Navitas and EPFL to show solid state transformer at APEC 2026
Navitas Semiconductor and the research centre École Polytechnique Fédérale de Lausanne (EPFL) will show a 250kW solid state transformer solution at APEC 2026 in San Antonio Texas 22-26 March 2026.
The Solid-State Transformer (SST) platform developed by the Power Electronics Laboratory of EPFL enables the grid architecture required by next-generation data centres, eliminating bulky low-frequency transformers while improving end-to-end efficiency.
The EPFL design uses single stage, modularised bridge rectifier SST topology for converting 3.3kV-AC to 800V-DC at 250 kW power and delivers enhanced performance and modularity. This is built using Navitas GeneSiC ultra-high voltage (UHV) 3300V and high voltage (HV) 1200V SiC Trench-Assisted Planar (TAP) MOSFETs and modules.
The SST demonstrator is developed as part of the Power Electronics Laboratory’s project HeatingBits, aiming to deploy and showcase the latest technologies inside the EPFL’s actual data centre.
“This engagement with EPFL demonstrates how next-generation medium-voltage power conversion can directly address the growing energy and thermal challenges inside AI data centres,” said Paul Wheeler, VP & GM of the SiC business unit at Navitas. “By combining our 3300V and 1200V GeneSiC MOSFETs and modules with a novel single-stage solid-state transformer architecture and advanced real-time control, we are enabling a scalable 800 V-DC distribution approach that delivers higher efficiency from the grid to the rack while creating new opportunities for heat reuse.”
“This novel solid-state transformer platform provides a galvanically isolated, flexible, scalable, and efficient interface between the medium-voltage AC grid and an 800 V-DC data centre architecture, while serving as a real-world experimental environment for advanced distributed control,” said Drazen Dujic, Associate Professor and Director of the Power Electronics Laboratory at EPFL. “By leveraging Navitas UHV and HV SiC MOSFETs portfolio, EPFL was able to optimize system performance for the highest efficiency and optimal design margins for system robustness and reliability."
