I keep hearing this meme that goes along the lines of “a Google search will use X amount of energy”, where X is often stated in a form of a scary number.
I think numbers are important.
According to one source a Google search is about 0.0003 kWh of energy, whereas a 3kW kettle running for one minute uses 3 x (1/60) = 1/20 = 0.05 kWh, which is 160 times as much (another piece uses an equivalent figure – Note 1).
On the UK grid, with a carbon intensity of approximately 300 gCO2/kWh (and falling) that would equate to 0.09 gCO2 or roughly 0.1 gCO2 per search. On a more carbon intensive grid it could be double this, so giving 0.2 gCO2 per search, which is the figure Google provided in response to The Sunday Times article by MIT graduate Alex Wissner-Gross (cited here), who had estimated 7 gCO2 per search.
If the average Brit does the equivalent of 100 searches a day, that would be:
100 x 0.0003 kWh = 0.03 kWh, whereas according to Prof. Mackay, our total energy use (including all forms) is 125 kWh per person per day in UK, over 4,000 times more.
But that is not to say the that the total energy used by the Google is trivial.
According to a Statista article, Google used over 10 teraWatthours globally in 2018 (10 TWh = 10,000,000,000 kWh), a huge number, yes.
But the IEA reports that world used 23,000 TWh in 2018. So Google searches would represent about 0.04% of the world’s energy on that basis, a not insignificant number, but hardly a priority when compared to electricity generation, transport, heating, food and forests. Of course, the internet is more than simply searches – we have data analysis, routers, databases, web sites, and much more. Forbes published findings from …
A new report from the Department of Energy’s Lawrence Berkeley National Laboratory figures that those data centers use an enormous amount of energy — some 70 billion kilowatt hours per year. That amounts to 1.8% of total American electricity consumption.
Other estimates indicate a rising percentage now in the low few percentage points, rivalling aviation. So I do not trivialise the impact of the internet overall as one ‘sector’ that needs to address its carbon footprint.
However, the question naturally arises, regarding the internet as a whole:
how much energy does it save, not travelling to a library, using remote conferencing, Facebooking family across the world rather than flying, etc., compared to the energy it uses?
If in future it enables us to have smarter transport systems, smart grids, smart heating, and so on, it could radically increase the efficiency of our energy use across all sectors. Of course, we would want it used in that way, rather than as a ‘trivial’ additional form of energy usage (e.g. hosting of virtual reality game).
It is by no means clear that the ‘balance sheet’ makes the internet a foe rather than friend to the planet.
Used wisely, the internet can be a great friend, if it stops us using planes, over-heating our homes, optimising public transport use, and so forth. This is not techno-fetishism, but the wise use of technology alongside the behavioural changes needed to find climate solutions. Technology alone is not the solution; solutions must be people centred.
Currently, the internet – in terms of its energy use – is a sideshow when it comes to its own energy consumption, when compared to the other things we do.
Stay focused people.
Time is short.
(c) Richard W. Erskine, 2019
I have discovered that messing about with ‘units’ can cause confusion. So here is an explainer. The cited article uses a figure of 0.3 Watt hours, or 0.3 Wh for short. The more commonly used unit of energy consumption is kilo Watt hours or kWh. As 1000 Wh = 1 kWh, so it remains true if we divide both sides by 1000: 1 Wh = 0.001 kWh. And one small step means 0.1 Wh = 0.0001 kWh. Hence, 0.3 Wh = 0.0003 kWh. If you don’t spot the ‘k’ things do get mighty confusing!