jj jashjacob.com
← field notes smart home 24 June 2026 5 min read

Six Months with the Aqara A100

129 days after installing it the batteries died and every way into my house died with them. Fingerprint, NFC, watch, app, all useless at once. What got me in was a powerbank ordered off Instamart and a USB-C port nobody ever reviews.

Everything else in this house fails politely. A sensor drops off the mesh and I find out days later. A light doesn’t turn on and I flip the switch like a caveman and move on.

The lock is not like that. The lock has exactly one failure mode that matters, and on 30 April I got the full experience. Standing outside my own front door, holding a phone and a watch and a fingerprint that were all completely useless, looking at a lock that had worked perfectly for four months and had simply run out of power.

The Aqara A100 went on the door on 22 December 2025. Six months in is the first time I have anything real to say about it, because six months is when it actually failed.

The setup

The A100 talks to the Aqara M3 hub like everything else in the house. On top of that it’s bound to Apple Home, with an Apple TV on the network acting as the home hub. That’s the part people miss when they buy a HomeKit compatible lock and wonder why half of it doesn’t work. The compatibility is in the lock, but automations and remote access need a hub sitting at home, and an Apple TV is the easiest way to have one that also does something else for a living.

The Apple TV side has been seamless. Genuinely zero issues in six months. It even plays nice with the Samsung SmartThings app, which I didn’t expect and haven’t seen anyone mention.

lock Aqara A100 Fingerprint · NFC · keypad · Home Key · USB-C failsafe
radio Aqara Hub M3 Zigbee 3.0 · Matter and Thread border router
home hub Apple TV Six months, zero issues · also talks to SmartThings

For the install I got a carpenter in. He knows doors, I know the lock, and those are different jobs. He handled the cutting and mounting, I read the manual beforehand and stood there explaining what the lock needed.

install

Don’t hand over the manual and leave. Most smart lock horror stories are misalignment, and misalignment happens in the twenty minutes you weren’t watching.

129 days

Installed 22 December, dead 30 April. 129 days on the first set of batteries.

first battery set 22 dec 2025 → 30 apr 2026
dec 10d
jan 31d
feb 28d
mar 31d
apr 29d
Lockout, 30 April 2026 ~1,500 unlock cycles · no warning I ever saw

That number needs context though, because this door gets opened a lot. 10 to 15 unlocks a day is normal here, and higher when there are guests or my parents are over. Over 129 days that’s somewhere around 1,500 cycles. The motor throwing the bolt is what eats the battery, not the radio. So battery life isn’t really a property of the lock, it’s a property of your household.

days per set, at ~1,500 cycles
4 / day
375 d
8 / day
188 d
12 / day
125 d
15 / day
100 d
20 / day
75 d
Derived from my own 129 days and ~1,500 unlocks. Do this maths with your own door.

A quiet door doing 4 unlocks a day would get over a year from the same set, which is probably exactly where the number on the box comes from. So don’t take 129 days from this post. Take roughly 1,500 unlocks per set and do your own maths.

caveat

These weren’t the batteries that came in the box. I bought the lock in Thailand, from the dotlife store, and binned the originals at the airport rather than explain loose cells to an airline. A lock motor is a demanding load and cheap cells show it. The second set is the same story, so if it also dies around 129 I’ll know the number is real. I’ll update this when it does.

The lockout

Here’s what I can’t tell you: whether the low battery warning ever fired. Maybe it did, somewhere in a stack of notifications. I never saw it. The first time the battery level entered my brain was when I was outside the house and the door wouldn’t open.

And here’s the thing I hadn’t thought through. Every entry method on this lock runs off the same battery. Fingerprint, NFC, keypad, app, watch. On paper that’s five options and it reads like redundancy. It isn’t. They’re all just different ways of asking the same motor to move, and when the motor has no power, none of them mean anything.

entry methodpowered byat 0%
Fingerprintlock batterydead
NFC taglock batterydead
Keypadlock batterydead
Apple Watch / iPhonelock batterydead
Aqara applock batterydead
Physical keynothinginside the house
USB-C emergency portany powerbankopened it
Entry methods audited at 0% battery, 30 April 2026
Five methods sharing one battery is one method wearing five hats.

The physical key would have worked. The physical key was inside the house, because nobody carries the key to a keyless lock. That’s not a discipline failure, that’s the whole reason you bought the thing. The key stops being part of your life about a week after installation, and pretending otherwise is how people end up with a false sense of backup.

What saved me is the emergency USB-C port on the outside of the lock. Feed it power and the lock wakes up and works normally, dead batteries or not. So I ordered a powerbank on Swiggy Instamart, waited outside my house, plugged it in, and the door opened in about ten seconds. Fingerprint read, motor turned, done.

Total damage: one wasted evening and a powerbank I didn’t need. No locksmith, no drilled cylinder, no drama. And the failsafe got tested for real, which honestly no review ever does, because on a review timescale it never comes up. It’s the most boring line on the spec sheet and on 30 April it was the only one that mattered.

If you’re choosing between smart locks and one has an external power port and the other doesn’t, that’s the whole comparison. Everything else is preference.

Who opens this door, and how

Six months in, the household has sorted itself into three methods, and the split is by person, not by feature.

whomethodstatus
MeWatch / iPhone tapdefault
GuestsNFC tagdefault
Everyone at homeFingerprintdefault
NobodyAqara appunused
Observed over six months · nobody uses the app

Notice what’s not on the list. Nobody uses the app to open the door. The app works fine, it’s just slower than every other option, and a two second delay is enough to permanently lose to a wrist tap. Friction beats features, every time.

What I’d tell someone before they buy

  • Buy the lock with external emergency power. Not a nice to have. It’s the difference between a wasted evening and a broken door.
  • Budget batteries per unlock, not per month. Around 1,500 cycles a set in my house. Multiply by your own door.
  • Swap on a date, not on a warning. The warning is one notification competing with a hundred others, and mine lost. A calendar entry at 110 days competes with nothing.
  • Don’t cheap out on the cells. And don’t bin the originals at an airport like I did.
  • Keep a powerbank in the bag, not a key in your pocket. The key is the backup that sounds sensible and nobody sustains. The powerbank is the one you’ll actually have on you.
  • Count your entry methods honestly. If they all run off one battery, you have one entry method.

Six months in

Would I install it again? Yes, same lock, same door, tomorrow. For 129 days this was the least interesting device in the house, and on a front door that’s the highest compliment there is. Then it failed, recoverably, with a way back in that cost me ten seconds and one online order. That’s a well designed failure.

I trust the lock more after the lockout than I did before it.

Measured, not assumed

129
days, first battery set
10-15
unlocks a day
1
lockout
10
seconds to open, once powered