Satisfying strong demand

Schneider Electric is introducing new, comprehensive solutions for AI data centers. These address the energy and sustainability challenges created by the strong demand for AI systems.

On the one hand, there is a new data center reference design developed in collaboration with Nvidia that supports high-density liquid-cooled AI clusters with a power of up to 132 kW per rack. The design is optimized for Nvidia's GB200 NVL72 and Blackwell chips and facilitates both infrastructure design and practical implementation, as it is based on proven and validated architectures that meet the challenges of large-scale liquid cooling.

Another new innovation is the Galaxy VXL UPS. The uninterruptible power supply solution is the most compact of its kind and saves 52 percent space compared to the industry average. It offers a very high power density and was specially developed for AI, data centers and large electrical loads. With a power density of up to 1042 kW/m² and a power output of 1.25 MW, this scalable, modular UPS offers efficient power in a small footprint.

Partnership with Nvidia

Schneider Electric's new data center reference design was developed together with NVIDIA to support liquid-cooled AI clusters while addressing the unique challenges of deploying liquid cooling in hyperscale, colocation and enterprise data center environments. It includes options for Coolant Distribution Units (CDUs) and direct-to-chip liquid cooling, as well as comprehensive, novel mechanical and electrical specifications to ensure energy-efficient and sustainable operation for future AI data centers. The design was developed using Schneider Electric software tools, including Ecodial and EcoStruxure IT Design CFD. It can be customized to the specific needs of AI workloads and helps users leverage the most sustainable and energy-efficient infrastructure designs for high-density applications.

"The future of accelerated computing and artificial intelligence requires speed and a stable foundation," said Jensen Huang, founder and CEO of NVIDIA. "Our collaboration with Schneider Electric enables customers to develop technological innovations on a stable and resilient infrastructure. Together, we are creating AI data centers that are specifically designed for accelerated computing and support the complex architectures required to bring digital intelligence to every business and industry."
End-to-end solutions for AI data centers

With the new solutions, Schneider Electric is demonstrating its commitment to developing sustainable, holistic and AI-enabled data centers. They are designed to enable data center owners and operators to build an energy-efficient, high-density infrastructure to run AI workloads as sustainably as possible. The tech group focuses on three core areas:

• Developing an energy strategy for the AI ​​era
• Implementing advanced infrastructures
• Consulting services on efficiency and sustainability
• An energy strategy for the AI ​​era

Schneider Electric helps companies secure and use renewable energy and optimize local power generation from sources such as wind, solar and hydrogen. The company offers services that help with site selection and create geographic analyses based on customers' deployment plans. On-site power generation is supported by AlphaStruxure. AlphaStruxure is a joint venture between Schneider Electric and the global investment company Carlyle. The company designs, finances, builds, owns, operates and maintains sustainable energy infrastructures, including microgrids, to provide its customers with rapid availability, emission reduction, reliability and resilience in energy supply.

Implementing advanced infrastructures

Schneider Electric has a comprehensive portfolio of high-density and energy-efficient infrastructure systems designed to meet a wide range of AI requirements with an output of more than 100 kW per rack. These include infrastructure components for data centers from the grid to the chip and from the chip to the chiller, AI-supported remote monitoring and energy management software, and digital services to optimize the lifecycle.

In this context, the new Galaxy VXL UPS is the latest addition to Schneider Electric's holistic portfolio of advanced infrastructure solutions. For the first time, this UPS offers

Customers a scalable and modular solution with 125kW/3U power modules in a 1.2 m2 footprint that can support up to 1.25 MW of critical load in one chassis and up to 5 MW with four units connected in parallel in just 4.8 m2.

In addition, Schneider Electric signed an agreement to acquire a majority stake in Motivair Corporation in October 2024 to expand its liquid cooling portfolio and strengthen its expertise in direct-to-chip liquid cooling and high-performance thermal solutions.

Efficiency and sustainability consulting services

Schneider Electric's Sustainability Consulting division helps customers achieve their decarbonization goals. This is achieved through tailored strategies, detailed emissions assessments and programs to integrate entire supply chains. More than 2,400 Schneider Electric experts in over 100 countries offer customers, among other things, data-based insights into their infrastructure with the EcoStruxure Resource Advisor.

"The impact of artificial intelligence on energy consumption and environmental impact is increasing at an unprecedented rate. It is crucial that we counteract this by finding new ways to decarbonize data centers and digital infrastructures," says Pankaj Sharma, Executive Vice President, Data Centers & Networks at Schneider Electric. "Schneider Electric aims to set new standards here, shape the future of AI while protecting the environment. A strategic approach is required from the network to the chip to the chiller and beyond."
Reducing the energy consumption of AI

Schneider Electric is committed to a science-based approach to reducing the energy consumption of AI. The focus is on the "Energy Intelligence for Sustainable AI" approach, which the French tech group understands as the integration of AI-driven applications into the data center infrastructure. This not only reduces the energy consumption of AI. It also opens up the possibility of using AI as a tool for more comprehensive decarbonization efforts and efficiency improvements in all industrial sectors.

"By 2027, data center electricity consumption is expected to account for 2.5 percent of global demand, while sectors such as buildings, manufacturing, transportation and energy account for the remaining 97.5 percent," said Sean Graham, research director for Cloud-to-Edge Data Center Trends at IDC. "While data centers are pursuing their own net-zero goals in light of their tremendous growth, the real sustainability promise lies in using artificial intelligence to decarbonize entire value chains across a wide range of industries. As Schneider Electric and NVIDIA show, long-term collaboration and innovation are key to driving efficiency and sustainability."