EcoFlow River 2 Max: our full review

It returns 439 Wh out of 512 when it is asked politely, but only 391 when it is pushed. Its efficiency depends far less on the socket chosen than on the speed at which the energy is drawn, and that is what nobody explains.

Our rating
3.8/5
Capacity and efficiency: 3.8 out of 5 (weighting 25%)Capacity3.8Output power: 3.8 out of 5 (weighting 25%)Power3.8Recharging: 4.4 out of 5 (weighting 15%)Recharging4.4Ports and control: 3.5 out of 5 (weighting 15%)Ports3.5Build and endurance: 3.7 out of 5 (weighting 12%)Build3.7Noise: 3.5 out of 5 (weighting 8%)Noise3.5
EcoFlow River 2 Max portable power station seen from the front, screen lit, two 230 V sockets and USB ports visible

The essentials

This 512 Wh station is judged on a curve, not on a single figure. It returns 439 Wh through its 230 V socket when around 200 W is asked of it, but only 391 Wh at 500 W, and barely more than 70 % of its capacity on very small loads. Used well, it is efficient; used badly, it loses a fifth of its runtime. The rest is mixed: the one-hour recharge is real, the build is careful, but the fan triggers erratically and four years on sale have brought no fix, X-Boost works by lowering the voltage rather than genuinely delivering 1000 W, and a night-time fault acknowledged in writing rules out leaving it as an unattended solar reserve. Out of stock at the manufacturer but discounted at resellers, it remains more interesting than the generation meant to replace it.

Strengths

  • 439 Wh genuinely returned through the 230 V socket in its good load range, or 86 % of the claimed capacity
  • Full charge in 58 minutes to 1 h 03, a promise kept
  • 500 W continuous held with margin: 522 W for thirty seconds, 721 W at peak
  • Careful modular build, LiFePO4 chemistry and a five-year warranty naming this model
  • No recall on this generation, where the previous one had three in Europe
  • 512 Wh in a single 6.1 kg case, which the following generation no longer offers

Limitations

  • Efficiency collapses on small loads: around 72 % at 50 W against 86 % at 200 W
  • 60 to 72 % efficiency from the wall to the socket, poor even for this format
  • Erratic fan trigger, sometimes audible with nothing plugged in, unchanged in four years on sale
  • X-Boost lowers the voltage below 170 V instead of delivering 1000 W, and disappears as soon as the station is on the mains
  • Night-time fault acknowledged in writing: with a solar panel connected, the station reads zero per cent at night
  • Only two 230 V sockets in Europe, and only one of them stays powered beyond 500 W
EcoFlow River 2 Max: see the full details and price →

What the 512 Wh are really worth, and why that depends on you

The 512 Wh printed on the case are a cell capacity, eight cells in series for 20 amp hours. It is not what comes out of the outputs, and on this machine the gap varies more than on any other in its category.

Through the 230 V socket, the energy recovered settles at 439 Wh when around 200 W is drawn. The same machine, the same socket, but at 500 W, gives back no more than 391 Wh. Forty-eight watt hours of difference, close to a tenth of the pack, purely as a function of the speed at which the energy is consumed.

439 Wh
returned on 230 V at 200 W
391 Wh
returned on 230 V at 500 W
431 Wh
returned on the 12 V output
Energy returned through the 230 V socket depending on the power drawn, set against the 512 Wh claimed: 439 Wh at 200 W, but only 391 Wh at 500 WClaimed512 WhAt 200 W439 WhAt 500 W391 Wh
Same socket, same machine: 48 Wh of difference depending on the power drawn, close to a tenth of the pack.

The bottom of the range shows up elsewhere too. Driving a halogen heater, so working very close to nominal power, the return falls to 402 Wh, or 78 % of the claimed capacity. Three figures, one lesson: the harder you pull, the less you get back.

The inverted bell curve, and what it changes in use

Here is the most useful point in this review, and the least known. Instinct says that a battery gives back more energy the more gently it is asked. On this machine that is wrong at both ends.

The efficiency of the 230 V output draws an inverted bell: around 72 % at 50 W, it peaks at 86 % around 200 W, then falls back towards 76 % at 500 W. The optimum sits in the middle, not at the bottom.

Put into practice: powering a 50 W appliance from this station costs around 20 % of its runtime, compared with the same watt hour drawn at 200 W. A night light left on until morning is the worst possible use; a hotplate for twenty minutes, one of the best.

The mechanism comes down to two effects pulling against each other. At the bottom of the range, the inverter draws power simply to exist, whatever is asked of it, and that fixed levy weighs all the more heavily as the load is light. At the top of the range, conduction losses take over, and they grow with the square of the current. Between the two lies a window where the machine works well.

