Home Assistant Energy Dashboard: Track Your Home's Power Usage (2026)

Disclosure: Some links in this article are affiliate links. We may earn a commission on qualifying purchases at no extra cost to you.

Most smart home dashboards show only the devices that are online; the Home Assistant Energy Dashboard instead shows what those devices are actually drawing from the wall — in real time and over the long term, without routing the data through a vendor cloud. For a local-control home, that visibility is the difference between guessing where power goes and knowing it.

Since Home Assistant 2025.12, the Energy Dashboard has outgrown its reputation as a historical kWh archive. It now accepts live power sensors alongside accumulated energy sensors, which adds real-time W/kW graphs and flow visualization on top of the long-term consumption, cost, and statistics views. Home Assistant 2026.6 then added battery state-of-charge display and custom names for grid, solar, battery, gas, and water sources. The underlying requirement never changed, though: the right entity metadata is what makes the dashboard work at all.

This guide covers the entity model first — the difference between power and energy, and why it matters more than any setup screen — then walks through a clean setup workflow, matches each measurement layer to local-first hardware, and closes with the automations that make live power data useful. The angle throughout is compatibility and local control: where data comes from, how it reaches Home Assistant, and whether a vendor cloud is in the path.

It is written for anyone running Home Assistant with at least one energy-reporting device, from first-sensor beginners adding a single smart plug to advanced users wiring clamp meters into a panel. No single device or protocol is required; the right stack depends on which loads need to be seen.

Quick Reference: Power vs Energy

If a device only exposes W/kW, an Integration / Riemann-sum helper can derive accumulated Wh/kWh from the discrete samples. The helper's sampling interval and integration method affect accuracy, so it is a fallback — a native cumulative energy entity is always preferable when the hardware provides one.

How the Energy Dashboard Works

The Energy Dashboard does not generate data. It aggregates entities that integrations already expose, which means most setup failures are entity problems, not dashboard problems. Two distinct quantities matter:

Energy (kWh / Wh) is accumulated usage over time. It drives historical consumption, cost, and long-term statistics. An eligible energy entity uses device_class: energy, reports a Wh or kWh unit, and carries state_class: total or state_class: total_increasing. Both are accepted for accumulated electricity; total is not wrong.

Power (W / kW) is instantaneous draw. It uses device_class: power, state_class: measurement, and a W or kW unit. Since Home Assistant 2025.12, power sensors can be configured alongside energy sensors, which is what enables live flow visualization rather than only historical totals.

A W-only device can still contribute accumulated energy. Home Assistant's Integration helper — often called the Riemann-sum helper — integrates the instantaneous samples into a Wh/kWh entity. The compromise is accuracy: Home Assistant is integrating discrete measurements, not a continuous signal, so the sampling interval and method influence the result.

The dashboard's source categories are grid consumption, grid export/return, solar production, home battery storage, gas, water, and individual devices. Once sources are configured, the dashboard generates automatically and individual cards can be shown or hidden independently. For accumulated energy, selecting a valid Wh/kWh entity is the whole game.

What You Need Before You Start

Four infrastructure prerequisites matter before touching the setup screens:

• A running Home Assistant instance, any supported install method. The dashboard is built in, accessible from Settings → Dashboards → Energy.
• The right radio, matched to your hardware. Zigbee metering devices need a ZHA or Zigbee2MQTT coordinator. Z-Wave devices need a Z-Wave stick running Z-Wave JS. Matter-over-Thread devices need both a Matter controller and a Thread Border Router. Wi-Fi devices such as Kasa or Shelly need no coordinator at all, but their integration path matters (see below).
• A clear picture of which measurement layer you need. Whole-home CT monitors reveal total demand and grid/solar flow; branch CTs reveal individual circuits; smart plugs reveal specific appliances. The most useful deployment often combines layers — Home Assistant's upstream-device hierarchy exists precisely for that nesting.
• An honest read on local versus cloud. "Home Assistant compatible" is not the same as "local." Shelly talks to Home Assistant over the LAN with a Local Push integration; Z-Wave JS is Local Push; TP-Link and IoTaWatt are polled locally; Zigbee, Z-Wave, and Thread are local radio stacks. Emporia's stock Vue firmware, by contrast, routes measurements through Emporia's cloud and exposes no local API.

