The Grid Edge: Where the Power System Meets Your Home

1. What the grid edge is

The classic power system is a one-way street. Large plants generate electricity, transmission lines carry it across regions, distribution lines deliver it to streets and buildings, and the meter on the side of your house is the finish line. Information flowed the same way: slowly. A utility typically learned what your home had used once a month, when the meter was read. Nothing at the edge generated power, stored it, or made decisions.

The grid edge is everything that has since appeared at that finish line. The working definition used across the industry covers the devices, sensors and software at the distribution level of the grid, on both sides of the meter, that can generate, store, shift or measure electricity. The generating and storing devices have a collective name, distributed energy resources (DERs): rooftop solar, home and community batteries, electric vehicles and their chargers, small wind and fuel cells, and controllable loads such as heat pumps, water heaters and pool pumps that can change when they run. Smart meters, home energy management systems and smart electrical panels are the sensing and control layer that ties them together.

Three changes flipped the edge from passive to active. Generation moved onto rooftops, so power now flows backwards through equipment designed for one-way delivery. Big home loads became schedulable: an EV, a heat pump water heater and a clothes dryer can all be told to run at 2 p.m. instead of 7 p.m. And the meter became a sensor. Advanced metering infrastructure (AMI) records use in intervals as short as a few minutes and reports it daily or faster, which is what allows a utility to bill by time of day and to verify that a home actually reduced load during a grid event. FERC's 2025 assessment of demand response and advanced metering counts 128.4 million advanced meters in operation in the United States in 2023, 76.8% of the 167.2 million meters in service.

For a homeowner, one feature of this shift matters more than the rest: the edge is the only part of the grid you can own. You cannot buy a share of a transmission line, but the battery in your garage, the panels on your roof and the charger by your driveway are grid assets in the technical sense, capable of services the system needs and, increasingly, is willing to pay for.

2. Why the edge is filling up now

Start with the peak problem. A power system is built for its worst hour, not its average one. Summer heat waves and winter cold snaps set the capacity that utilities must have on call, and that capacity sits idle for most of the year. FERC reports that demand response in the seven US wholesale markets reached 33,272 MW in 2024, enough to cover about 6.5% of those markets' combined peak demand of roughly 515 GW. The US Department of Energy's virtual power plant analysis frames the squeeze from the other side: the United States needs to add more than 200 GW of new peak capacity by 2030 as old plants retire and demand grows. Every megawatt of peak that flexible homes can absorb is a megawatt that does not need a new peaking plant, a new substation, or a bigger wire.

The devices arrived before the coordination did. NREL's assessment of rooftop solar technical potential found that suitable US rooftops could host 1,118 GW of solar capacity and generate 1,432 TWh a year, equal to 39% of national electricity sales, with small buildings alone accounting for about 65% of that potential. That is a ceiling, not a forecast, but it shows the scale of hardware that can physically sit at the edge. Electric vehicles add the second wave: the International Energy Agency counts more than 17 million electric cars sold worldwide in 2024, over 20% of all new cars, with US sales growing about 10% to more than one in ten cars sold. Each EV is one of the largest electrical loads in a home, and with a two-way charger it is also one of the largest batteries. Heat pumps do the same for winter: as heating electrifies, cold-climate demand peaks grow, and a heat pump that can pre-heat a house before the peak, then coast through it, becomes a grid resource rather than only a load.

Policy then started paying for coordination. FERC Order 2222, covered in section 6, opened wholesale markets to aggregated home and business devices. States began setting their own programs: FERC's 2025 assessment notes, for example, a Virginia law directing Dominion Energy to run a virtual power plant pilot with aggregations totalling up to 450 MW. And the economics argue for it at national scale. DOE concludes that deploying 80 to 160 GW of virtual power plants by 2030, two to five times the roughly 33 GW in place in its January 2025 update, could serve 10 to 20% of peak load and reduce annual grid costs by about $10 billion, because coordinating devices that households have already bought is cheaper than building equivalent conventional capacity.

