Gen AI to Fuel Data Center Growth

Global capacity of hyperscale data centers could witness a threefold increase over the next five to six years, thanks to the massive demand likely to be generated by generative artificial intelligence or Gen AI. This was the gist of a new report based on data generated by the Synergy Research Group (SRG).

According to the research group, there are more than 900 hyperscale data centers currently operating around the world today and more than 400 new ones are in various stages of planning and execution. Among them, the top three – AWS, Microsoft Azure and Google Cloud and possibly Meta have a strong pipeline for future growth in such data centers. 

These findings were based on SRG’s analysis of data center footprints across 19 hyperscale cloud service providers across the software-as-a-service, platform-as-a-service and infrastructure-as-a-service, social networking, eCommerce and gaming verticals. Per the report, there are 926 existing data centers with an additional 427 in the pipeline. 

AWS, Azure, Google set for robust growth

A report published by SDxCentral quoted SRG Chief Analyst and Research Director John Dinsdale to suggest that the top cloud service providers were well-positioned to keep up with the impact of GenAI technology and the resultant growth in demand for powerful data centers. AWS, Microsoft and Google were all planning for aggressive growth, he noted.

In fact, the research notes that fresh construction of hyperscale data centers weren’t the single option for meeting this growing demand. It predicts that there would be a degree of retrofitting of existing data centers to jack up capacity, especially in the case of operators with a large global presence and robust and long-term growth plans. 

The next step could be retrofitting data centers

Companies such as Equinix and Digital Realty could veer towards retrofitting as a means to meet the growing demand though several companies would be hampered by challenges such as their existing data center location and the facilities available when it comes to this option. Even something as mundane as power availability could hinder retrofitting programs. 

Put together, the new builds and retrofits could result in a threefold growth in global data center capacity by 2029-30, says the report by SRG, which also noted that the mix of hyperscale data centers continues to change, region by region, and owned versus leased, but in aggregate the total number of worldwide data centers has doubled in the last five years. 

The growing energy demand has solutions too

The impact of recent generative AI advances is not so much to increase the number of data centers – which will continue to grow by well over a hundred per year – but to substantially increase the amount of power required to run those data centers. As the number of GPUs in hyperscale data centers skyrockets, driven primarily by AI, so the power density of associated racks and data center facilities also need to increase substantially. This is causing hyperscale operators to rethink some of their data center architecture and deployment plans, it said.

According to Dinsdale, the hyperscale providers were ready to cover the GenAI impact on data center capacity, though there could be some under-representation of data center capacity requirements by the enterprises that use GenAI in its operations. The impact of excess computing capacity required for AI training and inference may not have been accounted for. 

On the question of increased power demand to run these additional data centers, the research notes that most of the hyperscalers were focusing on adding energy efficiencies to their existing data centers with renewable energy being the go-to option. Moreover, when it comes to retrofits, the focus would automatically shift to using more sustainable energy options. 

Synergy provides quarterly market tracking and segmentation data on IT and Cloud related markets, including vendor revenues by segment and by region.