By the year 2026, the consumption of electricity by AI data centers is projected to exceed 90 TWh annually. This staggering statistic, highlighted in the latest annual report by the International Energy Agency, sheds light on the significant impact AI is having on the infrastructure landscape. It is forcing data centers, utilities, and technology vendors to reevaluate how power is generated, distributed, and utilized.
At the recent Data Center World Power conference in San Antonio, Lancium co-founder and CEO Michael McNamara, along with David Holmes, global industries CTO and energy lead at Dell Technologies, delved into the rapidly evolving landscape of AI infrastructure. They discussed how power availability, grid integration, and rack-level engineering are now pivotal in enabling the growth of AI.
Holmes emphasized how the pace of change has transformed infrastructure into a critical global economic issue. He remarked, “Who would have imagined that data centers and power would become the focal points of the geopolitical growth of the global economy? But here we are.”
The exponential growth in demand for AI services is evident at Dell, where in the past fiscal year, they sold $9.8 billion worth of AI services. This number soared to $12.1 billion in the first quarter of 2025, with a backlog of $14.1 billion at the end of the quarter. The trajectory of this growth is nothing short of extraordinary.
While hyperscalers often dominate headlines, Holmes pointed out that there is a vast number of customers building their AI data centers, shaping the infrastructure landscape in new ways. This proliferation of private AI deployments is reshaping expectations for infrastructure, necessitating the adoption of cutting-edge technologies to operate contemporary, efficient data centers.
McNamara, involved in major projects like the Stargate development in Texas, highlighted the need for new data center designs to accommodate the substantial loads required to support AI clusters. With AI clusters drawing massive amounts of power, the limitations of the grid infrastructure come into play. McNamara stressed the importance of tight integration with grid operators to manage load swings and prevent chaos.
Looking ahead, flexibility emerges as a key requirement for the future of AI infrastructure. McNamara emphasized the role of battery storage in providing the grid with the flexibility needed to ensure reliability, manage power prices, and maintain resource adequacy.
Holmes discussed the challenges of cooling high-density AI systems and redesigning data center systems to maximize computational capacity while minimizing power consumption. Dell’s innovative approach includes enclosed rear-door heat exchangers, a hybrid of liquid cooling and in-rack containment, which significantly reduces cooling energy consumption.
The convergence of energy and compute in the realm of AI infrastructure is reshaping traditional paradigms and necessitating collaborative efforts to address the evolving needs of data centers. AI’s impact on power, design, and collaboration is ushering in a new era of innovation and transformation in the data center industry.
In the ever-evolving landscape of data centers, a new era of collaboration and innovation has emerged. As gigawatt-scale campuses become the norm, the traditional approach of planning in isolation is no longer viable. Grid operators, data center developers, vendors, and end users are now required to work hand in hand to design facilities that seamlessly integrate with utilities, adopt new operational models, and utilize the latest hardware technologies. The exponential growth in AI technology further emphasizes the need for continuous evolution and adaptation within the industry. McNamara highlights the importance of flexibility, capacity optimization, storage integration, and deep collaboration with the grid as essential elements for the future success of data centers. This shift towards a more interconnected and agile approach is revolutionizing the way data centers operate and interact with the broader energy ecosystem.
