SolarEdge and NVIDIA publish joint 800 VDC protection framework for AI data centers as DC powertrain testing advances

SolarEdge Technologies and NVIDIA today jointly published a white paper titled “800 VDC Protection and Grounding for AI Data Centers,” offering a technology-neutral protection and grounding framework for the high-voltage DC distribution architectures now taking shape across the AI data center industry. The release coincides with SolarEdge’s announcement that its medium-voltage to 800 VDC conversion stage is operating under load in its engineering laboratories.

The data center industry is increasingly moving toward 800 VDC DC distribution to meet the power demands of AI workloads — but the protection practices developed for AC environments don’t readily transfer to these converter-fed DC systems.

SolarEdge and NVIDIA release 800 VDC protection white paper

The white paper is designed to be technology-neutral. Any vendor can implement it, any operator can evaluate it, and any standards body can test it against common functions and measurable outcomes. That’s a deliberate choice — rather than locking the industry into a single vendor’s approach, the framework separates shared protection and grounding requirements from the specific design decisions any one company might make.

SolarEdge’s own architecture appears in the paper as one candidate implementation, offered to advance engineering discussion rather than set a standard.

SolarEdge also confirmed its medium-voltage to 800 VDC conversion stage is now running under load in its engineering labs. System-level validation of the full DC power path is underway. No commercial availability date has been announced.

Why existing AC protection practices do not apply to 800 VDC systems

Here’s the core problem the white paper is trying to solve: protection methods that work in AC data center environments don’t simply carry over to converter-fed DC architectures. The physics are different. Fault behavior is different. Personnel safety requirements need to be rethought from the ground up, not patched onto inherited assumptions.

The paper addresses this by evaluating grounding options and the criteria for choosing grounding impedance — two decisions with significant downstream effects on how faults are detected and cleared. Separating those common requirements from any single vendor’s implementation choices gives operators and standards bodies a shared starting point rather than a fragmented collection of proprietary approaches.

NVIDIA’s position is worth noting. The company supports an open, multi-vendor 800 VDC ecosystem and actively encourages multiple conversion topologies, grounding configurations, and protection technologies — not a single prescribed path.

Zone-based protection approach and candidate technical methods

The framework’s most concrete contribution may be its zone-based protection model. In the restricted facility zone — where access is controlled and personnel aren’t routinely present — the system should be able to locate a fault and keep operating. At the rack interface, where people actually work, any fault should trigger immediate interruption.

That distinction matters because protection doesn’t have to be uniformly aggressive across the entire power path. The response can be calibrated to where the risk actually is. The white paper presents candidate technical approaches including high-resistance midpoint grounding, solid-state transformers, and solid-state circuit breakers — options under evaluation, not universal requirements. Key protection functions identified in the framework include series arc detection, DC disconnection, and insulation monitoring.

SolarEdge’s own architecture appears in the paper as one candidate implementation, offered to advance engineering discussion rather than set a standard.

SolarEdge’s DC platform development status and prior experience

SolarEdge points to more than a decade of running arc detection, DC disconnection, and insulation monitoring in its photovoltaic products. That’s a meaningful track record at the component level — though the company is careful to frame it accurately. Component precedent and data center validation are different things, and the distinction matters when you’re talking about systems powering critical AI infrastructure.

The full data center platform under development includes a medium-voltage solid-state transformer, DC distribution with per-branch residual current measurement, solid-state interruption, insulation monitoring, and system-level monitoring. It’s a comprehensive stack, but still in development. None of these AI data center products are generally available yet. SolarEdge is explicitly inviting independent technical evaluation, welcoming operators, vendors, and standards bodies to test and challenge the framework — a posture consistent with the white paper’s technology-neutral framing.

What this means for the industry

The joint publication from SolarEdge and NVIDIA represents an early attempt to establish shared engineering ground before the industry fragments into incompatible approaches. Several things are worth keeping in mind as this effort develops.

The white paper is a framework for evaluation, not a finalized standard. Standards bodies, testing labs, and certification organizations haven’t weighed in yet, and that process will take time. SolarEdge’s lab milestone — the medium-voltage to 800 VDC conversion stage operating under load — is a development checkpoint, not a product launch, with system-level validation still in progress.

The zone-based protection model and the candidate technical approaches are presented as options, not prescriptions. High-resistance midpoint grounding, solid-state transformers, solid-state circuit breakers — these are possibilities the framework is designed to accommodate across multiple vendor implementations, not favor any single one. The protection functions at the heart of the framework have real-world precedent in solar applications. Whether that experience translates cleanly to the demands of AI data center infrastructure is precisely what ongoing validation is meant to determine.

Kelly Lippke

Kelly is an experienced writer with 15 years of experience exploring the big stories that shape our world, from tech breakthroughs and space exploration to climate, energy, and the fascinating quirks of science. She has a talent for turning complex ideas into sharp, memorable insights that stay with readers long after they’ve finished reading.