INNOSCIENCE (02577) provides a complete GaN power solution for 800 VDC power architecture, empowering the next generation of AI Factories.

INNOSCIENCE (02577) announced that NVIDIA will support the 800VDC power architecture. The 800VDC rack power architecture brings breakthrough progress to artificial intelligence data centers, enabling higher efficiency, higher power density, while reducing energy consumption and lowering carbon dioxide emissions. Similar to the upgrade from 400V to 800V in the electric vehicle industry, raising the rack voltage from 48V to 800V can reduce the current by 16 times, significantly reducing IR losses and minimizing the demand for copper materials. INNOSCIENCE is partnering with NVIDIA to support the 800VDC power architecture, ensuring the expansion of the new generation GPU roadmap. Traditional artificial intelligence systems based on 48V voltage are facing significant challenges - low efficiency, excessive copper consumption, with over 45% of total power consumed on cooling. Future artificial intelligence clusters (such as racks with over 500 GPUs) using old PSU power designs will have no space to accommodate computing units. The 800VDC architecture is the solution that supports the system's leap from kilowatt-level to megawatt-level. In addition to transitioning to the 800V rack power supply, this architecture also requires achieving ultra-high power density and ultra-high efficiency in the voltage conversion from 800V to 1V. Only gallium nitride power devices (GaN) can meet these stringent requirements simultaneously. To meet the power density requirements of 800VDC, the power switch frequency must be increased to nearly 1MHz to reduce the size of magnetic components and capacitors. The typical switch frequency of existing rack power supplies can reach up to 300kHz, and increasing it to 1MHz can reduce the size of magnetic cores by approximately 50%. INNOSCIENCE's third-generation gallium nitride technology has decisive advantages: On the 800V input side, INNOSCIENCE gallium nitride (GaN) can reduce 80% of driver losses and 50% of switch losses in each switch half-cycle compared to silicon carbide (SiC), resulting in an overall 10% reduction in power consumption. At the 54V output end, only 16 INNOSCIENCE gallium nitride devices are needed to achieve the same conduction losses as 32 silicon MOSFETs, doubling the power density and reducing driver losses by 90%. Compared to silicon MOSFETs in existing rack architectures, using gallium nitride in the low-voltage power conversion stage of 800VDC can reduce switch losses by 70% and increase power output by 40% within the same volume, significantly increasing power density. The gallium nitride-based low-voltage power stage can be expanded to support higher power GPU models, with improved dynamic response and reduced capacitor costs on the circuit board. As the industry's only full-stack gallium nitride supplier and leading gallium nitride IDM company, INNOSCIENCE is the only company to achieve mass production of gallium nitride from 1200V to 15V and can provide a complete chain solution from 800V to 1V. This makes INNOSCIENCE the only supplier capable of providing full GaN power solutions for all conversion stages to meet the evolution of future architectures to meet higher power demands. INNOSCIENCE's gallium nitride also leads in reliability. Its third-generation devices have undergone rigorous accelerated stress tests, including extended 2000-hour dynamic HTOL tests, high-temperature (175C) verification, and large-sample failure verification. Self-developed online dynamic resistance monitoring and long-term board-level stress tests ensure the high-performance work life of its data center products exceeds 20 years. As a global leader in gallium nitride IDM, INNOSCIENCE's third-generation gallium nitride devices have excellent fast switching characteristics, high efficiency, high power density, and outstanding reliability. By integrating the 800VDC power architecture with INNOSCIENCE's gallium nitride technology, artificial intelligence data centers will make the leap from kilowatt-level racks to megawatt-level racks, ushering in a new era of higher efficiency, higher performance, more reliable, and more environmentally friendly artificial intelligence accelerated computing.