Which apps drain a Mac's battery

Updated 30 September 2026 · every figure below is measured on two Apple silicon Macs and links to the page where the run is written out in full. Who measures this.

Activity Monitor gives each app an "Energy Impact" score with no unit. These are watts, measured with WattMate on battery, across more than a thousand windows of about 90 seconds each, and the spread between apps turns out to be about four hundredfold.

The measured cost, side by side

Two figures where there are two: the same app on two different Macs. The answer is not a property of the app alone. Chrome costs almost twice as much on the quiet machine as on the busy one, with a different person's tabs open on each. An earlier version of this paragraph put that down to less competition for the same watts on a quiet Mac. Our own check does not support it: on the same Mac, Slack and Telegram used more, not less, while Chrome was busy.

Claude's row is the exception to that reading: its two figures are one machine, not two. The desktop app and Claude Code are separate application bundles, so macOS and this table count them apart. They behave nothing alike.

The popular suspect is innocent

Ask which app is eating a Mac's battery and the usual answers are Slack and Chrome, in that order. Measured with WattMate, Slack sits at 0.01 W: the smallest number in the table, roughly a thirtieth of Telegram and a four hundredth of Chrome on the same machine. Quitting it gives back minutes you would not notice in a working day.

Chrome is a real cost. But even there, the arithmetic almost everyone runs is wrong: you cannot subtract an app's watts from the machine's and call the remainder savings, because the screen and simply being awake do not go away when the app does. On the quiet Mac that subtraction broke in 117 of 157 windows (three quarters of the time).

The one that is worth quitting

Xcode is the clearest case in the set: a typical 3.7 W, a middle half spanning 1.7 to 13.0 W, and a busiest window of 48.5 W. It is the only app here whose peak is bigger than most Macs draw in total. If a build is running, that is where the charge is going, and the page has the full 217-window run.

Questions people ask

Which app drains a MacBook battery the most?

Of the six measured here, Xcode: a typical 3.7 W and a busiest window of 48.5 W, which is larger than most Macs draw in total. Chrome is second at 2.3–4.1 W depending on the machine. But the honest answer for your Mac is your own list: Chrome alone cost almost twice as much on one of our Macs as on the other.

Does Slack drain a Mac's battery?

Measured across 153 of 157 windows on one Mac and 173 of 339 on another, Slack sat at 0.01 W, the cheapest app in the set, and roughly a four hundredth of Chrome on the same machine. It is the most frequently accused app in this category and the least deserving of it.

Why does Activity Monitor show nothing when the fan is loud?

Energy Impact is a relative score with no unit, sampled when you open the window: the app that spiked two minutes ago is no longer there to see. Watts measured over continuous windows of about 90 seconds keep the history the score throws away.

Can macOS tell which app used the GPU?

No. macOS has a total for graphics energy but no per-app counter for it, so any split between apps (ours included) is inferred from the driver's per-process GPU counter, which is not the same as time the GPU spent busy. The total is exact; the split is not.

Why per-app watts are hard, and who can print them

Chrome holds a median of 16 processes on one Mac and 19 on the other, peaking at 40: renderers, GPU helpers, the network service, each its own line in Activity Monitor. A tool that samples once a second misses everything that lived and died in between. WattMate credits that energy back through kernel resource coalitions, which is why its Chrome row is one number instead of forty. The measurement behind that, and why the obvious approach loses four fifths of the work, is in a separate write-up.

Graphics is the honest limit: macOS has a total for GPU energy but no owner for it, so the split between apps is inferred from the driver's per-process GPU counter, not measured. The full method, and where it stops, is in how we count.

What to do with a number like this

A watt figure alone still does not answer the question people actually have, which is how much longer the charge would last. WattMate names that in minutes before you quit anything, then measures the Mac again two minutes after you do and prints what really changed, or says plainly that the effect could not be isolated. That re-measurement is the part we have not found in any other Mac battery app.

Related: what macOS means by "Using Significant Energy" · whether quitting apps saves battery at all · how many watts a MacBook uses · Mac battery apps by the job they do · all the guides · the closest product to this, reviewed