Infineon Technologies has joined NVIDIA MGX AI Factory ecosystem to redefine how next generation AI data centers receive and manage power. By integrating Infineon’s power management solutions into NVIDIA’s MGX architecture and its 800 VDC power framework, the partnership offers an open, modular blueprint built for the agentic AI era.
These 800 VDC MGX compatible power racks give existing AI infrastructure a clear upgrade path, boosting compute performance and power density without a complete rebuild.
Infineon brings deep expertise in power conversion from grid to core, using the full spectrum of semiconductor materials like the silicon (Si), silicon carbide (SiC), and gallium nitride (GaN). This comprehensive approach speeds the shift to full scale 800 VDC architectures.
Its GaN technology operates at switching frequencies nearing 1 MHz, enabling ultra compact bus converters with industry leading efficiency.
Meanwhile, Infineon’s proprietary SiC JFET technology paired with dedicated control ICs delivers the ideal solution for protection and hot swap functionality on native 800 V server boards. The company’s power management stage converts 800 V down to 50 V, 12 V, or even 6 V.
As part of the NVIDIA MGX AI Factory ecosystem, The german semiconductor major will support the entire 800 VDC power conversion chain from input all the way to intermediate bus voltage and core voltage in MGX based systems.
By reducing conversion stages and bringing DC power closer to the rack, the solution improves efficiency, simplifies infrastructure, and enables higher density AI deployments.
About NVIDIA 800 VDC
NVIDIA’s 800 VDC MGX-compatible power racks help existing AI infrastructure increase compute performance and power density without waiting for full-scale 800V DC AI factories.
They provide an upgrade path for higher-density accelerated computing, enabling hybrid power architectures that protect current infrastructure investments while preparing AI factories for future workloads.
Leadership Comment
“As a member of NVIDIA’s ecosystem, Infineon is working with NVIDIA to redefine power delivery systems from the grid to the processor core, which is required for this next phase of AI innovation,” says Adam White, Division President Power & Sensor Systems at Infineon. “As AI models continue to grow in size and complexity, data centers must deliver dramatically more compute performance within the same physical, power, and cooling constraints. Combined with NVIDIA’s modular MGX architecture, Infineon’s power solutions significantly enhance energy-efficient power distribution across the entire data center power flow. We look forward to continuing our work with NVIDIA to bring more MGX-powered innovations to market.”
