Backup Power Reviews

Understand the evidence

Usable power-station capacity: what published tests actually tell us

Use measured energy from a similar load when it is available. Keep the test conditions attached to the number: a result at 1,000 W is not a reliable prediction for a router using 18 W.

A small evidence ledger

These measurements come from independent reviewers, not our own lab. They are deliberately shown separately. We do not average different loads, treat another region’s sockets as UK hardware, or turn a single result into a universal efficiency score.

Evidence for every shortlisted model: published tests and explicitly missing results
StationDelivered energyConditions and source
Anker C1000 Gen 2890 WhReported US result; the source mixes Anker specifications, so model identification is uncertain. OutdoorGearLab: shared US tests
BLUETTI Elite 100 V2890 WhReported in that US comparison; not evidence of an efficiency tie with the UK Anker. OutdoorGearLab: shared US tests
Anker C1000 Gen 2907 WhUK unit; 1,000 W heater load. The Technology Man: UK C1000 Gen 2 test
BLUETTI Elite 100 V2936 WhEuropean unit; 200 W load. Smartzone: European Elite 100 V2 test
BLUETTI Elite 100 V2903 WhEuropean unit; 1,800 W load. Smartzone: European Elite 100 V2 test
Anker C2000 Gen 2Just over 1,600 WhEuropean unit; roughly 200 W. Notebookcheck: European C2000 Gen 2 test
Anker C2000 Gen 2About 1,850 WhSame review; 1,800 W. Notebookcheck: European C2000 Gen 2 test
DJI Power 1000 V2880 WhReported US table entry only. Its prose and table disagree on output/generation; exact-model measurement is unverified. OutdoorGearLab test
EcoFlow RIVER 3 PlusNo verified usable-energy result in our evidence setManufacturer-rated storage: 286 Wh. This is not delivered AC energy. Manufacturer source
Anker SOLIX C300XNo verified usable-energy result in our evidence setManufacturer-rated storage: 288 Wh. This is not delivered AC energy. Manufacturer source
BLUETTI Elite 200 V2No verified usable-energy result in our evidence setManufacturer-rated storage: 2,073.6 Wh. This is not delivered AC energy. Manufacturer source
EcoFlow DELTA 3 Max PlusNo verified usable-energy result in our evidence setManufacturer-rated storage: 2,048 Wh. This is not delivered AC energy. Manufacturer source
Jackery Explorer 2000 Plus v2No verified usable-energy result in our evidence setManufacturer-rated storage: 2,048 Wh. This is not delivered AC energy. Manufacturer source

What we can reasonably conclude

OutdoorGearLab’s Anker table and prose disagree on output and sockets. We retain its reported numbers for traceability, but do not use them to rank efficiency or claim a tie. The UK and European tests below identify their loads more clearly. The separate UK and European measurements add context, but they do not prove one model is always more efficient. Our 1 kWh buying comparison also considers sockets, charging and the intended equipment.

The two C2000 readings illustrate why load matters. A fixed amount of power used by the station takes a larger share of the battery during a long, low-power discharge. A higher load may therefore extract more energy at the sockets while still exhausting the battery sooner in hours. There is no contradiction.

Use the right denominator

Rated capacity describes energy stored in the battery. Usable AC energy is what a particular test delivered at the mains sockets. Round-trip efficiency compares energy delivered with energy bought to recharge the station under a defined cycle. These quantities should not share one unqualified “efficiency” percentage.

If a test already measured 900 Wh delivered to its appliance, do not multiply that by 85% again before calculating runtime at the same load. Those conversion losses have already happened. If you instead start from rated capacity, you need assumptions about losses and overhead, clearly labelled.

How to use this when buying

First find the appliance’s demand and duration. Then look for evidence near that load. For a fridge, the compressor cycles; for a laptop, the workload and internal battery charge change. An average is useful for a budget but cannot establish starting power or continuous-output compatibility.

Where a matched test is missing, use a range and leave margin. The battery-size calculator exposes its assumptions. For small continuous loads, read why inverter overhead matters before paying for a bigger battery.

We would rather leave a missing measurement blank than fill it with a manufacturer runtime estimate presented as a test. If you have a reproducible reading, our contact page explains how to send a correction or source.

Can I compare every percentage in the table directly?

No. Different loads, voltage variants, cut-off settings and test methods affect the result. Use the conditions column and prefer measurements taken at a load similar to the one you plan to run.

Is a calculated runtime a published test?

No. A calculation is a planning estimate. We label it separately from a reviewer’s measured output and a manufacturer’s claim, and retain the source and conditions for each.

Correction, 16 November 2026: Flagged generation inconsistencies in the OutdoorGearLab Anker and DJI entries. Added the load-specific BLUETTI result from Smartzone.

Updated