Interview: the skeptic's view on Energy efficiency & demand response — what would change their mind

The cleanest megawatt is the one you never generate. According to the International Energy Agency, energy efficiency improvements and demand response programs could deliver the equivalent of 1,300 GW of avoided capacity globally by 2030—more than the entire installed power capacity of the United States today. Yet despite decades of advocacy and billions in investment, skeptics remain unconvinced that "negawatts" can truly substitute for megawatts when grid reliability hangs in the balance. This article synthesizes perspectives from grid operators, utility executives, and policy analysts who have questioned the scalability of demand response, examining their concerns honestly while presenting the evidence that might shift their stance.

Why It Matters

The stakes for demand response (DR) and energy efficiency have never been higher. FERC Order 2222, fully implemented across U.S. markets by late 2024, opened wholesale markets to distributed energy resource aggregations, creating a $3.2 billion addressable market for DR providers by 2025. Global spending on energy efficiency reached $560 billion in 2024, according to the IEA's World Energy Investment report, representing the largest single category of clean energy investment.

Grid flexibility needs are accelerating as variable renewable penetration increases. California's duck curve—characterized by midday solar oversupply and evening demand ramps of 15-18 GW within three hours—has become a defining challenge for grid operators. The Electric Reliability Council of Texas (ERCOT) now relies on 3.5 GW of demand response capacity during summer peaks, representing roughly 4% of total system capacity. European grids, facing the dual challenge of electrification and Russian gas supply disruptions, have turned to demand flexibility as both an emergency measure and a structural solution, with the EU targeting 160 GW of demand-side flexibility by 2030.

Emerging markets face even more acute grid flexibility gaps. India's Power System Operation Corporation reported 47 GW of peak demand unmet by available generation capacity in summer 2024, while Southeast Asian grids are projected to require $1.2 trillion in grid infrastructure investment by 2040—much of which could be deferred through strategic demand-side management.

Key Concepts

Demand Response Types: DR programs fall into three primary categories. Emergency DR provides contingency reserves activated during grid emergencies, typically compensated at scarcity pricing ($1,000-9,000/MWh). Economic DR shifts load based on price signals, allowing participants to reduce consumption when wholesale prices exceed contracted thresholds. Behavioral DR leverages nudges, information, and gamification to encourage voluntary load reduction during peak periods.

Negawatts vs. Megawatts: The concept of negawatts, coined by Amory Lovins in 1989, frames avoided energy consumption as functionally equivalent to generation. Critics argue this equivalence breaks down in practice—a natural gas peaker can deliver 200 MW for 12 hours straight, while a demand response curtailment may only sustain 50% of committed capacity beyond the first hour. Proponents counter that DR's value lies precisely in its deployment duration: most grid emergencies last 2-4 hours, matching DR's optimal performance window.

Building Efficiency Measures: Commercial and residential efficiency programs target HVAC systems (40% of building energy use), lighting (15%), and plug loads (20%). Deep retrofits achieving 30-50% energy reduction typically require $15-40/square foot investment with 5-12 year payback periods. Shallow retrofits—LED lighting, smart thermostats, air sealing—cost $2-8/square foot with 1-4 year paybacks.

Grid Flexibility Services: Beyond simple curtailment, demand-side resources increasingly provide frequency regulation, spinning reserves, and ramping services. Fast-responding loads like water heaters, refrigeration systems, and EV chargers can provide sub-second frequency response, earning capacity payments of $50-150/kW-year in organized markets.

Time-of-Use Pricing: TOU rates shift consumption by exposing customers to price differentials between peak and off-peak periods. Effective TOU designs feature peak-to-off-peak ratios of 3:1 or greater, with California's default TOU rates reaching 5:1 differentials during critical peak periods. Studies show 10-15% peak demand reduction from well-designed TOU programs, though customer fatigue and gaming behaviors can erode savings over 2-3 year periods.

Demand Response KPI Benchmarks

Metric	Skeptic Threshold	Industry Average	Top Performer
Event Response Rate	<60%	75-85%	>95%
Curtailment Persistence (hours)	<1 hour	2-4 hours	6+ hours
Verification Accuracy (M&V)	±30%	±15%	±5%
Customer Retention (annual)	<50%	65-75%	>90%
$/kW Acquisition Cost	>$200	$80-150	<$50
Capacity Factor (% of contracted)	<50%	70-85%	>90%
Grid Integration Latency	>30 min	5-15 min	<1 min
Coincident Peak Reduction	<5%	8-12%	>15%

What's Working and What Isn't

What's Working

Commercial DR Programs: Large commercial and industrial (C&I) facilities have proven the most reliable demand response participants. Companies like Target, Walmart, and data center operators regularly curtail 15-30% of facility load during grid emergencies, with event response rates exceeding 90%. The key success factor is alignment between operational flexibility and financial incentives—facilities with backup generation, thermal storage, or process flexibility can shift load without impacting core business operations.

