AI Power and cooling demands will raise the Data Center Physical Infrastructure market to $46B

Feb. 6, 2024
Dell’Oro Group states liquid cooling will produce $10 billion in revenues over the next five years.

A new Dell’Oro report revealed that the Data Center Physical Infrastructure (DCPI) market is set to grow at an 11 percent compound annual growth rate (CAGR) from 2023 to 2028 to over $46 billion.

The research firm said that the proliferation of accelerated computing to support AI and ML workloads has emerged as a significant DCPI market driver, significantly increasing data center power and thermal management requirements.

Lucas Beran, Research Director at Dell'Oro Group, said that AI will significantly influence the data center market in the future.

"DCPI vendors are increasing manufacturing capacity to support the expected scale of orders for purpose-built AI facilities,” he said. “Meanwhile, end users continue to plan, design and develop operational readiness for these facilities.” 

He added that “we remain a few quarters away from this materializing in a meaningful way, which is why I characterize 2024, or at the least the first half of the year, as the calm before the storm.”

Dell’Oro has forecasted that DCPI revenue growth will slow to a high single-digit rate in 2024 before accelerating to higher growth in 2025 through 2028.

Rising rack power density

As Data Center Physical Infrastructure (DCPI) vendors raise capacity to support purpose-built AI facilities, a key focus will be rack power density.

Dell’Oro estimates today’s average rack power density is around 15 kW/rack; AI workloads will require 60 – 120 kW/rack to support nearby accelerated servers.

“While this requires innovation and product development on the power distribution side, a bigger change is unfolding in thermal management – the transition from air to liquid cooling,” Beran said. “The emergence of accelerated computing propels liquid cooling to become a mainstream technology, with its presence likely in most greenfield facilities.”

He added, “We have raised our liquid cooling forecast, which is now surpassing $3 billion in 2028.”

Cloud colocation remains dominant

Cloud and Colocation service providers are expected to account for most of the growth during Dell’Oro’s forecast period. The research firm said that enterprise customer segment growth is forecast to be modest.

From a regional perspective, Asia Pacific (excluding China), North America and Europe, the Middle East, and Africa (EMEA) are forecast to grow at the fastest CAGRs during the forecast period. 

China, the Caribbean, and Latin America (CALA) are forecast to grow slower.

For related articles, visit the Data Center Test Topic Center.
For more information on optical components and suppliers, visit the Lightwave Buyer’s Guide.
To stay abreast of optical communications technology, subscribe to Lightwave’s Enabling Technologies Newsletter.

About the Author

Sean Buckley

Sean is responsible for establishing and executing the editorial strategies of Lightwave and Broadband Technology Report across their websites, email newsletters, events, and other information products.

Sponsored Recommendations

ON TOPIC: Innovation in Optical Components

July 2, 2024
Lightwave’s latest on-topic eBook, sponsored by Anritsu, will address innovation in optical components. The eBook looks at various topics, including PCIe (Peripheral...

PON Evolution: Going from 10G to 25, 50G and Above

July 23, 2024
Discover the future of connectivity with our webinar on multi-gigabit services, where industry experts reveal strategies to enhance network capacity and deliver lightning-fast...

The Journey to 1.6 Terabit Ethernet

May 24, 2024
Embark on a journey into the future of connectivity as the leaders of the IEEE P802.3dj Task Force unveil the groundbreaking strides towards 1.6 Terabit Ethernet, revolutionizing...

Advancing Data Center Interconnection

July 24, 2024
Data Center Interconnect (DCI) solutions provide physical or virtual network connections between remote data center locations. Connecting geographically dispersed data ...