One safety boundary belongs here too: clamp and DIN-rail monitors involve mains equipment and electrical panel work. That is a materially different risk level from plugging in a smart outlet. Follow local electrical code and use a qualified electrician where panel work is required.

Step 1 — Add Hardware and Verify Energy Entities

Add the hardware integration first, under Settings → Devices & services. The correct integration depends on the device family: native TP-Link or Shelly integrations for Wi-Fi hardware, ZHA or Zigbee2MQTT for Zigbee, Z-Wave JS for Z-Wave, ESPHome for flashed devices, Matter for Thread-based hardware, and the native IoTaWatt integration for its monitors.

Then verify the exposed entities under Settings → Tools → States. A device that measures power will usually expose multiple entities — live power in W, accumulated energy in Wh/kWh, voltage, current, and sometimes power factor. The dashboard only accepts the ones with the correct unit and state class, which is why an entity that appears in States but never in the Energy setup is almost always a metadata problem, not a missing device.

For historical energy and cost, select a valid Wh/kWh entity. If the integration only surfaces a W/kW measurement, create an Integration/Riemann-sum helper when accumulated energy is required — and confirm a native cumulative entity does not already exist before doing so, because converting power twice is a classic error.

Step 2 — Configure Grid, Solar, and Battery Sources

Open Settings → Dashboards → Energy and add sources in order: grid consumption plus grid return where applicable, solar production, battery charge and discharge, then gas and water if those exist in the deployment.

The cleanest reference for correct entity selection is the official IoTaWatt integration mapping, which documents exactly which calculated outputs belong where:

• MainsConsumption.wh → grid consumption
• MainsExport.wh → grid return / export
• Solar.wh → solar production

For solar inverters, prefer inverter-native cumulative/lifetime energy entities over reconstructing production from sampled instantaneous wattage. Enphase Envoy/IQ Gateway and Fronius integrations expose authoritative totals locally; the Enphase path can require Enphase Enlighten credentials to renew an authentication token even though normal telemetry is read over the LAN. That distinction between a local data path and zero cloud dependency matters throughout this setup.

One explicit trap: Home Assistant warns against summing multiple independently resettable utility_meter helpers into a single source. When their reset moments differ, the transient false total can pollute long-term statistics. Cumulative sources should be added separately, not fused into one helper.

Step 3 — Add Live Power Monitoring

Since Home Assistant 2025.12, power sensors are configurable alongside energy sensors, which upgrades the dashboard from a historical tool into a live one. Where the hardware exposes W/kW entities, add them for grid import and export, solar, battery, and individual devices. The result is real-time power graphs and flow visualization layered on top of the accumulated history.

Keep the two quantities separate in purpose. Live power entities drive automations — a load crossing a threshold, a pump drawing more than normal, a device falling to standby. Energy entities drive usage and cost over days, months, and years. Using a power sensor where an energy entity is expected, or the reverse, is the most common source of empty or absurd cards.

The individual-device cards also support usage ranking. On device-specific energy cards, the hide_compound_stats option keeps compound and upstream statistics from crowding the view when a circuit meter and its downstream appliance meters are both present.

Step 4 — Add Individual Devices and Upstream Hierarchies

Individual-device monitoring accepts smart plugs, smart relays, and any compatible power/energy entities. Add appliance-level meters here so specific loads become visible instead of disappearing into a whole-home total.

The upstream-device hierarchy is the mechanism that prevents double-counting when both a circuit meter and its downstream appliance meters exist. Declaring the breaker/circuit meter as the parent of the appliance-level meters removes the duplicated downstream consumption from the compound total. All related individual-device entities must be added before assigning the hierarchy, and the parent relationship is configured from the individual-device section.

If totals still look corrupted or an entity reports statistics errors, check Settings → Tools → Statistics. The most persistent causes of corrupted totals are not the dashboard itself, but incorrect state classes, double conversion of power into energy, summed resettable helpers, and unmanaged upstream/downstream overlap. For readers building a dedicated metering stack from scratch, the best smart plugs for local control and the Zigbee/Z-Wave plug breakdown cover coordinator-based options in detail.

Smart Plugs with Energy Monitoring — Local-First Picks

A smart plug is the fastest way to get real per-appliance data into Home Assistant, but the protocol and integration path decide whether that data stays local.

