Best Home Assistant Energy Monitors: A 2026 Buying Guide

Almost every "best energy monitor for Home Assistant" list is written for a US reader with a breaker panel and no smart meter data port. If you live in the Netherlands, Belgium, Germany, or most of the rest of Europe, that advice will talk you into a €150 clamp meter when a €30 dongle would have done the job better. This guide fixes that.

Below is what each of the popular options actually costs, how it talks to Home Assistant, how fast it updates, and where it makes sense. Prices are approximate and move around a lot — treat them as a ballpark for budgeting, not a quote.

1. Whole-Home vs Per-Device: You Need One of Each, Not Ten

Every energy monitoring setup answers one of two different questions, and people waste money by buying the wrong tool for the question they actually have.

Whole-home monitoring answers "what is this house drawing from the grid right now, and what did it cost today?" This is the number that drives your bill, your tariff comparison, your solar self-consumption calculation, and every peak-shaving automation you will ever write. You need exactly one of these, and it needs to be accurate and reliable.

Per-device monitoring answers "which appliance is responsible?" This is what turns a suspicious 300 W baseline into "the old freezer in the garage." You need a handful of these, and they can be cheap and approximate.

The mistake I see most often is people skipping the whole-home number and buying twelve smart plugs instead. Plugs cannot see your oven, your heat pump, your lights, or your EV charger, so the sum of your plugs will never match your bill. Get the grid total first. Add plugs for the three to six loads you actually intend to automate.

2. If You're in the EU, Start With Your P1 Port

This is the single most useful and most overlooked answer in home energy monitoring, and it is the reason US-written buying guides mislead European readers.

If your smart meter was installed in the last decade in the Netherlands, Belgium, Luxembourg, or increasingly Germany and Austria, it very likely has a P1 port — an RJ12 socket on the front of the meter that streams a plain-text telegram of your readings, typically once per second. Import, export, per-phase power, tariff registers, and on Dutch meters usually your gas reading too. Your utility already paid for the most accurate energy meter in your house. The port exists so you can read it.

What that means in practice:

• No electrician, no consumer-unit work. You plug a cable into a socket. There is nothing to install near live busbars and nothing to size incorrectly.
• Billing-grade accuracy. Your readings come from the meter the utility bills you against, so your dashboard and your invoice agree. Clamp meters are typically 1–2% off and drift with placement.
• Export and gas for free. Solar feed-in and gas consumption come in the same telegram. Most clamp installs need extra hardware for export direction and cannot see gas at all.
• Roughly €30. A tenth of a Sense, a fifth of a Shelly Pro 3EM install.

There are three common ways to read it. A HomeWizard Wi-Fi P1 Meter is the plug-and-play route: it has a documented local API and a first-party Home Assistant integration, so you can be up in five minutes with no YAML. A SlimmeLezer+ is the enthusiast route: a small ESP-based reader that runs ESPHome, so it appears in Home Assistant natively over the local API and you can change absolutely anything about it. A generic P1/DSMR USB dongle is the cheapest route: a cable from the meter to a USB port on your Home Assistant machine, read by the DSMR smart meter integration, which is fine if your server happens to live near the meter.

The honest headline: if you have a P1 port, a €30 dongle makes a €150 clamp-meter install pointless for most people. Clamps become worth it only when you want per-circuit detail the meter cannot give you.

3. The Comparison Table

Prices are approximate street prices in euros and exclude installation. "Update interval" is what you can realistically expect in Home Assistant, not a best-case marketing figure.

A few things worth pulling out of that table. The three P1 options are an order of magnitude cheaper than everything below them and update faster than most of it. Shelly's clamp devices are the best local option for anyone without a P1 port. IotaWatt is expensive and hard to buy but is the only many-channel device on this list that is genuinely local and open. And Sense is the only entry with no local API at all.

4. Per-Device Monitoring: When a €12 Plug Beats a €150 Clamp Meter

For anything with a plug on it, a metered smart plug wins on almost every axis. It costs a tenth as much as a clamp meter, installs in two seconds, needs no electrician, gives you a switch as well as a sensor, and reports a cumulative kWh total that drops straight into the Home Assistant Energy dashboard as an individual device.

Choose the protocol on this basis:

• Zigbee for anything you buy more than two of. It does not touch your Wi-Fi network, the mesh gets stronger with each mains-powered plug you add, and pairing is instant in ZHA or Zigbee2MQTT. Ikea, Sonoff, Aqara, and Third Reality all make metered plugs in this bracket.
• Wi-Fi (Shelly) when you want sub-second data or a switch you can hammer with automations. Shelly Plug S at ~€18 reports about once a second, which is fast enough to use as an automation trigger rather than just a log.
• Avoid cheap Tuya-cloud plugs unless you are certain you can flash them or run them via Local Tuya. A plug that needs a vendor server to report watts is a plug that stops working when that server does.

Two real limitations to know before you order ten of them. First, accuracy at low power is bad: most inexpensive plugs are unreliable below about 2–5 W, so they will not measure standby draw honestly. Second, they are current-limited. A plug rated 10–16 A cannot go on an oven, a hob, an instant water heater, or an EV charger — and this is exactly where a clamp meter earns its price. Hard-wired, high-power circuits are the legitimate use case for a Shelly 3EM in the consumer unit, with one clamp per circuit you care about.

5. What to Avoid

Four ways to waste money here, in rough order of how often I see them.

