Table of Contents
Jeremy Deutsch, president of APAC at Vantage Data Centers, told RCR Wireless News that the latest wave of AI applications for machine learning—covering both training and inference—is introducing new requirements for data center infrastructure
The rapid expansion of artificial intelligence workloads is changing how data centers are designed and deployed across Asia-Pacific, according to Jeremy Deutsch, president of Asia-Pacific at Vantage Data Centers.
In an interview with RCR Wireless News, Deutsch said the latest wave of AI applications for machine learning—covering both training and inference—is introducing new requirements for data center infrastructure.
“Obviously the latest wave in data centers has been the deployment of large amounts of artificial intelligence applications around the world for machine learning, for both training and for inference,” Deutsch said.
According to Deutsch, these workloads are affecting several aspects of facility design. “The density of the racks, the amount of power they’re drawing is higher,” he said.
He added that cooling technologies are also evolving as operators adapt facilities to support these requirements.
Another change relates to the scale of deployments needed to support AI systems. “Some of these sites are getting to the point where to do training or to do inference at scale, you need a much larger deployment,” he said.
Deutsch said these factors are influencing how the company designs its data center sites globally.
Power availability is also shaping new capacity. While AI deployments were initially concentrated in the United States, Deutsch said expansion is now accelerating across Asia-Pacific.
“We certainly see great opportunity in markets like Australia, in Japan, in Malaysia to deploy infrastructure to support this new wave of data centers,” he said.
Customer expectations have also shifted. “The first one I would say is speed,” Deutsch said. “Customers need to deploy this infrastructure today.”
He said that the timing of AI chip deployment has become critical. “When you purchase a chip, you want to be able to put that chip to use as quickly as possible,” he said.
Scale is another key factor. “So it’s no longer 5 and 10 megawatts over a number of years,” Deutsch said. “There is a big push to put larger clusters in place to make sure artificial intelligence deployments have the ability to scale.”
Cooling design has also been adjusted. “About four or five years ago, we made sure that the design of all of our sites enabled for direct to chip liquid cooling,” Deutsch said.
He added that Vantage Data Centers is reducing water use in its facilities. “We’ve made sure that as we’re looking at deploying our infrastructure, we’re doing so without using evaporative cooling,” he said.
Vantage operates across the Americas, Europe and Asia-Pacific, with a focus on data centers designed for cloud and AI workloads. In Asia-Pacific, the company is prioritizing markets including Australia, Malaysia, and Japan, with additional presence in Hong Kong and Taiwan.
Deutsch also pointed to regional demand drivers. “When you consider that globally, over 60% of the world’s population actually sits in the APAC region, you’re going to see more and more of these deployments coming,” he said.