The ranking of the outputs exists as well, but it is secondary. USB-C gives back the most, around 88 %, and the 12 V output around 84 %. Beware one received idea, though: the 12 V is not a direct tap on the pack. It is a step-down converter, capped at 126 W, which pays for its own conversion. At moderate load it does no better than the 230 V socket.

606 Wh at the wall to give back 439

The figure that decides the bill is not capacity but round-trip efficiency: how much has to go through the meter to get one watt hour back out of a socket?

Path of the energy: 606 Wh drawn at the wall for a full charge, 512 Wh stored in the cells, 439 Wh returned through the 230 V socketAt the wall606 WhStored512 WhReturned439 Wh
It takes 606 Wh at the meter to give back 439 through a socket. The rest goes in conversion, on the way in as on the way out.

It takes 606 Wh to refill the pack, and 439 come back out in the best case. Efficiency from the wall to the socket therefore settles between 60 and 72 % depending on the load applied. That is poor, and it is structural: the smaller a station, the more the conversion electronics, whose draw is roughly fixed, weigh against the pack they serve.

For comparison, the 1 kWh stations of the same brand run around 77 to 80 % on the same path. The difference comes not from build quality but from size.

What it draws while powering nothing

Two regimes coexist, and the gap between them turns on a single button.

3 to 6 W
230 V output switched off
12 to 25 W
230 V output on, powering nothing
20 h
to empty itself, in the worst case

With the 230 V output switched on and nothing plugged in, the machine absorbs between 12 and 25 W, inverter and fan included. Against 439 Wh available, that is enough to empty it in under a day in the worst case. With the output switched off, the draw falls to a few watts and the wait is counted in days.

The habit to build: switch off the 230 V output as soon as it is not in use. It is the one action that separates a reserve still full after several days from a battery flat the next morning.

X-Boost: the 1000 W are not what they seem

The spec sheet announces 500 W continuous and 1000 W with X-Boost. The second figure deserves an explanation that the manufacturer does not give.

X-Boost does not double the power available. It lowers the output voltage, from 240 V to under 170 V. A resistive appliance labelled 1000 W plugged into it draws little more than 500 W, and heats accordingly. That is not a defect, it is the very principle of the feature: it allows a too-hungry appliance to be plugged in, provided you accept that it works at half pace.

Three limits to know: X-Boost suits only resistive appliances, heating or cooking; it becomes unavailable as soon as the station is plugged into the mains, so never in backup use; and its ceiling depends on the grid, 1000 W in Europe but 750 W in Japan.

The demonstration in figures exists on the model just below it in the same range: a 400 W hotplate plugged in with X-Boost active draws 297 W. It works, but it heats less, and the water takes longer to boil. The manufacturer itself describes the feature as lowering the operating voltage of the appliance in order to reduce its consumption. The marketing, for its part, merely announces a power figure.

Another limit, rarely written down: beyond 500 W, only one of the two sockets stays powered. The station does not spread 1000 W across its two outputs.

Facing motors: one simple rule, the soft start

The 500 W continuous are held with margin: 522 W pass for thirty seconds before cut-off, and a peak of 721 W holds for one second. That is comfortable for the category.

But the real limit against a motor is not read off the label of the appliance. It depends on the way that appliance starts.

Behaviour observed with power tools
AppliancePowerResult
Sander310 WPasses, with a peak beyond 800 W
Orbital sander400 WPasses
Rotary hammer750 WPasses, provided the ramp-up is gradual
Angle grinder with no soft start720 WCuts out instantly

A 720 W angle grinder makes the station drop out where a 750 W rotary hammer passes. It is not the power that decides, it is the current surge at start-up.

The manufacturer acknowledges as much for a use it puts forward itself. Its manual warns that the station, plugged in normal mode into a fridge, may be switched off automatically because of the fluctuation in the electrical supply of the fridge, and recommends setting the output to never switch off in the app. In other words: the compressor can trip the very machine sold to back it up.

One hour, and that promise is kept

This is the field where the machine keeps its word without reservation. A full charge takes 58 minutes to 1 h 03, against the one hour announced.

58 min
for a full charge from the mains
600 W
actually absorbed, out of 660 announced
190 W
obtained from a 220 W panel

Charging slows at the end of the cycle, which is normal and desirable: 80 % in around forty minutes, then the last twenty per cent in about twenty more, at far lower power.

