The future of energy: How innovation and infrastructure are needed to respond to AI growth

In 2024, the International Energy Agency (IEA) estimated that data centers accounted for roughly 1.5 percent of global electricity demand. That number is expected to more than double by 2030, driven largely by the rise in AI infrastructure. To put this into perspective, that increase would be equivalent to Japan’s total electricity consumption today.

How will we meet this rising need for energy — not only from artificial intelligence (AI), but from other digital technologies and the electrification of industries, such as transportation and buildings?

While this challenge may seem daunting, there is reason for optimism. We are seeing numerous innovations and technologies paving the way forward, including advancements in more energy-efficient data centers, breakthroughs in liquid cooling, improved software models, increased energy security, and the exploration of alternative energy sources.

I recently had the opportunity to discuss the future of energy with Mary de Wysocki, SVP and Chief Sustainability Officer at Cisco; Adele Trombetta, SVP & GM, CX EMEA at Cisco; and Christopher Wellise, VP of Sustainability at Equinix, a Cisco customer that provides global digital infrastructure and colocation services. Here are some of the highlights from our conversation.

Q: How are customers adjusting their strategies in response to the energy picture right now?

Adele: Our customers and partners are driving strong demand for AI deployment to unlock its benefits and stay competitive. This trend spans across industries in both the public and private sectors. The pressure is undeniable, but the energy impact of AI, particularly generative AI, is also on our customers’ minds, especially as large language models (LLMs) are being deployed and trained at scale. Many of our customers across Europe, the Middle East and Africa (EMEA) have net-zero targets, so they need to manage the accelerating adoption of AI carefully. They are approaching this with sustainability in mind, making it a core part of strategy development rather than an afterthought. It’s about adopting AI while simultaneously managing its energy impact.

Mary: Interestingly, perspectives can vary depending on who you’re speaking with — whether it’s customers, partners, or suppliers. A common theme emerging, particularly in light of global dynamics, is resiliency. There is a clear focus on proactive investments to secure the energy needed for the future. It’s also about finding ways to move forward collaboratively and identifying opportunities for co-investment. The priorities are clear: we need growth, resilience, and a more sustainable approach.

Chris: We’re seeing several key themes across the customer landscape. Resiliency and reliability are top priorities, with customers focused on ensuring their applications run smoothly. Regulatory compliance is another major concern, especially in regions like the European Union (EU) with directives such as the EU Energy Efficiency Directive. Another request is for end-to-end solutions that optimize operations across the entire value chain as well as support sustainability reporting and regulatory requirements. As customers adopt hybrid multi-cloud environments, they are keen to optimize energy use across platforms and regions. Finally, partnerships are critical. Customers recognize the need to collaborate with suppliers, energy providers, and others to meet their goals and optimize energy use. For example, in the Cisco-Equinix partnership, 70% of devices connected to the Equinix fabric run on Cisco technology.

Q: We know data centers are the foundation for supporting the AI boom and managing its related energy needs. What are some technological advancements that are happening in the data center?

Mary: Designing products with energy efficiency in mind is a critical first step in delivering business outcomes and addressing sustainability. For example, Cisco’s Silicon One chip is engineered to be both energy-efficient and optimized for AI workloads, enabling customers to reduce power consumption while meeting the growing demands of modern networks and data-intensive applications. In addition to that, a foundational innovation for customers, partners, and suppliers is our Sustainability Data Foundation (SDF). It provides a single source of truth, offering the data needed to manage carbon footprints and progress toward net-zero goals. This information empowers technology leaders with the tools to better manage energy and drive sustainability.

Chris: Designing for efficiency is so important. Since 2021, we have required all new build sites to pursue LEED or an equivalent green building certification to demonstrate adherence to recognized sustainability best practices in design and construction. Data centers, built to last 20 to 30 years, require optimization in both design and operations. Innovation in cooling is especially important because cooling typically accounts for over half of energy consumption. Over 100 Equinix data centers are now enabled with access to liquid cooling technology, such as heat exchangers or direct-to-chip cooling. In the latter, a copper plate, fluid, and closed-loop system remove heat directly from the chip while using chemicals to prevent erosion and bio slimes. From a sustainability perspective, this concentrated heat becomes highly usable. For example, in Helsinki, Finland, heat from data centers warms over 10,000 homes, and during the last Summer Olympics, the aquatic center pools were heated by an Equinix data center. Additionally, AI-powered advanced software can create digital twins to optimize cooling parameters and reduce energy consumption for cooling.

Adele: Cisco’s new products now integrate both sustainability and security into the design process. Customers increasingly want to understand how we deploy, monitor, and optimize technology to address energy consumption, performance, and AI. According to a Gartner study, “By 2030, more than 70% of data centers will monitor sustainability metrics, up from approximately 10% today.” (source: Gartner®). Collaboration within the partner and customer ecosystem is key to modernization and efficient resource use. Coordinating diverse data — ranging from Cisco networking equipment to grid data, weather, location, and IT/OT systems — presents a complex but exciting challenge. With Splunk, we can streamline this process and generate the insights needed for effective information flow and optimization.

Q: How well is the global electricity infrastructure equipped to handle growing electricity demand?

Chris: Our primary challenge lies more in distribution than in supply, and the reasons for this vary by region. In the United States, aging infrastructure and complex policy and regulatory environments plays a role. In Europe, while there is rapid growth in renewable energy, integrating it effectively into the grid remains a challenge. In Asia, the situation is more diverse, with both rapid renewable energy expansion and a continued heavy reliance on fossil fuels. To tackle these issues effectively, it is crucial to address both distribution and supply simultaneously.

Mary: Generative AI requires significant energy, prompting the question: how do we ensure reliable access to the grid? In New York City, I see both opportunity and challenge in the grid. The U.S. grid, built mostly in the ’60s and ’70s, lacks reliability and resilience, with 70% of transmission lines over 25 years old. We see the potential of AI to help address major challenges, but its success depends on modernizing the grid and data centers. Industrial IoT can play a key role in creating smart, more secure grids that maximize available energy, support diverse energy sources and enable predictive maintenance.

Adele: We are partnering with customers eager for digital transformation, including energy companies supporting critical national infrastructure. Leveraging this shift, we are focusing on creating more sustainable solutions and we’re building smart grids that prioritize efficiency. While AI is still in its early stages, ongoing collaboration and partnership with utilities is vital to ensure flexibility and adaptability for their evolving needs.

Interested in learning more about the future of energy and the influence of AI? Join me in person as I lead a discussion about this topic at Cisco Live US in San Diego from June 8-12. The session will take place on Tuesday, June 10 from 2 to 2:30pm PT, and you can register here.

Source: Gartner, [10 Performance Metrics to Improve Data Center Sustainability], [Henrique Cecci, Autumn Stanish], [14 February 2025]

GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.

Share: