How to Size a Power Station for a Real Blackout: Japan's Load-List Method Applied to 2026 Units

Start with the outage, not the product

US marketing for these units counts phone charges. That number is useless, because a phone is the smallest thing you will plug in.

Japan’s preparedness culture starts from the other end: a specific outage length, a written list of loads, and only then a purchase. It is a better method, and the US now has the outage data to justify it. The EIA reports that in 2024 US electricity customers averaged about 11 hours without power — nearly twice the annual average of the preceding decade. Major events accounted for roughly 9 of those hours, against about 4 hours annually across 2014-2023, with Hurricanes Beryl, Helene and Milton alone responsible for 80% of the total. Non-major events sit steadily at around 2 hours a year.

So 24 hours is the realistic median-bad case and 72 hours is the storm case. Size for one of those two numbers.

Step 1: write the load list

Every row needs four values: running watts, surge watts, hours you will actually run it, and the watt-hours that produces. For the fridge, skip the running-watts argument entirely and use the FTC EnergyGuide label on the door — annual kWh divided by 365 gives you a measured daily figure. ENERGY STAR top-freezer models typically fall between 300 and 500 kWh/year, so roughly 0.8-1.4kWh per day at 24-hour duty.

Load	Running W	Surge W	Hours	Watt-hours
Fridge, door kept shut, cycled ~8h	400-800 (label method preferred)	1,200-1,600	8	~370
Fridge, continuous 24h	—	1,200-1,600	24	~1,100
CPAP, humidifier and hose off	~20-30	negligible	8	~70
CPAP, humidifier + heated hose	~55-70	negligible	8	~250
Router + ONT/modem	~10 + ~4-12	negligible	24	~430
LED bulbs, 3 x ~10W	~30	negligible	5	~150
20-inch box fan, high	~87	small	8	~690
Two phones, full charges	5-25	—	—	~40
Laptop	~65	—	3	~195

Two rows in that table usually surprise people. The router and ONT, at under 20W combined, cost more energy over a day than a CPAP night, because they run all 24 hours. And a single box fan can be the largest line on the list — larger than the fridge if you cycle the fridge properly.

Add it up. Cycled fridge, CPAP with humidifier, comms, lights, phones and laptop is about 1,435Wh for 24 hours. Add the box fan and you are at roughly 2,125Wh.

Step 2: the two numbers fail differently

Watt-hours decide how long you last. Continuous and surge watts decide whether the thing switches on at all — and they fail instantly, not gradually.

Compressors and pumps draw locked-rotor current at startup, typically 3-8 times their full-load amps, and a common rule of thumb puts a fridge’s starting watts at 3-4 times its running watts. That is why generator-sizing references quote 1,200-1,600 starting watts for a household fridge that runs at 400-800W. A power station whose continuous rating merely matches the running figure will trip on the first compressor cycle.

Every brand publishes a second, higher number for this. BLUETTI calls it Power Lifting, Anker calls it SurgePad, EcoFlow calls it X-Boost. Read the fine print. BLUETTI ships Power Lifting switched off by default and restricts it to pure resistive loads — its own examples are electric kettles, space heaters, hair dryers and electric blankets — while explicitly warning against devices with motors or compressors; the Elite 200 V2 manual is blunt that even with the feature on, “its actual output power is still 2,600W”. EcoFlow describes X-Boost as lowering output voltage so a higher-rated resistive appliance will still run. Either way, this class of feature suits resistive loads and is not a general answer for motor startup. The plain surge or peak rating is the one to check against your compressor.

Step 3: add the overhead everyone forgets

Inverters lose roughly 10-20% converting DC to AC, and a 1kWh-class station idles at about 5-15W with the AC output on — around 120-360Wh over a full day doing nothing.

Apply that to the list. The 1,435Wh no-fan case becomes roughly 1,650Wh plus about 240Wh of idle, so close to 1,900Wh. That is the finding that matters: a 2kWh unit is a one-day machine for a real load list, not a three-day one. For 72 hours you need solar, an expansion battery, or a much shorter list.

Step 4: now pick a unit

All figures below are manufacturer-published. Note that Jackery rates cycle life on the flagship LiFePO4 Explorers listed here to 70% capacity while the others use 80%, so the numbers are not directly comparable — the sibling guide covers that in detail. BLUETTI publishes no capacity endpoint at all for the AC180’s cycle figure, so that one compares to neither.

