www.industryemea.com

STMicroelectronics advances power solutions for 800 VDC architecture for next-generation AI factories

Solution leverages ST power chip portfolio by combining silicon carbide (SiC), gallium nitride (GaN),and silicon technologies with advanced custom design at both chip and package levels.

  www.st.com
STMicroelectronics advances power solutions for 800 VDC architecture for next-generation AI factories

STMicroelectronics, a global semiconductor leader serving customers across the spectrum of electronics applications, has unveiled a complete prototype of its new power delivery system as it develops new chip designs supporting VDC power architecture announced by NVIDIA for next-generation AI data centers.

The rapid growth of AI workloads is creating unprecedented power demands in data centers. Traditional 54 V power distribution systems, designed for kilowatt-scale racks, are insufficient for the needs of emerging megawatt-scale AI compute racks. To address this, NVIDIA is working with the ecosystem on the 800 VDC power architecture, which supports megawatt-scale compute racks while improving efficiency, reducing copper usage, and simplifying infrastructure.

STMicroelectronics is actively developing the essential chip technologies for 800 VDC architectures, leveraging its diverse power chip portfolio that combines silicon carbide (SiC), gallium nitride (GaN), and silicon technologies and advanced custom design at both chip and package level.

At OCP 2025, ST shared an important milestone with the development of a compact 12kW power delivery board, approximately the size of a smartphone. The 12kW GaN-based LLC converter, operating at an 800-volt input and a 1MHz switching frequency, has successfully completed full-power testing, delivering continuous 12kW output power with over 98% efficiency and a power density exceeding 2,600 W/in³ at 50V.

The solution addresses the fundamental challenges of power density, efficiency, thermal management, and reliability, which are essential for deploying megawatt-scale AI compute racks while reducing infrastructure complexity and total ownership costs. This marks a significant step toward the next generation of hyperscale AI data centers.

www.st.com

  Ask For More Information…

LinkedIn
Pinterest

Join the 155,000+ IMP followers