Solar, on the other hand, needs bringing back down to earth. The input accepts 220 W, and a panel of that rating delivers around 190 in good conditions. But with a 160 W panel in midwinter, a full charge took more than two days, where the sales page announces three hours.

A dead band to know about: below around 30 W of solar input, the station does not charge at all. Twenty-two watts for half an hour produced no gain, where seventeen watts over USB-C did. Small folding panels are useless on this machine.
The condition nobody publishes: to reach the hour announced, the manual calls for a wall socket of at least 10 amps, with no other appliance on the same circuit, and explicitly forbids power strips and extension leads. In an ordinary home, that condition is rarely met.

One last point, seasonal: charging is impossible below 0 °C, a normal protection for this chemistry. It resumes above 3 °C. Note that the electronics go by the real temperature of the cells, not that of the air: a station brought in from the cold takes time before it accepts any current.

46 dB at one metre, and a badly set fan

Only one sound level for this machine comes with both an instrument and a distance: 46 dB at one metre, during a mains charge at full speed. That is around ten decibels above the background noise of an office.

The 62 dB figure that circulates everywhere is a sales footnote, with no distance, no load and no instrument, repeated as though it were a result. It is comparable to nothing.

The real reproach is not the level anyway, it is the trigger. The threshold wanders between ten watts and three hundred, and the fan can be heard turning with nothing plugged in, on a cold machine. Over four years on sale, that behaviour has never been corrected.

It is not a mechanical problem: the fan is a three-wire model with speed feedback, so it can be controlled finely. The aggressive curve is a software choice. Swapping it for a model known to be quiet leaves temperatures unchanged: the fan is not the bad part, its control law is.

One observation confirms it: while charging, the ventilation follows the input power and not the temperature. A vigorous solar charge runs the fan flat out on a case that has stayed cold. The workaround exists and costs only time: reduce the charging speed in the app. The one-hour recharge is the noisy mode.

None of this appears anywhere in the manufacturer documentation, yet the noise is frankly high on fast charging, and night-time use is best avoided.

On backup: what the spec sheet promises and what the manual forbids

The changeover is announced under 30 milliseconds. It has to be said plainly: no figure exists for this model. The only one in the family belongs to the Pro version, at 20 ms.

What is established comes down to one sentence: a desktop computer, a storage server and a network switch all ride through the cut without flinching. But a UPS placed downstream detects the gap and switches to its own battery. The dip is therefore real, and measurable by sensitive equipment.

The manual sets the limit itself, and it is unusually frank. The function does not support zero millisecond changeover and must not serve a server or a workstation. Better still: if backup is the main use, it recommends connecting one single appliance, under 750 W, because the overload protection can trip at the very instant of the changeover. A backup station that risks cutting out at the moment of taking over is a limit worth knowing about beforehand.

A use explicitly discouraged: the manual rules out life-critical medical equipment, including continuous positive airway pressure machines. If that is the need, it calls for a medical UPS, not a portable station.

Two field reservations are added. The 230 V output does not always come back on by itself when the power returns, where the sales page promises automatic restoration. And while running on the mains, the machine can discharge then recharge in short cycles, fan included.

A series fault acknowledged in writing, and never fixed

This point has to be known before considering this station for unattended solar backup, because it targets exactly that use.

With a solar panel permanently connected, the machine reads zero per cent as night falls, then finds its real level again within seconds as soon as the sun or the mains comes back. It therefore behaves as though empty at precisely the moment it was being counted on.

What makes the case notable is the answer from the manufacturer. Evidence in hand, it acknowledged in writing that the behaviour is not limited to that one unit and may concern products of the series, and that a standard exchange does not guarantee it will be eliminated. No recall followed, no fix was published.

The workaround is simple, but it cancels the point of the arrangement: unplug the panel at night, or give up leaving the station to watch over things on its own.

Two sockets, not four: the trap of the European version

The European version of this station has two 230 V sockets. Four is often written. That is the count of the American version, whose far more compact sockets fit four across the same panel where two Schuko already take up all the room.

The error spreads all the more easily because the manufacturer publishes the socket count of its European version nowhere. Its manual writes AC output socket, in the singular and with no quantity, and its product pages say nothing about it. Added to that is an ambiguous sales phrase, four ways to recharge, which designates four supply modes and turns easily into four sockets under a hurried pen.

The rest of the connectivity is more generous: three USB-A ports sharing 24 W, a 100 W USB-C that works in both directions, a 126 W cigarette lighter socket and two DC outputs. The station accepts no expansion battery, however: its capacity is fixed.

