How much battery does Chrome actually use on a Mac?

Measured 14 August 2026 · 374 completed measurement windows across two Apple silicon MacBooks · every number below comes from those logs and is reproducible with the script at the bottom

Ask the internet and you get adjectives. Chrome is "a battery hog", Safari is "more efficient" — and not one page gives you a number, because macOS doesn't hand out one: Activity Monitor's Energy Impact column is a relative score with no unit. So we measured it.

The answer: 24 minutes on one Mac, 142 on the other

We measured Chrome on two Apple silicon MacBooks over 374 windows between them. The interesting result isn't a number — it's that the two machines disagree by almost six times about what quitting Chrome is worth:

Same browser, same operating system, six-fold difference in what quitting it buys you. And the reason has almost nothing to do with Chrome.

Why the same app is worth six times more on one Mac

On Mac 2 the whole machine used 7.9 W and Chrome was 4.1 of it — the browser was the workload. On Mac 1 the machine used 25 W, of which Chrome was 2.3: it was a rounding error next to compilers and tests. Quitting the browser on a machine that's busy doing something else changes very little, however hungry the browser is in absolute terms.

That's the sentence worth taking away from this page: an app's cost is only meaningful next to what else your Mac is doing. Which is exactly why every "Chrome uses X watts" claim — including the two above — is useless without your own machine's total beside it.

The spread inside one machine is bigger still

Look at the extremes: on Mac 1 Chrome ranged from 0.04 W to 24.5 W — a factor of six hundred, on one machine, in one week. On Mac 2, 0.07 to 17.8. "How much does Chrome use" doesn't have a single answer even for a single person. It depends on what's open in it right now, which is why a live measurement beats any published figure, ours included.

The arithmetic almost everyone gets wrong

Here's the trap, and we fell into it on the first draft of this page. Chrome's median on Mac 2 is 4.1 W and the machine's is 7.9 W, so quitting should leave 3.8 W. Except that Mac's floor — the screen and simply being awake — is 5.5 W. You cannot subtract an app and land beneath the floor.

We counted how often the naive subtraction breaks: 117 of 157 windows on Mac 2 — 75% of them — and 17 of 205 on Mac 1, where the floor is a smaller share of a bigger total. So the honest figure is computed per window with the floor as a lower bound, which is where the 24 and 142 minutes in the table come from rather than the four hours subtraction would have promised.

This is the arithmetic WattMate runs for every app in your list. And then it does the thing no arithmetic can replace: two minutes after you actually quit, it measures the Mac again and prints what really changed — or says plainly that the effect couldn't be isolated.

Sixteen to forty processes wearing one name

Chrome held a median of 16 processes on one Mac and 19 on the other, peaking at 40 — renderers, GPU helpers, the network service, extensions, each its own line in Activity Monitor. That's the practical reason per-app measurement is harder than it looks: a tab you close takes its process with it, and 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 comparison we can't honestly make

You'd expect a Chrome-versus-Safari table here. We don't have one: Safari accumulated 0.01 Wh on each machine across the whole period — neither of us really uses it. Publishing "Chrome uses a thousand times more than Safari" from that would be a lie made of true numbers. When we have a machine where both browsers do real work, we'll measure it and add it here.

What this does and doesn't prove

Two Macs, two people's tabs. Yours will differ again — and the two machines here already differ from each other more than most readers would guess. Treat these as worked examples of the method, not as your numbers.

What generalises is the shape, not the value: a browser is usually the largest single app on a Mac; its cost swings by two orders of magnitude with what's open; quitting it returns less than its own usage suggests, because the floor stays; and its worth depends on what else the machine is doing. Those four hold regardless of whose Mac it is.

Measure your own

Every number here comes from one script reading WattMate's own database:

python3 Scripts/app-energy-report.py <database> "Google Chrome"

It prints the median, the quartiles, the extremes and the process count — and refuses to do anything cleverer, because everything cleverer turned out to be wrong. Point it at your own database and you'll get your own row of this table. If you'd rather just see it on screen, the trial shows the same thing live, for whichever app is on top of your list — which, statistically, is probably this one.

Related: does quitting apps actually help? · healthy battery, short day · how we count