Cloud-only devices with no local API

If the watts have to travel to a company's server and back before Home Assistant sees them, you have bought a latency problem and a shutdown risk. Cloud data cannot reliably trigger automations, it disappears when your internet does, and the vendor can change or withdraw the API without asking you. This is the main reason Sense is a poor fit for Home Assistant, despite being a genuinely clever product: there is no local endpoint, and its headline feature — machine-learning device disaggregation — is a black box you cannot control. It decides what your "Device 7" is, it can take weeks to identify anything, it merges and splits devices on its own schedule, and you cannot correct it. Compare that with a €15 plug that simply knows it is measuring the dryer.

Anything requiring a subscription

A handful of energy monitors gate historical data, export, or API access behind a monthly fee. Walk away. You are buying a sensor, and the whole point of Home Assistant is that the history lives in your own database.

CT clamps sized wrong for your mains

The most common hardware mistake in clamp installs. A clamp has a rated current and a physical opening, and both have to fit. Buy a 50 A clamp for a 3×25 A supply and you will clip the peaks silently — your dashboard will look plausible and be wrong during exactly the events you care about. Buy a clamp whose jaw does not close around a 16 mm² main and you have an expensive paperweight. Check your main breaker rating and your cable diameter first, and size the clamp above your breaker, not above your average load. Also check the direction: a clamp fitted backwards will report export as import.

Buying per-device before whole-home

Already covered, but it is worth repeating because it is the expensive one. Ten plugs cost more than a P1 dongle and still cannot tell you what your house drew last night.

6. What I'd Buy, By Scenario

Apartment, single-phase, no solar

A P1 reader (HomeWizard or SlimmeLezer+) at ~€30, plus two or three metered Zigbee plugs for the dishwasher, the washing machine, and the entertainment corner. Total around €60–70. No P1 port? A single Shelly clamp device on the incoming feed, or just start with the plugs.

House with solar

P1 reader first — and this is where it really pays, because it gives you import and export from the same source, which is what makes self-consumption and feed-in figures actually reconcile. Your inverter integration supplies production. Add plugs or a clamp device on the loads you want to shift into sunny hours: dryer, dishwasher, hot water. Around €60–120 all in.

Three-phase house with an EV and a heat pump

P1 reader for the billing total, plus a Shelly Pro 3EM in the consumer unit for per-phase current. The second device is not redundant here: dynamic load balancing and phase-aware charging need per-phase amps at high resolution, which the meter telegram gives you more coarsely. Budget roughly €180–220 plus an electrician for the DIN-rail install. If you are in North America instead, this is the scenario where Emporia Vue or IotaWatt makes sense, with IotaWatt the better choice if you insist on local data.

Whichever you pick, the hardware is only half the job — the sensors still have to be wired into the Energy dashboard with the right device classes and state classes before any of it becomes useful. Our complete guide to Home Assistant energy monitoring is the natural next step once the box arrives. And if your goal is lower bills rather than prettier graphs, the whole-home sensor you just installed is the input for peak shaving with Home Assistant, which is where the monitoring starts paying for itself.

Frequently Asked Questions

What is the best energy monitor for Home Assistant?

If you live in the Netherlands, Belgium, or another country with a P1-equipped smart meter, the best whole-home monitor is a P1 reader such as the HomeWizard Wi-Fi P1 Meter or a SlimmeLezer+, at roughly €30–45. It reads the meter your utility already installed, needs no electrical work, is fully local, and updates about once a second. If you have no P1 port, a Shelly Pro 3EM in the consumer unit is the strongest option because it is local, three-phase capable, and supported by the core Shelly integration. For per-device monitoring, a metered smart plug is better value than any clamp meter.

Do I need a clamp meter if I already have a P1 dongle?

Usually not. A P1 reader already gives you total import, total export, and near-real-time power for the whole house, which is everything the Energy dashboard needs at the grid level. A clamp meter like a Shelly 3EM only becomes worth the extra cost and the consumer-unit install when you want per-circuit detail the meter cannot give you — separating a heat pump or an EV charger from the rest of the house, or getting per-phase current for load balancing.

Does the Sense energy monitor work with Home Assistant?

Sense data can be brought into Home Assistant, but every reading travels through Sense's cloud account first, so it stops working when your internet or their service goes down and it cannot be polled at local-network speed. The device-level breakdown also comes from machine-learning disaggregation that you cannot correct, retrain, or override, so entities appear and disappear on Sense's terms rather than yours. For a Home Assistant user who wants dependable, local, automatable data, Sense is a poor fit compared with a P1 reader or a Shelly clamp meter.

Can I use cheap Zigbee smart plugs for energy monitoring?

Yes, and for per-device monitoring they are usually the right answer. A metered Zigbee plug costs about €10–20, works locally through ZHA or Zigbee2MQTT, and reports both instantaneous power and a cumulative energy total that plugs straight into the Energy dashboard as an individual device. The limits are that cheap plugs are inaccurate below roughly 2–5 W, their reporting interval is often tens of seconds rather than seconds, and they are rated for around 10–16 A, so they cannot be used for an oven, a heat pump, or an EV charger.

How many energy monitors do I actually need?

Two kinds, not ten devices. One whole-home monitor gives you the grid total that every cost calculation, tariff comparison, and peak-shaving automation depends on. Then add per-device monitoring only for loads you intend to act on — for most homes that is three to six things: the EV charger, the heat pump or air conditioning, the dryer, the dishwasher, and perhaps a server or freezer. Monitoring a lamp you will never automate produces a nice graph and no savings.