LITEON Technology (2301) participated in the Open Compute Project (OCP) Global Summit 2026 held from October 12th to 15th in San Jose, California, USA, showcasing its next-generation AI factory infrastructure product portfolio centered around the 800 VDC architecture. The exhibits included power rack solutions that support the NVIDIA Vera Rubin platform, as well as integrated power and thermal management solutions designed for NVIDIA MGX and OCP ORv3 architectures, demonstrating their Grid-to-Chip technology portfolio.

800 VDC Promotes the Upgrade of Power Supply and Cooling Architecture for Megawatt-Level AI Infrastructure

LITEON demonstrated the 800 VDC Power Shelf of the integrated battery backup unit ( BBU), showcasing the next-generation 800 VDC power supply architecture for AI factories. To address the continuously increasing heat dissipation requirements of high-density AI computations, LITEON also introduced for the first time a 2.6MW Enterprise Cooling Distribution Unit ( CDU) from its strategic investment partner DCX Liquid Cooling Systems. This solution can provide up to 2.6MW of heat exchange capacity, with a temperature difference of nearly 3°C, and adopts a N+1 redundant pump design to maximize cooling efficiency while ensuring system reliability. It is suitable for high-density AI and HPC deployments. The site also showcased the next-generation HMI of this CDU as well as the integrated software layer, aimed at improving operational efficiency and making key system parameters clear at a glance.

For the NVIDIA Vera Rubin NVL72 platform, LITEON also presents an integrated infrastructure solution, which includes 95kW DC / DC Power Shelf and 280kW Liquid-to-Liquid ( L2L ) within the rack CDU. This solution is based on the NVIDIA MGX architectural design and combines 800 VDC power distribution, energy backup, and thermal management technologies to reduce transmission losses and heat generation, improve energy efficiency and power density, simplify the power supply infrastructure, and free up more rack space for high-density AI deployments.

"In the AI era, the competitiveness of data centers is no longer determined solely by the scale of computing power," said LITEON Cloud Infrastructure Platform, General Manager of Solution SBG and and. "Rather, it is measured by the token output per watt, as well as the efficiency of converting electricity into meaningful AI output. As the power density of AI platforms continues to increase, data centers must simultaneously optimize energy efficiency, heat dissipation performance, and deployment flexibility. High-efficiency power supply and thermal management architectures have become a key foundation for future AI Factory expansion."

ORv3 Power and thermal management solutions support scalability AI Infrastructure

LITEON also showcases an integrated AI power rack solution designed for the ORv3 architecture, integrating 180kW Power Shelf and 108kW Capacitor Backup Unit ( CBU ). With centralized power supply and supercapacitor backup capabilities, this solution can handle dynamic AI workloads, while also enhancing power supply stability and overall system reliability.

The ORv3 architecture has been optimized for liquid cooling compatibility, enabling higher density deployments and enhancing the flexibility of infrastructure. The key cooling modules demonstrated by LITEON include 120kW and 220kW In - Rack CDU, which can be integrated with advanced cooling technologies such as Direct-to-Chip. These solutions help to increase GPU density, enhance thermal performance, and provide a more flexible infrastructure design for the next generation of AI and HPC platforms. By integrating power supply, backup power, and thermal management systems, LITEON helps customers build efficient, reliable, and scalable AI infrastructures.

From power distribution at 800 VDC to power supply at GPU: Building a complete Grid-to-Chip energy architecture

As the power consumption of AI processors continues to rise, how to efficiently deliver power from the rack level to GPU has become a key challenge for the next generation of AI infrastructure. To address this challenge, LITEON has demonstrated its complete range of DC-DC power conversion products, including 800V Power Bricks, IBC Modules, 50V Power Bricks, and Voltage Regulator Modules ( VRM ). These products cover every aspect from 800V power distribution and server-level power conversion to chip-level voltage regulation, showcasing LITEON's comprehensive Grid-to-Chip energy architecture for future high-density AI computing platforms.

Among them, the 800V DC / DC power brick can be deployed on the HVDC distribution board ( PDB ) for downstream power distribution and step-down conversion. Multiple power brick can be connected in parallel to match different power requirements, and active current sharing helps to improve power supply stability and reliability. Subsequently, the intermediate bus converter ( IBC ) perform the next stage of voltage conversion, thereby improving overall power supply efficiency, reducing distribution losses, optimizing board-level space and system costs, and enhancing power supply stability in a multi-output architecture. Finally, the Voltage Regulator Module ( VRM ) provides a precise and stable operating voltage for the GPU, completing the "last mile" of power supply from the rack to the chip.

Strengthen the global ecosystem and support comprehensive AI infrastructure capabilities.

As the computing infrastructure of AI continues to evolve towards higher power density and greater energy efficiency, LITEON is expanding its capabilities in power transmission, thermal management, and system-level energy optimization. By extending its expertise from infrastructure power supply systems to comprehensive Grid-to-Chip energy management, LITEON is building a complete portfolio of AI infrastructure energy solutions.

LITEON continues to deepen cooperation with NVIDIA and a broader range of OCP ecosystems, leveraging their global manufacturing infrastructure and localized service capabilities. Through these close ecological partnerships, LITEON is helping customers meet the growing demands of the AI Factory era for high-performance, highly reliable, and highly scalable infrastructure, thereby accelerating the deployment and expansion of the next generation of AI data centers.

LITEON Technology 2026 OCP Global Summit Exhibition Information:

• Date: October 12 to 15, 2026
• Location: San Jose, USA McEnery Convention Center
• Booth: C51, Hall 2

SOURCE LITEON Technology