The grid’s hidden solar problem and how you can fix it at the grid edge

Anna Manukyan
6 hours ago
3 min read

The residential solar boom is a massive win for clean energy, but behind the scenes, it’s giving utility distribution planners and operators a serious run for their money. Ironically, the cleaner the grid gets, the harder it becomes to see what’s actually happening on it.

As traditional smart meters only measure net load, the leftover electricity a home consumes from the grid after using its own solar power, utilities often have incomplete visibility regarding the true relationship between residential electricity consumption and solar production. This "hidden load" masks true energy consumption, making distribution forecasting, feeder planning, and managing reverse power flows an operational guessing game.

Historically, fixing this meant throwing expensive hardware or clunky cloud processes at the problem. With the rollout of AMI 2.0, however, utilities finally have the infrastructure to solve this problem where it matters most: right at the grid edge.

NET2GRID EnergyAI® Edge PV is turning next-gen (AMI 2.0) smart meters into distributed solar sensors, seamlessly.

The challenge of measuring behind the meter solar

If a utility wants a clear picture of behind-the-meter (BTM) solar generation today, the options aren't great:

Physical submeters: Installing dedicated hardware at every solar home is too expensive to scale for most utilities and their customers.
Cloud-based processing: Uploading continuous streams or bulk batches of high-frequency data to the cloud chokes communication networks and spikes processing fees.
Outdated registries: Relying on paperwork means utilities rarely know when a system is upgraded.

Utilities need localized, autonomous intelligence that doesn't rely on customer Wi-Fi, third-party inverter APIs, outdated data registries, or constant cloud connectivity.

Enter NET2GRID EnergyAI®: Process at the source

Instead of the old "collect-and-upload" approach, NET2GRID’s application embeds lightweight deep learning models directly onto AMI 2.0 smart meters. This approach renders historical records completely unnecessary. Using a minimum of 1-minute active power data and local timezone offsets, the NET2GRID EnergyAI PV agent reconstructs a home's true gross load entirely on the edge.

Real-world value: From hidden solar to actionable grid intelligence

Once the gross load is reconstructed on the meter, utilities can use that information immediately for planning, operations, and DER management. Such intelligence immediately supports several critical operational use cases.

Grid visibility and operational awareness: Reveal actual feeder loading and identify locations where rooftop PV is driving reverse power flows.
Distribution system planning: Improve transformer loading estimates, hosting capacity studies, and network investment decisions using more accurate demand profiles.
Automated DER registry and asset intelligence: Detect PV installation upgrades without relying on customer paperwork or manual updates.
Future DER and flexibility programs: Create a trusted foundation for virtual power plants, demand flexibility, and DER orchestration by accurately separating load from generation.

Enabling predictive grid operations

As distributed energy resources continue to grow, utilities need intelligence that operates where the data is generated: at the grid edge.

By turning the humble smart meter into an active, edge-resident sensor, utilities can finally eliminate the net-load blind spot and build a more transparent, resilient network.

Performance validation

It’s one thing to promise edge intelligence; it’s another to prove it works when completely isolated from weather feeds, customer inputs, or inverter telemetry.

To validate the technology, NET2GRID tested the application against a massive real-world dataset spanning 300 homes and over 1,000 cumulative years of solar data, demonstrating strong performance: