Schneider Electric And NVIDIA Create Data Center Reference Designs To Support AI Technology
JAKARTA Schneider Electric, the energy and automation manager for digital transformation, says it has accelerated data center solutions that support Artificial Intelligence (AI) technology.
The company is collaborating with NVIDIA to create a new data center reference design that supports high-capacity AI clusters. This data center is optimized to make NVIDIA's GB200 NVL72 and Blackwell chips.
Schneider Electric explained that this data center reference design will facilitate planning and implementation with proven and validated architecture. In addition, this design can overcome the challenges of using liquid cooling on a large scale.
"Building a computing and AI future requires speed and a strong foundation," said NVIDIA CEO Jensen Huang. "Our collaboration with Schneider Electric allows customers to design world technology advances over stable and resilient infrastructure."
Huang added that this collaboration will create an AI data center that can accelerate the cloud and support a very important complex architecture. That way, this data center can bring digital intelligence to companies and industries.
VOIR éGALEMENT:
The reference design developed by Schneider Electric and NVIDIA will support AI clusters with liquid cooling. This data center is expected to overcome the challenges of large-scale data management, server space rental in third-party facilities, as well as data centers managed by the company for internal needs.
From this partnership, Schneider Electric said that this reference design will include options for liquid-to-liquid Coolant Distribution Units (CDUS) and liquid cooling directly onto the chip. In addition, this reference design is equipped with a comprehensive mechanical and electrical plan to ensure more efficient operations.
This data center reference design was developed using the Schneider Electric software, including the Ecodial and EcoStruxure IT Design CFD. This data center design has been adjusted to AI workload needs that support energy efficiency.