The statistics they do cite are often misleading, and are never very convincing. For example, comparing the energy use from streaming a video to riding a train or walking. In reality, all of those energy uses are negligible. In other words, they are comparing video streaming to the most efficient versions of transportation, and ignoring the sources of most transportation-related emissions. Run that comparison between video streaming and flying to a meeting in person!
It reminds me of an article I once saw funded by the meat industry which showed that the water requirements to produce a calorie of beef are lower than a calorie of lettuce (ignoring calorie-dense plants such as grains and legumes that are much more efficient).
This one is a doozy:
> This grand ‘experiment’ in shifting societal energy use is visible, at least indirectly, in one high-level fact set. By the first week of April, U.S. gasoline use had collapsed by 30 percent, but overall electric demand was down less than seven percent.
By the article’s own numbers, energy use actually went down! Then they somehow use this fact to prove that the digital future will use more energy?
Later, they make more strange comparisons, like datacenters to skyscrapers, when of course, most of the world’s energy use doesn’t come from skyscrapers. And they bizarrely compare datacenter energy use to electric car energy use, when again, electric cars obviously aren’t yet a major component of energy usage.
Overall, I feel this article is not fact-driven reporting. It’s twisting the numbers with the aim of making people who care about the environment misdirect their concerns toward software companies.
To make a more constructive recommendation - if you’re concerned about carbon emissions, start by looking at a breakdown of which industries are responsible for the most emissions!
The movement to the cloud is more energy efficient than companies running everything on-premises.
Usage of remote working tools will lower the carbon footprint of employees and workers.
How much these efficiencies will be canceled out by the mass movement to digital transformation will be obvious in ten years from now.
Meanwhile there is an increase in clean energy that will cancel out some of this growth in digital transformation.
For private users, how is downloading a song or movie again each time I want to watch it more energy efficient that having it on my hard drive?
For corporate users, "Cloud software" is to a large part replacing traditional desktop software. As such, the comparison is only partially with in-house data centers - it's also with software running exclusively on users' actual work PCs - which they would shut down after they have finished working.
Even if you compare client/server software, you leaving out the energy expended by routing the data over the internet. Even if the cloud service itself can use energy more efficiently inside their data center, routing uses energy that you wouldn't need for a server in the basement.
> How much these efficiencies will be canceled out by the mass movement to digital transformation will be obvious in ten years from now.
Which is probably too late.