TP-Link Kasa EP25 / EP25P4

Specs: Wi-Fi 2.4 GHz; 15 A / 1.8 kW maximum at 120 V; 1/6 HP motor rating; energy monitoring with consumption entities; UL listed with V-0 flame-retardant housing; Apple HomeKit support in addition to Kasa's wider voice-platform compatibility. Home Assistant's TP-Link integration polls the hardware locally after initial provisioning.

Budget Wi-Fi plug with energy monitoring.

Check Price on Amazon

Pros

• Polls locally in Home Assistant once provisioned
• HomeKit support alongside Kasa's wider ecosystem
• Four-pack configuration makes per-outlet monitoring an inexpensive add

Cons

• 2.4 GHz Wi-Fi only
• Initial provisioning can still involve vendor credentials or an app
• Not the pick for a strict no-account-ever deployment

Verdict: The strongest current Kasa choice for US readers who want Wi-Fi energy monitoring without a Zigbee or Z-Wave coordinator. It supersedes the aging KP115, which remains documented but sits at legacy-generation status.

Perfect for: Renters and Home Assistant beginners adding per-outlet energy data on an existing Wi-Fi network.

Aqara Smart Plug US (ZNCZ12LM)

Specs: Zigbee 3.0; 100–125 V AC; 15 A / 1,875 W maximum; energy monitoring with daily and monthly consumption records; overheat and overload protection; operating range −10°C to 40°C. Home Assistant ZHA can pair many standard Zigbee devices directly through a supported coordinator, replacing the proprietary Aqara hub where the device's cluster implementation is supported.

Budget-to-mid-range Zigbee energy plug.

Check Price on Amazon

Pros

• Direct ZHA / Zigbee2MQTT pairing without a vendor hub in typical HA setups
• Compact, low-power Zigbee end device
• Daily and monthly consumption records exposed

Cons

• Official Aqara ecosystem expects an Aqara hub
• Some features may require ZHA quirks per firmware and device handler
• Zigbee coordinator required — it is not a standalone Wi-Fi device

Verdict: A reliable Zigbee energy plug for homes that already run ZHA or Zigbee2MQTT. Compatibility per firmware is worth verifying rather than assuming every Aqara feature maps identically.

Perfect for: Zigbee mesh users who want local energy data without another Wi-Fi device or cloud account.

Zooz ZEN15 Power Switch 800LR

Specs: Z-Wave 800-series with Z-Wave Long Range support where the controller and region allow; 120 V / 60 Hz; 15 A / 1,800 W maximum; reports power, accumulated energy, and electrical measurements; ETL-certified; designed for heavier appliance use than many miniature plugs. Integrates locally through Home Assistant Z-Wave JS, which is classified Local Push.

Mid-range Z-Wave appliance plug.

Check Price on Amazon

Pros

• Fully local Z-Wave JS control — no cloud path for routine operation
• 800-series radio improves range and battery/telemetry behavior
• Configurable reporting thresholds and intervals
• Built with appliance loads in mind

Cons

• Requires a Z-Wave controller — not useful without one
• Aggressive power/voltage/current reporting can consume Z-Wave bandwidth
• Current retail availability is thinner than mainstream Wi-Fi plugs

Verdict: The strongest Z-Wave energy plug covered here, and the right shape for freezers, pumps, and other loads that exceed what a miniature Wi-Fi plug should handle. Tune telemetry to the automation need instead of setting every channel to report constantly.

Perfect for: Z-Wave homes monitoring appliances, especially where Z-Wave Long Range reaches detached loads.

Eve Energy (Matter)

Specs: Matter smart plug over Thread; US 120 V / 60 Hz; 15 A / 1,800 W maximum; measures electrical consumption; Thread Full Device, so the mains-powered hardware participates in the Thread mesh; requires a Matter controller and Thread Border Router; supports Home Assistant alongside Apple Home, Alexa, Google Home, and SmartThings.

Premium Matter-over-Thread plug.