3. The hardware at the edge

Rooftop solar and the smart inverter

A rooftop array is the most common way a home joins the edge. The component that matters for the grid is not the panel but the inverter, the box that converts DC from the roof into AC for the house and the street. Modern smart inverters can do far more than convert: they can adjust output, manage voltage on the local feeder, ride through brief grid disturbances instead of disconnecting, and accept instructions from the utility. On a street where many homes export at noon, those functions are what keep local voltage in range. Solar's weakness is timing: output peaks in the middle of the day while residential demand peaks in the evening, which is why solar at the edge is so often paired with the next device.

Home batteries

A home battery turns solar from a midday event into an evening resource, and gives a home without solar a way to buy power when it is cheap and use it when it is not. For the grid, a battery is the most useful edge device per kilowatt: it can discharge on command, at a set power level, for a predictable duration, and the result can be measured at the meter. For the owner it does three jobs in priority order that they choose: backup during outages, self-consumption of rooftop solar, and program participation. Sizing and cost questions are covered in our guide to whether home battery storage is worth it.

Electric vehicles, managed charging and V2G

Left unmanaged, home EV charging lands exactly on the evening peak. Managed charging is the simplest grid edge behaviour there is: the car still charges every night, but the session is scheduled into off-peak hours or paused briefly during an event. Bidirectional charging goes further. Vehicle-to-home (V2H) uses the car as backup power for the house; vehicle-to-grid (V2G) exports energy for the system under a program. An EV battery holds several times the energy of a typical home battery, so even a small number of exporting cars carries weight in a neighbourhood fleet. The current state of the hardware and standards is in our V2H bidirectional charging guide.

Smart meters: the sensing layer

None of the coordination above works without measurement. AMI meters record consumption in short intervals and communicate in both directions, which gives the system three things the old meter never could: time-of-use billing based on when power was actually used, near-real time outage detection, and a trusted record of what a home did during a demand response event, which is what payment is based on. The next generation, often called AMI 2.0, adds faster data and processing inside the meter itself. To see what your own meter already knows, start with our walkthrough on reading smart meter energy data.

Smart panels and controllable loads

The quiet workhorses of the grid edge are loads that can shift without anyone noticing: water heaters that heat an hour early, thermostats that pre-cool a house before a peak and let the temperature drift a degree or two during it, pool pumps, and heat pumps with thermal mass in the building behind them. A smart electrical panel brings these under one control point, measuring each circuit and switching loads within the limits the owner sets, which also helps older homes with small service capacity add an EV charger or heat pump without a service upgrade. Our smart panel deep dive covers how that control works in practice.

4. The control layer: how thousands of devices act as one

Hardware alone is just clutter. What makes the grid edge a system is a stack of control layers, summarised in the table below. Inside the home, a home energy management system (HEMS) or the smart panel decides how devices cooperate: charge the car after midnight, hold the battery for the evening peak, keep the freezer circuit live no matter what. Above the home, aggregators and utilities run distributed energy resource management systems (DERMS), platforms that forecast how much flexibility a fleet of homes can offer, send dispatch signals, and settle payments afterwards. The forecasts matter as much as the signals: a grid operator will only plan around flexible capacity it believes will show up.

Standards decide whether these layers can talk to each other. IEEE 1547, the interconnection standard for distributed resources in North America, is what requires modern inverters to support grid functions rather than simply switching off. IEEE 2030.5, adopted in California's Rule 21 as the Common Smart Inverter Profile, gives utilities and aggregators a shared language for talking to devices. OpenADR carries price and event signals for demand response. Where these are in place, a battery from one brand can join a program run by another company. Where they are not, owners get stuck inside one vendor's app, which is the single most common complaint in program reviews.

Control at the edge also keeps working when the wider grid does not. A home set up as a microgrid can island itself during an outage, running critical circuits from solar and battery while disconnected. That local intelligence, deciding in the house rather than in a distant control room, is the literal meaning of edge computing applied to power. Our guide to home microgrid design and islanding explains how the disconnection and resynchronisation work.

5. Demand response and virtual power plants

Demand response is the oldest grid edge service: reduce or shift consumption when the system is stressed, in exchange for payment or a better rate. It began with industrial customers and direct load control of air conditioners, and FERC's figures put wholesale market demand response at 33,272 MW in 2024. A virtual power plant is what happens when demand response grows up and merges with generation and storage. In DOE's definition, a VPP is an aggregation of distributed resources, rooftop solar with batteries, EVs and chargers, smart appliances and flexible building loads, coordinated to provide grid services at the scale of a traditional power plant: peak capacity, energy, and balancing services such as frequency response.