Smart Thermostat Integration: Connected thermostat programs have achieved scale that was unimaginable a decade ago. Google Nest, Ecobee, and Honeywell collectively manage over 50 million devices in North America, with major utilities like Duke Energy, ComEd, and PG&E enrolling 20-40% of connected thermostat customers in DR programs. Typical per-device curtailment of 0.7-1.2 kW during summer peaks aggregates to meaningful grid-scale resources—ComEd's Bring Your Own Thermostat program delivers 200+ MW of curtailable capacity.

Industrial Load Shifting: Energy-intensive industries including aluminum smelting, chlor-alkali production, and data centers have developed sophisticated load flexibility capabilities. Alcoa's Texas smelters historically provided 300+ MW of interruptible load to ERCOT, while hyperscale data centers increasingly treat computational workload scheduling as a grid flexibility asset. Microsoft's datacenter fleet has demonstrated 50+ MW of flexible capacity through workload migration and cooling system optimization.

What Isn't Working

Residential Participation Barriers: Despite smart thermostat success, broader residential DR participation remains stubbornly low. Only 5-8% of residential customers actively participate in utility DR programs nationally, with participation skewing heavily toward high-income households with smart home technology. Renters, multifamily building residents, and low-to-moderate income households face structural barriers including split incentive problems, lack of device ownership, and limited time to engage with complex program rules.

Measurement and Verification Challenges: The fundamental challenge of proving what didn't happen undermines DR credibility with grid operators and regulators. Baseline methodologies—calculating what consumption would have been absent curtailment—introduce ±15-30% uncertainty in typical implementations. Critics point to documented cases of "phantom DR" where payments flowed for curtailments that never materially occurred, including a 2023 CAISO market manipulation case involving inflated baseline claims.

Split Incentives: Building owners who pay capital costs rarely capture efficiency benefits that flow to tenants paying utility bills. This principal-agent problem affects 40% of commercial floor space and 35% of residential units nationally. Green lease provisions and property-assessed clean energy (PACE) financing address split incentives partially, but adoption remains below 5% of relevant building stock.

Key Players

Established Leaders

Enel X: The demand response subsidiary of Italian utility Enel manages 8.5+ GW of flexible capacity globally, making it the world's largest DR aggregator. Enel X's platform integrates C&I loads, distributed generation, and storage assets across 20+ countries.

CPower Energy Management: Acquired by LS Power in 2022, CPower manages 5.5+ GW of DR capacity across North American wholesale markets. The company specializes in complex industrial loads and has pioneered real-time automated dispatch capabilities.

Johnson Controls: Beyond its building automation systems business, Johnson Controls operates OpenBlue Energy Solutions, aggregating 2+ GW of commercial building flexibility for grid services. Their integration of efficiency and DR within building management systems creates persistent flexibility.

Schneider Electric: Through its EcoStruxure platform and AutoGrid acquisition, Schneider Electric provides end-to-end efficiency and flexibility solutions. The company manages 3+ GW of flexible capacity with particular strength in industrial and commercial segments.

Emerging Startups

Voltus: Founded in 2016, Voltus has grown to manage 3+ GW of distributed energy resources across all U.S. wholesale markets. The company's technology platform enables rapid customer onboarding and automated dispatch, reducing per-site acquisition costs below industry averages.

Leap: This virtual power plant platform aggregates residential and commercial distributed energy resources, focusing on California and Texas markets. Leap's API-first approach enables integration with any connected device, from smart thermostats to EV chargers.

OhmConnect: Targeting residential customers in California, OhmConnect gamifies demand response through real-time energy reduction challenges and cash rewards. The company claims 200,000+ active users and has delivered grid-scale curtailments during CAISO emergency events.

Uplight: Formed from the merger of Tendril and Simple Energy, Uplight provides white-label customer engagement platforms to 80+ utilities. Their behavioral demand response programs use personalized messaging and incentives to drive participation.

GridBeyond: This Irish company has expanded from European markets to North America, specializing in industrial load flexibility and battery storage optimization. GridBeyond's AI-driven platform optimizes across multiple revenue streams including frequency regulation and capacity markets.

Key Investors & Funders

Energy Impact Partners: This utility-backed VC has invested in multiple DR and efficiency startups including Uplight and CPower. Their portfolio companies have access to utility customer channels and pilot opportunities.

Congruent Ventures: A climate-tech focused VC with investments in Leap, GridBeyond, and other flexibility-focused startups. Congruent's thesis emphasizes grid-edge technologies enabling decarbonization.

Department of Energy Loan Programs Office: The LPO has deployed $4.5+ billion for grid modernization projects, including demand-side management and efficiency programs at scale.

California Energy Commission: CEC's Electric Program Investment Charge (EPIC) has funded $500+ million in DR technology development and demonstration projects, with particular emphasis on residential participation and equity.

