azure , Oracle and other large data users such as Meta . Many of these platforms are increasingly doing site sourcing and development in-house , leading to longerterm questions on the opportunities that developers and landlords will see from this significant tenant base over the next decade . In the meantime , the demand for space from these users outstrips both available colocation space as well as their self-perform programmes .
Market interest in Artificial Intelligence ( AI ) and Machine Learning ( ML ) deployments is now adding to and augmenting the demand for data centre space . With all the major hyperscalers , as well as independent enterprise tech and start-up users , rushing into the space , demand for compute and data storage capabilities has skyrocketed dramatically as a result . gigawatts of capacity for the purposes of their AI platforms across North America in 2023 . This demand will be bifurcated between training and inference facilities . While inference facilities will likely continue to be closely located to cloud regions , training facilities will likely not have as stringent latency requirements – opening up more markets with available power for evaluation . AI facilities will often require higher rack densities and therefore more intensive liquid cooling technologies .
In terms of sales of facilities , as many owners in the space are long-term operators with committed capital or , increasingly , hyperscalers through selfperformance operations , transactions in the sector are limited .
While transaction volumes have tempered somewhat from the past several years , sales of existing assets that do occur are most often part of a portfolio disposition , merger or acquisition .
In other cases , certain operators have sold or diluted their control of assets to help fund the development of new data centres . Given the relatively high cost of development , operators require significant capital to redeploy and expand their pipelines , particularly in the current high-demand environment .
A significant amount of capital activity in the space goes towards purchasing land , which has faced rising costs in recent years . As local land markets have reacted quickly to the arrival of new data centres , construction of these assets is often capital-intensive , with development budgets often running into the billions of dollars for data centres sized for hyperscale use .
As the demand for data centres continues to rise , fuelled by booming interest in AI and substantial funding , the sector grapples with highly constrained power availability and rising land costs . The challenges and opportunities in this evolving landscape underscore the importance of strategic planning , collaboration and adaptation for stakeholders in the data centre industry . �
Rumours have suggested that cloud platforms have already leased multiple
22 www . intelligentdatacentres . com