Contáctanos

800 VCC: Construyendo la arquitectura de energía para la infraestructura de IA de próxima generación a gran escala.

Publicado en
25 de agosto de 2026

From components to integrated systems

We contribute expertise across multiple layers of this ecosystem, spanning critical and embedded power, liquid cooling, rack integration, and advanced manufacturing.

Our new Rack de alimentación de 800 V CC, developed in collaboration with NVIDIA to support the NVIDIA Vera Rubin platform, provides a practical bridge between today’s facilities and next-generation AI infrastructure. Its disaggregated architecture moves power products out of the IT rack, freeing valuable space for compute while enabling existing data centers to support the higher power densities required for accelerated computing.

We have developed advanced intermediate bus converters (IBCs) and high-efficiency DC/DC modules optimized for 800 VDC architectures. These solutions convert 800 VDC to the lower voltages required by GPUs, CPUs, and memory devices while addressing the tight space and thermal constraints of next-generation AI factories. Our power portfolio also includes advanced power shelves supporting the NVIDIA GB300, NVIDIA GB200, Vera Rubin and Vera Rubin Ultra platforms, co-engineered with the chipmaker to meet the demanding efficiency and power requirements of accelerated computing.

We also take a broader, system-level approach to AI infrastructure with a portfolio encompassing high-density power distribution, megawatt-scale liquid-cooled racks, energy storage, and prefabricated modular solutions. Bringing these capabilities together enables power, compute, and cooling systems to be optimized as a whole, simplifying integration, improving performance, and accelerating deployment worldwide.

Additional solutions include:

  • UL 1973–certified capacitor energy storage systems that mitigate electrical disturbances caused by dynamic AI workloads
  • Modular, scalable coolant distribution units (CDU) for AI, HPC, and hyperscale applications
  • Prefabricated modular power infrastructure solutions that simplify installation and shorten build times by leveraging factory assembly, testing, and validation

From design to deployment

Arquitectura de alimentación de 800 VCC para infraestructura de IA de próxima generación a gran escala.

NVIDIA’s 800 VDC roadmap demonstrates how the transition can begin within today’s infrastructure and evolve alongside future AI factories. It provides multiple on-ramps for different stages of data center evolution:

  • An MGX-compatible 800 VDC power rack that brings 800 VDC to compute racks within existing AC facilities
  • A row power center that distributes 800 VDC across multiple racks via overhead busways
  • A DC SST (Solid State Transformer) that converts grid power directly to 800 VDC at the facility level

Realizing that roadmap requires complementing technology innovation with system-level integration, manufacturing and supply chain expertise, and close collaboration across the ecosystem to move solutions rapidly from design and validation to production.  We work alongside NVIDIA and hyperscalers to help bridge the gap, engineering and integrating Soluciones de infraestructura de TI, energía y refrigeración for deployment at scale. Together, these capabilities can help turn the industry’s 800 VDC roadmap into the infrastructure for the next generation of accelerated computing.

Learn more about the future of 800 VDC directly from NVIDIA in their Why Scaling AI Compute Performance Requires a New Power Architecture blog and 800 VDC Architecture: Industry Alignment & Execution white paper.