According to Utility Dive, U.S. electricity generation is projected to grow by 2.3% in 2025 due to surging demand from data centers and industrial customers, with residential electricity rates already increasing over 5% from July 2024 to July 2025. Investor-owned utilities plan to invest approximately $1.1 trillion in infrastructure from 2025 to 2029, up from $765 million in the previous five years, while 91 GW of new capacity is under construction and 488 GW planned or proposed. Experts including University of California, Berkeley Professor Severin Borenstein warn that infrastructure costs for meeting demand peaks—which typically occur during just 50 hours annually—drive up customer rates, but new policies encouraging demand flexibility could transform this dynamic. The analysis reveals that flexible demand strategies could enable U.S. power systems to accommodate over 98 GW of new data center demand with only 0.5% annual curtailment during peak periods. This emerging approach represents a fundamental shift in how utilities and regulators are approaching the collision between economic development and consumer protection.
Table of Contents
The Infrastructure Cost Trap
The core challenge facing utilities isn’t simply meeting growing electricity demand—it’s the economic inefficiency of building infrastructure that sits idle 99.4% of the year. Traditional utility planning has focused on ensuring reliability during the absolute highest demand periods, typically hot summer afternoons when air conditioning loads peak. This “gold-plating” approach means billions in transmission lines, substations, and generation capacity exist primarily to serve what Borenstein correctly identifies as roughly 50 annual hours of peak demand. The problem becomes exponentially worse with data center growth, as these facilities represent massive, constant loads that can push regional systems toward their physical limits. When utilities must build new infrastructure to accommodate these loads, the capital costs get rolled into rate bases that all customers pay for decades through higher bills.
Demand Flexibility Economics
The emerging solution lies in treating electricity demand as a dynamic resource rather than a fixed obligation. Modern data centers actually possess remarkable flexibility—their computing workloads can often be shifted across time zones or deferred during system stress without impacting end-user experience. Cloud providers already distribute loads globally; the innovation is coordinating this capability with grid needs. The economic case is compelling: shifting just 0.5% of annual consumption away from peak hours, as the Duke University research suggests, could avoid billions in infrastructure investment. This represents a fundamental rethinking of value—instead of viewing data centers as passive loads, utilities can treat them as grid assets that provide reliability services. The key insight is that the marginal cost of flexibility is often far lower than the marginal cost of new infrastructure.
Regulatory Transformation Required
Current utility regulation creates perverse incentives that work against flexibility solutions. Traditional rate-making rewards capital expenditure—utilities earn returns on infrastructure investments—while operational solutions like demand management typically get treated as expenses. This “build more, earn more” model must evolve toward performance-based regulation that rewards outcomes like reliability and affordability regardless of how they’re achieved. The 49 newly approved or proposed large-load tariffs across 30 states represent early recognition of this need, but deeper reform is essential. Regulators must establish clear frameworks for valuing avoided infrastructure costs and create mechanisms for utilities to recover investments in flexibility technologies. Without these changes, utilities will continue defaulting to traditional infrastructure solutions even when cheaper alternatives exist.
Technology Enablers and Barriers
The rapid growth of virtual power plants—up 33% from 2024 to 2025 according to Wood Mackenzie—demonstrates the scalability of distributed energy resource aggregation. However, significant technical and operational barriers remain. Most utilities lack the sophisticated analytics needed to optimize flexibility across their systems, and data accessibility issues hamper third-party participation. Initiatives like EPRI’s Open Power AI represent crucial steps toward standardizing data exchange while maintaining security. The real breakthrough will come when utilities can seamlessly integrate industrial-scale flexibility with residential distributed resources, creating a layered approach to peak demand management. This requires not just technology investment but cultural transformation within traditionally conservative utility organizations.
Competitive Landscape Implications
The push for flexibility is creating new market opportunities while reshaping relationships between utilities and their largest customers. Companies like Schneider Electric, Uplight, and Sparkfund are positioning themselves as essential partners in the flexibility ecosystem, developing the software and business models to monetize load flexibility. Meanwhile, hyperscale data center operators face increasing pressure to demonstrate grid-friendly operations as a condition of interconnection. The emerging dynamic creates strange bedfellows—utilities that traditionally viewed large customers as reliability risks now need their cooperation, while tech companies that could theoretically bypass utilities entirely are recognizing the value of partnership. This collaboration, as Elizabeth Cook notes, could become a “complete win-win” if properly structured.
Implementation Risks and Challenges
Several critical challenges could undermine the flexibility revolution. First, compensation mechanisms must be sophisticated enough to ensure participant engagement without becoming administratively burdensome. Second, reliability concerns persist—if too many flexibility resources fail to deliver during critical periods, the consequences could be severe. Third, the complexity of coordinating thousands or millions of distributed resources requires advanced coordination that many utilities lack. Finally, there’s the risk of regulatory fragmentation—if every state develops different rules and standards, the market for flexibility technologies may struggle to achieve the scale needed for cost-effective deployment. Success will require careful balancing of local needs with interoperable standards.
Future Outlook and Predictions
Within five years, demand flexibility will likely become a standard consideration in utility resource planning, with distributed capacity procurements competing directly with traditional infrastructure projects. We’ll see consolidation among DER aggregators as the market matures, and utilities will develop specialized business units focused exclusively on flexibility optimization. The most progressive states will move beyond voluntary programs to incorporate flexibility requirements into interconnection standards for large loads. Crucially, the valuation methodologies for flexibility will become more sophisticated, incorporating randomized control trial data to precisely quantify avoided costs. As stakeholders across the ecosystem recognize the mutual benefits, what begins as a cost-containment strategy could evolve into a fundamental rearchitecture of how we plan, operate, and value grid resources.
