The next phase of artificial intelligence may depend as much on volts as it does on algorithms. Vertiv, a global provider of digital infrastructure, has announced a major milestone in its collaboration with NVIDIA to develop 800 VDC power platform designs, intended to supply the vast electrical demands of next-generation AI data centres—often described as AI factories.
The move signals a fundamental shift in how power is distributed inside hyperscale computing facilities. Traditional 54 VDC systems, which have long been the standard for data centre racks, are proving inadequate as AI workloads expand from kilowatts to megawatts per rack. Vertiv’s new 800 VDC architecture is designed to deliver greater efficiency, scalability, and reliability, paving the way for the dense, synchronised computing clusters required to train and operate the world’s most advanced AI models.
The announcement marks the transition from concept to engineering readiness, with Vertiv’s 800 VDC portfolio expected to launch in the second half of 2026, timed to support NVIDIA’s planned rollout of its Rubin Ultra platforms in 2027.
Powering the new intelligence economy
As AI workloads evolve, the physical infrastructure that supports them is being redesigned from the ground up. Each new generation of NVIDIA platforms pushes the limits of power and cooling, forcing the industry to rethink long-standing electrical standards.
Vertiv and NVIDIA are working together on a scalable, integrated 800 VDC system that can combine high-efficiency rectifiers, DC busways, and rack-level DC-DC converters, all engineered for megawatt-scale operation. The collaboration also includes integrating energy storage solutions to stabilise power delivery and improve sustainability across the compute lifecycle.
“Larger AI workloads are reshaping every aspect of data centre design,” said Scott Armul, executive vice president of global portfolio and business units at Vertiv. “With the development of our 800 VDC platform designs, we’re translating our systems-level expertise into next-generation solutions that support the massive compute densities required for AI factories.”
From NVIDIA’s perspective, the new architecture is not a luxury but a necessity. “Powering the next generation of megawatt-scale AI factories requires a fundamental shift in power architectures,” said Dion Harris, senior director for HPC, Cloud and AI Infrastructure at NVIDIA. “NVIDIA and Vertiv are working closely together to develop the scalable and efficient power foundation needed to unlock the full potential of next-generation AI infrastructure.”
From telecom roots to AI factories
Vertiv’s 800 VDC approach is not without precedent. The company has decades of experience designing DC power systems for telecommunications and industrial markets, sectors that have long relied on high-voltage direct current for efficiency and reliability. What is new is the scale at which this technology will be deployed.
The company is already involved in the early design stages of several large AI factory projects, using its 800 VDC reference architecture as the basis of design for facilities expected to operate at gigawatt scale. These next-generation plants, purpose-built for AI model training and inference, will require stable and efficient power delivery to sustain the enormous computational demand.
Vertiv said its platform-level readiness extends beyond hardware to include global servicing capabilities. Managing such high-voltage systems safely and effectively is a core requirement for AI factory adoption, and Vertiv’s network of more than 4,000 field engineers provides the operational support that hyperscale operators depend on.
Engineering the foundations of intelligence
The collaboration underscores a broader transformation in the data centre ecosystem. As AI models become larger and more energy-intensive, the bottleneck is shifting from compute power to the infrastructure that feeds it. Power, cooling, and serviceability are emerging as the new strategic battlegrounds in AI’s industrial era.
Vertiv describes its 800 VDC platform as a holistic system, where every component, from rectifiers to converters to busways, operates as part of an integrated design. “We are engineering a scalable system where infrastructure parts interoperate as one,” said Armul. “This is about moving from vision to readiness and enabling the infrastructure necessary to power future-ready AI factories.”
The collaboration with NVIDIA, one of the defining technology partnerships of the AI age, reflects the convergence of hardware, infrastructure, and energy innovation. As AI becomes central to global economic growth, the companies building its physical foundations, those that can balance performance, efficiency, and sustainability, are shaping not only the future of computing, but the very architecture of intelligence itself.
