Pair your load with the right power.

What size power station do I need?

There is no “size”. There are three numbers — watts, surge watts and watt-hours — and a station can pass two of them and still fail you.

What the three numbers each decide

Watts decide whether your loads start at all, surge watts whether they start without tripping the inverter, watt-hours how long they keep going. The box usually advertises one of the three.

This quiz sizes all three separately, from a load list you pick in four taps — and it will tell you when the honest answer is “nothing here does that”. Watt-hours on a spec sheet are not the watt-hours you get: depth of discharge and inverter losses take a fifth to a third of them, and most brands publish neither figure.

Skip to the four questions ↓

Loading the plan you were sent…

Reading the configuration in your link and re-running the engine on it. Underneath this panel is the page's built-in WORKED EXAMPLE — not your shared plan; it is covered for a moment so it cannot be mistaken for your answer.

Your shared plan could not be loaded. Everything below is this page's WORKED EXAMPLE — not your shared plan. Reload, or use the controls to rebuild it.

Worked example

The fridge and the phones · 12 h · no budget cap

What your loads need

Full-size refrigerator + LED light bulb × 4 + Router + modem + Laptop computer

  • Running watts 103 W added up, all on at once
  • Start-up surge 334 W worst instant, not a steady draw
  • Energy for 12 h 1,141 Wh at 95 W averaged over the duty cycles

What that means you should buy

Tap a card to change which one the button below buys. The highlighted one is the pick with the strongest evidence behind it, which is not always the biggest or the cheapest.

Runtime on this set —

This brand publishes no usable capacity for this model, so no runtime is claimed for it — only what the load draws.

The cheapest record that clears every constraint for this set — your loads sit at 3.4% of its continuous output. Its runtime cannot be checked: this brand publishes no usable capacity, only a nameplate.

What it runs, and what it does not 21/26

21 of the 26 loads our library tags “Backup at home” — green or amber — run on this machine. Not the loads you ticked, and not the whole appliance library.

  • Air fryer runs it
  • Box fan tight
  • Chest freezer runs it
  • Coffee maker runs it
  • CPAP machine tight
  • Desktop PC tight
  • Electric blanket runs it
  • Electric kettle tight
  • Electric space heater (1500 W) runs it
  • Full-size refrigerator runs it
  • Gas furnace blower tight
  • Laptop computer runs it
  • LED light bulb runs it
  • LED TV runs it
  • Microwave oven tight
  • Mini fridge runs it
  • Mini-split heat pump runs it
  • Pancake air compressor too big
  • Portable air conditioner too big
  • Router + modem runs it
  • Starlink Mini runs it
  • Sump pump runs it
  • Washing machine runs it
  • Welder (stick / MIG) too big
  • Well pump too big
  • Window air conditioner too big

Runtime on this set 15 h

The cheapest record that both clears the constraints and can be SHOWN to last 15 h on this set — past the 12 h you asked for. That runtime is computed from a published usable capacity, not from the number on the box.

What it runs, and what it does not 19/26

19 of the 26 loads our library tags “Backup at home” — green or amber — run on this machine. Not the loads you ticked, and not the whole appliance library.

  • Air fryer runs it
  • Box fan tight
  • Chest freezer runs it
  • Coffee maker runs it
  • CPAP machine runs it
  • Desktop PC runs it
  • Electric blanket runs it
  • Electric kettle too big
  • Electric space heater (1500 W) runs it
  • Full-size refrigerator runs it
  • Gas furnace blower tight
  • Laptop computer runs it
  • LED light bulb runs it
  • LED TV runs it
  • Microwave oven tight
  • Mini fridge tight
  • Mini-split heat pump runs it
  • Pancake air compressor too big
  • Portable air conditioner too big
  • Router + modem runs it
  • Starlink Mini runs it
  • Sump pump too big
  • Washing machine tight
  • Welder (stick / MIG) too big
  • Well pump too big
  • Window air conditioner too big

View at BLUETTI Official Store (US) → Price checked 2026-08-01. Only records with a verified purchase link get a button; the rest carry none rather than a guessed one.

Not your setup? Four questions

These four questions need JavaScript. The worked example above already answers them for a home-backup list — and the runtime calculator will check any station against any appliance without them. The options below are pre-filled with the example's answers so you can see what was asked; they are not a choice you have made.

Question 1 of 4

Where will you use it?
What do you want to keep running?
For how long, without recharging?
What are you willing to spend?

Checking one machine you already own? The runtime calculator does that without this form →

The same four questions, answered for all three situations

Each row is a real run of this page's engine at build time, at the same 12 h target and no budget cap. Nothing here is a table anyone typed.

Situation Loads Running Surge Energy Highlighted pick Proven runtime
Backup at home The fridge and the phones 103 W 334 W 1,141 Wh BLUETTI Elite 200 V2 15 h
RV or van Fridge, lights, internet 194 W 249 W 1,911 Wh BLUETTI Elite 200 V2 9.4 h
Camping and trips Lights, phone, laptop 31 W 31 W 376 Wh BLUETTI Elite 100 V2 20 h

An em dash in the last column is not a zero. It means the record that best fits the other tests publishes no usable capacity, so no duration is claimed for it — 3 of 14 US records can carry one at all.

