AI data centres drive power trends at PCIM

The PCIM conference in Nuremberg, Germany, highlighted emerging power system technologies addressing demands from AI data centers, robotics, and e-mobility. Silicon carbide (SiC) with trench structures is reducing costs in AI data center devices, while gallium nitride (GaN) architectures are enhancing reliability, size, and weight efficiency for robotics. Renesas is scaling up GaN production eightfold to meet rising demand, with Pietro Scalia, head of power system architecture, predicting breakthroughs in 2027 and 2028. Toshiba Electronics Europe announced engineering samples of a 6 mm × 6 mm device integrating a 32-bit Arm Cortex-M23 microcontroller with MOSFETs for three-phase brushless DC (BLDC) motor drives. The AEC-Q100 TB9M040FTG targets automotive applications like electric valves and HVAC systems, offering compact solutions with built-in protection features such as undervoltage, overvoltage, and thermal shutdown detection. Rutronik introduced the RAK-GaN evaluation platform, enabling GaN-based motor control and power conversion testing. The platform uses Infineon’s PSOC Control microcontroller, featuring high-resolution PWM, fast analog-to-digital converters, and hardware accelerators for precise motor control. GaN components, though initially pricier than silicon MOSFETs, reduce cooling needs and material usage, lowering total system costs. The platform’s integration of GaN HEMTs with the microcontroller supports high-frequency switching, improving efficiency and power density in applications requiring dynamic control. These advancements reflect broader industry trends toward more compact, reliable, and energy-efficient power solutions across sectors.