Both my Nitecore NB20000s are tired after several years of use, and the capacity has quietly drained away, so my phone and camera batteries no longer reach day three. Time to replace the battery packs.
Replacing them meant weighing the three key factors: weight, capacity, and cost. The method turned out to be worth more complex than I thought.
My decision: I have bought an INIU pack for three-season use, and I am buying a heated Nitecore pack for winter. Two packs, not one, for reasons that come down to physics rather than preference.
Watt-hours, a Water Tank Analogy
Every power bank is sold on a milliamp-hour number. On its own it cannot be compared between packs, and it takes one piece of plumbing to see why.
Think of the battery as a water tank feeding a hose.
- Amp-hours (or milliamp-hours) are litres. How much water is in the tank.
- Volts are pressure. How hard the water is being pushed down the hose.
- Watt-hours are litres times pressure. The actual work the water can do when it comes out.
A tank holding 100 litres at high pressure will drive a water wheel for longer than a tank holding 100 litres at a dribble. Same litres, different useful output. Quoting litres alone tells you nothing until you also know the pressure.
That is exactly the position mAh puts you in. The mAh on the box is measured at the pressure inside the cells, which is somewhere around 3.6 to 3.85 V depending on the chemistry. Your phone needs 5 V or more, so the pack has to pump the water up to a higher pressure before any of it reaches your device, and running that pump costs some of the water.
Watt-hours are litres and pressure multiplied together, which makes them a single number you can compare across any pack, measured anywhere in the circuit:
watt-hours = volts × amp-hours
So a 20,000 mAh pack (20 amp-hours) at 3.85 V holds 77 Wh. The same 20,000 mAh at 3.6 V holds 72 Wh. Identical headline number, 5 Wh apart, because the pressure is different.
First move on any comparison: convert everything to watt-hours. If a spec sheet gives you mAh and nothing else, find the cell voltage and multiply. If it gives you mAh at 5 V, multiply by 5. Everything below is in watt-hours for that reason.
Two Different Watt-hour Figures, and Which One You Are Reading
Once everything is in watt-hours there is a second thing to line up, because packs are rated at two different points in the circuit.
Cell energy is what is stored inside the cells: volts times amp-hours, the figure you get from the headline mAh. Output energy is what comes out of the USB port. The two differ because the pack has to pump the cells' 3.6 to 3.85 V up to 5 V or higher, and running that pump uses some of the energy. The cells also self-discharge slowly and the safety electronics take a share.
Back to the tank analogy: cell energy is what is in the tank, output energy is what actually reaches the water wheel after the pump has taken its cut.
Nitecore publish both. They call the second one rated energy and quote it in mAh at 5 V, so multiply by 5 for watt-hours:
| Model | Cell energy | Rated energy at 5 V | Output as % of cells |
|---|---|---|---|
| NB20000 Gen 1 | 77 Wh | 13,500 mAh → 67.5 Wh | 88% |
| Carbo 20000 | 77 Wh | 13,500 mAh → 67.5 Wh | 88% |
| Summit 20000 | 77 Wh | 13,500 mAh → 67.5 Wh | 88% |
| NB20000 Gen 3 | 72 Wh | 10,100 mAh → 50.5 Wh | 70% |
Those are the numbers as published. Two things to note before using them.
The first is that the 88% appears identically on three physically different packs, and the Gen 3 is the one outlier at 70%. Typical real-world conversion for this kind of circuit sits in the 70 to 85% range, so the Gen 3's figure is the one in the expected band. That does not tell you the Gen 3 is a worse pack; it more likely means Nitecore changed how they derive the figure between generations. Ratings are only comparable when they were produced the same way, and across these four they were not.
The second is that INIU and Xtorm publish no watt-hour figure at all. INIU's spec table gives 20,000 mAh and nothing else, so the only figure available for them is cell energy, worked out from the mAh.
Where a manufacturer stays quiet, the output figure usually exists anyway, because some very helpful person on the internet has put the pack on a USB load tester, drained it flat at a fixed current and written down what came out. Review sites do it, and so do a good number of hobbyists with a cheap meter and a spreadsheet. That is where most of the real-world 70 to 85% range comes from, and it is often better evidence than the spec sheet, since it measures the same thing you care about: watt-hours out of the port.
