AI Hunger is Reshaping the Global Data Center Industry

In recent years, the data center industry has experienced remarkable growth, driven primarily by the expansion of AI. This phenomenon has resulted in unprecedented demand for specialized infrastructure, leading to the concentration of resources and the emergence of new operating models to meet the computational and energy requirements of AI workloads.

"They [data centers and AI] have grown together in many ways," says Ashley Sturm, Vice President of Marketing, Nautilus Data Technologies. "As new use cases emerged, data centers found ways to densify operations and deliver more power. AI found new ways to harness that power, putting the industry in a cycle of progress and increasing pressure."

The AI boom is driving investment across the sector. For example, Michael Intrator, CEO, CoreWeave, tells Reuters that the company plans to buy cryptocurrency mining company Core Scientific for US$9 billion to "accelerate its ability to deploy AI and HPC (high-performance computing) workloads at scale." Meanwhile, AI hardware developers such as NVIDIA estimate that reasoning model answers will require over a hundred times the computing resources of previous generations, substantially increasing the demand for AI data center capacity.

"Today, we are all faced with the fact that the very technology that data centers have helped develop is inevitably causing problems for data center operations and construction, forcing them to chart a new green path for themselves and for AI," says Sturm.

Mexico is following this trend. A Ciena and Censuswide survey of 1,300 data center decision makers in 13 countries indicated that 43% of new data center facilities in the country will be dedicated exclusively to supporting AI workloads. This percentage, on the margin of the global average, aligns Mexico with international AI infrastructure requirements, calling for the optimization of existing resources while developing new specialized facilities.

"Fifty-three percent of respondents believe that AI workloads will account for the largest demand for data center interconnection infrastructure in the next two to three years, surpassing cloud computing (51%) and big data analytics (44%)," according to Ciena and Censuwide.

These initiatives and expectations from the international and local market are transforming how data centers are designed and operated, prioritizing processing density and high-speed connectivity, says Moody’s. Thus, the industry faces the challenge of balancing high throughput volumes with fast response times for user queries, requiring, just in the front line, robust and scalable infrastructures.

Different Challenges for Different Markets

The expansion of AI-focused data centers involves market-specific challenges and trends. Mexico, for example, forecasts data center industry growth of 400% by 2030, equaling almost all of the country's existing nuclear power. This expansion faces challenges mainly around environmental considerations, particularly in terms of energy and water consumption, driving the search for more efficient cooling solutions, Animal Politico reports.

"The ones that are problematic are the hyperscale data centers, which are the ones that are being created right now," says Ana Gaitán, Lawyer, Red en Defensa de los Derechos Digitales, referring to the current "Big Tech" projects in the country: Amazon Web Services, Google Cloud, and Microsoft.

The Brazilian data center market, the largest in the region, is undergoing a strong expansion, with projections indicating growth from US$2.99 billion in 2024 to US$6.26 billion in 2029, reflecting a compound annual growth rate of 15.96%.  This growth trajectory, which anticipates a significant increase in interconnection bandwidth demand, faces specific challenges. These challenges include the pressing need for continued investment in energy infrastructure and regulatory complexities, both of which are critical to sustaining the anticipated market expansion and supporting advanced interconnection needs.

Chile, meanwhile, faces environmental concerns in relation to water use, which have affected large-scale projects. Google has reportedly restarted a US$200 million data center project in Santiago due to these problems, reports Data Center Dynamics. Although the Chilean government has made efforts to streamline permitting processes for data centers, focusing on sectoral procedures rather than extensive environmental impact assessments for new construction, public scrutiny of environmental footprints remains an important factor for investors and operators.

Colombia faces challenges related primarily to land acquisition and rising energy costs, which are critical to the development of new facilities, notes the Credence Research's Colombia Data Centre Market Overview study. Securing reliable and sufficient energy sources for new data center projects in the country is becoming increasingly difficult. These factors collectively impact the speed and cost of consolidation efforts, requiring strategic planning to overcome infrastructure constraints and ensure a stable operating environment.

Data center industry consolidation in North America and the European Union also faces distinct challenges. In North America, land availability has become a major obstacle to building new data centers, notes JLL's North America Data Center Report Year-end 2024. Beyond land, rising costs for construction, equipment, and critical materials, such as generators and transformers, further complicate the development of new facilities, impacting the pace of consolidation, reports CBRE.

The European Union also faces challenges related to energy availability and regulatory compliance. According to McKinsey & Company, the electricity consumption of data centers in the European Union will triple five years from now, requiring an increase in the supply of electricity from mostly low-carbon sources. Additionally, total data center computing load demand is expected to increase from 10GW now to 35GW by 2030, reports Reuters.

Established hubs such as Frankfurt, London, Amsterdam, Paris, and Dublin (often referred to as FLAP-D markets) suffer from significant network bottlenecks and long waiting times for connections, sometimes stretching up to ten or even thirteen years, notes Hardwarewartung. In addition, the changing EU regulatory landscape, which includes new directives on energy efficiency and data protection, introduces bureaucratic burdens and ambitious technical targets that require meticulous navigation by data center operators and investors, reports Bird & Bird.

These challenges highlight the need for data centers to adapt to evolving environments to ensure the sustainability and security of their operations, reports Bird & Bird: "Despite the challenges, the data center industry is expected to continue to grow tirelessly around the world to meet the enormous and growing demands of the digital economy and AI."