A grid event inside a VPP runs in five steps. First, the operator or aggregator forecasts a tight period, usually a hot late afternoon, a day or a few hours ahead. Second, enrolled devices receive a dispatch signal through the aggregator's platform. Third, the devices respond inside the limits each owner set: batteries discharge, EV charging pauses or slows, thermostats ease off, water heaters wait. Fourth, meter and device data verify what each home actually delivered against its normal baseline. Fifth, the owner is paid, as a bill credit, a per-event payment or a capacity payment spread over the season, depending on the program. The owner keeps an override in every mainstream program, and using it usually reduces the payment for that event rather than ending enrolment. The full mechanics are in our explainer on demand response programs and virtual power plants.

The scale is still early, which is precisely why DOE wrote a commercial liftoff report about it. Wood Mackenzie counted 37.5 GW of capacity enrolled in North American virtual power plants in its latest market figures, with California, Texas, New York and Massachusetts accounting for 37% of deployments, and residential customers making up just over 10% of VPP capacity in wholesale markets. Most of today's fleet is still commercial demand response and thermostats. The household battery and EV fleet is the growth edge, and the related business of selling grid services from a home battery is covered in our piece on VPP grid support from home batteries. The research frontier beyond today's programs, where homes trade energy with each other under automated rules, is covered in our explainer on transactive energy and grid edge trading.

6. Markets, money and the rules regulators are writing

A 10 kW home battery cannot bid into a wholesale electricity market built for power stations. Aggregation is the bridge, and in the United States the bridge is a federal rule. FERC Order 841 (2018) required wholesale markets to let energy storage participate. FERC Order 2222, issued on September 17, 2020, extended the same right to aggregations of distributed energy resources: regional grid operators must let aggregators pool small resources and bid them into energy, capacity and ancillary service markets, with a minimum aggregation size no larger than 100 kW and no minimum size at all for an individual device. One hundred kilowatts is a neighbourhood-scale number, as the worked example in the next section shows. Grid operators have been filing and phasing in their compliance plans since, at different speeds in different regions, so whether a home can reach wholesale markets today depends on the state and the operator. Most households still meet the grid edge through the retail door: utility demand response and battery programs, aggregator offers bundled with hardware, and state pilots and mandates of the kind FERC catalogued in Virginia.

Rates are the second rulebook. Time-of-use pricing pays edge behaviour every day rather than only during events: charge the car overnight, run the dishwasher after the peak, let the battery cover the expensive hours. Our time-of-use rate strategy guide works through the arithmetic. Export rates are the contested third rulebook. Classic net metering credited rooftop exports at the full retail rate; many jurisdictions have moved or are moving toward net billing, where exports earn less than imports cost, which pushes solar homes toward batteries and self-consumption. The state-by-state picture is in our net metering policy comparison.

Regulators care about the edge for three practical reasons. It can defer expensive concrete: if flexible load relieves a constrained feeder, a substation upgrade or a new peaker can be postponed or avoided, an approach planners call non-wires alternatives. It adds reliability exactly where failures hurt, on the local network during heat waves and storms. And it creates a cost allocation problem they cannot dodge: when solar and battery homes buy less energy, the fixed cost of the shared grid lands on fewer kilowatt-hours, many of them consumed by households that cannot afford edge hardware. Program design, export rates and fixed charges are where that argument is being settled, state by state.

7. What this means at your house

The numbers become clearer with one house and one event. The table below is an illustration built from stated assumptions, not a measurement: a summer peak event from 6 p.m. to 9 p.m., a home with a battery, an EV on a Level 2 charger, central air conditioning and an electric water heater, all enrolled and responding.

Add the steady items together and one responding house can swing its draw on the grid by roughly 13 to 14 kW during an event, before counting the water heater. Eight such houses acting together pass 100 kW, the largest minimum aggregation size FERC Order 2222 allows a grid operator to set. A thousand houses are a 13 to 14 MW resource, the size of a small peaking unit, assembled from equipment that was bought for entirely private reasons. That arithmetic is the whole grid edge argument in miniature.