Model	Capacity	Continuous	Surge / lift	Cycles	AC recharge
EcoFlow RIVER 3	245Wh	300W	600W	3,000 to 80%	60 min to 100%
EcoFlow DELTA 3	1,024Wh	1,800W	3,600W (X-Boost 2,200W)	4,000 to 80%	56 min to 100%
Anker SOLIX C1000	1,056Wh	1,800W	2,400W SurgePad	3,000 to 80%	58 min to 100%
Jackery Explorer 1000 v2	1,070Wh	1,500W	3,000W	4,000 to 70%	1 hr emergency, 1.7 hr standard
BLUETTI AC180	1,152Wh	1,800W	2,700W lifting	3,500+, no endpoint published	45 min to 80%
Jackery Explorer 2000 v2	2,042Wh	2,200W	4,400W	4,000 to 70%	1.3 hr emergency
BLUETTI AC200L	2,048Wh	2,400W	3,600W lifting	3,000+ to 80%	45 min to 80%
Anker SOLIX F2000	2,048Wh	2,400W	3,600W	3,000 to 80%	1.4 hr to 80%
BLUETTI Elite 200 V2	2,073.6Wh	2,600W	3,900W lifting	6,000+	1.1 hr to 80%

Against the worked list: the 1kWh tier covers CPAP, comms, lights and devices for a night with the fridge left alone. It does not cover a fridge plus a fan for 24 hours. The 2kWh tier does, once, with the fan rationed.

Check BLUETTI AC180 on Amazon

Check Anker SOLIX C1000 on Amazon

Check Jackery Explorer 2000 v2 on Amazon

CPAP: use your own label

CPAP is the one row where a generic number is not good enough. Draw varies enormously with settings: reported figures for a ResMed AirSense 11 run about 20-30W with humidifier and heated tube off, 35-50W at moderate humidity, and 55-70W with maximum humidifier plus heated hose, against a 24V/3.75A (90W) supply. A measured AirCurve 10 ASV consumed about 250Wh over ten hours with heat on, versus roughly 70Wh with it off — humidification roughly tripled the load.

Practical consequence: turning humidification off in an outage can triple your nights of runtime. Whether that is appropriate for you is a question for your equipment’s manual and your clinician, not a buying guide. Read the wattage printed on your own power supply and size from that.

Solar recharge, honestly

Panels do not deliver their rating. Real-world output typically lands at 70-85% of rated wattage, and angle, panel temperature, passing cloud and charge-circuit losses together commonly consume 15-25%. A 200W panel realistically yields around 160W in good sun.

Do the division: 2,048Wh at 160W is roughly 13 hours of genuinely good sun — the better part of two clear days, and post-storm days are rarely clear. Solar extends an outage; it does not reset one. If you want meaningful recharge, buy panel wattage close to the unit’s stated maximum solar input (500W on the AC180 and DELTA 3, 1,000W on the Elite 200 V2, 1,200W on the AC200L).

Check 200W portable solar panels on Amazon

What these cannot do

Central air conditioning, electric ovens, electric water heaters and most well pumps are 240V hard-wired loads well beyond a 1,800-2,600W portable inverter, with startup surges larger still. Space heaters and hair dryers will technically run on the 2kWh tier and will drain it in under an hour. Treat this category as food, medical, communications and lighting — nothing structural.

The safety argument that does not get marketed

CPSC required a label on portable generators reading that using one indoors can kill you in minutes, and portable generators accounted for more than 85% of non-fire CO deaths from engine-driven tools across 1999-2012. A battery station produces no exhaust and can sit next to the bed running a CPAP. That is not a feature bullet; it is the entire reason this category exists for indoor overnight use.

Summary

Write the list, total the watt-hours for 24 or 72 hours, add 10-20% conversion loss and idle draw, then check the surge rating against your fridge’s startup — in that order. Most households discover they need roughly 2kWh for one honest day, and that the box fan and the always-on router matter more than the phone charging everyone advertises. For brand-level differences, chemistry, cycle-life rating conventions and the Japan-versus-US output derating, see our companion guide to portable power stations in 2026. More gadget guides at gadget-hub.jp English.