Episode Summary
Executive Summary: This Frontier Forum explores how Arizona Public Service turned a small thermostat pilot into a 190 MW virtual power plant, and why VPPs are becoming core utility resources. Carrie Carn and Seth Frader-Thompson emphasize trust-building through repeated performance, customer-centric program design, resource diversity, and AI-enabled dispatch as load growth, heat, EVs, and data centers strain the grid.
Main Topics: APS’s VPP evolution from pilot to major resource (Priority: 5/5): Carrie Carn describes APS’s path from experimenting with smart thermostats to operating a near-200 MW VPP that now factors into grid planning and operations. Trust and reliability as the barrier to adoption (Priority: 5/5): Both speakers explain that utilities need repeated evidence, scale, and operational experience before treating VPPs like dependable resources. Customer-centric program design and participation (Priority: 4/5): APS stresses bill savings, customer choice, and continuous feedback as essential to enrolling and retaining participants year after year. Resource mix expansion beyond thermostats (Priority: 4/5): The discussion covers batteries, EVs, C&I demand response, behavioral demand response, and potential future resources like pool pumps and water heaters. Load growth, heat, and data centers driving urgency (Priority: 5/5): Rapid peak growth in Arizona and other regions is pushing utilities to look for flexible, fast-to-deploy capacity sources like VPPs. AI and automation for large-scale orchestration (Priority: 4/5): The speakers argue that managing millions of flexible devices will require AI to optimize dispatch while preserving customer experience and grid reliability. Cultural change inside utilities (Priority: 3/5): Seth argues the biggest innovation is not only technical but organizational: successful utilities iterate quickly, learn from setbacks, and are willing to try new models.
Key Arguments: APS proved VPP value by scaling from a modest thermostat program to a 190 MW portfolio that operations now trusts as a real asset. Repeated successful dispatch during extreme heat, especially during the pandemic-era Arizona summer, helped convert skepticism into confidence. VPPs are not just reliability tools; they also reduce customer costs by avoiding expensive market purchases and fuel costs. Utilities are increasingly facing load growth from population, heat waves, factories, and data centers, making flexible demand resources essential. A diversified VPP can do more than a single device class, enabling longer events and eventually 24/7 resource expectations. Customer participation can improve utility sentiment; APS says VPP participants are more enthusiastic and view the utility more favorably. Batteries and EVs create layered value when paired with customer economics, rate design, and utility dispatch rights. AI is necessary to coordinate a large number of distributed devices and optimize across grid needs, customer comfort, and reliability constraints. The biggest barrier to broader VPP adoption is not platform capability but utility familiarity, planning norms, and organizational culture. Utilities and regulators should treat VPPs as core infrastructure for decarbonization and load-growth management, not experimental side programs.
Data Points: APS customer participation: nearing 10% - APS says nearly one in ten customers participates in its VPP programs, far above national averages. National participation comparison: 2-3x national average - APS participation in Phoenix is described as roughly two to three times the national average. Current VPP size: nearing 200 MW - Carrie says APS’s virtual power plant is close to 200 MW today. Thermostat dispatch this summer: 163 MW - Thermostats delivered 163 MW of capacity during the summer mentioned in the conversation. Initial resource size: a few megawatts, then 20 MW - The program began small and had to prove itself to operations at a scale initially viewed as too minor to matter. Reliability of platform events: 99.9+% reliable - Seth cites tens of thousands of events showing very high reliability for the platform. Load growth by 2031: more than 13 GW additional peak - Arizona is expected to see a significant rise in peak demand by 2031. Peak demand increase: about 40% - The anticipated peak growth was described as roughly a 40% increase. Event duration evolution: from 3-4 hours to 8 hours - A diversified VPP mix can extend the length of dispatch events beyond short peak-shaving windows. Utility network scale: over 70 utilities - Energy Hub says it works with more than 70 utilities across North America.
Pivotal Quotes: "We almost put ourselves at a disadvantage by calling it a virtual power plant... it almost makes it sound like it's not a real power plant." — Carrie Carn: On terminology and why VPPs can sound less tangible than traditional generation despite their operational value. "I could show you data from literally tens of thousands of events showing that this thing is 99.9 plus percent reliable." — Seth Frader-Thompson: On the evidence utilities need before fully trusting VPPs in planning and operations. "Everyone should do a VPP." — Carrie Carn: Her closing message to skeptical utilities and regulators about adopting customer-side flexibility.
Implications: VPPs are moving from pilot programs to essential grid infrastructure. Utilities that build customer trust, diversify resources, and adopt AI-driven orchestration will be better positioned to handle load growth, heat, and decarbonization.