Innoscience is advancing its all-gallium-nitride (All-GaN) power conversion technology within the NVIDIA MGX ecosystem, aiming to support next-generation high-density AI systems. The company's 12kW 800V-to-48V design achieves approximately 99% peak efficiency and 98.2% full-load efficiency, significantly reducing operational costs for high-performance computing facilities. This technology reduces the number of synchronous rectifiers by 50% and supports a range of intermediate bus voltages, enhancing power delivery efficiency.
As AI workloads scale, traditional power delivery models face challenges in meeting rising demands. Innoscience's GaN technology, with its low on-resistance and high-frequency operation, is emerging as a key solution, enabling smaller components and improved thermal performance. This advancement is crucial for breaking through rack power density limits and accelerating the commercialization of AI systems. The NVIDIA MGX ecosystem is leveraging these innovations to drive modular and scalable AI infrastructure development.
Innoscience Enhances AI Power Efficiency with All-GaN Technology in NVIDIA MGX
면책 조항: Phemex 뉴스에서 제공하는 콘텐츠는 정보 제공 목적으로만 제공됩니다. 제3자 기사에서 출처를 얻은 정보의 품질, 정확성 또는 완전성을 보장하지 않습니다.이 페이지의 콘텐츠는 재무 또는 투자 조언이 아닙니다.투자 결정을 내리기 전에 반드시 스스로 조사하고 자격을 갖춘 재무 전문가와 상담하시기 바랍니다.
