NVIDIA는 Open Compute Project를 통해 Google 및 Microsoft와 파트너십을 맺고 800VDC 직류 기술을 추진 중입니다.

Every new generation of accelerated computing demands more from the underlying infrastructure, requiring greater compute performance, higher rack density, and more efficient, scalable power distribution. The primary bottleneck is no longer just wattage, but how effectively power travels from the grid to the GPU. Traditional alternating current systems require multiple conversion stages as racks grow denser, adding overhead and complexity that compound quickly at the high power levels next-generation AI compute requires. Distributing power at 800 volts direct current simplifies this path by reducing the number of conversion stages between the grid and the accelerator, ensuring more available power reaches the compute hardware. NVIDIA DSX reference designs are built to guide AI factories through a phased transition from today's alternating current infrastructure, through hybrid architectures, and into fully native 800 VDC facilities.

NVIDIA, Google, and Microsoft have been developing the 800 VDC architecture together through the Open Compute Project, publishing a joint white paper in March 2026 and the LVDC Solid-State Transformer Specification v0.3 in July 2026. More than 80 equipment manufacturers and infrastructure companies are already building products to this specification. Because most existing AI factories were designed around alternating current distribution, NVIDIA introduced an MGX-compatible 800 VDC power rack arriving in the second half of 2026. This hybrid architecture slots into operational facilities and delivers 800 VDC directly to compute racks within each row without requiring changes to the building's electrical system. Vladimir Troy, vice president of data center infrastructure at NVIDIA, stated, “800 VDC unlocks the compute performance and power density required for AI at scale. Through OCP, NVIDIA is working with more than 80 ecosystem companies to give AI factories a practical path forward — not just a future vision.” For site owners, this approach ensures that existing investments in land, power rights, and building infrastructure do not become stranded, while still opening the door to higher compute density.

The 800 VDC framework provides a clear roadmap to scale with implementation options at every stage of growth. Operators can begin with the near-term, hybrid-compatible power rack, advance to a row power center expected in 2027 that uses an overhead 800 VDC busway to distribute power across multiple rack rows at up to 2 megawatts per row, or adopt a DC power block for newly designed facilities that converts grid power directly to 800 VDC in a single step. The architecture specifications define common interfaces so that power hardware from different vendors can operate together within the same 800 VDC facility, allowing a supply chain to form around an open standard. These components are captured in the NVIDIA DSX reference designs, which give operators a system-level blueprint to connect power architecture, rack-scale computing, and facility infrastructure as they expand. With Wood Mackenzie projecting $9 trillion in global AI and data infrastructure investment through 2040, the facilities that successfully absorb that capital will be the ones that resolved their power architecture before compute demand outran what their infrastructure could deliver. NVIDIA, Google, and Microsoft continue collaborating with the broader ecosystem to ensure 800 VDC is ready when operators need it and that existing facilities have a viable path to adoption now. Readers are encouraged to consult the 800 VDC OCP blog, the NVIDIA 800 VDC white paper, and the NVIDIA DSX reference architecture guide for further information.