It comes with the same rule attached, though. A figure measured at 1 A and a figure measured at 3 A will not match, because a harder draw wastes more in the pump, so check the test current before you put two testers' numbers side by side. And check the tester is using the model you are buying, which with INIU's naming is not a given.
That leaves one column everybody can be put in, and one only Nitecore can. So the comparison has to run on cell energy, with the output figures noted separately rather than mixed in. Mixing the two is what makes spec sheets look contradictory.
Comparing on the Figures Every Pack Publishes
Three specs are available for all six packs and are directly comparable: mass, which you can weigh; cell energy in watt-hours, which follows from the cells inside; and price.
| Model | Weight | Cell energy | CHF | CHF per Wh | Max output |
|---|---|---|---|---|---|
| INIU P512 | 346 g (measured) [2] | ~72 Wh (not listed) | 38 | ~0.53 | 22.5 W |
| Xtorm Fuel Series 5 | 400 g [13] | 74 Wh [13] | 50 | 0.68 | 35 W |
| Nitecore NB20000 Gen 1 | 324.5 g [3] | 77 Wh [3] | 110 | 1.43 | 45 W |
| Nitecore NB20000 Gen 3 | 291 g [4] | 72 Wh [4] | 120 | 1.67 | 22.5 W |
| Nitecore Carbo 20000 | 295.5 g [5] | 77 Wh [5] | 160 | 2.08 | 20 W |
| Nitecore Summit 20000 | 315 g [6] | 77 Wh [6] | 170 | 2.21 | 20 W |
Two things fall straight out of that table.
The difference in weight is small and the difference in price is not. From the heaviest pack to the lightest is 109 g. From the cheapest to the dearest is 132 CHF. Going from the INIU to the Carbo buys you 50 g for 122 CHF, which is 2.44 CHF per gram saved. For comparison, I have paid that kind of rate for a shelter, where the weight saved is measured in hundreds of grams and the kit keeps you alive. Paying it for a battery is a harder case to make.
The Gen 3 is not the upgrade its position in the range implies. It sheds 33 g against the Gen 1 and halves the output, 45 W down to 22.5 W, and if you use two ports at once the total drops to 15 W [4]. If you charge a laptop off your power bank, the Gen 1 was the pack that did it and nothing in the current Nitecore lineup replaces it. The Gen 1 is discontinued, which is part of why I am here.
Output Power Is a Separate Spec from Capacity
Watt-hours tell you how much energy a pack holds. Watts tell you how fast it can deliver it, and the two are independent. The P512 is 22.5 W on a single port, dropping to 15 W total across more than one [1]. INIU also sell 100 W packs under similar model names, so it is worth reading the wattage off the specific model rather than the range.
For what I actually charge, that is fine. A phone, a watch, a headtorch, occasionally a camera battery. 22.5 W fills a phone from flat in a little over an hour, and on trail I am charging overnight or during a long lunch, not racing a departure gate. If you run a laptop or a drone, this is the spec that rules the INIU out, and you could look at the Xtorm's 35 W instead and accepting the 54 g.
Check it against your own kit. The cheapest pack in the table is cheap partly because of this.
Where the Physics Bites: Sub-zero Temperatures
Everything above is a question of value. Cold is a question of whether the pack works at all.
Lithium-ion cells lose capacity as they get colder, because the electrolyte thickens and ions move through it more slowly. This is temporary and reversible: warm the pack and the capacity comes back. But while it is cold, it is gone.
Battery University's figure is that a cell delivering 100% at 27 °C typically gives about 50% at −18 °C [7], and the peer-reviewed spread at −20 °C runs roughly 50 to 66% depending on chemistry [8]. Push further and there is almost nothing left: 18650 cells, the type inside most power banks, delivered about 5% of their room-temperature energy at −40 °C, and their power output fell by around 99% [9].
Half your battery capacity, at the temperature of an ordinary alpine winter night.