How a recommendation is actually decided

  • Your answers become a load list, not a guess

    Each preset is a set of appliance records from the library, with their own running watts, start-up surge and duty cycles, each carrying a source and a date. The engine adds them the way an inverter sees them — worst instant for surge, duty-weighted for energy — rather than summing label wattages.

  • Every record is judged, then ranked

    All 14 US records are run against your set. Anything that cannot start and keep the loads running is out; anything above 82% of a station's continuous output is marked tight rather than quietly recommended. What is left is ranked on verified price, and the rungs are: the cheapest that clears it, the cheapest whose runtime we can prove, and the cheapest that proves twice your target.

  • The highlight follows the evidence

    Only 3 of the 14 US records publish a defensible usable capacity, so only those can be shown to last a given number of hours. The highlighted card is the cheapest of those that covers your target. When none of them does, the highlight stays on the pick whose card states the shortfall — it never moves to a bigger machine whose duration we would be guessing at.

  • A nameplate is only allowed to say “no”

    Nameplate watt-hours never become a runtime here. They do one job: if a record's entire nameplate is below the energy your target needs, it is ruled out, because capacity loss only runs one way. The bar is 85% of nameplate — the friendliest usable fraction measured anywhere in this library, read from the records that publish both figures rather than chosen.

  • Some loads refuse to be multiplied

    An energy figure is watts × duty cycle, and that product only means something when both were measured the same way. For 3 of the 30 appliance records they were not — a nameplate ceiling against an observed run-time fraction — so those loads state no energy target at all rather than a plausible-looking wrong one. Their verdicts are unaffected: deciding whether something starts never used that product.

How these numbers are graded

Every field carries a source tier, a source URL, a verification date and a confidence grade, and a conclusion inherits the weakest of them — a verdict is exactly as good as its worst input, so the grades are a minimum, never an average. A figure computed from a manufacturer's own coefficients stays official but is capped at medium confidence and labelled as derived.

Coverage today, read from the library at build time rather than claimed: 14 US station records are ranked here, 3 of them can carry a runtime in hours, and 30 appliance profiles are available as loads. Prices come from a registry with a check date attached; where a record has no verified price it is not claimed to fit any budget. No lab measurement exists in the library yet; when one does it will be labelled as such, and nothing on this site claims one before then.

Questions people actually ask

What size power station do I need?
It is three numbers, not one. First, watts: your running loads added up — 103 W for the worked example above — because anything over 82% of a station's continuous output is already tight. Second, surge watts: the worst instant when a motor starts, 334 W here, which is what actually trips an inverter. Third, watt-hours: the energy your loads consume over the time you need them, 1,141 Wh. A station can be big enough on watts and far too small on watt-hours, and the label on the box usually tells you about neither.
Is a 1000W power station the same as a 1000Wh power station?
No. Watts are how much you can draw at once — the ceiling on what will start and keep running. Watt-hours are how much energy is stored — how long it lasts. A 1000 W ceiling over a 500 Wh battery runs a big load briefly; 300 W over 2000 Wh runs a small load for a long time. The quiz sizes both, separately, because a machine that fails either one fails you.
Why does the recommendation say it cannot prove how long some stations last?
Hours need a defensible usable-capacity figure, and only 3 of the 14 US records in the library have one — the rest publish a nameplate watt-hour number and nothing about depth of discharge or inverter efficiency. Dividing that nameplate by your load produces a confident-looking number that is wrong by 20–35%. So those records are still ranked and still recommended on the tests we can run — will it start, will it stay running, is it inside the budget — and the duration is simply left unproven rather than invented.
Why is the highlighted pick not always the cheapest one?
The highlight goes to the strongest evidence, not the lowest price. The first rung is the cheapest record that clears the electrical test; the highlighted one is the cheapest record that also has a published usable capacity and can be shown to last the duration you asked for. Where those are the same machine, you get one card with both labels. Where nothing can be proven to last long enough, the highlight stays on the pick whose card states the shortfall rather than moving to a bigger machine we would be guessing about.
Can it just tell me a number of watt-hours to buy?
It does — that is the energy figure in the requirement above — but a watt-hour target is not a shopping instruction. Two stations with the same nameplate deliver different amounts of usable energy, and only some brands publish enough to say how much. Where a record's entire nameplate is below the energy your target needs, the quiz rules it out: capacity loss only runs one way, so a nameplate that is already too small cannot become big enough. At the friendliest usable fraction anywhere in this library — 85% — that is the only thing a nameplate is allowed to tell you here.
Want to change one load at a time?

This page answers in four taps. When you want to pick individual appliances, type your own wattage or compare a specific station against a specific load — open the runtime calculator

Runs it · 15 h BLUETTI Elite 200 V2 · from $1,099

Check price →