NVIDIA Unveils Vision for “AI Factories” at AI Infrastructure Summit
At this week’s AI Infrastructure Summit in Silicon Valley, Ian Buck, NVIDIA’s VP of Accelerated Computing, introduced a bold new initiative to reshape the future of data centers. The plan: transform traditional compute facilities into fully integrated, AI-native factories — intelligent infrastructure purpose-built for the age of artificial reasoning.
From Data Centers to AI Factories
NVIDIA is leading the charge with a new Omniverse Blueprint — a reference architecture for building high-performance, energy-efficient infrastructure optimized for AI workloads. The company is working closely with a wide network of partners to realize this vision, spanning every layer of the technology stack — from building design and grid integration to power delivery, cooling systems, and orchestration.
“This is not just about chips and servers anymore,” Buck said. “We’re building a new class of industrial product — AI factories — and no one can do it alone.”
Designing Intelligence from the Ground Up
Central to the initiative is a shift in mindset: designing infrastructure and compute systems in tandem rather than in isolation. Traditional data centers are often built independently of the AI platforms they support, creating inefficiencies in power usage, cooling, and software integration. NVIDIA’s AI factories aim to reverse that — optimizing every watt of energy to directly serve intelligence generation.
This effort is deeply collaborative. Partners like Jacobs are serving as design integrators, unifying the digital and physical layers of infrastructure. The result? A fully simulated, system-level digital twin of the AI factory — one that integrates not just IT and compute, but also power systems, cooling, and even grid-level resources.
Expanded Ecosystem of Industrial Partners
NVIDIA is working with a growing list of industry leaders to bring this system-level vision to life:
- Siemens Energy is enabling on-premises, gigawatt-scale power delivery.
- GE Vernova is focused on power generation and rack-level electrification.
- Schneider Electric, Vertiv, and Siemens contribute expertise in cooling, resilience, and facility management.
An expanded ecosystem of infrastructure specialists — including Cadence, PTC, ETAP, phaidra.ai, Vertech, E Tech Group, and emeraldai — are also contributing to digital simulation, orchestration, and optimization.
Simulation-Powered Design with Omniverse and OpenUSD
Simulation is at the heart of NVIDIA’s strategy. With the Omniverse digital twin platform, partners can model entire AI factories before any physical construction begins. The platform uses OpenUSD, an open and extensible framework that supports accurate simulation of all infrastructure layers — power, cooling, networking, and beyond.
These digital twins enable:
- Early design optimization
- Real-time collaboration
- Ongoing operational management
By integrating tools like Cadence and ETAP, teams can simulate gigawatt-scale facilities, assess system performance, and identify bottlenecks — all before breaking ground.
Beyond the Walls: Grid-Scale Integration
AI factories don’t operate in a vacuum. They must connect to broader systems — power grids, water infrastructure, transportation networks — requiring deep coordination throughout the lifecycle. NVIDIA’s blueprint includes simulation-ready assets and APIs that allow all stakeholders to plug into a unified, composable system, enabling real-time collaboration from design to deployment to day-to-day operation.
This approach redefines how infrastructure is built — composable, resilient, and scalable by design.
Looking Ahead
NVIDIA’s Omniverse Blueprint for AI Factories is already in development, with key milestones expected in the coming year. When complete, it will offer developers, enterprises, and infrastructure providers an open, extensible framework to build and manage the next generation of AI-native compute environments.
Join the Movement
Discover what’s next for AI infrastructure at NVIDIA GTC in Washington, D.C. — a premier event featuring expert sessions, hands-on labs, and industry showcases from NVIDIA and its partners.