A grid edge readiness checklist

• Confirm you have a smart meter. In the US the odds are in your favour (76.8% penetration in 2023, per FERC), and your utility account page will say which meter type you have.
• Pull a month of interval data from your utility portal and find your home's real peak hours. Every later decision gets easier with that picture in hand.
• Check your rate options. If a time-of-use rate exists, compare it against your interval data before switching; flat-rate households with evening-heavy use can lose money on the wrong plan.
• List your utility's demand response and battery programs, including aggregator-run offers for your thermostat, charger or battery brand.
• Before buying hardware, check that the inverter, battery or charger is on the eligible device list of the program you want to join, and that it speaks a standard protocol such as IEEE 2030.5 where the program requires it.
• Read the event terms: how many events per season, the maximum duration, your override rights, and what happens to payments when you override.
• Keep a backup reserve. If your battery joins a program, set a minimum state of charge for outages that the program cannot dispatch below.
• Decide your data position. Interval meter data can reveal when a home is occupied. Read the program's data terms before enrolling, not after.

8. The friction points, stated plainly

Four problems slow the edge down, and they belong on the map alongside the benefits. The first is physical: a feeder or a street transformer built for one-way flow has a hosting capacity limit, and interconnection studies and queues decide how much solar and how many chargers a neighbourhood can add before upgrades are needed. The second is interoperability: devices from different brands still enrol through different apps and protocols, and a program change can strand a perfectly good battery outside a new platform. The third is trust and measurement. Paying for flexibility means agreeing on a baseline, what the home would have used, and baselines can be gamed; more connected devices also widen the grid's cyber attack surface, which is why utilities vet aggregator platforms hard. The fourth is equity. Edge hardware concentrates in owner-occupied homes with capital to spend. DOE's liftoff work names equitable access to distributed resources as a barrier in its own right, and programs that only reward households which already own batteries will widen the cost allocation problem described above rather than solve it. None of these are reasons the shift stops. They are the list of things the next round of rules, standards and program designs has to answer.

Sources and references

• US Department of Energy, Pathways to Commercial Liftoff: Virtual Power Plants (2023, updated January 2025). Source for the VPP definition, current US VPP capacity of roughly 33 GW, the 80 to 160 GW deployment target for 2030, the 10 to 20% share of peak load, the estimate of about $10 billion in annual grid cost savings, and the need for more than 200 GW of new peak capacity by 2030.
• Federal Energy Regulatory Commission, 2025 Assessment of Demand Response and Advanced Metering (December 2025, built on EIA Form 861 data). Source for 128.4 million advanced meters in operation in 2023 (76.8% of 167.2 million meters), wholesale demand response of 33,272 MW in 2024 (about 6.5% of RTO and ISO peak demand), and the Virginia VPP pilot example.
• Federal Energy Regulatory Commission, Order No. 2222 (September 17, 2020), summarised in the Sandia National Laboratories issue brief on energy storage and resource adequacy. Source for DER aggregation access to wholesale energy, capacity and ancillary service markets and the 100 kW cap on minimum aggregation size.
• National Renewable Energy Laboratory, Rooftop Solar Photovoltaic Technical Potential in the United States (2016). Source for the 1,118 GW and 1,432 TWh rooftop technical potential, equal to 39% of US electricity sales, with small buildings at about 65% of the total.
• International Energy Agency, Global EV Outlook 2025. Source for global electric car sales above 17 million in 2024 (over 20% of new cars) and US sales growth of about 10% to more than one in ten cars sold.
• Wood Mackenzie virtual power plant market data, as reported in pv magazine USA, October 2026. Source for 37.5 GW of capacity enrolled in North American VPPs, the four leading states' 37% share of deployments, and residential customers at just over 10% of wholesale market VPP capacity.

Keep going

• Demand response programs and virtual power plants for how enrolment, events and payments work in practice.
• Transactive energy and grid edge trading for the research frontier where homes trade energy with each other.
• Time-of-use electricity rate strategy to put a price signal to work before joining any program.
• Home Energy Rebate Index 2026 for the incentives that help pay for edge hardware, counted from the EnergyBS grants database.
• More EnergyBS research in the source-linked guide library.