As the world continues to face the challenges of climate change and environmental sustainability, Infineon Technologies AG is at the forefront of innovation, harnessing the power of all relevant semiconductor materials including silicon (Si), silicon carbide (SiC), and gallium nitride (GaN) to drive meaningful progress towards decarbonization and digitalization.
In its 2025 predictions – GaN power semiconductors, Infineon highlights that gallium nitride will be a game-changing semiconductor material revolutionizing the way we approach energy efficiency and decarbonization across consumer, mobility, residential solar, telecommunication, and AI data center industries. GaN provides significant benefits in end customers’ applications enabling efficient performance, smaller size, lighter weight, and lower overall cost. While USB-C chargers and adapters have been the forerunners, GaN is now on its way to reaching tipping points in its adoption in further industries, substantially driving the market for GaN-based power semiconductors.
”Infineon is committed to driving decarbonization and digitalization through innovation based on all semiconductor materials Si, SiC, and GaN,” said Johannes Schoiswohl, Head of the GaN Business Line at Infineon. “The relevance of comprehensive power systems will increase with GaN manifesting its role due to its benefits in efficiency, density, and size. Given that cost-parity with silicon is in sight, we will see an increased adoption rate for GaN this year and beyond.
Powering AI will be highly depending on GaN. The rapid increase of required computing power and energy demand in AI data centers will drive the need for advanced solutions capable of handling the substantial loads associated with AI servers. Power supplies that once managed 3.3 kW are now evolving towards 5.5 kW, with projections moving towards 12 kW or more per unit. By leveraging GaN, AI data centers can improve power density, which directly influences the amount of computational power that can be delivered within a given rack space. While GaN presents clear advantages, hybrid approaches combining GaN with Si and SiC are ideal for meeting the requirements of AI data centers and achieving the best trade-offs between efficiency, power density and system cost.
In the home appliance market, Infineon expects GaN to gain significant traction, driven by the need for higher energy efficiency ratings in applications like washing machines, dryers, refrigerators and water/heat pumps. In 800 W applications, for example, GaN can enable a two percent efficiency gain, which can help manufacturers achieve the coveted A ratings. According to Infineon, GaN-based on-board chargers and DC-DC converters in electric vehicles will contribute to a higher charging efficiency, power density, and material sustainability, with a shift towards 20 kW+ systems. Together with high-end SiC solutions, GaN will also enable more efficient traction inverters for both 400 V and 800 V EV systems, contributing to an increased driving range.
In 2025 and beyond, robotics will see widespread adoption of GaN supported by the material’s ability to enhance compactness, driving growth in delivery drones, care robots and humanoid robots. As robotics technology integrates AI advancements like natural language processing and computer vision, GaN will provide the efficiency required for compact, high-performance designs. Integrating inverters within the motor chassis eliminates the inverter heatsink while reducing cabling to each joint/axis and simplifying EMC design.
Infineon is further pushing investment in GaN research and development to overcome the challenges of cost and scalability. With the broadest product and IP portfolio, the highest quality standards, leading-edge innovations such as 300 mm GaN wafer manufacturing and bidirectional switch (BDS) transistors, the company is bolstering its leading role in driving decarbonization and digitalization based on all relevant semiconductor materials including gallium nitride.
Availability
Download the “2025 GaN predictions” ebook here.