Check Price on Amazon

Pros

• Thread Full Device strengthens the mesh rather than being a passive endpoint
• Local operation without an Eve cloud or account dependency
• Standards-first path across multiple controller ecosystems

Cons

• Requires a Matter controller and Thread Border Router — infrastructure non-negotiable
• Higher cost than Zigbee or Wi-Fi equivalents
• Energy clusters must actually be implemented on the specific firmware; verify per model

Verdict: The premium standards-first example for readers who already have a healthy Thread network. Its main disadvantage versus Zigbee or Wi-Fi plugs is ecosystem prerequisites and cost, not cloud dependence.

Perfect for: Established Thread networks where a Matter-native plug fits the existing infrastructure.

IKEA GRILLPLATS

Specs: Current IKEA US smart plug with energy-usage tracking; Matter-compatible over Thread; requires a Thread Border Router for Matter/Thread operation; can join IKEA DIRIGERA or other Matter ecosystems; US 120 V / 15 A / 1,800 W rating.

Budget Matter-over-Thread plug, sold through IKEA retail channels rather than Amazon's own inventory.

Check Price on Amazon

Pros

• Energy tracking at an aggressive price point from IKEA retail
• Matter-over-Thread positions it as the forward-facing successor to INSPELNING
• Joins DIRIGERA or other Matter ecosystems

Cons

• Requires Matter controller support plus a Thread Border Router — "Matter" does not mean "no hub/radio infrastructure"
• Not carried by Amazon; purchase runs through IKEA retail

Verdict: The better current IKEA choice over hunting for remaining INSPELNING inventory. The older INSPELNING remains useful mainly as a legacy compatibility example, including community reports of energy updates stalling after restarts until reconfiguration.

Perfect for: Budget-conscious Matter homes already running a Thread Border Router, or existing IKEA DIRIGERA users.

A note on Shelly plugs

Shelly plug hardware is repeatedly favored by Home Assistant users for LAN operation and zero reliance on a vendor cloud. The current Shelly Plug S MTR Gen3 examined here, however, is a European Type E/F unit rated 220–230 V — it is not a drop-in US recommendation. For US readers, Shelly's clamp-based monitors below are the relevant part of the lineup, and they carry the same Local Push integration.

Whole-Home and Circuit-Level CT Monitors

Clamp meters answer a different question from smart plugs. A mains CT reveals total demand and grid/solar flow; branch CTs reveal individual circuits. They require panel access and, in many cases, a qualified electrician — but they capture loads no plug can see.

Shelly Pro 3EM 120A v2

Specs: DIN-rail, single- or three-phase metering; three included 120 A CTs; measures up to 120 A per phase and neutral; four-quadrant measurement; Class B active-energy compliance to IEC 62053-21 based on manufacturer testing; accuracy to approximately ±1% across 2–120 A, ±2% from 1–2 A, ±5% below 1 A; active/apparent power and energy, power factor, and fundamental active/reactive energy; PV-ready and bidirectional; Wi-Fi and Bluetooth; MQTT and scripting; Matter upgrade capability with current firmware; approximately 45–60 days of retrievable one-minute records.

Premium local-first whole-home monitor.

Check Price on Amazon

Pros

• Native Home Assistant Shelly integration with Local Push — no reflash required
• Four-quadrant, bidirectional measurement suits solar and grid-export homes
• Three-phase support for 240 V / three-phase panels
• MQTT, local API, and scripting for advanced use

Cons

• Panel space, DIN-rail install, and mains wiring complexity
• Three channels of granularity, not Emporia-style 16-circuit coverage
• A matching Amazon US listing for the v2 is not consistently available; the legacy 120 A Pro 3EM is a different SKU

Verdict: The strongest local-first whole-home option in this guide for users who want LAN/Wi-Fi, three-phase metering, and an official native HA path without reflashing hardware.

Perfect for: Whole-home or three-phase monitoring where a DIN-rail install is acceptable and solar bidirectional metering matters.

Shelly EM Gen3

Specs: Current-generation clamp-based Shelly energy monitor; 50 A and 80 A CT configurations depending on SKU and setup; measures active/apparent power, energy, power factor, and related electrical values; documented ammeter accuracy around ±2% through much of the 1–50 A range, degrading below approximately 1 A; one-minute local data history; Wi-Fi and Bluetooth; MQTT, local RPC/API, and scripting; native Shelly integration is Local Push.

Mid-range CT monitor for single-circuit or dedicated-load measurement.

