The electricity grid was designed as a one-way system: large plants generate, wires deliver, meters count, and a monthly bill follows. That model is being pulled apart from both ends. Generation is moving to millions of small, variable sources such as rooftop solar and wind farms, and consumption is gaining large, controllable loads such as electric vehicles and heat pumps. Keeping supply and demand balanced every second across that landscape is impossible without knowing what every part of it is doing, and increasingly without being able to tell parts of it what to do.
That is what makes the smart grid an IoT problem. Meters, sensors, inverters, chargers and batteries become connected devices; utilities and aggregators become fleet operators; and the data and control loops look like any large IoT system, with the added constraint that getting it wrong can turn the lights off.
This guide covers the main layers: metering, grid monitoring, distributed resources, and the flexibility services that tie them together.
Smart metering: the foundation
A smart meter measures consumption at short intervals, typically every fifteen minutes to an hour, and reports it to the utility without a reader visiting. Many also report voltage and outage events and can be read, configured and in some markets disconnected remotely. The collective system of meters, communication network and head-end software is called advanced metering infrastructure.
Metering networks are among the largest IoT deployments in existence. Communication choices vary by country and utility: power-line communication over the supply cables, sub-GHz radio meshes based on standards such as Wi-SUN, cellular using NB-IoT or LTE-M, and combinations where a neighborhood concentrator aggregates meters and backhauls over cellular. Application-layer standards such as DLMS/COSEM define how meter data is structured and read, which lets utilities mix meter vendors.
For the utility, metering data replaces estimated bills, locates outages faster, reveals losses and theft, and gives voltage visibility at the edge of the network it never had. For the customer, interval data enables time-of-use tariffs and, where regulation allows, access to their own data for third-party services. Meters are also a security concern: a remotely controllable disconnect at scale is a critical infrastructure target, which is why metering components are among the categories treated strictly under the EU Cyber Resilience Act and why meter communications are end-to-end authenticated.
Monitoring the grid itself
Transmission networks have long been instrumented and operated through SCADA systems. Distribution networks, the medium- and low-voltage wires that reach homes and businesses, mostly were not, because there was nothing to control there. Now there is.
Sensors on distribution transformers, feeders and substations report load, temperature, voltage and fault events, giving operators a picture of where the network is stressed by solar export at midday or EV charging at night. Fault indicators on overhead lines shorten outage location from hours of driving to minutes. Substation automation under IEC 61850 brings standardized device models and communication to the equipment that switches and protects the network, which makes it manageable as a fleet rather than a collection of bespoke installations.
Wind and solar farms are themselves dense IoT sites: every turbine and inverter reports performance and health continuously, and condition monitoring of turbine drivetrains is one of the clearest predictive maintenance cases in any industry, since a crane visit to an offshore turbine is extraordinarily expensive.
Distributed energy resources as connected devices
Rooftop solar inverters, home and commercial batteries, EV chargers, heat pumps and smart thermostats are all network-connected by default, usually to their manufacturer’s cloud. Individually they are appliances. Collectively they are a power plant that can be dispatched, if someone can talk to them.
Standards exist for that conversation. IEEE 2030.5 and the SunSpec models define how inverters and chargers expose their state and accept setpoints; OpenADR defines how a grid operator or aggregator signals a demand response event to participating sites; OCPP governs communication between EV chargers and their management systems. Several markets now require new inverters to be controllable, so that grid operators can curtail export when the network is congested.
The architectural pattern is an aggregator platform sitting between the grid operator and thousands of devices, translating a market signal (“reduce 5 MW in this area for the next hour”) into individual instructions (“pause this charger, discharge that battery, raise that thermostat by one degree”), verifying delivery through metering, and settling payments. This is what a virtual power plant is: an IoT fleet management problem whose output is sold on an electricity market.
Flexibility: the use case that matters
The value of all this converges on flexibility: the ability to shift consumption or generation in time and place to match what the network can carry and what renewable supply provides. Charging EVs overnight when wind is abundant rather than at six in the evening when everyone gets home; pre-heating buildings before a price peak; discharging batteries during a shortfall; curtailing solar export when a feeder is full.
Flexibility is where the IoT layers meet. The meter proves what happened. Grid sensors say where flexibility is needed. Connected devices deliver it. Tariffs and markets provide the incentive. An energy management system in a building or home sits at the customer’s end, optimizing against prices and signals while respecting comfort and the owner’s rules.
The engineering constraints are specific. Control signals must be delivered reliably within minutes and acknowledged. Devices must fail safe: a charger that loses its connection should keep charging, not stop, and a heat pump should keep heating. Customer override must always be possible. And the data that makes optimization work, which includes occupancy patterns and vehicle usage, is personal and must be handled as such.
Buildings, industry and the demand side
Commercial buildings use building automation systems and HVAC controls that were connected long before “IoT” existed, speaking BACnet and Modbus. Adding submetering, occupancy sensing and cloud analytics turns them into flexible loads and reveals waste. Industrial sites with large motors, furnaces or cold stores can shift substantial load when the process allows, and energy monitoring within Industrial IoT deployments is often what makes that visible.
Security and resilience
Energy infrastructure is a national security concern, and its IoT layer inherits that. Device identities and signed firmware for meters and inverters, network segmentation between operational and corporate systems, IEC 62443 for operational technology, and strict control over who can issue commands at scale are the baseline. A fleet of remotely controllable inverters is a target precisely because it is useful, and the same aggregator that can reduce 5 MW on request could, if compromised, do it without request. Designs that rate-limit commands, require multi-party authorization for large actions, and keep local safe defaults are the answer.
Where to go next
- Smart cities: what IoT actually does for a city covers municipal utilities and street lighting, which overlap with grid-side deployments.
- Industrial IoT: a practical guide explains the operational technology context that substations and plants share.
- IoT security fundamentals covers the controls critical infrastructure depends on.
- Edge computing explained shows why local control loops must survive loss of connectivity.
Topics: energy, smart grid, smart meters, demand response, distributed energy