The ever-increasing need to exchange, compute, and store data -driven in large part by the rapid rise of artificial intelligence (AI), machine learning, and cloud computing – has led to a surge in hyperscale data center deployments worldwide. AI workloads, in particular, are significantly more power-intensive than traditional computing, accelerating the growth in both data center capacity and energy demand.
Today, data centers account for approximately 1–2% of global electricity consumption, and this figure is expected to rise substantially in the coming years. This trend presents both environmental and economic challenges, making energy efficiency a critical priority for the sustainable development of next-generation data center infrastructure. Two of the key obstacles are the high energy losses in power conversion and the size and cost constraints of existing power supply solutions.
Conventional power supplies rely on silicon (Si) power devices, which are approaching their physical performance limits and can no longer deliver meaningful gains in efficiency or power density. Gallium Nitride (GaN) power devices offer a compelling alternative, enabling a new generation of high-performance power systems that can overcome these limitations.
GaN technology significantly improves power conversion efficiency while enabling higher power density and system integration. GaN-based power units are more compact, allowing more processors -especially high-performance AI accelerators – and storage to be deployed within the same rack footprint, thereby increasing overall data center capacity. At the same time, improved efficiency translates directly into lower operational costs and reduced environmental impact. Industry estimates suggest that GaN-based power solutions can save large-scale data center operators over $100 million annually in energy costs while reducing carbon dioxide emissions by nearly 1 million metric tons.
NovaWave’s Smart-GaN™ technology further enhances these advantages by integrating the the high performance of GaN devices with the precise driving control, extensive protection, and reduced switching losses. This approach significantly boosts system efficiency and power density while reducing overall power supply size, energy consumption, and operating costs, making it particularly well-suited for the demanding requirements of AI-driven data centers.