There is a second, nastier problem. Charging a lithium cell below freezing plates metallic lithium onto the anode instead of storing it inside. That damage is permanent, it accumulates, and it degrades safety as well as capacity [10]. Work on cells cut open after ageing confirms plating becomes the main way they wear out below about 25 °C [11]. So plugging a cold pack into a solar panel or a hut socket the moment you arrive does permanent harm.
Two Rules That Follow, and They Apply to Any Pack You Own
Sleep with it. Into the quilt at night, or an inside pocket against your body during the day. A battery at body temperature delivers its rated capacity; the same battery in the lid of your rucksack at −15 °C delivers half.
Never charge a pack that is below freezing. Warm it first, in your hands or in your sleeping bag, and give it long enough to reach temperature rather than just feel less cold. A cold charge costs you capacity you never get back.
Those two habits are free and they matter more than which pack you buy. But they have a limit, which is where the heated packs come in.
The Winter Pack
The Nitecore Summit 20000 puts heating elements inside the case. Below 0 °C, whenever power is flowing in or out, temperature sensors switch the heating on and hold the cells warm enough to work. There is a separate pre-heating mode for when the pack is plugged into a power source in the cold. Nitecore rate it to work down to −40 °C, and state around 70% available capacity at −20 °C, against 10% or nothing at all for an ordinary pack [6].
Where that 70% figure came from: it is Nitecore's own, and I could find no independent test of the Summit at temperature. Nitecore's current product page lists only the −40 °C operating range and does not describe the heating system; the launch announcement that carried the detail is no longer online. So it sits in a different evidence class from the mass and price figures above, and I have kept it out of the comparison table for that reason. The mechanism itself is straightforward and well understood.
At 170 CHF and 315 g it is expensive and it is not the lightest. It is also the only pack in the table that solves a problem I have. Most of my winter days are ski tours and high camps where the pack lives in the cold whether I like it or not, and where a dead phone is a backup navigation problem rather than an inconvenience.
This is the same trade I make everywhere else in my kit. Functionality and safety outweigh weight. I do not carry the lightest sleeping system either.
The Decision
Three-season: INIU P512, 38 CHF.
It is the cheapest energy in the table by a wide margin, 0.53 CHF per watt-hour against 2.08 for the Carbo. Its 346 g sits 50 g above the lightest option and 54 g below the heaviest. It has three ports, it will charge your phone while it is charging itself, and it drops automatically to a trickle for small devices like a watch or earbuds, which the ultralight packs often lack [1][2]. The 22.5 W ceiling is enough for what I charge.
The argument for spending four times as much rests on the output-energy column, which is published by one manufacturer out of three and on two different bases within that one. On the figures that line up across all six packs, the premium buys 50 g.
One mark against it. CleverHiker measured an 8.2% drop in available power in cold conditions, one of the worse showings in their test group [2]. For a three-season pack that is acceptable. It also confirms it is not the pack for February.
Winter: the Nitecore Summit 20000, on the list for before November.
The cost per watt-hour is the worst in the table and I do not care, because the comparison is not against other power banks. It is against having no usable power at −20 °C.
Total: around 208 CHF for both, which is less than a single Carbo, and covers a temperature range no single pack in the table covers on its own.
One More Thing, If You Fly
The rules changed this year and the old advice is out of date. ICAO issued an addendum to the Technical Instructions on 27 March 2026, and it goes well beyond the familiar watt-hour limit [12]. Power banks for personal use must now travel in carry-on only, must not be recharged onboard, should not be used to charge a device onboard, and are limited to two per person.
That last one is the trap for anyone running the two-pack strategy I have just recommended. Two is the ceiling, not a suggestion, and if you were planning to also bring a spare you are over.
The 100 Wh limit still stands for carry-on, and every pack discussed here is comfortably under it. The 100 to 160 Wh band that used to be allowed with airline approval is being removed from the 68th edition of the IATA regulations in force from 1 January 2027, so treat 100 Wh as a hard ceiling from here on. Different countries are enforcing it at different speeds, so check your specific airline rather than assuming.
The Bottom Line
Four steps, and they work on any set of packs:
- Convert every pack to watt-hours. Volts times amp-hours. mAh on its own cannot be compared.
