ABB is partnering with NVIDIA to accelerate the development of next-generation, gigawatt-scale data centers designed to power the explosive growth of artificial intelligence workloads. The collaboration focuses on creating high-efficiency, scalable power solutions that will support NVIDIA’s planned 800-volt direct current (VDC) architecture – a key enabler for the massive 1-megawatt server racks expected in future AI facilities.
As data center power consumption continues to rise globally, with total demand forecasted to grow from 80 gigawatts in 2024 to about 220 gigawatts by 2030, the need for more efficient energy delivery systems has become critical. According to Dell’Oro Group, AI workloads will represent around 70% of that expansion, driving data centers toward fundamentally new approaches to power distribution and system design.
The ABB-NVIDIA collaboration will focus on developing power architectures that combine medium-voltage uninterruptible power supplies (UPS) with direct current power distribution at the server level. This configuration, using advanced solid-state electronics, promises to deliver higher efficiency and improved reliability for increasingly power-hungry AI systems. The goal is to streamline the energy path from grid to chip, reducing conversion losses and enabling higher-density computing environments.
“ABB is leading the development of key power distribution technologies that will define the next generation of data centers,” said Giampiero Frisio, President of ABB Electrification. “We have been early investors in cutting-edge UPS, DC, and solid-state electronics that will help data centers meet AI’s unprecedented power demands. Our collaboration with NVIDIA to advance 800 VDC architectures is part of our broader commitment to driving innovation in this dynamic sector.”
Solid-State Medium-Voltage UPS: HiPerGuard
NVIDIA, whose GPUs form the backbone of modern AI infrastructure, is also pushing for a shift toward higher-voltage power systems. “As AI demands continue to grow worldwide, data centers require new approaches to power distribution that improve efficiency and simplify design,” said Dion Harris, Senior Director for HPC, Cloud, and AI Infrastructure at NVIDIA. “Through this collaboration, NVIDIA and ABB are helping the industry move toward 800-volt architectures that will support the high-density AI infrastructure fueling the next wave of innovation.”
ABB’s extensive data center portfolio includes intelligent power management, digital monitoring, and backup systems that ensure uptime and optimize energy use. Nearly 40% of ABB’s electrification R&D is dedicated to innovations in data center technologies, including DC distribution, protection devices, and advanced cooling systems.
Among ABB’s recent innovations are the HiPerGuard – a solid-state medium-voltage UPS – which increases energy efficiency and power density in a smaller footprint, and the SACE Infinitus, the first IEC-certified solid-state circuit breaker designed to make DC distribution both practical and scalable.
By combining ABB’s power engineering expertise with NVIDIA’s leadership in AI hardware and data infrastructure, the partnership aims to deliver the electrical foundation for the next generation of AI-driven data centers – systems capable of supporting ever-larger models while minimizing their environmental and operational footprint.
Discover more from WIREDGORILLA
Subscribe to get the latest posts sent to your email.
