Unveiled during the Nvidia GTC in Paris, the partnership sees the two industry leaders join forces to co-develop advanced technologies across cooling, power, building management, and high-density rack infrastructure.
These solutions are designed to underpin the next generation of AI data centres—referred to as “AI factories”—that are capable of delivering the immense compute power needed for both AI training and inference at scale.
This strategic collaboration directly supports the EU’s “InvestAI” initiative, a €200 billion programme aimed at establishing Europe as a global hub for AI. The initiative includes plans for at least 13 AI factories and up to five AI gigafactories, designed to meet growing demand from enterprises, governments, and AI developers.
Speaking at the launch, Olivier Blum, CEO of Schneider Electric said: “Schneider Electric and NVIDIA are not just partners — our teams are driving advanced R&D, co-developing the infrastructure needed to power the next wave of AI factories globally.”
He added: “Together, we’ve seen tremendous success in deploying next-generation power and liquid cooling solutions, purpose-built for AI data centres. This strategic partnership allows us to further accelerate this momentum, pushing the boundaries of what’s possible for the AI workloads of tomorrow.”
Nvidia founder and CEO Jensen Huang echoed the sentiment, calling AI “the defining technology of our time” and highlighting the importance of AI factories as the “essential infrastructure that brings AI to every company, industry, and society”.
What is an AI factory?
The concept of the AI factory—first coined by Nvidia—has rapidly gained traction as enterprises and cloud providers seek specialised infrastructure to meet the soaring demands of generative AI, machine learning, and large language models.
Rob Bunger, a long-standing member of Schneider Electric’s CTO office and a key figure in the company’s AI initiatives, explained in an interview with Capacity: “AI factories are here to produce tokens—either from a training perspective or for inference workloads. They’re higher-density than traditional data centres, more homogeneous, and more likely to be liquid-cooled. That makes them more efficient in terms of space, power and thermal management.”
While traditional data centres have historically been built without specific knowledge of the workloads they will host, the rapid evolution of AI infrastructure has forced a change.
“You could build a facility without having any idea what was going in it. That has really changed,” said Bunger. “This collaboration allows us to couple very closely with Nvidia, to see into the future and prepare our solutions accordingly.”
Launch of new infrastructure solutions
As part of the announcement, Schneider Electric also launched a suite of AI-ready solutions, including the new EcoStruxure Pod and Rack Infrastructure. These prefabricated modular systems are designed to simplify and speed up the deployment of AI factories by integrating power distribution, liquid cooling, network cabling, and containment into a compact, scalable unit.
The new Open Compute Project (OCP)-inspired rack system is engineered to support Nvidia GB200 NVL72 platform, which features the modular MGX architecture. This integration positions Schneider as a key infrastructure player within both the Nvidia HGX and MGX ecosystems.
Bunger highlighted the significance of these innovations in tackling real-world deployment challenges: “We’ve been doing prefabricated data centres for a while—mostly around power skids. What’s new is prefabricating the entire IT cluster infrastructure, flat-packing it in a factory, and assembling it onsite. This massively reduces time-to-deployment.”
“The workforce to build these AI factories simply doesn’t exist in enough volume. So factory-built, modular infrastructure is absolutely essential to meeting demand,” he added.
Supporting the EU’s AI vision
The European Commission’s AI Continent Action Plan and InvestAI strategy were front and centre in the partnership announcement. Both aim to position Europe as a leader in the responsible and sustainable deployment of artificial intelligence.
With Schneider Electric headquartered in Europe and deeply rooted in sustainable infrastructure, Bunger said the company is uniquely placed to support the EU’s vision: “We’re excited to support Europe’s AI investment initiatives. Our acquisition of cooling specialist Motivair, and the consolidation of our Uniflair operations in Italy, put us in a great position to meet regional supply chain needs with local expertise.”
He also emphasised the benefits of tighter integration between IT systems and the underlying facility infrastructure—a trend that aligns with digital twin technologies and Nvidia’s Omniverse platform.
“We’re starting to see integration where IT workload performance data feeds into building and energy management systems. If there’s an issue—like a leak—the system can instantly isolate that rack and respond intelligently,” said Bunger. “This level of coordination between compute and facility layers is unprecedented.”
Sustainability and energy efficiency
Concerns about the energy consumption of AI infrastructure have been growing, but Schneider and Nvidia are addressing these head-on. AI factories, especially those using liquid cooling, are inherently more efficient than traditional air-cooled data centres.
Bunger pointed out: “A liquid-cooled AI factory is significantly more efficient. You can operate at warmer temperatures, use more economisation, and reduce PUE and overall energy use. Yes, these workloads require more energy—but they use it more effectively.”
He added that grid readiness and sustainable power sourcing remain key challenges, especially in regions still scaling up their renewable capacity.
“The industry is scrambling to find the right sites with adequate grid connection and the right energy mix. In the U.S., for instance, we’re seeing temporary gas turbines being deployed until the grid catches up. It’s a hurdle, but one we’re actively working to overcome.”
Looking ahead
The partnership between Schneider Electric and Nvidia marks a new era of convergence between compute and infrastructure—one that Bunger describes as less of a partnership and more of a “combined R&D team”.
“We maybe overshare on both sides,” he joked. “They show us what’s coming, even if it’s not final. We do the same. It allows us to move fast, to be ready before the next chip launches, before the next workload shift.”
That speed and agility could prove critical as AI continues to evolve, disrupting industries and generating new, unanticipated applications.
“There are AI use cases we haven’t even imagined yet,” Bunger said. “Being closely tied with Nvidia allows us to pivot quickly when the next wave comes—whether that’s denser chips, new workloads, or entirely new industries adopting AI at scale.”
In a world where infrastructure can no longer afford to lag behind compute, the Schneider-Nvidia collaboration serves as a blueprint for how integrated innovation can power the AI revolution—not just in Europe, but globally.
RELATED STORIES
Schneider Electric boosts data centre efficiency with new service
Khazna and Nvidia join forces to advance AI infrastructure across MEA
Schneider Electric’s Alison Matte on improving business performance and sustainability
