According to Panduit, artificial intelligence (AI) is rapidly transforming data center infrastructure.

Guatemala, May 26, 2025. The adoption of AI in business and technology is increasing energy demands and operational challenges in data centers. According to Benjamín Granados, manager for North Central America at Panduit, technologies such as graphics processing units (GPUs) and high-speed networks have accelerated this transformation.

Panduit projects that by 2025, 10% of electricity consumption in data centers will be dedicated to AI-related tasks. Globally, a McKinsey study forecasts that by 2027, AI energy use could double the total energy consumption of US data centers recorded in 2022, highlighting the need for more efficient and sustainable infrastructure.

Economic impact and technological adoption.

Granados indicated that AI will have a significant impact on the global economy. According to PwC, it will contribute approximately US$1.6 trillion to global GDP by 2030, driven by sectors such as automation, e-commerce, and cybersecurity.

In parallel, the data center GPU market is expected to grow strongly through 2030, with technology companies and cloud providers leading the demand. Although AI currently runs mostly in hybrid clouds, a migration to edge computing is anticipated, which allows information to be processed near its source, reducing latency and optimizing efficiency in sectors such as manufacturing, transportation, and healthcare.

Energy and cooling: key challenges

Operating AI in data centers presents a significant energy challenge. Panduit has developed solutions such as high-density PDUs capable of handling up to 415V, enabling the efficient powering of high-capacity racks.

Regarding cooling, Granados highlighted advanced technologies such as direct-to-chip cooling, rack systems, rear heat exchangers, and immersion solutions. According to Mordor Intelligence, these technologies will experience substantial growth in adoption by 2032 due to their thermal efficiency.

High-performance networks

AI requires high-capacity networks. Granados recommended the Leaf-Spine architecture, which improves connectivity between GPUs and reduces latency. Furthermore, the use of base-8 and base-16 fiber cabling allows for handling the large volume of data generated by AI models.

Tools like RapidID™ help organize and manage cabling, reducing installation times by up to 50% and minimizing errors.

Data centers for the digital future

The physical infrastructure must also evolve. Reinforced cabinets weighing up to 3,500 pounds, adapted raised floors, and robust electrical systems are required. Liquid cooling, for example, can reduce server energy consumption by up to 92%, making it an effective alternative to traditional methods.

Granados concluded that this transformation is strategic: "AI is redefining the data center ecosystem. At Panduit, we develop solutions that allow companies in Guatemala and the region to operate more efficiently, scalably, and sustainably."

///

About Panduit

Panduit is a leading global manufacturer of high-quality electrical and network infrastructure and connectivity solutions. Operating from our headquarters in Tinley Park, Illinois, USA, and 112 locations worldwide, we drive innovation through strategic investments in R&D and the development of groundbreaking products, while providing seamless global support and service. Since 1955, our commitment to our customers and partners has remained unwavering. Together with them, we create exceptional solutions that support their businesses and benefit them and the world. Panduit is creating the connections that matter. To learn more, visit www.panduit.com .