- Check which point in the circuit each figure describes, cell energy or output at the port, and never put the two in the same column.
- Compare only figures produced the same way. Where a measured weight exists, use measured for every pack; where it does not, use published for every pack.
- Then divide price by watt-hours and see what the premium is buying, which in this category is a surprisingly small number of grams.
Then buy for the conditions rather than for the spreadsheet. The differences between these packs are smaller than the 50% you lose to a cold night, which means the temperature the pack lives at matters more than which pack it is.
Train hard. Fuel right. Pack smart. Go further.
References
- INIU. INIU Carry P512 Power Bank 22.5W, 20000mAh product page. Weight 313.5 g claimed, 22.5 W max single-port output, 15 W multi-port, 3 ports, low-current mode. No watt-hour figure published. iniushop.com
- CleverHiker. INIU 20000 PD Power Bank Review. Independently measured 12.2 oz (~346 g); confirms pass-through charging and automatic low-current mode; reports an 8.2% drop in available power in cold conditions. cleverhiker.com
- Nitecore. NB20000 user manual. 324.5 g; 20,000 mAh at 3.85 V = 77 Wh; rated energy 13,500 mAh at 5 V; 45 W max on USB-C OUT1; carbon-fibre-reinforced polymer frame, IPX5. manualslib.com / nitecore.co.uk
- Nitecore. NB20000 Gen 3 product page. 291 g ±5 g; 20,000 mAh at 3.6 V = 72 Wh; rated energy 10,100 mAh at 5 V; 22.5 W max per port, 15 W total. nitecore.com
- Nitecore. Carbo 20000 product page. 295.5 g ±5 g; 77 Wh; rated energy 13,500 mAh at 5 V; 20 W USB-C, total output capped at 5V/3A; unibody carbon fibre shell, IPX5. nitecore.us
- Nitecore. Summit 20000 product page and product manual. 315 g ±5 g; 77 Wh; operating ambient minimum −40 °C. The self-heating description and the "~70% available capacity at −20 °C" figure come from Nitecore's launch announcement and manual, reproduced by retailers; the original announcement URL now returns 404 and the current product page does not mention heating. No independent test located. nitecore.us
- Battery University. BU-502: Discharging at High and Low Temperatures. A cell delivering 100% at 27 °C typically delivers ~50% at −18 °C; loss is temporary and recovers on rewarming. batteryuniversity.com
- Lithium-ion batteries for low-temperature applications. Journal of Power Sources, 2023. Capacity retention at −20 °C approximately 50–66% depending on chemistry, additives and C-rate. sciencedirect.com
- Nagasubramanian G. Electrical characteristics of 18650 Li-ion cells at low temperatures. Journal of Applied Electrochemistry. 2001;31:99-104. Energy density fell from ~100 Wh/l at 25 °C to ~5 Wh/l at −40 °C; power from ~800 W/l to under 10 W/l. doi:10.1023/A:1004113825283
- Battery University. BU-410: Charging at High and Low Temperatures. Charging below freezing plates metallic lithium on the anode, causing permanent degradation in performance and safety. batteryuniversity.com
- Waldmann T, Wilka M, Kasper M, Fleischhammer M, Wohlfahrt-Mehrens M. Temperature dependent ageing mechanisms in Lithium-ion batteries: A Post-Mortem study. Journal of Power Sources. 2014;262:129-135. Below ~25 °C the dominant ageing mechanism switches to lithium plating. doi:10.1016/j.jpowsour.2014.03.112
- IATA. Guidance to Operators: Changes Regarding Power Banks, 31 March 2026. ICAO Technical Instructions addendum effective 27 March 2026: carry-on only, no recharging onboard, no more than two per person, individual short-circuit protection. The 100–160 Wh with-approval band will not carry into the 68th edition DGRs (in force 1 January 2027). iata.org (PDF)
- Xtorm. Fuel Series 5 20,000 mAh (FS5201). 400 g and 35 W per the manufacturer spec table; 74 Wh per Swiss retailer Digitec/Galaxus. Xtorm publish no watt-hour figure themselves. xtorm.eu / galaxus.de
