Does deleting email really help with climate change? (part one)

At my day job it is apparently Green Week in which we're encouraged to consider our impact on the natural environment. The IT department has taken it upon itself to promote "digital decluttering" as a vital way to decrease your carbon footprint. Apparently we need to delete old emails and files to save the planet. Ever the sceptic, I asked to see the receipts on that claim. Receipts were not forthcoming.

Being a long-term professional pedant I took it upon myself to try to get to the bottom of this claim.

I should perhaps point out first that I'm not in any way attempting to undermine the science of the climate crisis. Everyone who's not a fringe nutcase knows that human activity is leading to climate change and that the release of carbon dioxide into the atmosphere is one of the biggest causes. I just want to understand whether my twenty year archive of Gmail is making any difference to that.

The basic claim is something like this: storing data requires servers, servers require power, (some) power generation releases carbon dioxide; more data means more servers therefore deleting data means fewer servers and less CO2 going into the environment. On the face of it, it sounds reasonable and has a certain "truthiness" to it.

According to my IT department, deleting 1GB of emails can save between three and seven kilowatt-hours per year. Let's check those numbers. Backblaze are a cloud data storage company who regularly publish information on the hard drives they are using in their data centres. We can probably assume that the hardware they use is typical of the hyperscale hosting companies. In their 2025 report, the hard drives they've most recently added are Seagate 16TB SATA drives (model number ST16000NM000J). Seagate's data sheet for this drive specifies an idle power consumption of 5.3 watts. Multiple that by 24 and we get 127 watt-hours per day, and by 365 to get 46.428 kilowatt-hours in a year. That's for a drive that holds 16 terabytes of data. We are only talking one gigabyte (a gigabyte is one one-thousandth of a terabyte - give or take, don't harass me about the 1024 vs 1000 debate).

My back of the envelope maths suggests a gigabyte of unused data sitting unloved on a data centre hard drive is actually responsible for about 0.0029 kilowatt-hours of power per year. Even at the Seagate drive's maximum power consumption of 9.4 watts we're still two orders of magnitude out from the original claim.

We could assume that the data centre will store at least one other copy of our data as a backup against a hard drive failing. That's still only going to make a marginal difference and not get us anywhere near the 3-7 kWh claim.

Maybe I'm being too literal in just looking at the actual physical storage medium. Maybe IT are talking about the data centre itself. The sum total power used to run every server, air conditioning system, security gates, hell let's include the microwave in the staff kitchen too; all of it divided by the total number of gigabytes stored. The first number is certainly big but so is the second number.

Getting out the back of our envelope again and checking some more of Backblaze's blog posts we can assume sixty 16TB drives per 2kW server and add 50% overhead for cooling etc, which gives us 3.125 watts per terabyte and 1.14 kilo watt-hours. That's starting to get us within sight of the ballpark but it's massively overstating the power requirement. Those servers aren't going to be pulling the maximum rated watts of their power supply unless the data centre manager is a pyromaniac.

So far I've been talking about the power required to store data. The elephant in the room of this supposedly green pro-tip is the power required to delete data. In order to search for old emails or dig out those with big attachments requires processing. Whether that's on a clunky old PC or a modern fast, efficient virtual machine it still needs a little bit of electricity and that's power that wouldn't be needed if the data was just allowed to gracefully grow old on a hard drive. By encouraging people to prune their email folders the well-meaning IT department is actually causing more power consumption and a slightly larger carbon footprint.

Coming up in part two: just how much carbon dioxide is released by emailing and where do those widely quoted numbers come from?