NVIDIA, Google, and Microsoft are advancing 800 VDC high-voltage direct current power distribution to address AI infrast

Advanced AI computing requires dramatic improvements in infrastructure, particularly in how power reaches compute hardware. Traditional power delivery converts alternating current multiple times as it travels from the grid to accelerators, creating inefficiencies that compound at the massive power levels next-generation AI demands. By distributing power at higher voltage using direct current, 800 VDC reduces conversion stages and delivers more available power directly to compute.

NVIDIA, Google, and Microsoft have collaborated on the 800 VDC architecture 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 developing products to this specification.

Most existing AI factories use AC distribution and cannot be easily retrofitted. NVIDIA is addressing this with an MGX-compatible 800 VDC power rack arriving in the second half of 2026, creating a hybrid architecture that integrates 800 VDC compute infrastructure into facilities built on AC systems without requiring changes to building electrical systems. This approach preserves existing investments in land, power rights, and building infrastructure while enabling higher compute density.

The 800 VDC roadmap provides multiple entry points. The power rack offers a near-term hybrid path for existing facilities. The row power center, expected in 2027, centralizes power distribution for full rack rows using overhead 800 VDC busway, supporting up to 2 megawatts per row. The DC power block, designed for new facilities, converts grid power directly to 800 VDC in a single step, enabling direct medium-voltage conversion at massive scale.

Vladimir Troy, vice president of data center infrastructure at NVIDIA, stated that 800 VDC unlocks the compute performance and power density required for AI at scale, and through OCP collaboration, NVIDIA is giving AI factories a practical path forward. Wood Mackenzie projects 9 trillion dollars in global AI and data infrastructure investment through 2040. Facilities that resolve their power architecture before compute demand exceeds infrastructure capacity will be positioned to capture that investment.