Retrofitting an AC facility to 800 VDC is technically possible but rarely economic. Greenfield beats retrofit on 10-year TCO by 30-50 percent under typical assumptions. Retrofit only makes sense in three narrow cases; most operators should scope a satellite greenfield instead.
Native 800 VDC UPS collapses 4 conversion stages to 2 and connects the battery directly to the DC bus. End-to-end efficiency climbs 4-6 points. BMS complexity increases sharply. LFP is displacing NMC in data centre UPS applications faster than in the EV market.
Grounding topology at 800 VDC is a Stage-1 architecture decision that interacts with arc-fault protection and insurance. TN-S vs IT-System has different scaling properties at hyperscale. Insulation-resistance monitoring runs into engineering limits above 100 MW connected load.
DC arcs cannot self-extinguish at zero crossings. 800 VDC arc-flash risk is manageable with the right protection design; Schneider Electric August 2026 study shows incident energy comparable to AC when architecture, capacitor placement, and fault-clearing behaviour are specified correctly.
The AI data centre power train is six stacks, not one. Where each stage lives, who builds it, what is standardised. Reference map for architects speccing new AI campuses.
The trade press writes about 800 VDC as one architecture. There are two, they behave differently in retrofit, and hyperscalers are choosing between them right now.
Capacity is not why data centres are going DC. Four forces are pushing the transition: density economics, grid constraints, GPU synchronicity, hyperscaler strategy.