The promise of Virtual Power Plants (VPPs) is compelling: aggregate thousands of distributed energy resources (DERs) - batteries, thermostats, EV chargers - and dispatch them like a traditional power plant. Collect capacity payments, arbitrage energy markets, provide grid services. Turn residential batteries into revenue-generating assets.
But there's a problem. Most VPPs today operate at roughly 50% of their modeled capability. The gap between what a VPP should deliver and what it actually delivers creates a double economic penalty that undermines the entire value proposition.
The VPP Maturity Model: EnergyHub's Framework
EnergyHub introduced a VPP Maturity Model that categorizes VPP programs across five levels - from basic demand response to sophisticated, multi-program optimization. The model highlights a critical insight: most VPPs are stuck at Level 2 or 3, unable to progress to the grid-grade reliability required at Level 4 and 5.
Why? Because the infrastructure required to execute grid dispatch isn't there. Intelligence - forecasting, optimization algorithms, AI-driven decision-making - can tell you what to do. But execution infrastructure determines whether you can actually do it reliably, at scale, in real time.
The Double Economic Penalty
When a VPP underperforms, it doesn't just miss revenue - it gets penalized twice:
-
Underperformance vs. Capacity Commitment If you bid 10 MW into a capacity market but only deliver 5 MW during a dispatch event, you don't get paid for 5 MW. You get penalized for the shortfall. Programs like CAISO's Resource Adequacy or PJM's capacity market don't reward partial performance - they discount it aggressively.
-
Grid Operator Discounting After repeated underperformance, grid operators and market administrators apply a "discount factor" to future bids. If your VPP historically delivers 50% of its modeled capacity, the ISO will only count 50% of your nameplate capacity going forward. This shrinks your revenue potential across all programs.
The result: a VPP that should generate $2M/year in grid services revenue might only capture $500K - and lose credibility with utilities, ISOs, and regulators in the process.
Why VPPs Underperform: The Execution Gap
The maturity gap isn't caused by bad forecasting or poor optimization. Modern AI and machine learning models are excellent at predicting load, solar generation, and market prices. The bottleneck is execution infrastructure:
-
No Real-Time Telemetry: Many VPPs rely on 24-hour-delayed data from utility meters. You can't respond to a 5-minute dispatch signal if you don't know the current state of your fleet.
-
Partial Fleet Response: Without edge-based control, VPPs depend on internet connectivity, cloud latency, and device firmware that wasn't designed for grid services. When a dispatch signal goes out, 30-50% of devices might not respond - due to connectivity issues, firmware bugs, or customer overrides.
-
Cloud-Dependent Dispatch: Dispatching from the cloud introduces latency (hundreds of milliseconds to seconds). For frequency regulation or fast frequency response, that's too slow. Real-time grid services require local, deterministic control.
-
OEM Firmware Limitations: Battery inverters and smart thermostats are designed for energy management, not grid compliance. OEM firmware schedules capacity. It doesn't operate assets with the millisecond-level precision required for ancillary services.
The common thread: VPPs today have intelligence (the brain) but lack reliable execution (the nervous system).
The Path to Grid-Grade VPPs: Execution First
To close the VPP maturity gap, the industry needs to shift focus from more intelligence to better execution. That means:
-
Edge-First Control Deploy execution infrastructure at the site level - gateways, edge OS, local controllers - that can execute dispatch signals in real time, even during internet outages.
-
Real-Time Telemetry Move from 24-hour meter data to sub-second telemetry. If you can't see what your fleet is doing right now, you can't control it reliably.
-
Decoupled Intelligence and Execution Separate the "what to do" (optimization logic, forecasting) from the "how to do it" (deterministic control). This allows OEMs to keep building hardware, utilities to keep running optimization models, and a dedicated execution layer to ensure reliable dispatch.
This is the principle behind AERA (Adaptive Energy Response Architecture) - Molecule Systems' execution architecture. Intelligence can come from anywhere: your own algorithms, a third-party aggregator, utility dispatch signals. But execution has to be local, fast, and deterministic.
Why This Matters for OEMs, Operators, and Utilities
For Battery OEMs: If you're deploying thousands of residential batteries, VPP revenue is a key differentiator. But if your VPP partner can only deliver 50% performance, your customers lose money - and your brand takes the hit. Execution infrastructure is the difference between a VPP program that works and one that doesn't.
For Fleet Operators and Developers: You're betting on VPP revenue to finance projects. Underperformance doesn't just reduce ROI - it kills deal economics. Pay-for-performance contracts (common in California's ELRP and demand response programs) only pay for verified performance. If you can't prove real-time delivery, you don't get paid.
For Utilities and CCAs: You're being asked to integrate VPPs as grid resources. But if VPPs can't deliver reliable capacity, they're a liability, not an asset. The maturity gap is why utilities still hesitate to count VPPs as firm capacity in resource planning.
The Bottom Line
The VPP maturity gap isn't a forecasting problem or a market design problem. It's an execution problem. The industry has spent a decade building smarter algorithms. Now it's time to build the infrastructure that makes those algorithms work in the real world.
At Molecule Systems, we build that infrastructure. Edge-based, deterministic, protocol-agnostic execution for VPPs that need to perform like power plants - not prototypes.
Because if a grid operator can't distinguish your VPP from a conventional power plant, it shouldn't be treated any differently.
Ready to close the VPP maturity gap? Ask our AI how Molecule Systems' execution infrastructure delivers grid-grade VPP performance - or schedule a demo to see it in action.