Battery draining after the macOS 27 update?

Updated 20 September 2026 · macOS 27 shipped on 14 September; our first machine has been on it since the 16th. What's measured so far is below, and it grows as the nights come in

Every autumn follows the same script: a new macOS ships, Macs update, and for a few days the battery seems worse. Some of that is real and temporary. Some of it is real and permanent: an app that hasn't caught up with the new system. And some of it is coincidence wearing the update as a disguise. The difference is measurable, and this page is about telling the three apart.

macOS 27, announced as Golden Gate at WWDC in June 2026, shipped on 14 September after a public beta that began on 13 July. It is the first macOS release for Apple silicon only, which happens to be the only hardware WattMate measures.

If you're on the beta right now

Then you're in a different situation from everyone else reading this, and it's worth separating what's known from what gets repeated:

• The rebuild below still applies, and applies again after every beta you install: each one re-indexes, re-caches and recompiles. On a monthly beta cadence you may simply never leave the settling period.
• A debug profile, if you installed one for Feedback Assistant, genuinely adds logging, and that part Apple documents. If you're not actively filing reports, removing it is the cheapest thing you can do.
• "Betas use more battery because they log everything" is the thing everyone says and nobody sources. We couldn't find an Apple statement to that effect, and we haven't measured it ourselves, so we're not going to assert it. It may well be true; it isn't demonstrated.

What we can say is that the arithmetic below doesn't care which build you're on. A watt is a watt, and an app that got more expensive after an update looks the same whether the update was a beta or a release.

The first days are supposed to be expensive

Right after an upgrade the system rebuilds itself: Spotlight re-indexes the disk, photo analysis walks your library again, caches are rebuilt, apps recompile their shader and code caches on first launch. This is real energy used on real work, and it ends. On most Macs the rebuild is a matter of a day or two of ordinary use.

The catch: macOS doesn't show you this work in units you can check. Activity Monitor's Energy Impact is a number without a unit, and Apple's own documentation calls it a relative measure. So "it's just indexing, give it a day" stays a guess, and three weeks later the same guess is covering for an app that actually broke.

WattMate measures power by app (real watts, not a score) and keeps history. A post-update spike you can watch shrink day by day is the system settling. A row that is as expensive on Thursday as it was on Monday is not.

How to tell whether the update is innocent

1. Wait out the rebuild, but watch it: don't take it on faith

If the expensive work is the system settling, it shrinks. Leave the Mac plugged in overnight once or twice, use it normally, and look at the same list each day. Shrinking rows are the rebuild doing its job. If nothing has shrunk after three days, stop waiting: the update is no longer the suspect, something specific is.

2. Find what got more expensive, by name

The comparison that matters is before versus after, and Activity Monitor doesn't remember yesterday. WattMate keeps its measurements, so "this app got hungrier after the update" is a check, not a feeling. Installed it only after updating? Then you won't have the before, but you will have today against tomorrow, and that answers the question that actually matters: is it settling, or is it staying?

3. Check the night

Updates change what happens while the lid is closed: indexing and analysis prefer to run when you're not looking. macOS keeps a power log of every wake-up and, most of the time, the name of the service that asked for it, which almost nobody reads. WattMate does. On the Mac we measured, across six days of that log, 204 of 206 scheduled wake-ups carried a name, and half of the total wake time belonged to Spotlight, the indexer.

4. Then decide, one quit at a time

When a specific app stays expensive, quit it while on battery and let the measurement answer. Two minutes after you quit, WattMate re-measures the Mac and prints what actually changed, or says honestly that the effect couldn't be isolated. That receipt is the difference between "I think the update ruined my battery" and "Teams got 2 W more expensive on macOS 27, quitting it gives me the minutes back, and I'm filing that bug".

When it isn't the update at all

Three look-alikes worth ruling out before blaming Golden Gate. An app shipped its own update the same week: the timing points at Apple, the watts point at the app. Battery age picked this month to show. Health is a slow curve, but people check it right after updates, so updates get the blame. Or the workload changed: a new project, a new call app, an external display. In all three cases the ranked list settles the argument faster than a forum thread does.

What we've measured so far, and what we haven't

Two days in, on one Mac (a 14-inch MacBook Pro with M3 Max, updated from 26.6.2 on 16 September). Not measured yet: what macOS 27 costs overnight or per day. That takes nights on battery, and we'd rather post nothing than a number from one evening. If a page tells you today how many watts macOS 27 costs, ask where the number came from.

Measured already: macOS 27 changed two things underneath every battery app, ours included. The chip's CPU energy counters, which macOS 26 updated continuously, now update in bursts a few minutes apart. On our machine the counter sat still for three minutes of load, then jumped by 768 joules at once, about 3.5 W of average draw arriving in a single reading. Any app that splits power per app from those counters at a 10-second rhythm now gets lumps instead of a stream. WattMate 1.3.1 notices the stillness, stops treating the counter as a measurement and falls back to the model it used before those counters existed; the Measurements page says so. Per-app watts, the total draw, the quit check and the receipt are not affected, because they come from the per-process energy counter and the battery's own gauge, and both still work on 27.

The second change: macOS 27 moved the battery's capacity readings to a different place in the system registry. WattMate 1.3 on macOS 27 lost its battery health page to that and ended every quit with "couldn't isolate the effect"; 1.3.1, out 16 September, reads the new layout.

Correction, 17 September: the first version of this page said macOS 27 had dropped the battery temperature reading altogether. We had that wrong. The reading did not vanish; it moved, together with the raw capacity readings, into a child entry of the same registry node (AppleSmartBatteryPack), and we had only looked at the parent. 1.3.1 does not read that entry, so on 27 it shows no temperature row and works out battery health from the wear-adjusted capacity rather than the measured one; 1.4 reads it, the row is back, and health is back on the measured figure. The same probe showed the graphics counter still advancing every few seconds while the CPU counter moves in bursts, so 1.4 keeps the graphics share measured on 27 as well.

And one limit we'll keep either way: macOS reports charge in whole percent, so the price of a single overnight wake-up can't be measured honestly. We say how often the Mac woke and for how long, not a made-up cost per wake.

Questions people ask

Is it normal for the battery to be worse right after updating?

For the first day or two: yes, partly. The system rebuilds indexes and caches, and that work is paid for in battery. It should visibly shrink day by day. Worse-and-staying-worse after three days is not the rebuild.

Should I quit apps to save battery after the update?

Only the ones that measurably matter. Most apps cost close to nothing in the background, and quitting them buys you nothing back. Measure first: quit the one expensive row (not everything), and check what the quit actually returned.

Does WattMate work on macOS 27?

Yes, from 1.3.1. Versions 1.3 and earlier run on macOS 27 but show no battery health page and end every quit with "couldn't isolate the effect", because macOS 27 moved two battery readings; 1.3.1 fixes both, and the app offers the update itself. What 1.3.1 does not show on 27, the graphics share of the app pool and battery temperature, is back in 1.4, see the correction above. WattMate requires macOS 14 or later on Apple silicon.

Related: where the charge goes with the lid closed · what "using significant energy" actually means · why the percentage jumps and drops · all the guides · who measures this