Summary of "Accelerated Innovation: Liquid Cooling Solutions | Schneider Electric"

Summary of “Accelerated Innovation: Liquid Cooling Solutions | Schneider Electric”

This video presents an in-depth overview of the evolving landscape of data center cooling solutions, driven by the rapid growth and increasing demands of AI computing. Schneider Electric, in partnership with Motive Air, highlights the critical role of liquid cooling technologies in supporting high-density AI workloads and the next generation of data centers.

Key Technological Concepts and Product Features

1. AI Growth and Data Center Challenges

AI adoption is accelerating, driving exponential growth in GPU deployment and compute power needs.
Modern AI servers can house up to 16 GPUs, consuming over 20 times the power and generating 20 times the heat of standard servers.
Rack densities have surged from approximately 20 kW to over 100 kW, with future racks expected to reach 240 kW and beyond, making power delivery and cooling critical bottlenecks.
Cooling can consume up to 40% of a data center’s power budget, emphasizing the need for efficient thermal management.

2. Liquid Cooling as a Necessity

Direct chip liquid cooling (single-phase) is now the standard for dense AI clusters; two-phase direct chip cooling is anticipated as chips become hotter and larger.
Liquid cooling is up to 3,000 times more efficient than air cooling by removing heat directly at the chip level.
Adoption of liquid cooling solutions has doubled two years consecutively and continues to grow strongly.

3. Schneider Electric and Motive Air Partnership

Schneider Electric acquired Motive Air, combining Schneider’s global infrastructure, software, and service expertise with Motive Air’s 15+ years of advanced liquid cooling technology.
Motive Air has cooled six of the world’s top 10 supercomputers, including systems at Argonne, Oak Ridge, and Lawrence Livermore National Labs.
Their cooling solutions support rack densities up to 400-500 kW and coolant distribution units (CDUs) scalable beyond 2 megawatts.

4. End-to-End Liquid Cooling Portfolio

Schneider Electric offers a comprehensive portfolio covering hardware, software, and services for new or retrofit AI data centers and AI factories. Key components include:

Coolant Distribution Units (CDUs): Core of liquid cooling, directly managing heat at the GPU level, scalable from 105 kW to 2.5 MW. Certified for NVIDIA’s latest hardware.
Heat Dissipation Units (HDUs): For facilities lacking chilled water infrastructure, HDUs reject heat into the data center air, supporting up to 132 kW cooling per unit in a compact footprint.
Rear Door Heat Exchangers: Handle residual heat (10-30%) from non-liquid cooled components, combining air and liquid cooling efficiently.
Closed Loop Air-Cooled Chillers: Environmentally friendly, reduce water usage, and integrate free cooling to lower operational costs.
Technology Cooling System (TCS) Loops: Precisely engineered piping systems delivering coolant at optimal temperature, pressure, and flow for maximum GPU performance. Emphasis on filtration and cleanliness for reliability.

5. Software and Services

Schneider Electric’s EcoStruxure software platform provides continuous monitoring, control, and management of liquid cooling systems, including leak detection and operational optimization.
Over 50 years of service experience and a global network of trained technicians ensure reliable deployment, commissioning, and maintenance worldwide.
The company is training over 600 cooling field technicians and partners globally to support scaling AI infrastructure.

6. Modular Data Center Solutions

Prefabricated modular pods engineered for high-density racks and liquid cooling readiness enable fast deployment and scalability.
Collaboration with Compass Data Centers exemplifies rapid, flexible AI data center deployment with hybrid liquid-air cooling.

7. Co-Development and Industry Partnerships

Strong collaboration with NVIDIA, Dell, Microsoft, Google, Digital Realty, Equinix, and others ensures solutions are aligned with evolving server architectures and AI workloads.
Schneider Electric and NVIDIA have co-developed AI data center reference designs, providing blueprints for liquid-cooled AI clusters up to 142 kW per rack and integrated power/cooling management systems.
These reference designs enable digital twins and advanced AI factory management, optimizing infrastructure performance and reliability.

8. Manufacturing and Testing Excellence

Schneider Electric operates multiple global production facilities, recently adding a 2 MW testing capacity in Buffalo, NY.
Rigorous in-house testing ensures every product meets thermal, mechanical, and cleanliness standards before shipment.
This approach reduces field failures and accelerates time to market, improving ROI for data center operators.

Key Insights and Recommendations

AI’s impact on data center design is profound and ongoing; operators must urgently invest in next-generation infrastructure to keep pace.
Liquid cooling is no longer optional but essential for AI workloads due to heat density and power requirements.
Partnerships across semiconductor, computing, electrical, and cooling sectors are vital to drive innovation and cost-effective deployment.
Smaller operators and enterprises need support from local partners and managed service providers to adopt AI infrastructure effectively.
A holistic, end-to-end approach combining hardware, software, and services maximizes reliability, efficiency, and scalability.

Main Speakers / Sources

Andrew Bradner, Senior Vice President, Cooling Business, Schneider Electric
Rich Whitmore, CEO, Motive Air (now part of Schneider Electric)
Vlad (AI and infrastructure expert featured in the introduction)
References to partnerships and insights from NVIDIA, Equinix, Dell, Microsoft, Google, Digital Realty, Compass Data Centers, and other industry leaders.

This video serves as both a guide and a detailed product overview for data center operators, developers, and AI infrastructure planners, emphasizing Schneider Electric’s leadership in liquid cooling solutions designed to meet the demanding future of AI computing.