This page was machine-translated and may differ from the original. View original
Infineon Joins NVIDIA MGX AI Factory Ecosystem… Collaborating on AI Data Center Power Architecture
Support for 800VDC-based power conversion, intended to support high-density AI infrastructure
Infineon Technologies has joined NVIDIA's MGX AI Factory ecosystem to address the power architecture of next-generation AI data centers. The two companies plan to cooperate to secure high-performance, high-density infrastructure centered on an 800VDC-based power architecture.
Infineon announced on the 29th that it has joined NVIDIA's MGX AI Factory ecosystem. This collaboration aims to improve the power supply structure of data centers in response to the increasing demand for AI computing, and Infineon supports the MGX architecture and 800VDC power systems through its power management solutions.
Infineon provides power conversion technology based on various semiconductor materials, including silicon (Si), silicon carbide (SiC), and gallium nitride (GaN). The company explained that this enables the integrated design of power flow extending from the power grid to the processor.
Specifically, it includes the implementation of a small bus converter based on high-frequency switching using GaN technology, server board protection and hot-swapping support through SiC JFETs and control ICs, and step conversion of 800V input power to 50V, 12V, 6V, etc. This combination of technologies is applied to support the transition to an 800VDC power structure.
Infineon announced that it is pursuing step reduction by supporting the entire power conversion process, extending from the intermediate bus to the core voltage, within MGX-based systems. The explanation is that this allows for the direct supply of DC power to adjacent rack sections, aiming to improve power efficiency and reduce infrastructure complexity.
NVIDIA's MGX architecture is a platform that supports AI factory design based on an open modular structure. The 800VDC compatible power rack is known as a structure that supports the transition to high-density computing in the future while increasing computational performance and power density even in existing infrastructure environments.
The company stated that as AI models scale up, higher performance is required within the same space and power constraints, and that improving power distribution efficiency is emerging as a key element in data center design.
본 기사에 대한 정정·반론·추후보도 청구는 보도 청구 안내를, 그간 게재된 보도문은 정정·반론보도 모아보기를 참고해 주세요.















