Why Telcos are Targeting Faster AI Data Centre Rollout

SK Telecom is working with server manufacturer Supermicro and energy and infrastructure specialist Schneider Electric on a joint solution for AI data centres.

The partnership centres on a modular approach that aims to speed up data centre construction while easing supply bottlenecks linked to AI computing.

Telcos already operate large distributed facilities for mobile networks and enterprise services, and the rise of AI introduces new requirements around computing density and energy demand. The growth of AI workloads places pressure on data centre infrastructure that supports network services.

The three companies are planning a pre-fabricated modular architecture that combines computing hardware with power and cooling infrastructure before deployment. The design turns the data centre into a repeatable module that can be delivered to a site and installed with less on-site assembly.

As AI data centres rely on dense clusters of GPU servers and require high levels of electrical power and cooling capacity, faster and simpler construction of data centres can ease workloads.

Modular design for faster deployment

Traditional data centre construction uses a steel-reinforced concrete model. However, in this structure the building shell is completed first, and server racks, electrical systems and cooling equipment are installed afterwards.

This process works for many facilities but can extend deployment timelines. Equipment delivery delays can also slow the project because key infrastructure components arrive separately and require installation on site.

The model introduced by the three companies integrates AI computing servers and supporting infrastructure inside a single pre-manufactured module. Each unit includes computing hardware together with power distribution and cooling components.

The developers assemble the equipment before the module reaches the site, as companies aim to shorten construction schedules and avoid delays linked to on-site installation. This modular design offers a repeatable building block that expands capacity when demand increases for AI workloads.

SK Telecom, Supermicro and Schneider Electric state that the approach improves cost efficiency compared with conventional construction.

Telco expertise meets AI infrastructure

Each partner supplies a different tool across the data centre ecosystem.

SK Telecom contributes operational experience through its AIDC business. 

Supermicro supplies the server infrastructure that runs AI workloads inside the modular system, as the company focuses on GPU-optimised computing platforms built for machine learning and high-performance computing.

Cenly Chen, Chief Growth Officer at Supermicro, says: “Supermicro is excited to partner with SK Telecom to bring data centres online faster than ever before.

“This new integrated solution will leverage Supermicro's high-performance, GPU-optimised servers tailored to customer workloads. We look forward to helping organisations meet their growing data centre needs with this latest technology.”

The GPU-optimised servers provide the computing layer inside the modular architecture. 

This type of computing infrastructure supports AI functions like network optimisation and also supports data analysis platforms used across mobile and broadband services.

Min Yong Ha, Head of SK Telecom’s AIDC Business, says: “Through collaboration with global leaders in the AIDC business, we are advancing a total solution based on a pre-fabricated modular model.

“Building on this initiative, we aim to proactively address the AIDC deployment needs of global hyper-scalers while further strengthening our cost competitiveness.”

Power and cooling for AI workloads

AI processing places greater strain on power and cooling infrastructure than conventional enterprise computing as high-density GPU servers produce large amounts of heat and require stable electrical distribution.

Schneider Electric contributes engineering expertise covering mechanical and electrical systems, which form the backbone of a modern data centre because they manage power supply and thermal control.

Andrew Bradner, Senior Vice President at Schneider Electric, says: “In the era of AI, the true measure of competitiveness lies in how fast and sustainably organisations can deliver high-performance infrastructure.

“Through this collaboration, we are introducing an integrated AI DC model based on a pre-fabricated modular design, empowering customers to lower carbon emissions, eliminate supply bottlenecks, and operate high-density AI workloads with greater resilience and efficiency.”

Within the modular architecture, Schneider Electric designs power distribution and cooling systems that connect directly to the computing hardware inside each module. This allows the module to support the thermal and electrical demands of dense GPU clusters while performing its operations.