Siemens and NVIDIA unveil DSX Vera Rubin reference design for AI factories
Редакция ЦОД:Главное
Siemens and NVIDIA have expanded their industrial AI and AI-factory infrastructure collaboration with the NVIDIA DSX Vera Rubin Reference Design for the US and European markets. Siemens’ role spans electrical distribution, protection, automation, controls and a digital twin from grid connection to the AI rack.
- Siemens and NVIDIA introduced the NVIDIA DSX Vera Rubin Reference Design for the US and European markets.
- Siemens’ scope covers electrical distribution, protection, automation, controls and a digital twin from grid connection to the AI rack.
- The announcement concerns a reference design, not a confirmed operational AI-factory deployment.
- Siemens claims its AI-oriented electrothermal workflow can reduce multi-tool setup to under five minutes and enable 50 times more design exploration than manual methods.
- Integrated reference architectures could make interoperability between AI-compute platforms and facility infrastructure increasingly important.
Siemens announced on 22 October 2026 that it had expanded its partnership with NVIDIA in industrial AI and AI-factory infrastructure. Together with ecosystem partners, the companies introduced the NVIDIA DSX Vera Rubin Reference Design for the United States and Europe.
The design positions Siemens as the infrastructure provider across the electrical chain, including distribution, protection, automation and control systems from the grid interconnection through to the AI rack. Siemens is also providing a digital twin intended to support the facility’s design, construction and operation as an integrated system.
For data-center developers and operators, the significance is the integration of accelerated-computing architecture with the engineering systems that supply, protect and manage high-density AI loads. Such reference designs could simplify standardisation across power and facility engineering and shorten design workflows, although the announcement does not identify any deployed sites or commercial orders.
Siemens said its AI-oriented electrothermal workflow can configure multiple engineering tools in less than five minutes, compared with several days, and enable 50 times more design exploration than manual methods. These are company-reported performance claims and have not been independently verified in the supplied material.
The partnership reflects a broader shift toward integrated reference solutions in which compute platforms, power infrastructure, controls and digital modelling are treated as interconnected components. For infrastructure suppliers, this could increase the importance of compatibility with AI-computing platforms and of supporting project development from the grid connection to the rack.
The reference architecture links accelerated computing with the electrical and digital-engineering layers required for AI factories. It could support greater standardisation and faster facility design for high-density AI loads, while underscoring the growing importance of infrastructure interoperability with AI-compute platforms.
- Siemens — Press(первоисточник) / АнглийскийПервоисточник →
Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.