Vertiv and NVIDIA Collaborate to Enhance AI Token Production Rates by 30% Within Year One

Enhancing AI Token Production: Vertiv and NVIDIA's Innovative Approach

In a groundbreaking move at NVIDIA's GTC event in Berlin, Vertiv Holdings Co. teamed up with NVIDIA to unveil a revolutionary AI factory approach. This collaboration is set to redefine the rapid configuration and deployment of AI-powered data centers, potentially increasing cumulative AI token output by up to 30% in just the first year of operations. Such advancements are crucial for businesses seeking efficiency in their AI initiatives.

Accelerating Deployment With Pre-Engineered Designs

Traditionally, establishing a robust AI infrastructure involves several sequential processes like engineering power systems and cooling models, which could lead to complexity and delays. The partnership between Vertiv and NVIDIA introduces a streamlined process that starts with pre-engineered AI factory reference architectures. This allows teams to configure infrastructure tailored to specific site requirements quickly.

By integrating the NVIDIA Vera Rubin DSX AI Factory reference design with Vertiv's SimReady power and cooling assets and Dassault Systèmes' 3DEXPERIENCE platform, companies can achieve a 10x reduction in the time required to transition from requirements to a validated AI factory solution. For instance, tasks that typically take about 30 days for validation can now be completed in merely 3 days.

Factory-Built Infrastructure: A Game Changer

The innovative approach doesn't just stop at configuration. Once the AI factory design is validated, Vertiv's prefabricated solutions take precedence, allowing more assembly and system testing to occur in the factory rather than on-site. With initiatives like the Vertiv OneCore hybrid-built data center and the Vertiv SmartRun overhead infrastructure, the deployment time can drastically reduce from 90 days to merely 45 days.

According to the modeled outcomes, this rapid deployment could potentially shorten the total timeline from project conception to the operational phase by around 10 weeks. Consequently, companies can achieve a significant boost in their token output while maintaining the same installed capacity.

Case Study: Polar Data Centers in Norway

A prime example of this initiative in action is Polar Data Centers, which is adopting this deployment model at its facility in Herøya, Norway. The AI data center, which initially spans 41 MW and is scalable to an impressive 100 MW, aims to utilize Vertiv's infrastructure solutions, including their prefabricated power and thermal modules. By adjusting the assembly of infrastructure away from the installation site, Polar aims to minimize on-site complexities and accelerate service availability to its customers.

Giordano Albertazzi, the CEO of Vertiv, emphasized the importance of speed in the AI industry, stating, "An AI factory earns its return in tokens, and every day spent configuring infrastructure or integrating it on-site is output that never comes back. Success depends on how quickly operators can move from requirements to productive AI capacity. "

Vladimir Troy, Vice President of AI Infrastructure at NVIDIA, added that the integration of systems right from the design stage is essential. This co-designed and co-simulated infrastructure ensures smoother function and efficiency as the design transitions from a virtual framework to physical implementation.

Looking Ahead

As this innovative model continues to unfold, stakeholders can expect robust demonstrations from Vertiv and NVIDIA at industry events to showcase how AI factory designs can be effectively implemented. Attendees will learn about the integration of facility requirements that underpin infrastructure configuration alongside practical insights into overcoming deployment challenges in real-world scenarios.

Further, Vertiv's emphasis on providing training and resources through its SimReady Central Library aims to assist teams in efficiently integrating validated equipment, thus streamlining operational processes.

As the world increasingly pivots toward AI-driven solutions, approaches like that of Vertiv and NVIDIA could set a new standard for efficiency and output in the tech industry, marking a shift toward faster, more integrated, and most importantly, productive AI environments.