Check Price on Amazon

Pros

• Local HA integration with no cloud path for operation
• Flexible CT sizing for a single circuit or dedicated load
• MQTT and scripting match the rest of the Shelly ecosystem

Cons

• One to two channels, not whole-home branch coverage
• Accuracy degrades under roughly 1 A
• Panel installation raises the same safety considerations as any clamp meter

Verdict: The right Shelly family for readers who need clamp-based monitoring of a single circuit or load rather than a switchable appliance plug. Installation in or around a panel is a different safety level from installing a smart outlet.

Perfect for: Monitoring one large circuit, a subpanel, or a dedicated appliance with a CT clamp.

Emporia Vue 3

Specs: Whole-home and circuit-level CT energy monitor; up to three 200 A main CT channels; up to sixteen 50 A branch channels; supports single-phase two-wire, split-phase three-wire, and three-phase four-wire Wye (Delta unsupported); maximum 264 V AC line-to-neutral sensing; Wi-Fi 2.4 GHz with 10/100 Ethernet; approximately ±2% CT accuracy in manufacturer testing; UL 61010 and UL 2808 certified. Stock firmware is cloud-based: Vue sends measurements to Emporia's cloud, and no local device API is available on Vue Gen 1, 2, or 3.

Premium high-density circuit monitor — cloud architecture out of the box.

Check Price on Amazon

Pros

• Sixteen branch channels plus mains in one enclosure — unmatched granularity at this integration level
• UL certifications and documented accuracy
• Ethernet option avoids Wi-Fi-only limitations

Cons

• Stock firmware routes data through Emporia cloud with no local API
• Truly local Home Assistant use requires a community ESPHome flash, which Emporia does not support
• Flashing hardware changes the architecture and carries its own compatibility and firmware risk

Verdict: The most circuit granularity per enclosure, but the wrong architecture for a local-first home out of the box. Power users repeatedly accept the ESPHome conversion because the hardware itself is attractive — that decision belongs to the reader, not presented as a manufacturer feature.

Perfect for: Owners willing to flash community ESPHome firmware, or who accept cloud monitoring for maximum branch coverage.

IoTaWatt Open WiFi Electricity Monitor

Specs: Purpose-built local energy monitor with 14 CT inputs; external AC voltage-reference transformer; multi-circuit, multi-phase, and bidirectional/solar configurations; local browser-based configuration and data handling; official Home Assistant integration exposing both input CTs and calculated IoTaWatt Outputs; Local Polling IoT class.

Premium local-from-factory multi-channel monitor, sold through specialist channels.

Check Price on Amazon

Pros

• Local-native from the factory — no cloud, no reflash
• Fourteen CT inputs cover mains plus many individual circuits
• Official HA integration documents exact grid/solar/export outputs

Cons

• Specialist direct purchase; Amazon availability is not reliable
• Steeper setup than a consumer plug-in device
• Panel work and CT placement still require care

Verdict: The cleanest local-native multi-circuit monitor in this guide. When local-from-day-one matters more than the lowest purchase cost, IoTaWatt is the recurring recommendation in Home Assistant communities.

Perfect for: Local-control purists and solar/net-metering homes that want documented HA integration and no cloud path.

CircuitSetup Expandable 6-Channel ESP32 Energy Meter

Specs: DIY/open-hardware multi-channel energy-meter platform; main board provides six CT channels; expansion boards scale to as many as 42 CT inputs; ATM90E32-family metering ICs; exposes active/reactive/apparent power, energy, power factor, frequency, and related measurements; ESP32/ESPHome local operation; Wi-Fi or Ethernet depending on build.

Mid-range DIY platform for high-channel-count builds.

Check Price on Amazon

Pros

• Very high channel ceiling for the enclosure cost
• ESPHome-native, fully local, no vendor cloud
• Open hardware with an active project ecosystem

Cons

• Demands substantially more electrical and calibration knowledge than a commercial CT monitor
• Final accuracy depends on CT choice, calibration, voltage reference, and wiring — the IC's best-case spec is not a system guarantee
• Stock fluctuates; it is a specialist build, not a consumer pick

Verdict: Fills the niche between commercial CT monitors and fully home-built metering. It is the right move for an experienced user who wants many circuits, ESPHome, and no cloud — and is willing to own the calibration.