A machine at the end of its run, and what that is worth

Worth knowing before buying: this station is out of stock across every manufacturer store, in France as elsewhere in Europe, at the full price of 399 €. No end-of-line announcement has been published, no date, no designated replacement. It is a withdrawal in fact, not in law: the maker has stopped supplying its own channel and lets resellers run down their stock.

The consequence is a fragmented market. In France the floor price for a new unit settles around 360 €, without this being a dated promotion: it is a structural end-of-life low. In Italy, by contrast, the rare offers go above list price, for want of official distribution.

What reassures: the five-year warranty names this model, with shipping covered, and the station never had an expansion battery, so there is no dependence on an accessory on its way out. What worries: no spare parts programme is published, and one clause voids the warranty if the unit goes more than six months without being either charged or discharged, which is precisely the backup use it is credited with.

A word on the following generation, because it lights up the trade-off. The River 3 range offers no equivalent single block: its largest one-piece model tops out at 286 Wh, and the 572 Wh advertised correspond to an assembly of two blocks, for 449 € and 8.2 kg. Against that, the discounted River 2 Max offers 512 Wh in a single 6.1 kg case, for around a hundred euros less. What it really loses is not capacity: it is the backup changeover, three times slower, and quiet running.

Who it suits, and who it does not

It suits anyone looking for 500 Wh they can carry for a weekend, a van or a work site, and who will ask the machine for power in its good range, between one and three hundred watts. It also suits anyone who wants to recharge fast: one hour for a full charge remains a solid argument in 2026.

It does not suit anyone who wants to power small loads over time, where its efficiency collapses. Nor anyone looking for silence. Nor anyone counting on leaving it as an unattended solar reserve, because of the acknowledged night-time fault. Nor anyone who wants to protect a server, which the manual explicitly forbids.

In one sentence: a good work and weekend station, quick to recharge and solidly built, whose efficiency range has to be understood and whose fan has to be accepted, at an end-of-life price that makes it today more interesting than the generation meant to replace it.

One last figure to sum up the trade-off. At the price observed, the claimed watt hour comes to around 0.70 €. The equivalent assembly from the following generation charges 0.79 € for it, with two blocks and two kilos more. The outgoing generation is, for once, the best buy in its own range.

The product and its alternatives

EcoFlow River 2 Max

EcoFlow River 2 Max

See price (16)

512Wh, 500W and only 6.1 kg: the compact station that recharges 0 to 100% in one hour and powers 80% of your essential devices thanks to X-Boost.

View product
EcoFlow River 2

EcoFlow River 2

See price (10)

The fastest charging portable power station in its class: fully recharged in one hour, the River 2 follows you from campsite to home backup.

View product
EcoFlow River 2 Pro

EcoFlow River 2 Pro

See price (9)

Powerful, quick to recharge and generous with sockets, the 768 Wh EcoFlow River 2 Pro follows you from camping to home-office backup.

View product
EcoFlow River 3 Plus

EcoFlow River 3 Plus

£219.00 (14)

Compact, quiet and fully recharged in just one hour, the EcoFlow River 3 Plus delivers reliable power wherever you happen to need it.

View product
Anker Solix C300

Anker Solix C300

See price (22)

288Wh, two mains sockets and a 140W USB-C port in an ultra-compact, quiet body (25 dB): the small station that even recharges a laptop.

View product

Frequently asked questions

It depends on the power drawn, and the spread is wide. Through the 230 V socket, around 439 Wh come back when 200 W is drawn, but only 391 Wh at 500 W, and around 72 % of capacity on very small loads. The 12 V output gives back around 431 Wh.

Because the inverter draws a roughly fixed power as soon as it is switched on, whatever is asked of it. Set against a 50 W load, that levy weighs very heavily; set against 200 W, it becomes negligible. Beyond that, conduction losses take over.

No. X-Boost lowers the output voltage below 170 V, so that a resistive appliance labelled 1000 W draws little more than 500. The feature lets you plug it in, not run it at full power. It is also unavailable as soon as the station is connected to the mains.

Two on the European version. The four sockets often quoted are those of the American model, whose far more compact sockets fit four across the same panel. Beyond 500 W, only one of the two stays powered.

Not advisable. With a panel connected, the station can read zero per cent as night falls and find its real level again as soon as the sun comes back. The manufacturer acknowledged in writing that this behaviour is not limited to one isolated unit, and no fix has been published.

The rating in detail

Our rating
3.8/5
Capacity and efficiency 3.8
Output power 3.8
Recharging 4.4
Ports and control 3.5
Build and endurance 3.7
Noise 3.5
EcoFlow River 2 Max: see the full details and price →