Examples

1. California Flex Alert System: During August 2024's record heat wave, CAISO issued Flex Alerts requesting voluntary conservation as grid reserves dropped below 3%. Real-time tracking showed 2,100 MW of demand reduction during peak hours—enough to avoid rotating outages. Smart thermostat programs contributed 500+ MW, demonstrating that aggregated residential response can provide grid-scale resources during genuine emergencies. Post-event analysis revealed 85% of enrolled thermostat customers participated, though average per-device curtailment was lower than modeled.

2. ERCOT's Emergency Response Service: Texas created the Emergency Response Service (ERS) category following Winter Storm Uri, procuring 3,500 MW of firm demand response capacity with financial penalties for non-performance. The program achieved 96% performance during 2024 summer emergencies, with industrial loads providing the majority of response. ERS pricing of $75-150/kW-year has attracted significant private investment in load flexibility infrastructure.

3. UK Demand Flexibility Service: National Grid ESO launched the Demand Flexibility Service in winter 2022-23 as an emergency measure following the Ukraine energy crisis. The program paid consumers £3/kWh to reduce usage during 12 stress events, delivering 2.3 GW of curtailment. Notably, aggregator participation exceeded individual customer signups by 4:1, suggesting that intermediated participation models outperform direct utility enrollment for emergency DR.

Action Checklist

• Conduct facility-level load disaggregation to identify curtailable versus critical loads, targeting 20%+ of peak demand as flexible capacity
• Evaluate wholesale market participation opportunities through aggregators, comparing capacity payments across ISO/RTO markets
• Implement sub-metering for major loads (HVAC, lighting, plug loads) to enable accurate M&V baselines
• Negotiate green lease provisions that align tenant and landlord incentives for efficiency investments
• Deploy smart thermostats with utility program enrollment, targeting 1 kW curtailment per device
• Assess industrial process flexibility for load shifting, particularly electrolytic, thermal, and computational workloads
• Develop internal dispatch protocols specifying curtailment sequences, personnel responsibilities, and override conditions
• Establish contractual relationships with DR aggregators, comparing acquisition costs, revenue share, and performance requirements

FAQ

Q: Can demand response really substitute for generation capacity during extended grid emergencies? A: For typical grid emergencies lasting 2-4 hours, well-designed DR programs achieve 75-90% of contracted capacity. Extended events exceeding 6 hours challenge DR sustainability, particularly for comfort-sensitive loads like HVAC. The solution is portfolio diversity—combining industrial interruptible loads (high sustained capacity) with residential smart thermostat programs (high participation, limited duration) and battery storage (high reliability, limited energy).

Q: How do skeptics evaluate measurement and verification accuracy? A: M&V credibility depends on baseline methodology, meter granularity, and independent verification. The most credible programs use customer-specific baselines with weather normalization, interval metering at 15-minute or finer resolution, and third-party verification by ISO/RTO market monitors. Skeptics should demand ±10% verification accuracy before treating DR as equivalent to firm generation.

Q: What would change a skeptic's mind about residential demand response? A: Three proof points would shift skeptical perspectives: (1) demonstrated 90%+ response rates during actual grid emergencies, not just test events; (2) multi-hour sustained curtailment exceeding modeled capacity; and (3) retention rates above 80% over 3+ year program periods. Current residential programs are approaching these thresholds but haven't consistently achieved them.

Q: How do emerging markets differ from developed market DR implementation? A: Emerging markets face higher grid instability (frequent outages, voltage fluctuations) that actually increases DR value proposition, but lack the advanced metering infrastructure required for M&V. Successful emerging market DR typically starts with large industrial customers having sophisticated metering and evolves toward residential segments as smart meter rollouts complete. India's Open Access framework and Brazil's Free Consumer market provide regulatory models for emerging market DR.

Q: What role does artificial intelligence play in improving DR performance? A: AI and machine learning improve DR across three dimensions: baseline accuracy (reducing M&V uncertainty from ±30% to ±10%), dispatch optimization (maximizing revenue across multiple grid service products), and customer engagement (personalizing incentives and messaging to increase participation). However, AI cannot solve fundamental physical limitations—if a customer has no flexible load, no algorithm can create curtailable capacity.

Sources

• International Energy Agency. "World Energy Investment 2024." IEA Publications, June 2024. https://www.iea.org/reports/world-energy-investment-2024

• Federal Energy Regulatory Commission. "FERC Order 2222: Participation of Distributed Energy Resource Aggregations in Regional Markets." FERC, September 2020.

• Electric Reliability Council of Texas. "ERCOT Demand Response Programs: 2024 Performance Report." ERCOT, October 2024.

• Lawrence Berkeley National Laboratory. "A Meta-Analysis of Demand Response: Experimental Results from 170 Studies." LBNL, 2023.

• Hledik, Ryan, et al. "The National Potential for Load Flexibility: Value and Market Potential Through 2030." Brattle Group for FERC, 2019.

• California Independent System Operator. "2024 Summer Loads and Resources Assessment." CAISO, May 2024.

• Rocky Mountain Institute. "The Economics of Demand Flexibility." RMI, 2023.

• European Commission. "Action Plan for the Digitalisation of the Energy Sector." EU Publication Office, 2024.