Exelon CEO on if our utilities can handle the AI boom | Term Sheet

Key Concepts

• AI & Energy Demand: The significant and rapidly increasing energy demands posed by the growth of Artificial Intelligence and data centers.
• Grid Infrastructure Buildout: The necessity of substantial investment and expansion of the existing power grid to accommodate AI’s energy needs.
• Ramp-Up Period: The understanding that AI energy demand will not be immediate, allowing for phased infrastructure development.
• Transmission Security Agreements (TSA): Financial mechanisms used by utilities to ensure commitment from large energy consumers (like data centers) to their projected load, protecting other customers from cost burdens.
• Community Engagement: The critical importance of early and ongoing dialogue with local communities regarding infrastructure projects and their impact.
• Affordability & Equity: The challenge of balancing the needs of large AI consumers with the need to maintain affordable energy prices for all customers.
• Permitting Reform: The need for streamlined regulatory processes to accelerate infrastructure development.
• Net Zero Emissions: Exelon’s commitment to achieving net-zero emissions by 2050.

The Future of AI and the Power Grid: A Conversation with Calvin Butler, CEO of Exelon

Introduction

This podcast episode features a discussion with Calvin Butler, CEO of Exelon, one of the largest public utility providers in the United States, regarding the challenges and opportunities presented by the burgeoning AI industry’s energy demands. The conversation, recorded at Brainstorm AI, centers on the preparedness of the US power grid, the importance of proactive planning, and the need for collaboration between utilities, AI companies, and communities.

I. The Scale of the Challenge: Anticipating AI’s Energy Needs

Butler emphasizes that while the grid is currently prepared to some extent, a “significant buildout in infrastructure” is essential to meet future demands. He stresses that this demand will be sequential, not instantaneous, allowing for strategic planning and phased implementation. He provides a concrete example: Exelon anticipates adding over 19 gigawatts (GW) of new load in Illinois alone by 2030. To contextualize this, he notes that ComEd, the fourth largest utility in the nation, currently has a peak load of 23 GW. This means Exelon is planning to add nearly an entire utility’s worth of load within the existing ComEd system. He acknowledges the challenges of scaling infrastructure – the need for supplies, skilled labor, and land acquisition – but expresses confidence in the industry’s ability to meet the need, given sufficient time. The planning horizon is estimated to be 10-20 years minimum.

II. Doing it Right: Safety, Affordability, and Community Engagement

Butler outlines three core principles for successfully integrating AI’s energy demands: safety, reliability, and affordability. He emphasizes that no customer should be disproportionately impacted by the needs of another, particularly large data centers. “Done right” means thoughtful policy considerations and proactive community engagement. He advocates for early conversations with communities regarding the siting of data centers and transmission lines, seeking their input on environmental impact, tax structures, and overall community benefit. He aims for approximately 90% alignment with community stakeholders, recognizing that 100% consensus is often unattainable. He highlights the importance of partnering with communities, rather than imposing solutions upon them. This approach is rooted in his experience in government affairs, recognizing the passion and localized nature of community concerns.

III. The Local Nature of Energy & Customer Relationships

Butler stresses that energy is fundamentally local. Customers have a direct and personal connection to their local utility, particularly as energy is a basic necessity. He notes the enduring brand recognition of utilities like ComEd, citing examples of customers recalling childhood jingles. This connection necessitates a strong sense of responsibility and a commitment to serving all customers equitably, regardless of their economic status. He illustrates this point by referencing the need to serve both affluent and low-income communities in Washington D.C. with equal consideration.

IV. AI’s Impact on the Workforce: Retraining and Upskilling

Butler views AI as an enabler, not a replacement, for human workers. He cites the example of smart meters replacing meter readers, but emphasizes that Exelon successfully retrained those employees for new roles, avoiding job losses. He believes AI will similarly change existing jobs but also create new opportunities. The key to successful adaptation is change management and a commitment to continuous learning. He prioritizes slowing down and explaining the “why” behind changes to employees, ensuring they understand the benefits and have opportunities for growth. He contrasts this approach with the “move fast and break things” mentality often prevalent in Silicon Valley, arguing that utilities cannot afford disruptions to essential services.

V. The Importance of Early Engagement & Financial Security

Butler emphasizes the need for AI companies to engage with utilities early in the planning process. He describes the process when an AI company requests 5 GW of power: identifying a suitable site, determining infrastructure needs, establishing a timeframe, and implementing a Transmission Security Agreement (TSA). A TSA requires a financial commitment (letter of credit or cash deposit) from the AI company to protect other customers if the projected load doesn’t materialize. He also highlights the importance of cluster studies to assess the impact of new load on the overall grid. He notes a shift in the industry, with AI companies now proactively seeking collaboration with utilities, a change from the “11th hour” approach of the past.

VI. Addressing Misconceptions & Policy Needs

Butler clarifies a common misconception: the portion of a customer’s utility bill directly attributable to Exelon is often less than 50%, with the majority stemming from supply costs. He explains that rising energy prices are largely driven by supply-side factors, not utility actions. He identifies permitting reform as a critical policy need to accelerate infrastructure development, advocating for bipartisan support to streamline the process and reduce delays. He warns against waiting for a crisis to address these issues, urging proactive planning and investment.

VII. The Future Outlook & Data Needs

Butler’s biggest uncertainty regarding the future of AI and energy is the accuracy of growth projections. He emphasizes the need for reliable data to inform infrastructure planning. He believes that if growth projections are significantly lower than anticipated, it will allow for more efficient and targeted investment. He stresses the importance of collaboration and a partnership-based approach to navigate the challenges and opportunities presented by AI.

Conclusion

The conversation with Calvin Butler underscores the critical need for proactive planning, community engagement, and policy reform to prepare the US power grid for the demands of AI. While confident in the industry’s ability to meet the challenge, Butler emphasizes the importance of accurate data, collaborative partnerships, and a commitment to affordability and equity for all customers. The key takeaway is that successfully integrating AI’s energy needs requires a long-term, strategic approach that prioritizes both innovation and responsible stewardship of essential infrastructure.