Perfect for: Experienced Home Assistant tinkerers building a 12-to-42-channel local monitor.

Reading Your Utility Meter Directly

When the utility meter exposes a readable interface, direct reading can be cleaner than adding another measurement layer — it captures the value closest to the billing meter.

Home Assistant Glow is an open-source ESPHome project that reads the blinking pulse LED present on many electricity meters. It does not require opening the meter or tapping its circuitry. An ESP32 is the recommended controller, and the user configures the meter's pulse constant — typically printed as imp/kWh; 1,000 impulses/kWh is a common default but must match the physical meter, because entering the wrong constant creates a correspondingly large scale error. It is a DIY build, not a standardized retail product, and remains actively maintained.

frient Electricity Meter Interface 2 LED is the battery-powered Zigbee alternative: a Zigbee 3.0 optical/LED interface that counts utility-meter pulses with configurable scaling from roughly 50 to 10,000 pulses/kWh, two AA batteries rated around two years under stated conditions, and IP22 rating. Zigbee device support and pulse-scaling should be verified for the exact coordinator and integration before purchase.

WattWächter Plus is a niche but notable example of where direct smart-meter optical access is heading: Home Assistant gained a native integration in 2026.7 for reading compatible meters through SML/OBIS, exposing total consumption, grid feed-in, active power, phase voltage/current, frequency, and power factor over a local HTTP API with no cloud involved. It is most relevant to SML/OBIS meter markets rather than as a mainstream US recommendation.

The protocol caveat that applies across all of these: ZHA does not support devices that only use the Zigbee Smart Energy (ZSE) profile, and "Zigbee metering" is not interchangeable with "Zigbee Smart Energy." Standard plugs exposing electrical measurement clusters work through ZHA; ZSE-only utility hardware does not.

Common Mistakes and How to Fix Them

Automations That Use Live Power Data

Live power entities are where the Energy Dashboard turns from a monitoring tool into an automation engine. These patterns recur across Home Assistant communities because they act on thresholds, not history:

• Washer/dryer completion: trigger when live power falls below a threshold for several minutes — the classic end-of-cycle signal.
• Phantom-load detection: alert or switch off devices that remain above a nighttime standby baseline.
• High-demand alerts: trigger when total household load exceeds a defined threshold.
• Solar-surplus control: start a discretionary load such as an EV charger, water heater, or dehumidifier when grid export exceeds a threshold.
• Tariff shifting: run flexible loads during low-rate periods using scheduled automations tied to tariff windows.
• Anomaly detection: alert when a refrigerator, pump, server, or HVAC unit starts consuming materially more or less than its normal pattern.
• Consumption verification: compare the sum of measured circuits against mains CT totals to identify unmonitored loads.

The 2026-era nuance is consistent across all of them: automations use live power entities in W, while long-term usage and cost analysis uses energy entities in Wh/kWh. The same reporting-frequency rule applies on the automation side — capture meaningful wattage changes, not every watt, or the mesh pays the cost. For deeper automation patterns, the Home Assistant automations guide covers the mechanics, and the dashboard ideas article shows how to surface the results.

Frequently Asked Questions

Conclusion

The Home Assistant Energy Dashboard is no longer just a historical kWh archive. With 2025.12's live power support and 2026.6's battery and naming additions, it is the closest thing to a real-time local energy monitor that ships inside Home Assistant — provided the entity model is right. The recurring theme of every troubleshooting case is the same: W and kWh are different, state classes matter, and summing helpers or overlapping layers quietly corrupts totals.

Hardware choice is where local control is won or lost. Zigbee and Z-Wave plugs, Shelly's LAN monitors, IoTaWatt, and ESPHome builds all keep data in the home. Emporia's Vue 3 offers the best circuit density but starts cloud-based, and it only becomes local through an unsupported firmware conversion. The right move is to pick hardware by data path first and channel count second, then match the measurement layer — whole-home, circuit, or appliance — to the questions that matter.

For most readers, the practical starting point is a four-pack of energy-monitoring plugs on the existing protocol, then expanding to clamp monitors only where whole-home or branch visibility is worth panel work. The best smart plugs for local control, the energy-monitoring plug comparison, and the Zigbee vs Z-Wave vs Matter breakdown all cover the hardware decision in more depth.

→ Check the current price on Amazon