Hyperscalers like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP) are in the midst of a data center expansion spree that will see them control nearly two-thirds of all data center capacity worldwide by the end of this decade, which is eight times the capacity they controlled in 2017.

Synergy Research Group (SRG), in a new report, found that hyperscalers now control more than 1,000 total large data centers around the globe, which accounts for 41% of the worldwide capacity of all data centers. Just over half of that hyperscaler capacity is from own-build, owned data centers, with the remaining portion from leased facilities.

This expansion has allowed hyperscalers to now outpace worldwide on-premises data center capacity, which sits at 37% of the total. The analyst firm noted that was “in stark contrast” to six years ago, when nearly 60% of data center capacity was in on-premises facilities.

That delta will continue to grow over the next several years, with SRG forecasting hyperscalers will control 60% of worldwide data center capacity by 2029, compared to just 20% for on-premises locations.

“In 2012, enterprises spent 12 times as much on their data center hardware and software as they did on cloud infrastructure services, while today they spend three times more on cloud services than they do on their own data center infrastructure,” SRG chief analyst John Dinsdale wrote. “Add to that the huge growth in [software-as-a-service] and consumer-oriented digital services such as social networking, e-commerce, and online gaming, and the result is the burgeoning growth in hyperscale data centers.”

AI pushing data center demands

Dinsdale added that enterprises are also planting more of their gear in colocation facilities, which is reducing their overall on-premises data center capacity and is a practice expected to continue to increase as enterprises tap more into artificial intelligence (AI) services.

“The rise of generative AI technology and services will only exacerbate those trends over the next few years, as hyperscale operators are better positioned to run AI operations than most enterprises,” Dinsdale wrote.

Research firm ISG noted in a recent report that the average large enterprise is planning to nearly double its number of AI-enabled applications by the end of this year. This will see that average grow from 250 applications that were AI-enabled at the end of 2023, to 488 AI-enabled applications by the end of 2024.

This transition could also impact overall data center-based emissions.

John Mennel, managing director at Deloitte Consulting, recently explained in a contributed article to SDxCentral that cloud-based hyperscaler infrastructure generates up to 60% less emissions than in an owned data center.

“Why? Hyperscalers have made large sustainability investments of their own and they apply a greater percentage of their energy use directly to data handling,” Mennel wrote.

Dell’Oro Group reported that a portion of these data center investments are being steered toward advanced energy management technologies that will be needed to deal with a surge in AI-related data center processing and storage demands.

“I believe AI is an evolution, not a revolution. To date, most accelerated computing deployments have been in facilities not fully designed to support them,” Lucas Beran, research director at Dell’Oro Group, explained. “The next leg-up in the evolution of AI is accelerated computing deployed in facilities with purpose-built power and cooling infrastructure.

“In many cases, these [data center physical infrastructure, DCPI] orders have already been placed with vendors. In turn, DCPI vendors have been making manufacturing investments to support increased capacity, which is expected to materialize in meaningful DCPI market growth over the next 18 months. As deployments of accelerated computing and purpose-built physical infrastructure align, AI workload performance stands to benefit.”

Hyperscalers continue to invest in data centers

Hyperscalers have increasingly hinted at this need to increase their AI-connected investments.

“We are – and we’ve talked about now for quite a few quarters – we are constrained on AI capacity,” Microsoft CFO Amy Hood said during Microsoft’s latest earnings call. “And because of that … we’ve … signed up with third parties to help us as we are behind with some leases on AI capacity. We’ve done that with partners who are happy to help us extend the Azure platform, to be able to serve this Azure AI demand. And you do see us investing quite a bit as we’ve talked about in builds, so that we can get back in a more balanced place.”

However, they also remain cognizant of how they manage that investment.

Amazon CEO Andy Jassy during the company’s second-quarter earnings call acknowledged that it’s important to have enough capacity in place to support growing AI-infused data center traffic, but “if you actually deliver too much capacity, the economics are pretty woeful and you don’t like the returns of the operating income.”

“The reality right now is that while we’re investing a significant amount in the AI space and in infrastructure, we would like to have more capacity than we already have today,” Jassy added. “I mean, we have a lot of demand right now. And I think it’s going to be a very, very large business for us.”

ABI Research noted in a report that this AI-fueled investment spree will also angle toward “large and mega-sized colocation facilities.” The firm noted that 28% of total worldwide data centers currently fit this size definition, but that “number will grow to 43% by 2030 as companies build larger data centers that can accommodate AI/generative AI workloads and other data-hungry applications.”